Automated detection of sick pigs based on machine vision

An automated detection method for sick pig behavior was proposed after analyzing the disadvantages of traditional observation methods for abnormal behavior of pigs. An embedded detecting system based on computer vision and ARM (Advanced RISC (Reduced Instruction Set Computer) Machines) was designed to monitor the pig excretion behavior. The improved moving objecting detection method and symmetrical pixel block image identification algorithm is designed to recognize the suspected sick pigs. The relevant pictures are taken and sent to the surveillance center through GPRS (General Packet Radio Service) networks after finding the suspected sick pigs. The experiment results for 10 Yorkshire pigs showed that the detection accuracy is about 78.38%. The designed method and monitoring system will be helpful for improving production automation in modern pig farm.