No full-text available
To read the full-text of this research,
you can request a copy directly from the author.
Spatial and Temporal Bilateral Filter for Infrared Small Target Enhancement
Abstract
This paper presents a spatial and temporal bilateral filter (BF) to detect target trajectories, by extracting spatial target information using a spatial BF and temporal target information using a temporal BF. Background prediction when it is covered by targets is the key to small target detection. In order to apply the BF to a small target detection field for this purpose, this paper presents a novel spatial and temporal BF with an adaptive standard deviation to predict spatial background and temporal background profiles, based on analysis of the blocks surrounding a spatial and temporal filter window. In order to discriminate between the edge or object regions with a flat background and the target region spatially and temporally, spatial and temporal variances of the blocks surrounding the filter window are calculated in a spatial infrared (IR) image and temporal profile. The spatial and temporal variances adjust standard deviations of the spatial and temporal BF. Through this procedure, spatial background and temporal background profiles are predicted, and then small targets can be detected by subtracting the predicted spatial background (and temporal background profile) from the original IR image (and original temporal profile) and multiplying spatial and temporal target information. To compare existing target detection methods and the proposed method, signal-to-clutter ratio gain (SCRG) and background suppression factor (BSF) are employed for spatial performance comparison and receiver operating characteristics (ROC) is used for detection-performance comparison of the target trajectory. Experimental results show that the proposed method has a superior target detection rate and a lower false-alarm rate.
... Furthermore, in digital imaging systems, the image processing modules are required to increase the quality of an output video. These image processing modules implement advance image processing techniques like non-uniformity correction [12] , features enhancement [13] , noise removal [14] and spatial filtering [15] . In conventional ROICs, the image processing enhancements are performed off-chip due to the area and power limitations, which reduces the overall performance of an imaging systems [6] . ...
... The sum of integration time and dispatch time is known as frame time. Finally, the raw image is given to the image processing module for further enhancements through various image processing algorithms [12][13][14][15] . Once the image processing is performed, the data is available in a standard digital video format (e.g. ...
... The most important factors, which are considered in on-chip image processing, are area, power, processing speed and performance (image quality). Fig. 6 shows that the image processing block in the proposed DROIC implements four dedicated image processing algorithms: Bad pixel removal [14] , Non-uniformity correction [12] , Image enhancement [13] and Spatial image filtering [15] . ...
Readout Integrated Circuits (ROICs) are used to process the signal from a detector array in amaging systems. The important performance factors for ROICs, among others, include image resolution, power consumption and reconfigurability. However, the existing ROICs target only a single or few performance parameters at a time. Furthermore, the off-chip image processing in state-of-the-art methods increases the parasitic effects. The on-chip image processing solutions have also been proposed previously, but either with a low resolution or analog processing. To overcome these issues, this article presents a high resolution and low power configurable Digital Readout Integrated Circuit (DROIC) with on-chip image processing. The proposed DROIC has been designed for a digital pixel array of 1K × 1K with 10 um pixel pitch, using 65 nm CMOS process.
... Infrared (IR) imaging is widely used in military, medical, and industrial field since it is able to create a visual with otherwise non-visible wavelength band to the human eye [1,2]. Recently, IR imaging systems, such as IR search and track (IRST), IR countermeasures (IRCM), and directional IR countermeasure (DIRCM) etc, have been used for developing and evaluating military IR imaging systems. ...
... Additionally, for this purpose, target detection methods for detecting small targets or moving objects in IR images are used, because missiles are usually included in backgrounds of IR images. Since the IR-guided missiles are common in modern warfare, various spatial [25][26][27][28] and temporal [1,29,30] target-detection methods have been proposed for detecting missiles or small targets. The IR image tracker uses IR images of LWIR-band or MWIR-band. ...
... µm for the SWIR, 3~5 µm for MWIR, and 8~12 µm for LWIR, respectively. Figure 1 shows the spectral radiance according to the temperature for a black body in the wavelength, obtained from Planck's law of Equation (1). And the radiance intensity according to temperature obtained through Equation (2) for the three IR wavelength bands (SWIR, MWIR, and LWIR) are shown in Figure 2. In case area of a target seen in an IR detector is A p and the solid angle of the detector at the target is Ω, the output voltage V det from the detector is given by: ...
Military infrared (IR) imaging systems utilize one or more IR wavelength-bands, among short wavelength IR (SWIR), middle wavelength IR (MWIR), and long wavelength IR (LWIR) band. The IR image wavelength-band conversion which transforms one arbitrary IR wavelength-band image to another IR wavelength-band image is needed for IR signature modeling and image synthesis in the IR systems. However, the IR wavelength-band conversion is very challenging because absorptivity and transmittance of objects and background (atmosphere) are different according to the IR wavelength band and because radiation and reflectance characteristics of the SWIR are very different from the LWIR and MWIR. Therefore, the IR wavelength-band conversion in this paper applies to only IR targets and monotonous backgrounds at a long distance for military purposes. This paper proposes an IR wavelength-band conversion method which transforms one arbitrary IR wavelength-band image to another IR wavelength-band image by using the surface temperature estimation of an object and the error attenuation method for the estimated temperature. The surface temperature of the object is estimated by an approximated Planck’s radiation equation and the error of estimated temperature is corrected by using the slope information of exact radiance along with the approximated one. The corrected surface temperature is used for generating another IR wavelength-band image. The verification of the proposed method is demonstrated through the simulations using actual IR images obtained by thermal equipment.
... Bae [47] proposed a spatial and temporal bilateral filter (BF) intended for the identification of small trajectories in his study. This approach functions through the extraction of spatial target details via a spatial BF, and temporal target data via a temporal BF. ...
... Background-detection-based [26][27][28][29][30][31][32][33][34][35][36][37][38] Classical computer-vision-based [39][40][41][42][43][44][45][46][47][48][49] Deep-learning-based [50,52,[54][55][56][57][58][59][60] ...
Object detection and tracking are vital in computer vision and visual surveillance, allowing for the detection, recognition, and subsequent tracking of objects within images or video sequences. These tasks underpin surveillance systems, facilitating automatic video annotation, identification of significant events, and detection of abnormal activities. However, detecting and tracking small objects introduce significant challenges within computer vision due to their subtle appearance and limited distinguishing features, which results in a scarcity of crucial information. This deficit complicates the tracking process, often leading to diminished efficiency and accuracy. To shed light on the intricacies of small object detection and tracking, we undertook a comprehensive review of the existing methods in this area, categorizing them from various perspectives. We also presented an overview of available datasets specifically curated for small object detection and tracking, aiming to inform and benefit future research in this domain. We further delineated the most widely used evaluation metrics for assessing the performance of small object detection and tracking techniques. Finally, we examined the present challenges within this field and discussed prospective future trends. By tackling these issues and leveraging upcoming trends, we aim to push forward the boundaries in small object detection and tracking, thereby augmenting the functionality of surveillance systems and broadening their real-world applicability.
... The conventional temporal domain method assumes that the background can be regarded as still in consecutive frames [24]. The moving target, which usually changes significantly, can be extracted by subtracting the background in consecutive frames [13], [14]. ...
... The range of the s and t for the optimal gradient mask size is calculated by (21) to (24). where and denote the small target size. ...
Infrared (IR) small target detection has recently been widely used in civil and military applications. In IR small target detection, IR images probed at long distances are easily disturbed by complex backgrounds, light changes, and other noises, which makes it pretty challenging. Most existing literature focuses on low false alarms and high detection accuracy rather than computational efficiency. We propose a fast and reliable IR small target detection method that deals with spatial and temporal domains based on the Hadamard product for spatial-temporal matrices (HPSTM). In the spatial domain, a weighted tri-layer window is used to convolute the IR image and obtain the gradient matrices. In the temporal domain, three interval frames were used instead of conventional adjacent frames and contrast matrices were extracted. Finally, the two spatial-temporal matrices use the Hadamard product to multiply and a simple threshold to transfer into a binary image for target detection. We also proposed the optimal mask size selection (OMSS) method, which can adjust the optimal tri-layer window size in the spatial domain to obtain the best detection result. The experimental results show that the proposed HPSTM method has a high detection accuracy and fast running time compared with other methods, indicating that the proposed HPSTM method is reliable and suitable for real-time IR small target detection.
... Small targets are extracted by background difference algorithm. For example, max-mean/median filter [6], morphological filter [7], top-hat transform [8], two-dimensional least mean square filtering (TDLMS) [9] and bilateral filter (BF) [10] are all background predictions or modeling. Yet, these filtering algorithms are poor for complex 2 of 11 backgrounds in the face of strong clutter interference. ...
