Figure - available from: Plant Phenomics
This content is subject to copyright. Terms and conditions apply.
Dynamic color transform in a standard object detection pipeline. The input image x is first transformed to x′ by the DCT network, where x′=αx+β. Then, x′ is sent to detection network for computing losses ℓcls and ℓloc. The losses are used to update the DCT and the detection network.
Source publication
Wheat head detection can measure wheat traits such as head density and head characteristics. Standard wheat breeding largely relies on manual observation to detect wheat heads, yielding a tedious and inefficient procedure. The emergence of affordable camera platforms provides opportunities for deploying computer vision (CV) algorithms in wheat head...
Citations
... Recent studies on growth event detection mainly concentrated on using deep learning to count and detect/segment buds [8][9][10] and detect wheat heads [11,12] from plant images. Several researches [13,14] carried out budding detection and organ tracking at the same time on time-series plant images, but the task usually restricted on potted rosette plant or standard plant (Arabidopsis), and usually put strong restriction on imaging direction (vertically imaging). ...
... The study by Han and zang et al. 2,24,25 used GWHD2021 for wheat spike detection and got good results. Liu et al. 1 proposed Dynamic Color Transformation (DCT). The DCT model changes the color channel of the input image, which can significantly reduce false positives and improve the detection performance.It was used on the YOLOV4 26 network and obtained a average domain accuracy(ADA) of 69.5% on the GWHD2021 dataset, which is the runner-up of GWC2021. ...
Wheat is a crucial crop worldwide, and accurate detection and counting of wheat spikes are vital for yield estimation and breeding. However, these tasks are daunting in complex field environments. To tackle this, we introduce RIA-SpikeNet, a model designed to detect and count wheat spikes in such conditions. First, we introduce an Implicit Decoupling Detection Head to incorporate more implicit knowledge, enabling the model to better distinguish visually similar wheat spikes. Second, Asymmetric Loss is employed as the confidence loss function, enhancing the learning weights of positive and hard samples, thus improving performance in complex scenes. Lastly, the backbone network is modified through reparameterization and the use of larger convolutional kernels, expanding the effective receptive field and improving shape information extraction. These enhancements significantly improve the model’s ability to detect and count wheat spikes accurately. RIA-SpikeNet outperforms the state-of-the-art YOLOv8 detection model, achieving a competitive 81.54% mAP and 90.29% R². The model demonstrates superior performance in challenging scenarios, providing an effective tool for wheat spike yield estimation in field environments and valuable support for wheat production and breeding efforts.
... For the feature-level method regarding wheat head variation in size, implemented an EfficientDet and a BiFPN-based approach to tackle the occlusion-robust wheat head detection challenge. Regarding the image-level method, Liu et al. (2022) presented their technique for reducing false negatives in wheat head detection by applying color transformation. Hartley and French (2021) examined the impact of using synthetic data generated by a generative adversarial network (GAN) with domain augmentation to supplement the original dataset. ...
... Additionally, because there were wheat heads of different size and scale, we chose the EfficientDet-5 model since it employs Bi-FPN, a bidirectional feature pyramid network that can be improved by fast normalization and feature fusion, thus allowing easy multi-scale feature fusion, making it a good algorithm for multi-scale object detection because of the weights added for every input by the BiFPN. Many scholars, such as Hartley and French (2021) and Liu et al. (2022), have shown that using domain variation and color transformation of GWHD can better improve the model's adaptability and robustness. The results indicated that FDA with AABG had better performance than the base model, FDA with RGF and even better mAP performance than the former. ...
... Leveraging deep learning, particularly object detection technology, has proven to be a potential method for addressing this issue [11]. Deep learning can autonomously extract and learn features from raw data, reducing the need for manual feature engineering. ...
