CA-YOLOv5: Detection model for healthy and diseased silkworms in mixed conditions based on improved YOLOv5

International journal of agricultural and biological engineering, Journal Year: 2023, Volume and Issue: 16(6), P. 236 - 245

Published: Jan. 1, 2023

The accurate identification and localization of diseased silkworms is an important task in the research disease precision control technology equipment development sericulture industry. However, existing deep learning-based methods for this are mainly based on image classification, which fails to provide location information silkworms. To end, study proposed object detection-based method identifying locating healthy Images mixed were collected using a mobile phone, category each silkworm labeled LabelImg as labeling tool construct dataset detection. Based one-step detection model YOLOv5s, ConvNeXt-Attention-YOLOv5 (CA-YOLOv5) was designed large kernel with depth-wise separable convolution (7×7 dw-conv) ConvNeXt adopted expand receptive fields channel attention mechanism ECANet added enhance capability feature extraction. Experiments showed that mean average (mAP) values CA-YOLOv5 reached 96.46%, 1.35% better than achieved via YOLOv5s. At same time, overall performance significantly state-of-the-art models, such Single Shot MultiBox Detector (SSD), CenterNet, EfficientDet, even improved YOLOv5 lightweight backbone, like SENet-YOLOv5 MobileNet-YOLOv5. results can basis positioning development. Keywords: detection, YOLOv5; conditions, mechanism, DOI: 10.25165/j.ijabe.20231606.7854 Citation: Shi H K, Xiao W F, Zhu S P, Li L B, Zhang J F. CA-YOLOv5: Detection conditions YOLOv5. Int Agric & Biol Eng, 2023; 16(6): 236–245.

Language: Английский

---

An Intelligent Grazing Development Strategy for Unmanned Animal Husbandry in China

Drones, Journal Year: 2023, Volume and Issue: 7(9), P. 542 - 542

Published: Aug. 22, 2023

Grazing is the most important and lowest cost means of livestock breeding. Because sharp contradiction between grassland ecosystem livestock, has tended to degrade in past decades China; therefore, ecological balance been seriously damaged. The implementation grazing prohibition, rotational development a large-scale breeding industry have not only ensured supply animal husbandry products, but also promoted restoration ecosystem. For industry, welfare cannot be guaranteed due narrow crowded space, thus, production usually lower competitiveness than grazing. Disorderly leads crises; however, intelligent can ensure welfare, fully improve products. Under urbanization, workforce engaged pastoral areas gradually lost. Intelligent methods need developed popularized. This paper focuses on grazing, reviews grass remote sensing aerial seeding, wearable monitoring equipment UAV robots, summarizes elements, exploring new direction automatic management with robot at this stage.

Language: Английский

---

Using the MSFNet Model to Explore the Temporal and Spatial Evolution of Crop Planting Area and Increase Its Contribution to the Application of UAV Remote Sensing

Drones, Journal Year: 2024, Volume and Issue: 8(9), P. 432 - 432

Published: Aug. 26, 2024

Remote sensing technology can be used to monitor changes in crop planting areas guide agricultural production management and help achieve regional carbon neutrality. Agricultural UAV remote is efficient, accurate, flexible, which quickly collect transmit high-resolution data real time precision agriculture management. It widely monitoring, yield prediction, irrigation However, the application of faces challenges such as a high imbalance land cover types, scarcity labeled samples, complex changeable coverage types long-term images, have brought great limitations monitoring cultivated changes. In order solve abovementioned problems, this paper proposed multi-scale fusion network (MSFNet) model based on input feature series further combined MSFNet Model Diagnostic Meta Learning (MAML) methods, using particle swarm optimization (PSO) optimize parameters neural network. The method applied crops tomatoes. experimental results showed that average accuracy, F1-score, IoU optimized by PSO + MAML (PSML) were 94.902%, 91.901%, 90.557%, respectively. Compared with other schemes U-Net, PSPNet, DeepLabv3+, has better effect solving problem ground objects image samples provides technical support for subsequent technology. study found change different was closely related climatic conditions policies, helps use realization

Language: Английский

---

Egg production monitoring in commercial laying cages via the StrongSort-EGG tracking-by-detection model

Computers and Electronics in Agriculture, Journal Year: 2024, Volume and Issue: 227, P. 109508 - 109508

Published: Oct. 1, 2024

Language: Английский

---

Human-UAV interactive perception: Skeleton-based iterative perspective optimization algorithm for UAV patrol tracking of large-scale pedestrian abnormal behavior

Applied Soft Computing, Journal Year: 2024, Volume and Issue: 167, P. 112467 - 112467

Published: Nov. 16, 2024

Language: Английский

---