No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Edge Detection Based on Wavelet Analysis with Gaussian Filter
Abstract
In this paper an edge detection algorithm base on wavelet transform with Gaussian filter was proposed. In this algorithm original images are firstly converted into gray images and then each pixel was analyzed using wavelet transform to find the local maximum of the gray gradient of each pixel along the phase angle direction and compared with a given threshold value, through which real edge can be kept and fake ones will be eliminated. In the computation of local maximum, the gray gradients computed in eight directions, which can improve precision of edge detection. After the investigation of influence of filter length, scale and threshold value on the edge detection the proposed algorithm is validated by the comparison with N.L. Fen?ndez-Garc?a’s Minimean and Minimax methods for 100 real color images. The extraction result is more close to the real image which indicates the algorithm is effective and can be used to extract edges in different research areas.
... Gaussian filters exhibit the same degree of smoothness in all directions, an advantage when dealing with images with an unknown edge direction (Fude et al., 2008). However, the filter processes the whole image uniformly, which can create dissimilarities due to the non-linear distribution of image intensities (Seddik et al., 2014). ...
Detecting incipient damages in structures is an important task for establishing a new structural health monitoring (SHM) process, and it has a positive impact on reducing costs and increasing the safety of equipment and structures. In particular, in Brazil there are several large concrete-based structures, such as viaducts and bridges, for which detecting small damages is a relevant challenge, mainly due to their dimensions and limited access for inspection. However, the development of unmanned aerial vehicles (UAVs) allows one to simplify the procedure in different engineering services, including to take aerial photographs. Aerial photographs can be used to evaluate the structure and establish new SHM processes. However, this type of procedure involves an accurate image post-processing to detect the damage minimizing false alarms. Thus, this article presents an investigation on the use of different filters applied to aerial photographs obtained from a quadcopter-type UAV. The photographs are obtained in a controlled flight by a pilot on the ground and they are computationally evaluated using an algorithm written in Python. The proposed SHM process involves obtaining images of the structure in the undamaged condition and with a typical damage to be detected. Different filters are evaluated, such two of them are spatial-based and a third one is a frequency based filter. The analyses are performed considering concrete structures with crack-type damages, and the influences of different factors, such as flight speed, distance between UAV and structure, lighting conditions, among others, are evaluated. The results include developing a framework for aerial images analysis, as well as the definition of a process to obtain the photographs from an UAV, considering the preparation and flight stages. The effect of the filter mask dimension is investigated by considering different strategies to apply on the image edges. The results demonstrate that spatial filters are very simple to be considered and they provide interesting results. The proposed approach contributes to increasing the safety of concrete structures, with the potential for generating technological products and services.
... The Gaussian filter is a linear smoothing filter that is often used to eliminate Gaussian noise and has good elimination for noise obeying a normal distribution [49]. The basic idea of a Gaussian filter is that by discretizing the Gaussian function of images, the value of the Gaussian function of discrete points is adopted as weights corresponding to these discrete points, and a weighted average in some fields is achieved for each pixel of the image to eliminate noise [50]. Gaussian kernel convolution is achieved by applying a convolution layer to the processing of images [51]. ...
Recognizing coal and coal gangue is an important part of the coal industry and is mainly conducted via human sorting at present. Consequently, considerable manpower is needed, which adds a burden to enterprises and results in low efficiency. As an important branch of artificial intelligence, deep learning has been widely applied in many fields, especially in machine vision and voice recognition, its performance is greatly improved compared with the performances of traditional learning methods, and it also has good a transfer learning ability. This paper proposed an improved YOLOv4 algorithm as a classic deep learning method for the intelligent and highly accurate recognition of coal and coal gangue. Compared to other algorithms, YOLOv4 has a better anchor value by applying cluster analysis to different data sets, a good anti-interference ability due to using the Laplacian operator and Gaussian filter to reduce the impacts from mine dust and shock and acquires richer detailed information by increasing the number of layers of the feature pyramid. The experimental results show that compared with the other four algorithms of YOLOv4, YOLOv3, SSD and Faster-RCNN, the improved YOLOv4 proposed in this paper exhibits better detection accuracy, a better detection speed and robust performance.
