Search

Copyright of Journal of Henan Agricultural Sciences is the property of Editorial Board of Journal of Henan Agricultural Sciences and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

The rapid development of deep learning has promoted the research progress in the field of visual object detection. In metal surface irregular small defect object detection, the traditional metal surface defect detection methods were not effective in detecting small and irregular defects. Therefore, how to improve the detection accuracy of small irregular defects on metal surfaces is a hot and difficult research problem. In this study, we propose a deep learning-based method for detecting small irregular defects on metal surfaces. Firstly, we make our own infrared laser diode metal base surface defect dataset to fill the gap of the dataset. Secondly, a new shallow feature extraction layer based on You Only Look Once (YOLO) v5s (an improved YOLO model) is designed for detecting irregular small object defects in the metal surface defect dataset. Thirdly, an attention mechanism was introduced into the network model to enhance the network's ability to extract features of small target defects. The improved model was trained and evaluated on the laser diode metal base surface defect detection dataset, and the results showed that the accuracy of the improved algorithm improved by 3.8% over the original detection algorithm. The detection accuracy also achieves very significant results when compared with other excellent object detection algorithms. [ABSTRACT FROM AUTHOR]

To improve the detection ability for small defects on the surface of the metal base of an infrared laser sensor, master the fluctuation and distribution of product quality, and form closed-loop control of production and quality improvement, the advanced You Only Look Once (YOLO) v5s (an improved YOLO model) object detection algorithm was further improved in this study. Specifically, the same-scale feature fusion part that is easily ignored in the structure was strengthened to enhance the network detection performance. In deep learning object detection, the propagation of information between features in the neural network is important. It was often represented by the pyramid features in the neck part of the model to enhance feature fusion. First, this study proposed a cross-convolution feature strengthening connection method combining the backbone and neck, which shortened the path of information propagation and improved the semantic information between feature pyramids. Then, the concat module of the original network was improved, and a new enhanced feature concat module was proposed to enhance the fusion of features at the same scale. The attention modules implemented by combining the convolutional block attention module were integrated into the concat module to enable the network to learn the weights of each channel independently, enhance the information transmission between features, and improve the detection performance of deep learning small objects. Lastly, the K-means + + algorithm was used to optimize the self-made Metal Base dataset of Infrared Laser Sensor (ILS-MB) and generate a new anchor box suitable for small objects in this dataset to improve the matching degree of target objects. With a small increase in computational cost, the improved YOLO v5s algorithm enhanced the accuracy by 3.8% on the ILS-MB dataset and achieved a very significant effect compared with other state-of-the-art detection methods. [ABSTRACT FROM AUTHOR]

To improve the precision of intelligent sugarcane harvesting, and to meet the requirements of high precision and low complexity for use with embedded devices, a lightweight model called MobileNet v2-YOLO v5s for the real-time detection of sugarcane stem nodes in complex natural environments was developed. In this study, images of sugarcane stem nodes in a complex natural environment were collected and a dataset containing 12,600 images was constructed using a data extension process. The MobileNet network was introduced to replace the backbone of the YOLO v5s algorithm and the improved algorithm was used to train the MobileNet-YOLO v5s sugarcane stem node identification model. In experiments aiming to verify the advantages of the lightweight model, MobileNet v2-YOLO v5s achieved the best combination of high precision and low complexity. Its AP was decreased by only 0.8%, while its complexity was reduced by 40% compared to YOLO v5s. It also had a fast detection speed of 4.4 ms on a Dell workstation P7920. Therefore, 11 other models were selected for comparative experiments to demonstrate the superiority of MobileNet v2-YOLO v5s. Finally, TensorRT accelerated optimization tests, execution tests, and real-time detection tests were performed on Jetson Nano. The results showed that the optimised MobileNet v2-YOLO v5s outperformed YOLO v5s in terms of identification, lightweight and detection speed on embedded devices. Overall, MobileNet v2-YOLO v5s model meets the requirements of embedded devices and can provide a visual identification method for intelligent sugarcane harvesting.

