基于对称像素块识别的病猪行为监测系统

Sick Pig Behavior Monitor System Based on Symmetrical Pixel Block Recognization

针对传统人工观察方法不能及时发现群养猪患病的缺点,设计一种病猪行为自动监测系统。该系统基于高级RISC微处理器平台,对群养猪的排泄行为进行24h监控,采用改进的运动目标检测算法和基于像素块对称特征的图像识别算法定位具有异常行为的疑似病猪,将报警图像通过GPRS网络传送至监控中心。实验结果表明,该系统具有较好的实时性和监测效果,通用性强。

This paper designs an automatic detection system aiming at the disadvantage that traditional manpower observation method can not f＇md pigs sick in time. The system monitors the pigs excretion behavior by 24 h based on advanced RISC Microprocessor platform. It orients suspected sick abnormal pigs by using an improved moving object detection algorithm and image recognization algorithm based on symmetrical characteristic of pixel block, the alarming image is send to the monitor center through GPRS. Experimental results show that the system has well real-time and monitor effect, and its commonality is good.