... For any scale s ∈ {1, 3, . . . , s max }, saliency map of REC and RBLM can be calculated according to Equations (4) and (10) which are normalized to the range of [0, 1].The linear normalization equation is shown in Equations (11) and (12). r s (x, y) and q s (x, y) are normalized to obtain r s (x, y) and q s (x, y) in s scale. ...
Aiming at a thorny issue, that conventional small target detection algorithm using local contrast method is not sensitive for residual background clutter, robustness of algorithms is not strong. A Gaussian fusion algorithm using multi-scale regional patch structure difference and Regional Brightness Level Measurement is proposed. Firstly, Regional Energy Cosine (REC) is constructed to measure the structural discrepancy among a small target with neighboring cells. At the same time, Regional Brightness Level Measurement (RBLM) is constructed utilizing the brightness difference characteristics between small target and background areas. Then, a brand new Gaussian fusion algorithm is proposed for the generated saliency map in multi-scale space to characterize the overall heterogeneity in original infrared small target and local neighborhood. Finally, a self-adapting separation algorithm is adopted with the objective to obtain a small target from background interference. This method is able to utmostly restrain background interference and enhance the target. Extensive qualitative and quantitative testing results display that the desired algorithm has remarkable performance in strengthening target region and restraining background interference compared with current algorithms.
... The problem associated with traditional mean filter is that the edge is smoothened in addition to that the noise is averaged out. Bilateral filter is a popular edge preserving filter which is used for image denoising applications [1,8,15,17,18]. Bilateral filter is a combination of domain filter and range filter. Domain filtering averages image values with weights that fall off with distance. ...
... The remainder of the paper is organized as follows: Gaussian Mixture Model is detailed in section 2, bilateral filter is explained in section 3, the details of the (1) proposed method are explained in section 4, the experimental results are given in section 5 and conclusions are given in section 6. ...
... Because of the movement (jitter) of the infrared observation platform or the change of the imaging background, it is difficult to obtain the accurate infrared background by sequential detection methods [5][6][7], because the infrared small targets are easily mistaken for background and vice versa. In this case, the single frame detection methods have received a great attention recently, and are valid for infrared small target detection with static or changing backgrounds [8][9][10]. ...
... By combining spatial and temporal information together, a target detection method was introduced using spatial bilateral filter and temporal cross product, which are respectively used to extract the spatial target information and the features of temporal profiles [6]. Subsequently, a spatial-temporal bilateral filter was presented to detect target trajectories, by extracting spatial and temporal target information simultaneously [7]. As discussed above, pixel-wise detection methods use a local region or a temporal profile to extract target information under the assumption that the target location has conspicuous discontinuity with the nearby background. ...
In this paper, a novel infrared target co-detection model combining the self-correlation features of backgrounds and the commonality features of targets in the spatio-temporal domain is proposed to detect small targets in a sequence of infrared images with complex backgrounds. Firstly, a dense target extraction model based on nonlinear weights is proposed, which can better suppress background of images and enhance small targets than weights of singular values. Secondly, a sparse target extraction model based on entry-wise weighted robust principal component analysis is proposed. The entry-wise weight adaptively incorporates structural prior in terms of local weighted entropy, thus, it can extract real targets accurately and suppress background clutters efficiently. Finally, the commonality of targets in the spatio-temporal domain are used to construct target refinement model for false alarms suppression and target confirmation. Since real targets could appear in both of the dense and sparse reconstruction maps of a single frame, and form trajectories after tracklet association of consecutive frames, the location correlation of the dense and sparse reconstruction maps for a single frame and tracklet association of the location correlation maps for successive frames have strong ability to discriminate between small targets and background clutters. Experimental results demonstrate that the proposed small target co-detection method can not only suppress background clutters effectively, but also detect targets accurately even if with target-like interference.
... They are not suitable when both space debris and stars appear as point-like objects. Recently, many algorithms have also been proposed to detect dim and small objects in an IR image background, including max-mean and max-median filters [19], genetic algorithm [20], morphology (top-hat) [21], bilateral filter and temporal cross product [22], modified top-hat transformation [23], spatiotemporal difference-of-Gaussians filters [24], the directional support value of Gaussian transformation [25], and the multiscale gray difference weighted image entropy [26,27]. These algorithms are very effective for detecting point-like objects in a continuous IR background image. ...
... In addition, another different max-mean method proposed by Deshpande et al. [19] (in both 5 × 5 and 3 × 3 windows) and the top-hat method [21] are also compared with the time-index filtering method. The receiver operating characteristic (ROC) results [22] for the five sequential images in the first experiment are shown in Fig. 9. ...
We present a high-accuracy, low false-alarm rate, and low computational-cost methodology for removing stars and noise and detecting space debris with low signal-to-noise ratio (SNR) in optical image sequences. First, time-index filtering and bright star intensity enhancement are implemented to remove stars and noise effectively. Then, a multistage quasi-hypothesis-testing method is proposed to detect the pieces of space debris with continuous and discontinuous trajectories. For this purpose, a time-index image is defined and generated. Experimental results show that the proposed method can detect space debris effectively without any false alarms. When the SNR is higher than or equal to 1.5, the detection probability can reach 100%, and when the SNR is as low as 1.3, 1.2, and 1, it can still achieve 99%, 97%, and 85% detection probabilities, respectively. Additionally, two large sets of image sequences are tested to show that the proposed method performs stably and effectively.
... To further improve the detection performance under complicated background, some new detection methods based on spatial-temporal information have been proposed. Such as, a moving point target detection method based on three-dimensional spatiotemporal anisotropic diffusion model is presented in [22]; Chen et al. [23] proposed a novel spatial-temporal detection method based on bi-dimensional empirical mode decomposition and time-domain differential filtering; Bae [24] introduced a spatial and temporal bilateral filter for target detection, which uses spatial bilateral filter to extract spatial target information and applies temporal bilateral filter to extract temporal target information. These methods are more effective compared with the traditional detection methods. ...
... In this case, wavelet transform cannot efficiently represent the small moving target. Coincidentally, spectral graph wavelet (SGWT) [24] has a better flexibility, scale-invariance and local adaptivity than those of wavelet transform, which can efficiently represent isotropic structures and incomplete isotropic structures. Thus, in this paper, it is selected as the W T dictionary to efficiently and completely extract target features. ...
Infrared small moving target detection is one of the crucial techniques in infrared search and tracking systems. This paper presents a novel small moving target detection method for infrared image sequence with complicated background. The key points are given as follows: (1) since target detection mainly depends on the incoherence between target and background, the proposed method separate the target from the background according to the morphological feature diversity between target and background; (2) considering the continuity of target motion in time domain, the target trajectory is extracted by the RX filter in random projection. The experiments on various clutter background sequences have validated the detection capability of the proposed method. The experimental results show that the proposed method can robustly provide a higher detection probability and a lower false alarm rate than baseline methods.
... Track a and Track b belong to the same object if the following is true 22) where N is the number of points on the track and D is the maximum allowable distance (20 in our case). Q i and Q i ′ are two points on Track a and Track b , respectively. ...
... To evaluate the performance of the proposed method relative to some other methods, six image sequences, S1, …, S6 with different characteristics as shown in Table 1 are selected. Fig. 2 shows the results of simulations of the proposed method relative to the other techniques, such as, Max-Mean [7], Bae's method [22], Dong's method [23] and two other MHT-based methods, namely, He's method [24] and Blostein's method [10]. The speed of the target in image sequence S1 and S2 is between [0, 1] pixel/frame in which all of the above algorithms are applicable. ...
Dim object tracking in a heavy clutter environment is a theoretical and technological challenge in the field of image processing. For a small dim object, conventional tracking methods fail for the lack of geometrical information. Multiple hypotheses testing (MHT) is one of the generally accepted methods in target tracking systems. However, processing a tree structure with a significant number of branches in MHT has been a challenging issue. Tracking high-speed objects with traditional MHT requires some presumptions which limit the capabilities of these methods. This study presents a hierarchal tracking system in two levels to solve this problem. For each point in the lower-level, a multi objective particle swarm optimisation technique is applied to a group of consecutive frames to reduce the number of branches in each tracking tree. Thus, an optimum track for each moving object is obtained in a group of frames. In the upper-level, an iterative process is used to connect the matching optimum tracks of the consecutive frames based on the spatial information and fitness values. The experimental results show that the proposed method has a superior performance in relation to some common dim object tracking methods over different image sequence data sets.