Wheat spike detection is crucial for estimating wheat yields and has a significant impact on the modernization of wheat cultivation and the advancement of precision agriculture. This study explores the application of the DETR (Detection Transformer) architecture in wheat spike detection, introducing a new perspective to this task. We propose a high-precision end-to-end network named WH-DETR, which is based on an enhanced RT-DETR architecture. Initially, we employ data augmentation techniques such as image rotation, scaling, and random occlusion on the GWHD2021 dataset to improve the model’s generalization across various scenarios. A lightweight feature pyramid, GS-BiFPN, is implemented in the network’s neck section to effectively extract the multi-scale features of wheat spikes in complex environments, such as those with occlusions, overlaps, and extreme lighting conditions. Additionally, the introduction of GSConv enhances the network precision while reducing the computational costs, thereby controlling the detection speed. Furthermore, the EIoU metric is integrated into the loss function, refined to better focus on partially occluded or overlapping spikes. The testing results on the dataset demonstrate that this method achieves an Average Precision (AP) of 95.7%, surpassing current state-of-the-art object detection methods in both precision and speed. These findings confirm that our approach more closely meets the practical requirements for wheat spike detection compared to existing methods.
... They first extracted wheat color information from the acquired wheat RGB images to eliminate interference areas and then used the AdaBoost algorithm to train a classifier to locate ear objects [8]. The dynamic color transform (DCT) model utilizes the idea of dynamic color transformation to mitigate the influence of background interference by modifying the color channels of the input RGB wheat ear image, thereby achieving precise ear detection [9]. Fernandez-Gallego et al. removed low-frequency and high-frequency interference elements from wheat ear images under natural light conditions using filters and the Find Maxima algorithm. ...
Wheat ear counting is crucial for calculating wheat phenotypic parameters and scientifically managing fields, which is essential for estimating wheat field yield. In wheat fields, detecting wheat ears can be challenging due to factors such as changes in illumination, wheat ear growth posture, and the appearance color of wheat ears. To improve the accuracy and efficiency of wheat ear detection and meet the demands of intelligent yield estimation, this study proposes an efficient model, Generalized Focal Loss WheatNet (GFLWheatNet), for wheat ear detection. This model precisely counts small, dense, and overlapping wheat ears. Firstly, in the feature extraction stage, we discarded the C4 feature layer of the ResNet50 and added the Convolutional block attention module (CBAM) to this location. This step maintains strong feature extraction capabilities while reducing redundant feature information. Secondly, in the reinforcement layer, we designed a skip connection module to replace the multi-scale feature fusion network, expanding the receptive field to adapt to various scales of wheat ears. Thirdly, leveraging the concept of distribution-guided localization, we constructed a detection head network to address the challenge of low accuracy in detecting dense and overlapping targets. Validation on the publicly available Global Wheat Head Detection dataset (GWHD-2021) demonstrates that GFLWheatNet achieves detection accuracies of 43.3% and 93.7% in terms of mean Average Precision (mAP) and AP50 (Intersection over Union (IOU) = 0.5), respectively. Compared to other models, it exhibits strong performance in terms of detection accuracy and efficiency. This model can serve as a reference for intelligent wheat ear counting during wheat yield estimation and provide theoretical insights for the detection of ears in other grain crops.
... Wheat crops provide food for approximately 30% of the global population [1,2] and are directly linked to everyone's food security [3,4]. To ensure sustainable wheat production, breeders must continuously assess their organic traits to identify and capitalize on high-yielding varieties [5,6]. In agricultural evaluation and management, wheat head status and density during growth stages are commonly used to estimate yield [7], and the number of wheat heads per unit area directly reflects quality [8,9]. ...
Accurate real-time observations of wheat head growth are crucial for effective agricultural management. However, the dense distribution of wheat heads often leads to severe overlap in imagery. Existing target detection algorithms face challenges in overcoming this problem, rendering them ineffective for real-time field computations using portable devices. Therefore, this study proposes a lightweight you-only-look-once (YOLO) model with a simplified structure and a more powerful attention mechanism. A limitation of the traditional YOLO model is its complex structure: it requires a substantial number of parameters, and its accuracy is unsatisfactory. We remove the modules designed for large targets and reduced the number of detection heads from three to two. Moreover, we add an improved feature pyramid network to the neck, resulting in improved parameter count and accuracy over traditional YOLO methods. To improve inferencing, we replaced the spatial pyramid pooling (SPP) module with a simplified SPP-fast type. Finally, a large separable kernel attention and wise intersection-over-union method are introduced to integrate the attention mechanisms, and we replace the loss function to improve the discriminative capabilities of the model. Experimental results on the Global Wheat Head Dataset demonstrates a 53% reduction in memory usage, a 27% decrease in computational load, and a 5.2 frames per second increase in detection speed over extant methods. The proposed model also achieves 3.9, 2.1, and 1.3% improvements in terms of precision, recall, and mean average precision, respectively, even with its light weight and portability.
