In order to develop an underwater sea cucumber collecting robot, it is necessary to use the machine vision method to realize sea cucumber recognition and location. An identification and location method of underwater sea cucumber based on improved You Only Look Once version 5 (YOLOv5) is proposed. Due to the low contrast between sea cucumbers and the underwater environment, the Multi-Scale Retinex with Color Restoration (MSRCR) algorithm was introduced to process the images to enhance the contrast. In order to improve the recognition precision and efficiency, the Convolutional Block Attention Module (CBAM) is added. In order to make small target recognition more precise, the Detect layer was added to the Head network of YOLOv5s. The improved YOLOv5s model and YOLOv5s, YOLOv4, and Faster-RCNN identified the same image set; the experimental results show improved YOLOv5 recognition precision level and confidence level, especially for small target recognition, which is excellent and better than other models. Compared to the other three models, the improved YOLOv5s has higher precision and detection time. Compared with the YOLOv5s, the precision and recall rate of the improved YOLOv5s model are improved by 9% and 11.5%, respectively.

The proposed method for tire specification character recognition based on the YOLOv5 network aimed to address the low efficiency and accuracy of the current character recognition methods. The approach involved making three major modifications to the YOLOv5 network to improve its generalization ability, computation speed, and optimization. The first modification involved changing the coupled head in YOLOv5 to a decoupled head, which could improve the network’s generalization ability. The second modification proposed incorporating the C3-Faster module, which would replace some of the C3 modules in YOLOv5’s backbone and head and improve the network’s computation speed. Finally, the third modification proposed replacing YOLOv5’s CIoU loss function with the WIoU loss function to optimize the network. Comparative experiments were conducted to validate the effectiveness of the proposed modifications. The C3-Faster module and the WIoU loss function were found to be effective, reducing the training time of the improved network and increasing the mAP by 3.7 percentage points in the ablation experiment. The experimental results demonstrated the effectiveness of the proposed method in improving the accuracy of tire specification character recognition and meeting practical application requirements. Overall, the proposed method showed promising results for improving the efficiency and accuracy of automotive tire specification character recognition, which has potential applications in various industries, including automotive manufacturing and tire production.

The Efficient Convolution Operators for Tracking (ECO) algorithm has garnered considerable attention in both academic research and practical applications due to its remarkable tracking efficacy, yielding exceptional accuracy and success rates in various challenging contexts. However, the ECO algorithm heavily relies on the deep learning Visual Geometry Group (VGG) network model, which entails complexity and substantial computational resources. Moreover, its performance tends to deteriorate in scenarios involving target occlusion, background clutter, and similar challenges. To tackle these issues, this study introduces a novel enhancement to the pedestrian tracking algorithm. Specifically, the VGG network is substituted with a lightweight MobileNet v2 model, thereby reducing computational demands. Additionally, a Double Attention Networks (A2-Net) module is incorporated to augment the extraction of crucial information, while pre-training techniques are integrated to expedite model convergence. Experimental results demonstrate that the C-ECO algorithm achieves comparable accuracy and success rates to the conventional ECO algorithm, despite reducing the model size by 27.96% and increasing the tracking frame rate by 46.11%. Notably, when compared to other prevalent tracking algorithms, the C-ECO algorithm exhibits an accuracy of 82.20% and a success rate of 64.72%. These findings underscore the enhanced adaptability of the C-ECO algorithm in complex environments, offering a more lightweight model while delivering superior tracking capabilities.

As an international standard pressure-measuring instrument, the absolute pressure piston manometer’s working medium is gas, so the actual working process will be affected by many internal uncertainties and external disturbances, leading to its long stability time and poor performance. In this paper, a fuzzy linear active disturbance rejection control strategy (FLADRC) for absolute pressure piston manometers is proposed to address these problems. First, the characteristics of the main components are analyzed according to the actual working principle of the system to establish a theoretical model of the controlled system. Second, the corresponding linear active disturbance rejection controller (LADRC) is designed according to the model. The principle of fuzzy control is introduced to adaptively adjust the controller parameters of the LADRC in real time, which improves the disadvantages of the LADRC parameters, which are difficult to rectify and have poor immunity to disturbances due to fixed parameters, and the stability of the control method is subsequently demonstrated. Finally, a simulation model is built in the Simulink environment in MATLAB, and three different pressure operating points are selected for the corresponding experiments to make a comparative analysis with Kp, PID, and LADRC. The results show that FLADRC enables the absolute pressure piston manometer to achieve better stability and greater immunity to disturbances. This also verifies the effectiveness and feasibility of the control strategy in practical engineering applications.

One of the essential techniques for using underwater robots to fish sea cucumbers is that the robots must track sea cucumbers using computer vision technology. Tracking underwater targets is a challenging task due to suspension, water absorption, and light scattering. This study proposed a simple but effective algorithm for sea cucumber tracking based on Kernelized Correlation Filters (KCF) framework. This method tracked the head and tail of the sea cucumber respectively and calculated the scale change according to the distance between the head and tail. The KCF method was improved on three strategies. First of all, the target was searched at the predicted position to improve accuracy. Secondly, an adaptive learning rate updating method based on the detection score of each frame was proposed. Finally, the adaptive size of the histogram of the oriented gradient (HOG) feature was used to balance the accuracy and efficiency. Experimental results showed that the algorithm had good tracking performance. Keywords: visual tracking, correlation filters, kernelized correlation filters, sea cucumber, scale estimation, underwater DOI: 10.25165/j.ijabe.20231603.4503 Citation: Wei H L, Kong X Z, Zhai X Y, Tong Q, Pang G B. Visual tracking for underwater sea cucumber via correlation filters. Int J Agric & Biol Eng, 2023; 16(3): 16(3): 247–253.

To solve the chip location recognition problem, this paper proposes a lightweight E-YOLOv5 based chip detection algorithm based on the You Only Look Once version 5 (YOLOv5s) algorithm. For the problem of the difficult distinction between chip detection points and light spots, a simulated exposure algorithm is used to process part of the training set images to enhance model robustness; the existing model network is complex, and EfficientNet, a lightweight feature extraction network, is introduced to reduce the model size; for the problem of imprecise model recognition due to small detection points, Selective Kernel Neural Network (SKNet) module is introduced into EfficientNet is introduced to enhance the feature extraction ability of the model and improve the training efficiency, and Efficient Intersection over Union Loss (EIoU_Loss) is used as the loss function to reduce the false recognition rate. Experiments show that the algorithm in this paper improves by 3.85% and 3.92% in precision, recall rate, 28.89% in loss value, nearly 20% in model size and training time, and 46.67% in image processing speed on CPU compared with YOLOv5s. The experimental results show that the proposed algorithm outperforms other algorithms and is able to distinguish and identify chip locations precisely and stably.