... In recent years, the rapid development of Convolutional Neural Networks (CNN) [5][6][7][8][9] has enabled their transfer to object detection tasks in computer vision, making it feasible to push the accuracy limits of target detection in infrared scenarios [10][11][12][13][14][15][16]. Compared with traditional detection methods based on filtering [17,18], contrast [19,20], and rank [21,22], deep neural networks offer strong robustness, high detection accuracy, and broad adaptability, showcasing impressive advantages for infrared target detection tasks. ...
Infrared imagery surpasses the limitations of visible light images and finds widespread applications in fields such as military reconnaissance and security surveillance. Recent studies on infrared target detection aim to preserve local features and global representations to the greatest extent. However, compared to visible light images, infrared images exhibit inherent challenges such as insufficient texture information and coarse boundaries, which introduce new difficulties to this research. To address these issues, this paper introduces additional information cues from the perspective of enriching feature map information. Specifically, we propose a multidomain feature fusion object detector (MFFOD), whose backbone feature extraction network consists of a convolutional branch and a fast Fourier transform branch. This hybrid domain representation enables the extraction of both domain-specific information and global high-frequency and low-frequency information with minimal computational overhead. Furthermore, in the intermediate layers of the network, we have carefully designed a feature injection module that enables comprehensive interaction between channel features and spatial features within a single feature map. Experimental results demonstrate that MFFOD achieves average detection accuracies of 88.97%, 90.32%, and 99.25% on three significant infrared scene datasets, outperforming existing target detection methods. We hope that this general detection algorithm will provide a robust reference for future infrared target detection research.
... The first one is the spatial domain filtering, in which, the input infrared image is processed using local kernels to enhance the target area. Maxmean [12], max-median [12], bilateral filtering [13], morphological operators [14], two dimensional least mean square [15] are some instances of this sub-category. The second one refers to processing in the transformation domain. ...
Infrared small target detection (IRSTD) is a challenging task in computer vision. During the last two decades, researchers' efforts are devoted to improving detection ability of IRSTDs. Despite the huge improvement in designing new algorithms, lack of extensive investigation of the evaluation metrics are evident. Therefore, in this paper, a systematic approach is utilized to: First, investigate the evaluation ability of current metrics; Second, propose new evaluation metrics to address shortcoming of common metrics. To this end, after carefully reviewing the problem, the required conditions to have a successful detection are analyzed. Then, the shortcomings of current evaluation metrics which include pre-thresholding as well as post-thresholding metrics are determined. Based on the requirements of real-world systems, new metrics are proposed. Finally, the proposed metrics are used to compare and evaluate four well-known small infrared target detection algorithms. The results show that new metrics are consistent with qualitative results.
... Nevertheless, there are still large room for our baseline method to improve. Additionally, IOD may benefit the study of infrared dim small objection detection [2], ground-glass nodule diagnosis in medical [54], Synthetic Aperture Radar (SAR) detection [83] and some partially-occluded targets in RGB images [86], which also have the similar characteristic of indistinct boundary. We consider to extend the IOD-Video dataset to multispectral bands in the future, which will make it possible to distinguish the specific substance of detected objects. ...
We endeavor on a rarely explored task named Insubstantial Object Detection (IOD), which aims to localize the object with following characteristics: (1) amorphous shape with indistinct boundary; (2) similarity to surroundings; (3) absence in color. Accordingly, it is far more challenging to distinguish insubstantial objects in a single static frame and the collaborative representation of spatial and temporal information is crucial. Thus, we construct an IOD-Video dataset comprised of 600 videos (141,017 frames) covering various distances, sizes, visibility, and scenes captured by different spectral ranges. In addition, we develop a spatio-temporal aggregation framework for IOD, in which different backbones are deployed and a spatio-temporal aggregation loss (STAloss) is elaborately designed to leverage the consistency along the time axis. Experiments conducted on IOD-Video dataset demonstrate that spatio-temporal aggregation can significantly improve the performance of IOD. We hope our work will attract further researches into this valuable yet challenging task. The code will be available at: \url{https://github.com/CalayZhou/IOD-Video}.
... Different wavelengths (NIR, MIR, and FIR) of IR radiation have many different applications. The sources of IR introducing great technological advancements in imaging, thermal imaging, motion detection, gas analyzing, monitoring, and environmental health analysis, etc. IR imaging is widely used in the military, medical, scientific, and industrial fields, since it is able to create a visual with an otherwise non-visible wavelength band to the human eye [110,111]. Recently, multispectral, and hyperspectral imaging systems based on the IR spectrum, have been used for developing and evaluating most agricultural and food processing operations, such as tracking and estimating the quality of agricultural and food products. ...
In recent decades, the quality and safety of fruits, vegetables, cereals, meats, milk, and their derivatives from processed foods have become a serious issue for consumers in developed as well as developing countries. Undoubtedly, the traditional methods of inspecting and ensuring quality that depends on the human factor, some mechanical and chemical methods, have proven beyond any doubt their inability to achieve food quality and safety, and thus a failure to achieve food security. With growing attention on human health, the standards of food safety and quality are continuously being improved through advanced technology applications that depend on artificial intelligence tools to monitor the quality and safety of food. One of the most important of these applications is imaging technology. A brief discussion in this chapter on the utilize of multiple imaging systems based on all different bands of the electromagnetic spectrum as a principal source of various imaging systems. As well as methods of analyzing and reading images to build intelligence and non-destructive systems for monitoring and measuring the quality of foods.
... The MIT-DB includes 48 half-hour recordings with 360 Hz, and the heartbeat is classified and labeled into 15 types. The types of heartbeats belong to five classes as shown in Table 1 [50,51]. The abbreviated representation of the ECG signal interval is used for the experiment. ...
The characteristics or aspects of important fiducial points (FPs) in the electrocardiogram (ECG) signal are complicated because of various factors, such as non-stationary effects and low signal-to-noise ratio. Due to the various noises caused by the ECG signal measurement environment and by typical ECG signal deformation due to heart diseases, detecting such FPs becomes a challenging task. In this study, we introduce a novel PQRST complex detector using a one-dimensional bilateral filter (1DBF) and the temporal characteristics of FPs. The 1DBF with noise suppression and edge preservation preserves the P- or T-wave whereas it suppresses the QRS-interval. The 1DBF acts as a background predictor for predicting the background corresponding to the P- and T-waves and the remaining flat interval excluding the QRS-interval. The R-peak and QRS-interval are founded by the difference of the original ECG signal and the predicted background signal. Then, the Q- and S-points and the FPs related to the P- and T-wave are sequentially detected using the determined searching range and detection order based on the detected R-peak. The detection performance of the proposed method is analyzed through the MIT-BIH database (MIT-DB) and the QT database (QT-DB).
... The performance of computer vision-based technology depends on the imaging quality. Therefore, image enhancement is an important technology in computer vision tasks, such as person counting, crowd analysis, re-identification, target segmentation and recognition [1][2][3][4], which is widely used in public safety, military and other visual applications. ...
Computer vision-based crowd understanding and analysis technology has been widely used in public safety due to the rapid growth of population and the frequent occurrence of various accidents. Improving imaging quality is the key to improve the performance of crowd analysis, density estimation, target recognition, segmentation, and detection in computer vision tasks. Due to the complex imaging environment such as fog and low illumination, some images taken in outdoor environment often have the problems of colour distortion, lack of details, and the poor imaging quality, which affect the subsequent visual tasks. To improve the imaging quality and visual effect, an adaptive colour restoration and detail retention-based method is proposed for image enhancement. First, to overcome the problem of colour distortion caused by low illumination and fog, a multi-channel fusion based adaptive image colour restoration method is proposed. To make the enhancement result more consistent with human observation, the detail retention-based method is applied to enhance the details. Experimental results demonstrate that the authors' results are effective and outperform the compared methods both in visual and objective evaluations. © 2021 The Authors. IET Image Processing published by John Wiley & Sons Ltd on behalf of The Institution of Engineering and Technology
... Enhancement methods based on mapping processing, such as histogram equalization(HE) [1], contrast-limited adaptive histogram equalization(CLAHE)[2], plateaus histogram equalization(PHE) [3], are fast and simple, but the enhancements to the details are not obvious and the noises are overenhancement [4]. Enhancement methods based on hierarchical processing, usually layer the images through guided filter(GF) [5] or bilateral filter(BF) [6], and then uses Gamma calibration [7] or HE methods for different layers. In addition to the above two types of detail enhancement methods, there are also complex methods such as LEPF [8], Retinex-MSR [9], and integrated learning [10]. ...