... Finally, it is widely recognized that the bottleneck of applying AI techniques to plant research is high-quality training datasets, making ML/DL models difficult to be trained and verified effectively [10,23]. Unlike the openly available Global Wheat Head Detection (GWHD) dataset [28] where a large and diverse dataset of labeled RGB images of wheat spikes has been made available to public, an open rice panicle detection dataset still does not exist for researchers to develop and benchmark their rice panicle detection models. Furthermore, much research still has not made training data and detailed method implementation fully available [29], which is key to facilitate plant researchers and breeders to accelerate ongoing research to catch up with the pace of global climate changes. ...
Rice ( Oryza sativa ) is an essential stable food for many rice consumption nations in the world and, thus, the importance to improve its yield production under global climate changes. To evaluate different rice varieties’ yield performance, key yield-related traits such as panicle number per unit area (PNpM ² ) are key indicators, which have attracted much attention by many plant research groups. Nevertheless, it is still challenging to conduct large-scale screening of rice panicles to quantify the PNpM ² trait due to complex field conditions, a large variation of rice cultivars, and their panicle morphological features. Here, we present Panicle-Cloud, an open and artificial intelligence (AI)-powered cloud computing platform that is capable of quantifying rice panicles from drone-collected imagery. To facilitate the development of AI-powered detection models, we first established an open diverse rice panicle detection dataset that was annotated by a group of rice specialists; then, we integrated several state-of-the-art deep learning models (including a preferred model called Panicle-AI) into the Panicle-Cloud platform, so that nonexpert users could select a pretrained model to detect rice panicles from their own aerial images. We trialed the AI models with images collected at different attitudes and growth stages, through which the right timing and preferred image resolutions for phenotyping rice panicles in the field were identified. Then, we applied the platform in a 2-season rice breeding trial to valid its biological relevance and classified yield production using the platform-derived PNpM ² trait from hundreds of rice varieties. Through correlation analysis between computational analysis and manual scoring, we found that the platform could quantify the PNpM ² trait reliably, based on which yield production was classified with high accuracy. Hence, we trust that our work demonstrates a valuable advance in phenotyping the PNpM ² trait in rice, which provides a useful toolkit to enable rice breeders to screen and select desired rice varieties under field conditions.
... Thanks to the rapid development of image acquisition equipment and artificial intelligence algorithms, counting the harvest organ based on object detection models has been proved to be a promising artificial alternative, which has been applied to various field objects such as soybeans [1], wheat ears [2][3][4], rice panicles [5,6], fruits [7][8][9][10][11][12], etc. For example, gulzar et al. [11,12] carry out a series of work focus on the fruits classification based on deep learning. ...
Counting the soybean pods automatically has been one of the key ways to realize intelligent soybean breeding in modern smart agriculture. However, the pod counting accuracy for whole soybean plants is still limited due to the crowding and uneven distribution of pods. In this paper, based on the VFNet detector, we propose a deformable attention recursive feature pyramid network for soybean pod counting (DARFP-SD), which aims to identify the number of soybean pods accurately. Specifically, to improve the feature quality, DARFP-SD first introduces the deformable convolutional networks (DCN) and attention recursive feature pyramid (ARFP) to reduce noise interference during feature learning. DARFP-SD further combines the Repulsion Loss to correct the error of predicted bboxse coming from the mutual interference between dense pods. DARFP-SD also designs a density prediction branch in the post-processing stage, which learns an adaptive soft distance IoU to assign suitable NMS threshold for different counting scenes with uneven soybean pod distributions. The model is trained on a dense soybean dataset with more than 5300 pods from three different shapes and two classes, which consists of a training set of 138 images, a validation set of 46 images and a test set of 46 images. Extensive experiments have verified the performance of proposed DARFP-SD. The final training loss is 1.281, and an average accuracy of 90.35%, an average recall of 85.59% and a F1 score of 87.90% can be achieved, outperforming the baseline method VFNet by 8.36%, 4.55% and 7.81%, respectively. We also validate the application effect for different numbers of soybean pods and differnt shapes of soybean. All the results show the effectiveness of the DARFP-SD, which can provide a new insight into the soybean pod counting task.