Since shape is one of the important low level features of any Content Based Image Retrieval (CBIR) system, this paper proposes a new edge based shape feature representation method with multiresolution enhanced orthogonal polynomials model and morphological operations for effective image retrieval. In the proposed method, initially the orthogonal polynomials model coefficients are computed and reordered into multiresolution subband like structure. Edge image is then obtained by utilizing the two level adaptive thresholds and local maxima of the gradient in horizontal, vertical, diagonal and anti-diagonal directions. The approximate shape boundary of the image is recovered with morphological operations. Then the Pseudo Zernike moment based global shape features, which are invariant to basic geometric transformations, are extracted. The obtained features are termed as global shape feature vector and are used for retrieving similar images with Canberra distance metric. The efficiency of the proposed method is experimented on a subset of standard Corel, Yale and MPEG-7 databases and the results are compared with existing techniques.
Stock fluctuations anomaly increase the uncertainty and investment risk in the stock market, is an important element in financial research. In this paper, wavelet modulus maxima method is used in the detection of abnormal stock analysis. It is obtained based on the irregular sampling in the multi-scale wavelet transform. It overcomes the localized limitation about traditional Fourier analysis in time and frequency domains. Experimental results show that the wavelet modulus maxima method can not only depict the position of the point mutation in the signals but also capture the singular points of the stock unusual fluctuations quickly and accurately.
Stock market has gradually become an absolutely necessary part of financial market in China. The trend analysis and forecasting of stock prices become key topics in investment and security, which have great theoretical significance and application value. In this paper, the wavelet modulus maxima method is proposed for the abnormal detection of the stock market. The abnormal points detected by wavelet modulus maxima are replaced by the new interpolation points which will be used as an important index of Kalman algorithm to predict stock. The experimental results show that the proposed method can predict the stock data with higher credibility than Kalman algorithm. Therefore, the proposed method can reduce the investment risk and plays an important role in the economic development and financial building.
This paper describes a computational approach to edge detection. The success of the approach depends on the definition of a comprehensive set of goals for the computation of edge points. These goals must be precise enough to delimit the desired behavior of the detector while making minimal assumptions about the form of the solution. We define detection and localization criteria for a class of edges, and present mathematical forms for these criteria as functionals on the operator impulse response. A third criterion is then added to ensure that the detector has only one response to a single edge. We use the criteria in numerical optimization to derive detectors for several common image features, including step edges. On specializing the analysis to step edges, we find that there is a natural uncertainty principle between detection and localization performance, which are the two main goals. With this principle we derive a single operator shape which is optimal at any scale. The optimal detector has a simple approximate implementation in which edges are marked at maxima in gradient magnitude of a Gaussian-smoothed image. We extend this simple detector using operators of several widths to cope with different signal-to-noise ratios in the image. We present a general method, called feature synthesis, for the fine-to-coarse integration of information from operators at different scales. Finally we show that step edge detector performance improves considerably as the operator point spread function is extended along the edge.
Magnetic resonance spectral images provide information on metabolic processes and can thus be used for in vivo tumor diagnosis. However, each single spectrum has to be checked manually for tumorous changes by an expert, which is only possible for very few spectra in clinical routine. We propose a semi-supervised procedure which requires only very few labeled spectra as input and can hence adapt to patient and acquisition specific variations. The method employs a discriminative random field with highly flexible single-side and parameter-free pair potentials to model spatial correlation of spectra. Classification is performed according to the label set that minimizes the energy of this random field. An iterative procedure alternates a parameter update of the random field using a kernel density estimation with a classification by means of the GraphCut algorithm. The method is compared to a single spectrum approach on simulated and clinical data.
Edges are important cues defining coherent auditory objects. As a model of auditory edges, sound on- and offset are particularly suitable to study their neural underpinnings because they contrast a specific physical input against no physical input. Change from silence to sound, that is onset, has extensively been studied and elicits transient neural responses bilaterally in auditory cortex. However, neural activity associated with sound onset is not only related to edge detection but also to novel afferent inputs. Edges at the change from sound to silence, that is offset, are not confounded by novel physical input and thus allow to examine neural activity associated with sound edges per se. In the first experiment, we used silent acquisition functional magnetic resonance imaging and found that the offset of pulsed sound activates planum temporale, superior temporal sulcus and planum polare of the right hemisphere. In the planum temporale and the superior temporal sulcus, offset response amplitudes were related to the pulse repetition rate of the preceding stimulation. In the second experiment, we found that these offset-responsive regions were also activated by single sound pulses, onset of sound pulse sequences and single sound pulse omissions within sound pulse sequences. However, they were not active during sustained sound presentation. Thus, our data show that circumscribed areas in right temporal cortex are specifically involved in identifying auditory edges. This operation is crucial for translating acoustic signal time series into coherent auditory objects.
