Computers and Electronics in Agriculture, Год журнала: 2024, Номер 227, С. 109508 - 109508
Опубликована: Окт. 1, 2024
Язык: Английский
Computers and Electronics in Agriculture, Год журнала: 2024, Номер 227, С. 109508 - 109508
Опубликована: Окт. 1, 2024
Язык: Английский
Animals, Год журнала: 2023, Номер 13(21), С. 3411 - 3411
Опубликована: Ноя. 3, 2023
Pig counting is an important work in the breeding process of large-scale pig farms. In order to achieve high-precision identification conditions pigs occluding each other, illumination difference, multiscenes, and differences number imaging size, also reduce parameters model, a algorithm improved YOLOv5n was proposed. Firstly, multiscene dataset created by selecting images from several different farms enhance generalization performance model; secondly, Backbone replaced FasterNet model calculations lay foundation for be applied Android system; thirdly, Neck optimized using E-GFPN structure feature fusion capability Finally, Focal EIoU loss function used replace CIoU improve model's accuracy. The results showed that AP 97.72%, parameters, amount calculation, size were reduced 50.57%, 32.20%, 47.21% compared with YOLOv5n, detection speed reached 75.87 f/s. has better accuracy robustness complex house environments, which not only ensured but as much possible. Meanwhile, application system developed based on truly realized practical technology. could easily extended field livestock poultry counting, such cattle, sheep, geese, etc., widely value.
Язык: Английский
Процитировано
6International journal of agricultural and biological engineering, Год журнала: 2023, Номер 16(6), С. 236 - 245
Опубликована: Янв. 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.
Язык: Английский
Процитировано
6Drones, Год журнала: 2023, Номер 7(9), С. 542 - 542
Опубликована: Авг. 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.
Язык: Английский
Процитировано
5The Visual Computer, Год журнала: 2023, Номер 40(5), С. 3825 - 3842
Опубликована: Сен. 5, 2023
Язык: Английский
Процитировано
4Computers and Electronics in Agriculture, Год журнала: 2024, Номер 227, С. 109508 - 109508
Опубликована: Окт. 1, 2024
Язык: Английский
Процитировано
1