Millimeter wave radars in frequency-modulated continuous wave (FMCW) systems are widely used in the field of noncontact life signal detection; however, large errors still persist when determining the distance dimension of the target to be measured with the radar echo signal. The processing of the signals in the target environment is blind. We propose a method of using binocular vision to lock the distance dimension of the radar life signal and to determine the target distance by using the principle of the binocular camera parallax method, as this reduces the influence of the noise in the environment when determining the distance dimension of the target to be measured. First, the Yolo (you only look once: unified, real-time object detection) v5s neural network is used to call the binocular camera to detect the human body, where the resolution of the single lens is 1280 × 1200, and the DeepSORT (deep simple online real-time tracking) algorithm is used to extract the features of the target and track and register them. Additionally, the binocular vision parallax ranging method is used to detect the depth information of the target, search for the depth information in the range-dimensional FFT (frequency Fourier transform) spectrum of the radar echo signal, and take the spectral peak with the largest energy within the search range to determine it as the target. Then, the target is measured, the range gate of the target is determined, and the life signal is then separated through operations such as phase information extraction, unwrapping, and filtering. The test results showed that this method can be used to directionally separate and register corresponding life signals in a multiliving environment. By conducting an analysis using the Pearson correlation coefficient, we found that the correlation between the breathing frequency collected using this method and a breathing sensor reached 84.9%, and the correlation between the heartbeat frequency and smart bracelet results reached 93.6%. The target range gate was locked to separate and match the life signal. [ABSTRACT FROM AUTHOR]

Fruit detection is a popular key technology in agricultural automatic picking. Aiming at the problems of perishable grapes at maturity, different ripening conditions, complex background of grape orchards and changeable lighting conditions, a lightweight improved detection and recognition method based on YOLO v5s algorithm was proposed. Firstly, Efficientnet-v2 network was used as the backbone of feature extraction and integrated the no dimensionality reduction local cross-channel interaction mechanism, which greatly reduced the size of the model and the amount of parameters on the premise of ensuring the accuracy, and speeded up the model reasoning speed. Secondly, in order to further compensate for the accuracy loss caused by model simplification, coordinate attention mechanism was introduced at the key position of model feature fusion to strengthen the attention to the target, improved the model􀆳s ability to detect dense targets and resist complex background interference, and ensured the comprehensive performance and reliability of the algorithm. The experimental results show that the average accuracy of the improved algorithm is 98. 7%, the average detection speed is 0. 028 s, and the model size is only 12. 01 MB. Compared with the improved algorithm, the accuracy is increased by 0. 41%, the detection speed is increased by 22%, and the model is reduced by 13. 2%. In the orchard scene image detection test, the proposed algorithm can well detect the grape fruit and identify its status, and has strong adaptability to different environmental impacts, which provides a reference for the development of automatic picking technology. [ABSTRACT FROM AUTHOR]

Intelligent detection and localization of mature citrus fruits is a critical challenge in developing an automatic harvesting robot. Variable illumination conditions and different occlusion states are some of the essential issues that must be addressed for the accurate detection and localization of citrus in the orchard environment. In this paper, a novel method for the detection and localization of mature citrus using improved You Only Look Once (YOLO) v5s with binocular vision is proposed. First, a new loss function (polarity binary cross-entropy with logit loss) for YOLO v5s is designed to calculate the loss value of class probability and objectness score, so that a large penalty for false and missing detection is applied during the training process. Second, to recover the missing depth information caused by randomly overlapping background participants, Cr-Cb chromatic mapping, the Otsu thresholding algorithm, and morphological processing are successively used to extract the complete shape of the citrus, and the kriging method is applied to obtain the best linear unbiased estimator for the missing depth value. Finally, the citrus spatial position and posture information are obtained according to the camera imaging model and the geometric features of the citrus. The experimental results show that the recall rates of citrus detection under nonuniform illumination conditions, weak illumination, and well illumination are 99.55%, 98.47%, and 98.48%, respectively, approximately 2–9% higher than those of the original YOLO v5s network. The average error of the distance between the citrus fruit and the camera is 3.98 mm, and the average errors of the citrus diameters in the 3D direction are less than 2.75 mm. The average detection time per frame is 78.96 ms. The results indicate that our method can detect and localize citrus fruits in the complex environment of orchards with high accuracy and speed. Our dataset and codes are available at https: //github.com/AshesBen/citrus-detection-localization. [ABSTRACT FROM AUTHOR]

