Real-time Monitoring System for Suspected Diseased Pigs Based on Abnormal Behavior Analysis
Considering the shortcomings of the method of manually observing abnormal behavior of pigs to identify suspected diseased pigs,a real-time monitoring system based on behavior analysis is designed to detect the suspected dis-eased pigs.By deploying a telephoto lens and a wide-angle camera at the top of the pigsty food room,as well as a sec-ondary sensor,the dietary behavior of live pigs can be monitored for 24 hours.By improving the motion target detection algorithm and image recognition technology,the non-contact monitoring of diseased pigs and suspected diseased pigs with abnormal behavior is achieved,and logs are generated and data is recorded.The warning module will transmit image in-formation to the monitoring center through a network server,which can effectively detect abnormal behavior of pigs to in-tervene and treat sick pigs earlier.After training the model recognition,a total of 50 suspected diseased pigs are found.After testing,the actual number of diseased pigs is 49,and the detection accuracy of the system is 99%,basically achie-ving the expected design goals.
NETL
NSTL
万方数据