... Velumani et al. [35] adopted a FasterRCNN detection model to estimate the maize plant density. Liu et al. [36] applied the dynamic color transform networks for the wheat head detection based on YOLOV4. In general, image processing algorithm based on deep learning has been proven to be effective in the field of agriculture and also provided a feasible and powerful method for this study. ...
Verticillium wilt is one of the most critical cotton diseases, which is widely distributed in cotton-producing countries. However, the conventional method of verticillium wilt investigation is still manual, which has the disadvantages of subjectivity and low efficiency. In this research, an intelligent vision-based system was proposed to dynamically observe cotton verticillium wilt with high accuracy and high throughput. Firstly, a 3-coordinate motion platform was designed with the movement range 6,100 mm × 950 mm × 500 mm, and a specific control unit was adopted to achieve accurate movement and automatic imaging. Secondly, the verticillium wilt recognition was established based on 6 deep learning models, in which the VarifocalNet (VFNet) model had the best performance with a mean average precision ( mAP ) of 0.932. Meanwhile, deformable convolution, deformable region of interest pooling, and soft non-maximum suppression optimization methods were adopted to improve VFNet, and the mAP of the VFNet-Improved model improved by 1.8%. The precision–recall curves showed that VFNet-Improved was superior to VFNet for each category and had a better improvement effect on the ill leaf category than fine leaf. The regression results showed that the system measurement based on VFNet-Improved achieved high consistency with manual measurements. Finally, the user software was designed based on VFNet-Improved, and the dynamic observation results proved that this system was able to accurately investigate cotton verticillium wilt and quantify the prevalence rate of different resistant varieties. In conclusion, this study has demonstrated a novel intelligent system for the dynamic observation of cotton verticillium wilt on the seedbed, which provides a feasible and effective tool for cotton breeding and disease resistance research.
... It has a slow detection speed and a small density of fungal data. Representative single-stage networks are SSD [19], RetinaNet [20], and YOLO [21] series models. These networks treat object detection as a regression problem, with high real-time performance. ...
Rapid and accurate detection of pathogen spores is an important step to achieve early diagnosis of diseases in precision agriculture. Traditional detection methods are time-consuming, laborious, and subjective, and image processing methods mainly rely on manually designed features that are difficult to cope with pathogen spore detection in complex scenes. Therefore, an MG-YOLO detection algorithm ( M ulti-head self-attention and G host-optimized YOLO ) is proposed to detect gray mold spores rapidly. Firstly, Multi-head self-attention is introduced in the backbone to capture the global information of the pathogen spores. Secondly, we combine weighted Bidirectional Feature Pyramid Network (BiFPN) to fuse multiscale features of different layers. Then, a lightweight network is used to construct GhostCSP to optimize the neck part. Cucumber gray mold spores are used as the study object. The experimental results show that the improved MG-YOLO model achieves an accuracy of 0.983 for detecting gray mold spores and takes 0.009 s per image, which is significantly better than the state-of-the-art model. The visualization of the detection results shows that MG-YOLO effectively solves the detection of spores in blurred, small targets, multimorphology, and high-density scenes. Meanwhile, compared with the YOLOv5 model, the detection accuracy of the improved model is improved by 6.8%. It can meet the demand for high-precision detection of spores and provides a novel method to enhance the objectivity of pathogen spore detection.