Millimeter wave radars in frequency-modulated continuous wave (FMCW) systems are widely used in the field of noncontact life signal detection; however, large errors still persist when determining the distance dimension of the target to be measured with the radar echo signal. The processing of the signals in the target environment is blind. We propose a method of using binocular vision to lock the distance dimension of the radar life signal and to determine the target distance by using the principle of the binocular camera parallax method, as this reduces the influence of the noise in the environment when determining the distance dimension of the target to be measured. First, the Yolo (you only look once: unified, real-time object detection) v5s neural network is used to call the binocular camera to detect the human body, where the resolution of the single lens is 1280 × 1200, and the DeepSORT (deep simple online real-time tracking) algorithm is used to extract the features of the target and track and register them. Additionally, the binocular vision parallax ranging method is used to detect the depth information of the target, search for the depth information in the range-dimensional FFT (frequency Fourier transform) spectrum of the radar echo signal, and take the spectral peak with the largest energy within the search range to determine it as the target. Then, the target is measured, the range gate of the target is determined, and the life signal is then separated through operations such as phase information extraction, unwrapping, and filtering. The test results showed that this method can be used to directionally separate and register corresponding life signals in a multiliving environment. By conducting an analysis using the Pearson correlation coefficient, we found that the correlation between the breathing frequency collected using this method and a breathing sensor reached 84.9%, and the correlation between the heartbeat frequency and smart bracelet results reached 93.6%. The target range gate was locked to separate and match the life signal.

The rapid and accurate detection of apple fruitlets before fruit thinning is important for the realization of early yield estimation and automatic fruit thinning. However, factors such as a complex growth environment, uncertain illumination, and the clustering and occlusion of apple fruitlets, especially the extreme similarities between fruitlets and backgrounds, make it difficult to effectively detect apple fruitlets before thinning. The overall goal of this study was to develop an accurate apple fruitlet detection method with small model size based on a channel pruned YOLO V5s deep learning algorithm. First, using transfer learning, a YOLO V5s detection model was built to detect apple fruitlets. To simplify the detection model and ensure the detection efficiency, a channel pruning algorithm was used to prune the YOLO V5s model. The pruned model was then fine-tuned to achieve rapid and accurate detection of apple fruitlets. The experimental results showed that the channel pruned YOLO V5s model provided an effective method to detect apple fruitlets under different conditions. A recall , precision , F 1 score, and false detection rate of 87.6%, 95.8%, 91.5% and 4.2%, respectively, were achieved; the average detection time was 8 ms per image; and the model size was only 1.4 MB. The performance of our method outperformed seven methods in comparison, indicating that our method simplified the model effectively on the premise of ensuring the detection accuracy. Our method provides a reference for the development of portable mobile fruit thinning terminals, and it can be used to help growers optimise their orchard management. • A method for rapid and accurate apple fruitlet detection before fruit thinning is proposed. • The method is based on a state-of-the-art deep learning approach. • The model size of our method is only 1.4 MB and the F 1 score is 91.5%. • The proposed method effectively detected apple fruitlets under a variety of different conditions. [ABSTRACT FROM AUTHOR]

Pig behavior recognition and tracking in group-housed livestock are effective aids for health and welfare monitoring in commercial settings. However, due to demanding farm conditions, the targets in the pig videos are heavily occluded and overlapped, and there are illumination changes, which cause error switches of pig identify (ID) in the tracking process and decrease the tracking quality. To solve these problems, this study proposed an improved DeepSORT algorithm for object tracking, which contained three processes. Firstly, two detectors, YOLOX-S and YOLO v5s, were developed to detect pig targets and classify four types of pig behaviors including lying, eating, standing, and other. Then, the improved DeepSORT was developed for pig behavior tracking and reducing error changes of pig ID by improving trajectory processing and data association. Finally, we established the public dataset annotation of group-housed pigs, with 3600 images in a total from 12 videos, which were suitable for pig tracking applications. The advantage of our method includes two aspects. One is that the trajectory processing and data association are improved by aiming at pig-specific scenarios, which are indoor scenes, and the number of pig target objects is stable. This improvement reduces the error switches of pig ID and enhances the stability of the tracking. The other is that the behavior classification information from the detectors is introduced into the tracking algorithm for behavior tracking. In the experiments of pig detection and behavior recognition, the YOLO v5s and YOLOX-S detectors achieved a high precision rate of 99.4% and 98.43%, a recall rate of 99% and 99.23, and a mean average precision (mAP) rate of 99.50% and 99.23%, respectively, with an AP.5:.95 of 89.3% and 87%. In the experiments of pig behavior tracking, the improved DeepSORT algorithm based on YOLOX-S obtained multi-object tracking accuracy (MOTA), ID switches (IDs), and IDF1 of 98.6%,15, and 95.7%, respectively. Compared with DeepSORT, it improved by 1.8% and 6.8% in MOTA and IDF1, respectively, and IDs had a significant decrease, with a decline of 80%. These experiments demonstrate that the improved DeepSORT can achieve pig behavior tracking with stable ID values under commercial conditions and provide scalable technical support for contactless automated pig monitoring.

