Method and device for detecting abnormal behaviors of live pigs based on unsupervised clustering

• Main Authors: BI YANZHEN, GAO JUNFENG, LI YUAN, SUN TIANDONG, LI JIAO, MO ZHE, KIYOTAKE, TIAN BINGCHUAN, LI JIANI, LONG XIAOBO, ZHOU XIAOHONG, ZHU HAO, ZHANG YU, WU YANG, SHANG GUANGYONG, XIAO SHENGCAI, MENG XING
• Format: Patent
• Language: Chinese; English
• Published: 23.07.2024
• Subjects: CALCULATING; COMPUTING; COUNTING; PHYSICS

The invention provides a live pig abnormal behavior detection method and device based on unsupervised clustering, and relates to the field of animal husbandry phenotype recognition detection, and the method comprises the steps: recognizing a live pig target in an image frame in a video stream through employing a target detection algorithm; using a target tracking algorithm to track the motion trail of each live pig target to obtain the trail information of each live pig target, and extracting to obtain trail data; according to the trajectory data, performing clustering analysis on the trajectory data by using a clustering algorithm to obtain an analysis result; according to an analysis result, judging whether the live pig has an abnormal behavior; and marking the abnormal behavior on an image frame in the original video stream to obtain a visual prediction result. The invention provides a live pig abnormal behavior detection method and device based on unsupervised clustering, and solves the problem that an ex