The reactor in three-phase fluidized bed is a close reaction device, which contact area is greatly, has high production efficiency, and has a wide-ranging application in the reactions such as mineral oil hydrogenization splitting and so on. The air bubble sizes and the distributing with the moving velocity upward in the fluidized bed, will decided the effective passing area and the reaction efficiency. The image identifying and analyzing system, by means of gray degree values, can be used to identified and analyzed the air bubbles state in the reactor, such as the distributing, the moving, the gathering and blowing out of the air bubbles.
This paper introduces some applications in image processing with wavelets, such as edge detection, pattern recognition and image segmentation. Edge detection includes contour extraction and background removal for an illustration of gray - level, extraction of tumor region in color image and enlargement of type size of Chinese character. As for recognition of Chinese character or character at first, we can reduce the dimensionality of two dimensional patterns to one dimensional sign, then do wavelets decomposition for these one dimensional patterns, at last compute fractal dimension of each sub - pattern, so feature vectors can be obtained and used to recognition. The characters after affine transformation can also be recognized by a dyadic wavelet affine invariant function. At last we introduce two means of image segmentation. One is based on energy analysis, another is singularity analysis.
This paper highlights the development of a new ensemble correlation, multiphase flow PIV technique. The particular approach discussed in this paper is a straightforward extension of single-phase PIV and one of the major advantages of the technique is that it employs a single CCD camera. This velocimeter measures the motion of the liquid phase and the bubbles using an ensemble correlation procedure of successive recordings of bubble–tracer patterns. Moreover, the ensemble correlation technique enables the simultaneous measurement of the concentration of tracer particles and bubbles. The performance of the PIV technique is studied using synthetic bubble–tracer patterns. The performance is gauged in terms of the amplitude of the displacement correlation peak pertaining to the bubbles and in terms of the valid detection probability of that same peak. Finally, the newly developed PIV technique is tested by studying experimentally the two-phase flow in a pseudo-two-dimensional bubble column.
Whole field velocity and point temperature and surface heat flux measurements were performed to characterise the interaction of a single rising ellipsoidal air bubble with the free convection flow from a heated flat surface immersed in water at different angles of inclination. Two thermocouples and a hot film sensor were used to characterise heat transfer from the surface, while a time-resolved digital particle image velocimetry technique was used to map the bubble induced flow in a plane parallel to the surface. Heat flux fluctuations, preceding and following the bubble passage, were shown to correlate with the variation in both local flow velocities and fluid temperatures. The largest increases in heat transfer were recorded when both flow and temperature effects combined to enhance the convective cooling simultaneously. Such conditions were shown to be most likely met when the block was inclined at 45°, thus forcing the bubble to slide closer to the heated surface and hence to the thermal boundary layer.
Two new methods are proposed to automatically generate consensus ground truth for real images: Minimean and Minimax methods. These methods and a version of the Yitzhaky and Peli method have been used to provide ground truth for the comparison of edge detection techniques. The developed experiments have revealed that the Minimean consensus method is suitable for the comparison of edge detectors because its results are equivalent to those obtained with artificial or manual ground truth.
This research uses the object extracting technique to extract – the thumb, index, middle, ring, and small fingers from the hand images. The algorithm developed in this research can find the precise locations of the fingertips and the finger-to-finger-valleys. The extracted fingers contain many useful geometry features. One can use these features to do the person identification. The orientations of the index, middle and ring fingers are found in this research. Image rotating, image shifting, and image interpolating techniques are used to align different persons’ fingers. Image subtraction is used to exam the difference of two fingers.
This research uses the object extracting technique to extract the index, middle and ring fingers from the hand images. The algorithm developed in this research can find the precise locations of the different fingers' fingertips and the finger-to-finger-valleys. After finding the positions of the fingertips and finger-valleys, the index, middle and ring fingers can be extracted from the hand images by using morphological technique. The extracted index, middle and ring fingers contain many useful geometry features. One can use these features to do the person's identification. The orientations of the index, middle and ring fingers are found in this research. Image rotating, image shifting, and image interpolating techniques are used to align different persons' index, middle and ring fingers. Image subtraction is used to exam the difference of two index, middle and ring finger images. In this research so far only use the index, middle and the ring fingers as the features to identify different persons.