[Display omitted] • First method to recognize the medicinal parts of herbs using hyperspectral technology. • Recognition of rhizome of red ginseng based on spectrum and image information. • Using image enhancement techniques to expand the number of samples. • Both PLS-DA and YOLOv5 played well in recognizing the different parts of red ginseng. • This work provided a theoretical basis for the intelligent recognition of herbs. Red ginseng is a widely used and extensively researched food and medicinal product with high nutritional value, derived from steamed fresh ginseng. The components in various parts of red ginseng differ significantly, resulting in distinct pharmacological activities and efficacies. This study proposed to establish a hyperspectral imaging technology combined with intelligent algorithms for the recognition of different parts of red ginseng based on the dual-scale of spectrum and image information. Firstly, the spectral information was processed by the best combination of first derivative as pre-processing method and partial least squares discriminant analysis (PLS-DA) as classification model. The recognition accuracy of the rhizome and the main root of red ginseng is 96.79% and 95.94% respectively. Then, the image information was processed by the You Only Look Once version 5 small (YOLO v5s) model. The best parameter combination is epoch = 30, learning rate = 0.01, and activation function is leaky ReLU. In the red ginseng dataset, the highest accuracy, recall and mean Average Precision at IoU (Intersection over Union) threshold 0.5 (mAP@0.5) were 99.01%, 98.51% and 99.07% respectively. The application of spectrum-image dual-scale digital information combined with intelligent algorithms in the recognition of red ginseng is successful, which provides a positive significance for the online and on-site quality control and authenticity identification of crude drugs or fruits. [ABSTRACT FROM AUTHOR]

Intelligent detection and localization of mature citrus fruits is a critical challenge in developing an automatic harvesting robot. Variable illumination conditions and different occlusion states are some of the essential issues that must be addressed for the accurate detection and localization of citrus in the orchard environment. In this paper, a novel method for the detection and localization of mature citrus using improved You Only Look Once (YOLO) v5s with binocular vision is proposed. First, a new loss function (polarity binary cross-entropy with logit loss) for YOLO v5s is designed to calculate the loss value of class probability and objectness score, so that a large penalty for false and missing detection is applied during the training process. Second, to recover the missing depth information caused by randomly overlapping background participants, Cr-Cb chromatic mapping, the Otsu thresholding algorithm, and morphological processing are successively used to extract the complete shape of the citrus, and the kriging method is applied to obtain the best linear unbiased estimator for the missing depth value. Finally, the citrus spatial position and posture information are obtained according to the camera imaging model and the geometric features of the citrus. The experimental results show that the recall rates of citrus detection under non-uniform illumination conditions, weak illumination, and well illumination are 99.55%, 98.47%, and 98.48%, respectively, approximately 2-9% higher than those of the original YOLO v5s network. The average error of the distance between the citrus fruit and the camera is 3.98 mm, and the average errors of the citrus diameters in the 3D direction are less than 2.75 mm. The average detection time per frame is 78.96 ms. The results indicate that our method can detect and localize citrus fruits in the complex environment of orchards with high accuracy and speed. Our dataset and codes are available at https://github.com/AshesBen/citrus-detection-localization., Competing Interests: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest., (Copyright © 2022 Hou, Zhang, Tang, Zhuang, Tan, Huang, Chen, Wei, He and Luo.)