摘要
针对目前死淘鸡目标检测研究较少,高精度检测算法体积大难以部署至移动式设备等问题,提出一种基于YOLOv4的轻量化死淘鸡目标检测算法。采集大规模蛋鸡养殖工厂笼中死淘鸡图片,建立目标检测数据集;在算法中引入MobileNetv3主干提取网络与深度可分离卷积来降低模型体积;并在最大池化层前添加自注意力机制模块,增强算法对全局语义信息的捕获。在自建数据集中的试验结果表明,改进算法在死淘鸡目标检测任务中有更高的准确度,其mAP值与召回率分别达到97.74%和98.15%,模型大小缩小至原算法的1/5,在GPU加速下帧数达到77帧/s,检测速度提高1倍,能够满足嵌入式部署需求。
Aiming at the problems that there are few studies on dead chicken target detection and the large size of the high-precision detection algorithm makes it difficult to deploy to mobile devices,a lightweight dead chicken target detection algorithm based on YOLOv4 is proposed.Firstly,the team collects images of dead chickens in cages from large-scale egg production plants to build a target detection dataset.Then,MobileNetv3 backbone extraction network with depth-separable convolution is introduced in the algorithm to reduce the model size.Finally,a self-attentive mechanism module is added before the maximum pooling layer to enhance the algorithm s capture of global semantic information.Experimental results in a self-built dataset show that the improved algorithm has higher accuracy in the dead pheasant target detection task,with mAP values and recall rates of 97.74%and 98.15%respectively.The model size is reduced to 1/5 of the original algorithm,and the frame rate reaches 77 frames/s under GPU acceleration,doubling the detection speed and meeting the requirements of embedded deployments.
作者
漆海霞
李承杰
黄桂珍
Qi Haixia;Li Chengjie;Huang Guizhen(School of Engineering,South China Agricultural University,Guangzhou,510642,China;National Precision Agriculture International Joint Research Center for Aerial Pesticide Application Technology,Guangzhou,510642,China;Guangdong Provincial Laboratory of Lingnan Modern Agricultural Science and Technology,Guangzhou,510642,China)
出处
《中国农机化学报》
北大核心
2024年第5期195-201,共7页
Journal of Chinese Agricultural Mechanization
基金
广州市科技项目(20212100026)。