... A variety of methods with desirable results have been proposed for the infrared target detection. Bae et al. [3] put forward a spatial and temporal bilateral filter (BF) to detect target. Though this method seems theoretically effective for many types of different backgrounds, it is hard to extract the target between target background areas and target region in the actual image processing. ...
... Addressing this problem, a large number of background estimation methods are proposed, such as bilateral filter (BF) methods. 10,11 Unfortunately, edges are easy to be smoothed out in the process of filtering. Some methods dedicate to detect the direction of edges and use it to accomplish background estimation. ...
Infrared small target detection is widely applied in lots of practical applications, but due to the complicated edges in practical scenarios, most existing detection algorithms usually lead to many false alarms and cannot detect the target accurately. Addressing this problem, a novel edge-preserving background estimation method based on maximum patch similarity was proposed in this article. At first, we will propose an improved local adaptive contrast measure to suppress the pixel-size electronic noises. Then, maximum patch similarity with minimum improved local adaptive contrast measure can be utilized to preserve the edge in the estimated background. Finally, we can obtain target image by filtering the background image from original image and use adaptive threshold segmentation to detect the small target in our target image. It is shown from experiments that our proposed method has better detection results in diverse infrared images, improving signal-to-clutter ratio gain and background suppression factor of the images significantly and efficiently.
... Wavelet based algorithms [14] design a group of filters which are matched to a point spread function (PSF) at different scales by choosing a mother wavelet similar to PSF. Unfortunately, they are quite time-consuming and the false alarm rates are always high [15]. ...
In order to detect both bright and dark small moving targets effectively in infrared (IR) video sequences, a saliency histogram and geometrical invariability based method is presented in this paper. First, a saliency map that roughly highlights the salient regions of the original image is obtained by tuning its amplitude spectrum in the frequency domain. Then, a saliency histogram is constructed by means of averaging the accumulated saliency value of each gray level in the map, through which bins corresponding to bright target and dark target are assigned with large values in the histogram. Next, single-frame detection of candidate targets is accomplished by a binarized segmentation using an adaptive threshold, and their centroid coordinates with sub-pixel accuracy are calculated through a connected components labeling method as well as a gray-weighted criterion. Finally, considering the motion characteristics in consecutive frames, an inter-frame false alarm suppression method based on geometrical invariability is developed to improve the precision rate further. Quantitative analyses demonstrate the detecting precision of this proposed approach can be up to 97% and Receiver Operating Characteristic (ROC) curves further verify our method outperforms other state-of-the-arts methods in both detection rate and false alarm rate.
... IR imagers have been used in a variety of efforts aimed at the general problem of detecting the presence of small [1] or anomalous [2,3] thermal targets. Other recent works addressing the "target agnostic" case include the bilateral filtering approach of Bae [4], the high-pass "template" filtering method of [5], and the multi-scale, wavelet-based approach of Wei et al. [6]. ...
This work documents the performance of a recently proposed generalized likelihood ratio test (GLRT) algorithm in detecting thermal point-source targets against a sky background. A calibrated source is placed above the horizon at various ranges and then imaged using a mid-wave infrared camera. The proposed algorithm combines a so-called “shrinkage” estimator of the background covariance matrix and an iterative maximum likelihood estimator of the point-source parameters to produce the GLRT statistic. It is clearly shown that the proposed approach results in better detection performance than either standard energy detection or previous implementations of the GLRT detector.
... So far, many methods have been proposed for detection of targets in the IR images with desired results in some sets of IR images. Some conventional methods are max-mean and maxmedian [1], morphology [2], sparse representation [3], bilateral filter [4], least mean squares [5] and other methods [6,7]. In addition, Wang et al. [8] have used cubic facet model for detection of small targets [9]. ...
Detection of small targets in an infrared (IR) image with high reliability is very important for defence systems. Small targets in an IR image are defined as salient features which attract the attention of human visual system. In this study, a robust method for detection of small targets in an IR image is proposed based on HV attention. In this method, first, the Gaussian-like feature maps are extracted from the original image. Then, saliency maps (SMs) are created based on pulsed discrete cosine transform, in which the target is salient and background clutter is suppressed. Finally, to increase the contrast between target and background clutter and to raise robustness of this method against false alarms, SMs are fused adaptively. Experiments are carried out on the data set including real-life IR images with small targets as well as various and complicated backgrounds. Qualitative and quantitative assessments show that the proposed method can detect small targets in IR image with high reliability and is more effective compared with other methods based on HV attention. Therefore, it can be used in many applications for detection of small targets in IR image with minimum false alarms.
... Space-based remote sensors pay much attention to radiances from the top of the atmosphere since they often constitute the strong interferences [1,2] or act as the measurement objects [3] in the imaging chain. Therefore it is fundamental to carry out computations of atmospheric radiative quantities to support both system design [4] and processing algorithm development [5] in remote sensing applications. ...
A new optimization technique is proposed to generate the k-interval parameters including the quadrature weights and their corresponding equivalent absorption coefficients when using correlated k-distribution method for computations of spectrally integrated atmospheric upwelling radiances in infrared absorption bands. The main feature of this technique is the introduction of the upwelling contribution function (UCF) both in expressing the atmospheric upwelling radiance and in constructing the merit function for fitting, so that the optimization for the band radiance smoothly turns into the optimization for the band UCF of each atmospheric sublayer. The performance of the proposed technique is illustrated by comparing with the Δlog(k) technique and the Gauss–Legendre quadrature technique at three different band settings in both the US standard atmosphere and the tropical atmosphere, and the line-by-line results are treated as the reference. Results show that there are less relative errors by the proposed optimization technique than by the other two techniques under the same number of k-intervals. In most cases, the computational efficiency can be improved by more than 10% under the same precision requirement in upwelling radiance calculations, and even 80% in some preferable conditions.
Deep learning is transforming most areas of science and technology, including electron microscopy. This review paper offers a practical perspective aimed at developers with limited familiarity. For context, we review popular applications of deep learning in electron microscopy. Following, we discuss hardware and software needed to get started with deep learning and interface with electron microscopes. We then review neural network components, popular architectures, and their optimization. Finally, we discuss future directions of deep learning in electron microscopy.
Recognition of small moving targets in space has become one of the frontier scientific researches in recent decade. Most of them focus on detection and recognition in star image with sidereal stare mode. However, in this research field, few researches are about detection and recognition in star image with track rate mode. In this paper, a novel approach is proposed to recognize the moving target in single frame by machine learning method based on elliptical characteristic extraction of star points. The technical path about recognition of moving target in space is redesigned instead of traditional processing approaches. Elliptical characteristics of each star point can be successfully extracted from single image. Machine learning can achieve the classification goal in order to make sure that all moving targets can be extracted. The experiments show that our proposed approach can have better performance in star images with different qualities.
Small target detection in infrared (IR) images has been widely applied for both military and civilian purposes. In this study, because IR images contain sparse and low-rank features in most scenarios, we propose an optimal IR patch-image (OIPI) model-based detection method to detect small targets in heavily cluttered IR images. First, the OIPI model was generated based on a conventional IR image model using a novel optimal patch size and sliding step adaptive selection algorithm. Secondly, the sparse and low-rank features of IR images were extracted and fused to generate an adaptive weighted parameter. Thirdly, the adaptive inexact augmented Lagrange multiplier (AIALM) algorithm was applied in the OIPI model to solve the robust principal component analysis (RPCA) optimization problem. Finally, an adaptive threshold method is proposed to segment and calibrate targets. Experimental results indicate that the proposed algorithm is capable of detecting small targets more stably and accurately, compared with state-of-the-art methods.
In view of the difficult detection of infrared dim point target from complex backgrounds, a multiscale homogeneous feature (MHF) based multispectral image fusion detection method is proposed in this paper. Inspired by the local contrast measure (LCM), we extract two local statistical features from the perspective of the homogeneity of gray difference distribution to characterize local structure of the infrared point target. Based on these two local features, we obtain the MHF map that can effectively highlight the potential point targets and suppress the backgrounds simultaneously. For the pixel-size electronic noise (PSEN) and some similar local structures to the point target, the multispectral image fusion detection is a positive way to alleviate these interferences and promote the robustness of the dim point target detection. Experimental results on six real scenarios and synthetic scenarios demonstrate that the proposed method not only works more stably for different target sizes and brightness, but also can achieve superior detection performance compared with the state-of-art detection methods.
