基于YOLO算法的鱼群探测监控系统

A fish detecting and monitoring system based on YOLO algorithm

针对水下鱼群实时检测与管理,基于YOLO的鱼群检测算法以及WCF技术构建了鱼群探测系统。系统分数据采集与控制、深度学习图像处理单元以及软件系统三个模块。首先,系统利用双目高清摄像头采集水下实时视频并进行视频传输;然后,利用YOLO算法对输入的视频图像进行鱼群检测;最后,利用Html与WCF技术对视频检测的结果进行展示与存储。在昆山市淀山湖水下环境中进行鱼群检测,可以实时检测鱼群数量并上传鱼群的截图,然后根据历史数据实现数据可视化显示,为研究该水域鱼群分布以及活动规律提供参考依据。

For the underwater fish real-time detection and management,we have constructed a fish detection system with the fish detection algorithm based on YOLO and WCF technology.The system consists of three modules:data acquisition and control,deep learning image processing unit and software system.Firstly,we used the binocular HD camera to collect real-time underwater video and transmit it to the system.Then,the real-time YOLO algorithm was used to detect the fish video images.Finally,the Html and WCF technology were used to display and store the results of detection.The fish detecting and monitoring system was carried out in Dianshan Lake of Kunshan.The number of fishes can be detected in the real time while the screenshot of fishes can be uploaded.The data visualization can be realized according to the historical data,which provides data for the investigation of the fish distribution and activity in this area.