To establish reasonable safety concepts for the realization of commercial liquid-metal fast breeder reactors, it is indispensable to demonstrate that the release of excessive energy due to re-criticality of molten core could be prevented even if a severe core damage accident took place. Two-phase flow due to the boiling of fuel-steel mixture in the molten core pool has a larger liquid-to-gas density ratio and higher surface tension in comparison with those of ordinary two-phase flows such as air-water flow. In this study, to investigate the effect of the recirculation flow on the bubble behavior, visualization and measurement of nitrogen gas-molten lead bismuth in a rectangular tank was performed by using neutron radiography and particle image velocimetry techniques. Measured flow parameters include flow regime, two-dimensional void distribution, and liquid velocity field in the tank. The present technique is applicable to the measurement of velocity fields and void fraction, and the basic characteristics of gas-liquid metal two-phase mixture were clarified.
In this study, an on-line concentration method which coupled cation-selective exhaustive injection (CSEI) sweeping technology with microemulsion electrokinetic chromatography (MEEKC) was used to detect and analyze several tobacco alkaloids (nornicotine, anabasine, anatabine, nicotine, myosmine and cotinine) that are commonly found in various tobacco products. First, the effects of microemulsion compositions (oil, cosurfactant and solution pH) were examined in order to optimize the alkaloid separations in conventional MEEKC. The pH value and the injection length of basic plug were found to be the predominant influences on the alkaloid stacking. This optimal CSEI sweeping MEEKC method provided approximately 180- to 540-fold increase in detection sensitivity in terms of peak height without any loss in separation efficiency when compared to normal MEEKC separation. Furthermore, this proposed CSEI sweeping MEEKC method was applied successfully for the detection of the minor alkaloids nornicotine, anabasine and anatabine in tobacco products.
This paper presents a database containing `ground truth' segmentations produced by humans for images of a wide variety of natural scenes. We define an error measure which quantifies the consistency between segmentations of differing granularities and find that different human segmentations of the same image are highly consistent. Use of this dataset is demonstrated in two applications: (1) evaluating the performance of segmentation algorithms and (2) measuring probability distributions associated with Gestalt grouping factors as well as statistics of image region properties
Image Recognition for Light-Scattering of Air Bubbles in Water. Chinese Optics & Optoelectronic Technology
- Li Fei
- Xu De-Sheng
- Liu Qi-Zhong
- Xia Min
- Yang Kecheng
The Medicine Image Edge Detection Based on
- Tang Huang Hongqiong
- Tianhao
Huang Hongqiong Tang Tianhao., 2007, The
Medicine Image Edge Detection Based on
Spline Wavelet Transform. Microcomputer
Information. 23(4-3), 313-314.
Detection Benchmark and Dataset Available from: <http://www.eecs.berkeley
- Berkeley Segmentation
- Boundary
The Algorithm of Applying Wavelet Transform to Obtain Cancer Cell's Edge. Microcomputer Information(Control and Automation)
- Lin Changyi
Lin ChangYi., 2007, The Algorithm of
Applying Wavelet Transform to Obtain
Cancer
Cell's
Edge.
Microcomputer
Information(Control and Automation), 23(2-1),
270-272.
A Method of Coding TV Signals Based on Edge Detection
- B Julez
Julez, B., 1959, A Method of Coding TV
Signals Based on Edge Detection. Bell
System Technology. 38(4), pp.1001-1020.
Image Recognition for Light-Scattering of Air Bubbles in Water
- Li Fei
- Xu De-Sheng
- Liu Qi-Zhong
- Xia Min
Li Fei, Xu De-sheng, Liu Qi-zhong, Xia Min,
Yang Ke-cheng., 2006, Image Recognition
for Light-Scattering of Air Bubbles in Water.
Optics & Optoelectronic Technology, 4(6),
25-27.
Multiscales Edges Detection Based On Bubble Wavelet
- Jia Tianxu
- Zheng Nanning
- Jia Tianxu
- Zheng Nanning
Jia Tianxu, Zheng Nanning., 1996,
Multiscales Edges Detection Based On
Bubble Wavelet. ACTA ELECTRONICA
SINICA, 24(4), 117-119.