The multi-scale low rank method is now a novel technology for the target detection in an image, in which the pattern of the sub-image block partition plays a crucially important role for the performance of the target detection. In this paper, we propose an efficient multi-scale low rank method for the infrared small target detection. The main work is that we systematically analysis the principle of how to select the optimal parameters such as the shape, size and number of the image block to obtain an efficient decomposed combinatorial form which is appropriate for the infrared small target detection. Then, we propose an efficient strategy to determine the optimal iterative times adaptively for ADMM’s loop iteration process, by which the more accurate results can be obtained and the running speed is improved. Finally, we fuse the infrared small target images with three optimal scales to preserve the small target information. The experimental results show that the presented strategy on the selection of optimal parameter and adaptive iterative method can guarantee the effectiveness of small target extraction and save the running time.
This paper presents a fast detection method for infrared small dim target based on region proposal. Detection is divided into two parts: region proposal determination and target precise position extraction. First, potential target positions are extracted by a novel direction detection method based on stationary wavelet transform (SWT). Region proposal can be achieved by self-cross validation of the sequence of potential target positions. Based on the region proposal, a predictor-corrector algorithm is applied to obtain target precise position of current frame. Background prediction is acquired by a unified multi-scale temporal profile (UMTP) model with adaptive fusion and corrected by the 3σ criterion. Results show that the proposed method can preserve almost exactly the background while eliminating target pixels and has the ability to cope with multiple targets simultaneously. Based on the region proposal, real-time processing can be achieved.
It is always a challenging task to detect an infrared (IR) small target with high-detection rate, low false alarm rate, and high detection speed since the target usually has a small size and dim gray value, and the background is usually complex. Local contrast, which is based on the human visual system, has been proved efficient for IR small target detection, but existing local contrast algorithms are either difference-form or ratio-form and cannot enhance true target and suppress background simultaneously. In addition, most of them are pixel-level algorithms and require a huge amount of calculations. A simple but efficient method named subblock-level ratio-difference joint local contrast measure (SRDLCM) is proposed; it can enhance real small target and suppress complex background simultaneously. In addition, SRDLCM is calculated for each subblock but not each pixel, so its calculation amount can be reduced significantly. The experimental results on seven real IR sequences and one single-frame image dataset show that the proposed algorithm can achieve a good detection performance in detection rate and false alarm rate with a good robustness, and the time consumed for a single frame is only less than 0.04 s. © 2018 Society of Photo-Optical Instrumentation Engineers (SPIE).
Infrared small targets can easily be submerged in complex backgrounds in single frame infrared small target detection, the edges usually cause high false alarms and lead to erroneous detection results. In this paper, a novel edge-preserving background estimation method based on most similar neighbor patch is proposed to attenuate this problem. First, we make the best of structural features in infrared image and introduce an improved local adaptive contrast measure (ILACM) to measure the patch similarity. Then most similar neighbor patch with maximum patch similarity can be utilized to realize edge-preserving background estimation model. At last, we can obtain target image by eliminating estimated background from original image. Experiments and comparisons with state-of-the-art methods show that our method has better background estimation performance in diverse infrared images and improves SCR values of the images significantly.
Visual enhancement for infrared small dim targets is a standing problem in infrared image processing. Existing approaches cannot enhance the target well and suppress the background simultaneously, especially for targets which are so faint that they are hardly visible. This paper proposes a novel real-time visual enhancement algorithm for infrared small dim targets in video by introducing temporal cues. In this work, Dynamic Programming Algorithm (DPA) is used to detect the target’s trajectory in the video and the target is enhanced through energy accumulation along the trajectory. The shape prior of the small dim target is adopted for background suppression and adaptive merging. Experimental results on real infrared small dim target videos indicate that the proposed algorithm can improve the visual quality of these types of images notably, especially for cases in which the target is hardly visible. In addition, the proposed algorithm takes on average 8.35 ms to process a image, and thus meets the needs of real-time applications.
In surveillance and early warning systems, the enhancement of targets is a very important stage for the high reliability detection and tracking in Infrared images with complex backgrounds. In order to enhance small targets in an Infrared image and suppress the background clutter, consequently increasing the contrast between them, this paper proposes a method using a model for the target area with a three-layer patch-image model and based on the difference between the variance of the layers in the neighboring areas of the investigated pixel. Results of the experiments indicate that the proposed method is quite effective on the enhancement of small targets as well as suppression of the background clutter in IR images with a minimum false alarm rate. This is realized while the runtime of the proposed method is minimal compared to other commonly used methods, which makes it effective to be used in real time applications.
Infrared small target tracking plays an important role in applications including military reconnaissance, early warning and terminal guidance. In this paper, an effective algorithm based on the Singular Value Decomposition (SVD) and the improved Kernelized Correlation Filter (KCF) is presented for infrared small target tracking. Firstly, the super performance of the SVD-based algorithm is that it takes advantage of the target's global information and obtains a background estimation of an infrared image. A dim target is enhanced by subtracting the corresponding estimated background with update from the original image. Secondly, the KCF algorithm is combined with Gaussian Curvature Filter (GCF) to eliminate the excursion problem. The GCF technology is adopted to preserve the edge and eliminate the noise of the base sample in the KCF algorithm, helping to calculate the classifier parameter for a small target. At last, the target position is estimated with a response map, which is obtained via the kernelized classifier. Experimental results demonstrate that the presented algorithm performs favorably in terms of efficiency and accuracy, compared with several state-of-the-art algorithms.
The special characteristics of slowly moving infrared targets, such as containing only a few pixels, shapeless edge, low signal-to-clutter ratio, and low speed, make their detection rather difficult, especially when immersed in complex backgrounds. To cope with this problem, we propose an effective infrared target detection algorithm based on temporal target detection and association strategy. First, a temporal target detection model is developed to segment the interested targets. This model contains mainly three stages, i.e., temporal filtering, temporal target fusion, and cross-product filtering. Then a graph matching model is presented to associate the targets obtained at different times. The association relies on the motion characteristics and appearance of targets, and the association operation is performed many times to form continuous trajectories which can be used to help disambiguate targets from false alarms caused by random noise or clutter. Experimental results show that the proposed method can detect slowly moving infrared targets in complex backgrounds accurately and robustly, and has superior detection performance in comparison with several recent methods. © 2016, Journal of Zhejiang University Science Editorial Office and Springer-Verlag Berlin Heidelberg.
In the research operation of Digital Signal Processing (DSP) and Digital Image Processing (DIP), one of the most essential obstacles is the image denoise algorithm by the reason of a very large demand of high quality noise-free images therefore there are many image denoise algorithms have been invented in the time of two decades. Bilateral filter is one of the most impressive and feasible algorithms, which is usually applied for denoise propose, but the performance of the Bilateral filter is substantially bank on three parameters: spatial variance, radiometric variance and window size. Consequently, this paper investigates the performance influence impact of spatial variance, radiometric variance, window size for the Bilateral Filter in the denoise propose. In the denoise experiment, Bilateral filter (BF) is applied on three noisy standard images under five Gaussian noise power levels and the best results in the PSNR prospective point of view from deniose algorithm is picked. Moreover, an optimal value of three parameters: spatial variance, radiometric variance, window size, which make the performance of Bilateral filter the highest PSNR, are extensively investigated for each types of tested images and each noise powers.
Small dim target tracking is an active and important research area in image processing and pattern recognition. Recently, there has been an emphasis on the development of algorithms based on spatial domain Constant False Alarm Rate (CFAR) detection. This paper presents a novel algorithm for detecting and tracking small dim targets in Infrared (IR) image sequences with low Signal to Noise Ratio (SNR) based on the frequency and spatial domain information. Using a Dual-Tree Complex Wavelet Transform (DT-CWT), a CFAR detector is applied in the frequency domain to find potential positions of objects in a frame. Following this step, a Support Vector Machine (SVM) classification is applied to accept or reject each potential point based on the spatial domain information of the frame. The combination of the frequency and spatial domain information demonstrates the high efficiency and accuracy of the proposed method which is supported by the experimental results.
The intensive emission of earth limb in the field of view of sensors contributes much to the observation images. Due to the low signal-to-noise ratio (SNR), it is a challenge to detect small targets in earth limb background, especially for the detection of point-like targets from a single frame. To improve the target detection, track before detection (TBD) based on the frame sequence is performed. In this paper, a new technique is proposed to determine the target associated trajectories, which jointly carries out background removing, maximum value projection (MVP) and Hough transform. The background of the bright earth limb in the observation images is removed according to the profile characteristics. For a moving target, the corresponding pixels in the MVP image are shifting approximately regularly in time sequence. And the target trajectory is determined by Hough transform according to the pixel characteristics of the target and the clutter and noise. Comparing with traditional frame-by-frame methods, determining associated trajectories from MVP reduces the computation load. Numerical simulations are presented to demonstrate the effectiveness of the approach proposed.
A novel dim small target detection algorithm based on the nonsubsampled contourlet transform (NSCT) and the singular value decomposition (SVD) is proposed in this paper, which is to improve the performance of the dim small target detection under the complex sky cloud background. Firstly, the original infrared image is decomposed with the SVD, and several different numbers of the singular value for reconstruction is chosen to analyze the application of the SVD to the image. The complex sky cloud background in the infrared target image is predicted by choosing a certain number of the singular value to reconstruct the image, and it is subtracted from the original image to suppress the background and enhance the target signal. Secondly, to use the scale and the direction information of the image, the residual image is decomposed by the NSCT into several high-pass directional subbands and a low-pass subband. Thirdly, the SVD filtering is utilized again to those directional subbands to eliminate the noise and the residual background. And the low-pass subband is modified by the local mean removal method. Finally, the refined subbands are reconstructed by the inverse NSCT to fulfill the dim small target detection. The experimental results demonstrate that the proposed algorithm has better subjective vision and objective numerical indicators, and can acquire a better performance of the target detection.
Target detection is of great importance both in civil and military fields. Here a new moving target detection approach is proposed, which employs a nonlinear adaptive filter to remove large fluctuations on temporal profiles that are produced by evolving clutters. Initially, this paper discusses the temporal behaviors of different pixels in infrared sequences. Then, the new nonlinear adaptive filter that is a variation of the median-modified Wiener filter is given to extract pulse signals on temporal profiles that relate to moving targets. Next, the variance of each temporal profile is estimated by segmenting each temporal profile into several segments to normalize the amplitude of the pulse signals. Finally, the proposed approach is tested via two infrared image sequences and compared with several conventional target detection algorithms. The results show our approach has a high effectiveness in extracting target temporal profiles amidst heavy and slowly evolving clutters.
Target detection algorithms based on temporal profiles have been showing excellent performance in detecting moving small targets in infrared image sequences. However, the temporal noise presented in infrared image sequences may change the form of temporal profiles, causing a degradation of detection performance of target detection algorithms that are performed on temporal profiles. To address the issue, a temporal noise suppression approach is proposed. First, we will show the effect of temporal noise on target detection algorithms based on temporal profiles. Then, a temporal noise suppression approach is proposed and its impact on temporal profiles is analyzed. Next, the execution procedure of the proposed approach is given. Finally, the proposed approach is evaluated and compared with several conventional target detection algorithms based on temporal profiles. The results show that our approach can significantly improve the detection performance of target detection algorithms based on temporal profiles.
The edges in infrared image will cause false alarms in small target detection. So a novel edge-preserving background estimation method is proposed in this paper for single frame small target detection. First, we propose a novel background estimation method based on semi-supervised learning, and the Graph Laplacian regularization is utilized in this model to preserve accurate edges in estimated background image. Then, the bilateral kernel is utilized to realize background estimation method. At last, edge-preserving estimated background is eliminated from original image to get difference image which is used as foreground to detect the small target. The experiment results demonstrate that our proposed method can achieve edge-preserving background estimation significantly and efficiently, and get better small target detection results.
Bilateral filter (BF) is a nonlinear filter for sharpness enhancement and noise removal. The BF performs the function by the two Gaussian filters, the domain filter and the range filter. To apply the BF to infrared (IR) small target detection, the standard deviation of the two Gaussian filters need to be changed adaptively between the background region and the target region. This paper presents a new BF with the adaptive standard deviation based on the analysis of the edge component of the local window, also having the variable filter size. This enables the BF to perform better and become more suitable in the field of small target detection Experimental results demonstrate that the proposed method is robust and efficient than the conventional methods.
The function of DIRCM (directed infrared countermeasures) jamming is to cause the missile to miss its intended target by disturbing the seeker tracking process. The DIRCM jamming uses the pulsing flashes of infrared (IR) energy and its frequency, phase and intensity have the influence on the missile guidance system. In this paper, we analyze the DIRCM jamming effect of the spin-scan reticle seeker. The simulation results show that the jamming effect is greatly influenced by frequency, phase and intensity of the jammer signal.
This paper presents a new bilateral filter (BF) using the adaptive standard deviation based on the target similarity index (TSI) for small target detection. At first, the threshold value of pixel's TSI in window decides whether any pixel is a potential small target or not. The TSIs of the potential small target pixels are mapped to standard deviations of domain and range filters of the BF by linear mapping and the two standard deviations increase in proportion to the TSI for blurring the small targets. For the more blurred small targets, the filter size according to the TSI is increased. Experimental results demonstrate that the proposed method is more robust and efficient than the conventional methods.
We propose a method for automatic target detection and tracking in forward-looking infrared (FLIR) image sequences. We use morphological connected operators to extract and track targets of interest and remove undesirable clutter. The design of these operators is based on general size, connectivity and motion criteria, using spatial intraframe and temporal interframe information. In a first step, an image sequence is filtered on a frame-by-frame basis to remove background and residual clutter and to enhance the presence of targets. Detections extracted from the first step are passed to a second step for motion-based analysis. This step exploits the spatiotemporal correlation of the data, stated in terms of a connectivity criterion along the time dimension. The proposed method is suitable for piplined implementation or time progressive coding/transmission, since only a few frames are considered at a time. Experimental results, obtained with real FLIR image sequences, illustrating a wide variety of target and clutter variability, demonstrate the effectiveness and robustness of the proposed method.
Mathematical morphology techniques, such as white top-hat (WTH) and new white top-hat (NWTH) transformation have been researched for small target detection application in the infrared (IR) images. However, its target enhancement performance still depends on its SE size set by users. To solve this problem, we propose a recursive multi-structuring elements (multi-SEs) NWTH method with an automatic decision mechanism of the SE size. The proposed method based on the NWTH transformation updates the multi-SEs by calculating candidate target-to-clutter ratio gain (CTCRG) of the NWTH images by the multi-SEs. Through a recursive procedure, final multi-SEs is automatically selected, and then small targets can be detected in a summation image of NWTH images by the final multi-SEs.
An infrared (IR) image synthesis method is proposed for the synthesis of a real IR background and modeled IR target, used as IR signatures, as well as a band-transformation between short wave IR (SWIR), middle wave IR (MWIR), and long wave IR (LWIR) in an IR imaging system simulation. IR target images are created by the RadThermIR software, an IR signature prediction software. Individual radiances for IR signatures, corresponding to the max/min temperatures of a real IR background and modeled IR target image, are calculated with Planck’s law. First, an IR background of an arbitrary wavelength band is transformed to one of the other wavelength bands with the temperature-radiance characteristics. And then, after adjusting the gray levels of the arbitrary IR target signatures based on their radiances for the wavelength band of the transformed IR background, these IR target and background signatures can be synthesized as one image for a specific wavelength band. The experimental results show that the modeled IR target images, such as a modeled helicopter and F16, can be synthesized on the IR background images of three IR wavelength bands. And we confirmed that IR background images of the three IR wavelength bands can diversely be synthesized with the modeled IR targets as the setting temperature of the target and background, the target distance, and the field of view (FOV) arbitrarily.
In directed infrared countermeasures (DIRCM), the purpose of jamming toward missiles is making missiles miss the target (aircraft of our forces) in the field of view. Since the DIRCM system directly emits the pulsing flashes of infrared (IR) energy to missiles, it is more effective than present flare method emitting IR source to omni-direction. In this paper, we implemented a reticle seeker simulation tool using MATLAB-SIMULINK, in order to analyze jamming effect of spin-scan and con-scan reticle missile seeker used widely in the world, though it was developed early. Because the jammer signal has influence on the missile guidance system using its variable frequency, it is very important technique among military defense systems protecting our forces from missiles of enemy. Simulation results show that jamming effect is greatly influenced according to frequency, phase and intensity of jammer signal. Especially, jammer frequency has the largest influence on jamming effect. Through our reticle seeker simulation tool, we can confirm that jamming effect toward missiles is significantly increased when jammer frequency is similar to reticle frequency. Finally, we evaluated jamming effect according to jammer frequencies, by using correlation coefficient as an evaluation criterion of jamming performance in two reticle missile seekers.
This paper starts with a review of the status of presently available IRST sensors and continues with the problems that they are facing, even after more than 25 years of development. The problems arise from uncertainty in target signatures, atmospheric effects, background clutter and different types of hardware and processing shortcomings. Yet IRST has a promising future thanks to increase in knowledge on target/background signatures, improved optics and electronics sensor fusion and algorithm development. Furthermore IRST has good perspective due to the increased number of applications for various platforms and scenarios. The perspective grows with the increased performance for lower price. It is important to consider here IRST as part of the platform system and to endeavor for close integration with other system components.
In this paper, we approach the problem of point target detection in infrared image sequences by investigating the application of the continuous wavelet transform to temporal pixel profiles. Observations made on the resulting time-scale images suggest that a multiscale temporal filtering algorithm is suitable for this application. Such an algorithm would exploit the fact that those clutter pixels that are the cause of false alarms when using a medium scale filter, respond to fine and coarse scale filters differently than the target pixels. To demonstrate the effectiveness of a multiscale approach, a first-cut 3-scale approach is tested on actual infrared image sequences featuring targets of opportunity and evolving cloud clutter. Results indicate that the 3-scale approach exhibits improved clutter suppression and target detection, when compared to a single scale filtering approach.
Clutter rejection is often an essential task in applications involving the detection and identification of small targets, making the choice of a clutter rejection algorithm extremely important if such a system is to perform as desired. Many different clutter rejection algorithms have been developed by various groups seeking to address this problem; however, as the performances of the algorithms are often very scenario dependent, selecting an appropriate algorithm for a given application usually requires thorough testing and performance analysis. This paper describes the methodology and results of a study done on clutter rejection algorithms for a system involving a staring IR camera mounted on an airborne platform. The purpose of this system is to detect the use of ordnance on a battlefield and then determine what type of ordnance was used. The clutter rejection algorithm needed to be real-time and possible to implement in hardware. The algorithms chosen for testing included 17 spatial filters and 4 temporal filters, along with two different types of thresholding (spatially fixed and spatially adaptive). Appropriate datasets for testing were created using a combination of real ordnance data taken by the IR camera, and clutter backgrounds from MODIS Airborne Simulator. Several different metrics were chosen to assist in algorithm performance evaluation. The final algorithm selection was based both on computational complexity and algorithm performance.
This paper deals with the problem of detection and tracking of low observable small-targets from a sequence of IR images against structural background and non-stationary clutter. There are many algorithms reported in the open literature for detection and tracking of targets of significant size in the image plane with good results. However, the difficulties of detecting small-targets arise from the fact that they are not easily discernable from clutter. The focus of research in this area is to reduce the false alarm rate to an acceptable level. Triple Temporal Filter reported by Jerry Silverman et. al., is one of the promising algorithms in this are. In this paper, we investigate the usefulness of Max-Mean and Max-Median filters in preserving the edges of clouds and structural backgrounds, which helps in detecting small-targets. Subsequently, anti-mean and anti-median operations result in good performance of detecting targets against moving clutter. The raw image is first filtered by max-mean/max-median filter. Then the filtered output is subtracted from the original image to enhance the potential targets. A thresholding step is incorporated in order to limit the number of potential target pixels. The threshold is obtained by using the statistics of the image. Finally, the thresholded images are accumulated so that the moving target forms a continuous trajectory and can be detected by using the post-processing algorithm. It is assumed that most of the targets occupy a couple of pixels. Head-on moving and maneuvering targets are not considered. These filters have ben tested successfully with the available database and the result are presented.
We introduce a spatial and temporal target detection method using spatial bilateral filter (BF) and temporal cross product (TCP) of temporal pixels in infrared (IR) image sequences. At first, the TCP is presented to extract the characteristics of temporal pixels by using temporal profile in respective spatial coordinates of pixels. The TCP represents the cross product values by the gray level distance vector of a current temporal pixel and the adjacent temporal pixel, as well as the horizontal distance vector of the current temporal pixel and a temporal pixel corresponding to potential target center. The summation of TCP values of temporal pixels in spatial coordinates makes the temporal target image (TTI), which represents the temporal target information of temporal pixels in spatial coordinates. And then the proposed BF filter is used to extract the spatial target information. In order to predict background without targets, the proposed BF filter uses standard deviations obtained by an exponential mapping of the TCP value corresponding to the coordinate of a pixel processed spatially. The spatial target image (STI) is made by subtracting the predicted image from the original image. Thus, the spatial and temporal target image (STTI) is achieved by multiplying the STI and the TTI, and then targets finally are detected in STTI. In experimental result, the receiver operating characteristics (ROC) curves were computed experimentally to compare the objective performance. From the results, the proposed algorithm shows better discrimination of target and clutters and lower false alarm rates than the existing target detection methods.
In this paper, we introduce an edge directional 2D least mean squares (LMS) filter for small target detection in infrared (IR) images. Generally, the 2D LMS filter functions as a background prediction to apply to IR small target detection field. In order to accurately predict background objects as well as regions covered by small targets, the proposed 2D LMS filter take full advantage of edge information of prediction pixels corresponding to surrounding blocks around current filter window. And, to adjust adaptively its step size in the background and small target region, the adaptive region-dependent nonlinear step size is calculated by using the variance of the prediction pixels of the surrounding blocks. This prediction structure and adaptive step size of the proposed 2D LMS filter is applied to the background region including objects such as cloud edge and small target region differently. Through this way, the proposed 2D LMS filter predicts the background excluding small targets. Then, by subtracting the predicted background from the original IR image, small targets can be extracted. Experimental results show that the proposed 2D LMS filter has stronger target extraction and better background suppression ability compared to the existing 2D LMS filters.
In this paper, we propose a novel TDLMS filter using a sub-sampling mask and step-size index for small target detection in infra-red (IR) imagery. The concept of the proposed TDLMS is to utilize the sub-sampling mask of the weight-matrix having the structure reducing effects of the target pixels in order to predict the background exactly. And the nonlinear step size using the luminance rate (LR) and contrast rate (CR) is used and then background is predicted finally by using the two-dimensional Gaussian distance (TDGD). Experimental results show that the proposed method exhibits higher detection rates in comparison to the conventional TDLMS filter.
A new small-target detection method in forward-looking infrared images (FLIR) is proposed. The goal is to identify target locations, with low false alarms, in thermal infrared images of a natural battlefield. Unlike previous approaches, the proposed method deals with small-target detection in low-contrast images, and it is able to determine centers of targets accurately. The method comprises three distinct stages. The first stage is called center-surround difference, whose function is to find salient regions in an input image. In the second stage, local fuzzy thresholding is applied to the region of interest that is chosen from the result of the first step. The extracted binary objects are potential targets that will be classified as valid targets or clutters. Finally, using size and affinity measurements, the potential targets are compared with target templates to discard clutters in the third stage. In the experiments, many natural infrared images are used to prove the effectiveness of the proposed method. The proposed method is compared to previously reported approaches to verify its efficiency.
In this paper, we propose a new criterion to estimate the quality of infrared small target images. To describe the criterion
quantitatively, two indicators are defined. One is the “degree of target being confused” that represents the ability of infrared
small target image to provide fake targets. The other one is the “degree of target being shielded”, which reflects the contribution
of the image to shield the target. Experimental results reveal that this criterion is more robust than the traditional method
(Signal-to-Noise Ratio). It is not only valid to infrared small target images which Signal-to-Noise Ratio could correctly
describe, but also to the images that the traditional criterion could not accurately estimate. In addition, the results of
this criterion can provide information about the cause of background interfering with target detection.
Bilateral filter (BF) performs sharpness enhancement and noise removal by using two Gaussian filters, the domain filter in
spatial domain and the range filter in intensity domain. To apply the BF to infrared (IR) small target detection, the standard
deviation of the two Gaussian filters need to be changed adaptively between the background region and the target region. This
paper presents a new BF for small target detection with the adaptive standard deviation based on the analysis of the edge
component, also having the variable filter size. At first, threshold of pixel edge components for four directions decides
whether any pixel is potential small targets or not. For the proposed BF operation for the potential small target pixels,
its edge component is mapped to two standard deviations of the domain filter and the range filter in the proposed BF by mapping
function. When the BF comes to a target region, the two standard deviations increase in proportion to the edge component to
blur the small targets. To further blur the small targets, the filter size of the BF also increases by its edge component.
This enables the BF to perform better and become more suitable in the field of small target detection Experimental results
demonstrate that the proposed method is more robust and efficient than the conventional methods.
KeywordsSmall target detection-Bilateral filter-Edge component
In this paper, we introduce an efficient TDLMS filter, using the new weight structure and nonlinear step size for small target
detection within infra-red (IR) imagery. A new TDLMS filter that can efficiently detect a small target in IR imagery is proposed.
The concept of the proposed filter is to utilize the new weight matrix having the structure reducing effects of the target
pixels in order to predict exactly the background. The nonlinear step size utilizing the block statistics is used and background
estimation is calculated finally by using the Gaussian distance map. Experimental results show that the proposed method exhibits
higher detection rates and lower false alarm rates in comparison to the conventional TDLMS filter.
TDLMS is a general adaptive filter algorithm and when applied to infrared small target detection, traditional structure and
implementation of TDLMS may cause some problems in this field. This paper presents a new TDLMS filter structure and implementation
incorporating neighborhood analysis and data fusion, which is capable of acquiring and analyzing more information from the
vicinity of the target, leading to a more prominent detection result. This enables TDLMS filter to perform better and become
more suitable in the field of small target detection. Experiments showed the efficiency of the proposed algorithm.
In this paper, we implement a reticle seeker missile simulator on MATLAB-SIMULINK to analyze the jamming effect of the spin-scan and conscan reticle seeker. The DIRCM (Directed Infrared Countermeasures) system uses the pulsing flashes of infrared (IR) energy and its frequency and intensity have influence on the missile guidance system. Our simulation results show that jamming effect is indicated significantly when jammer frequency and reticle frequency are similar and present a 3D trajectory of missile motions by jamming.
This paper presents a novel algorithm for detecting and tracking dim moving point target in IR image sequence with low SNR. Original images are preprocessed using temperature non-linear elimination and Top-hat operator, and then a composite frame is obtained by reducing the three-dimensional (3D) spatio-temporal scanning for target to 2D spatial hunting. Finally the target trajectory is tracked under the condition of constant false-alarm probability (CFAR). Based on the experimental results, the algorithm can successfully detect dim moving point target and accurately estimate its trajectory. The algorithm, insensitive to the velocity mismatch and the changes of statistical distribution of background or noise, is adaptable to real-time target detection and tracking.
This paper presents a new small target detection method using cross product of temporal pixels based on temporal profiles in infrared (IR) image sequences. Temporal characteristics of small targets and various backgrounds are different. A new algorithm classifies target pixels and background pixels through hypothesis testing using the cross product of pixels on temporal profile and predicts the temporal backgrounds based on the results. Small target pixels are detected by subtracting the predicted temporal background profile from the original temporal profile. For performance comparison between the proposed method and the conventional methods, the receiver operating characteristics (ROC) curves were computed experimentally. Experimental results show that the proposed algorithm has better discrimination of target and clutter pixels and lower false alarm rates than conventional methods.
To improve the performance of a top-hat transformation for infrared small target enhancement, a new class of top-hat transformation through structuring element construction and operation reorganization is proposed. The structuring element construction and operation reorganization are based on the property of the infrared small target image and thus can greatly improve the performance of small target enhancement. Experimental results verified that it was very efficient. (C) 2008 SPIE and IS&T. [DOI: 10.1117/1.2955943]
The clinical performance of a laboratory test can be described in terms of diagnostic accuracy, or the ability to correctly classify subjects into clinically relevant subgroups. Diagnostic accuracy refers to the quality of the information provided by the classification device and should be distinguished from the usefulness, or actual practical value, of the information. Receiver-operating characteristic (ROC) plots provide a pure index of accuracy by demonstrating the limits of a test's ability to discriminate between alternative states of health over the complete spectrum of operating conditions. Furthermore, ROC plots occupy a central or unifying position in the process of assessing and using diagnostic tools. Once the plot is generated, a user can readily go on to many other activities such as performing quantitative ROC analysis and comparisons of tests, using likelihood ratio to revise the probability of disease in individual subjects, selecting decision thresholds, using logistic-regression analysis, using discriminant-function analysis, or incorporating the tool into a clinical strategy by using decision analysis.
Sensitivity and specificity are the basic measures of accuracy of a diagnostic test; however, they depend on the cut point used to define "positive" and "negative" test results. As the cut point shifts, sensitivity and specificity shift. The receiver operating characteristic (ROC) curve is a plot of the sensitivity of a test versus its false-positive rate for all possible cut points. The advantages of the ROC curve as a means of defining the accuracy of a test, construction of the ROC, and identification of the optimal cut point on the ROC curve are discussed. Several summary measures of the accuracy of a test, including the commonly used percentage of correct diagnoses and area under the ROC curve, are described and compared. Two examples of ROC curve application in radiologic research are presented.
In accordance with the characteristics of the small target against sea and sky background, a small target detection algorithm based on multi-scale mutual energy cross is proposed in this paper. In order to determinate potential region in which the targets lie, sea-level line first is detected by using direction gradient operator. Then on the basis of wavelet decomposition, we define the multi-scale mutual energy cross function to eliminate the intense infrared clutter. The precise position of the small target is finally found by region-growth technique. Experimental results show that the algorithm proposed has better performance with respect to detection probability of single frame and less computation complexity. It is an effective small infrared target detection algorithm against sea and sky background.
Bilateral filtering smooths images while preserving edges, by means of a nonlinear combination of nearby image values. The method is noniterative, local, and simple. It combines gray levels or colors based on both their geometric closeness and their photometric similarity, and prefers near values to distant values in both domain and range. In contrast with filters that operate on the three bands of a color image separately, a bilateral filter can enforce the perceptual metric underlying the CIE-Lab color space, and smooth colors and preserve edges in a way that is tuned to human perception. Also, in contrast with standard filtering, bilateral filtering produces no phantom colors along edges in color images, and reduces phantom colors where they appear in the original image
In this paper, we present the adaptive bilateral filter (ABF) for sharpness enhancement and noise removal. The ABF sharpens an image by increasing the slope of the edges without producing overshoot or undershoot. It is an approach to sharpness enhancement that is fundamentally different from the unsharp mask (USM). This new approach to slope restoration also differs significantly from previous slope restoration algorithms in that the ABF does not involve detection of edges or their orientation, or extraction of edge profiles. In the ABF, the edge slope is enhanced by transforming the histogram via a range filter with adaptive offset and width. The ABF is able to smooth the noise, while enhancing edges and textures in the image. The parameters of the ABF are optimized with a training procedure. ABF restored images are significantly sharper than those restored by the bilateral filter. Compared with an USM based sharpening method-the optimal unsharp mask (OUM), ABF restored edges are as sharp as those rendered by the OUM, but without the halo artifacts that appear in the OUM restored image. In terms of noise removal, ABF also outperforms the bilateral filter and the OUM. We demonstrate that ABF works well for both natural images and text images.
This paper addresses the problem of detecting small, moving, low
amplitude objects in image sequences that also contain moving nuisance
objects and background noise. We formulate this problem in the context
of a hypothesis testing procedure on individual pixel temporal profiles,
leading to a computationally efficient statistical test. The technique
assumes we have reasonable deterministic and statistical models for the
temporal behavior of the background noise, target, and clutter, on a
single pixel basis. Based on these models we develop a generalized
likelihood ratio test (GLRT) and perfect measurement performance
analysis, and present the resulting decision rule. We also propose a
parameter estimation technique and compare its performance to the Cramer
Rao bound (CRB). We demonstrate the effectiveness of the technique by
applying the resulting algorithm to real world infrared (IR) image
sequences containing targets of opportunity. The approach could also be
applicable to other image sequence processing scenarios, using
acquisition systems besides IR imaging, such as detection of small
moving objects or structures in a biomedical or biological imaging
scenario, or the detection of satellites, meteors or other celestial
bodies in night sky imagery acquired using a telescope
Track-Before-Detect (TBD) approaches have important applications such as the detection of small, dim targets in a long-range Infrared Detection and Tracking (IRDT) system. The movement model of target limited by TBD is extended to that of target with constant acceleration in this work. An algorithm based on linear variant-coefficient difference equation for moving target indication is proposed. Moreover, based on parametric models of target and background, this paper presents an analysis of its optimal signal-to-noise ratio (SNR) gain versus target and background characteristics as well as the sensitivity of this gain to mismatch.
The Statistical Evaluation of Medical Tests for Classification and Prediction Fig. 11. Target trajectory results of IR sequences: (a) original image, detection results of (b), Max-mean method, (c) Max-median method, (d) Zhang's method, (e) Tzannes's method and (f)
- M S Pepe
M.S. Pepe, The Statistical Evaluation of Medical Tests for Classification and Prediction, Oxford University Press, New York, 2003. Fig. 11. Target trajectory results of IR sequences: (a) original image, detection results of (b), Max-mean method, (c) Max-median method, (d) Zhang's method, (e) Tzannes's method and (f) proposed spatial and temporal BF.