A method for identifying abnormal behavior of forest musk deer

By using an improved YOLOv8-Pose model and pressure distribution sensor array, the problems of background interference and limb joint localization in the detection of abnormal behavior of musk deer were solved, realizing accurate identification and non-invasive monitoring of abnormal behavior of musk deer, and improving detection accuracy and ease of operation.

CN122454640APending Publication Date: 2026-07-24CHENGDU UNIV OF INFORMATION TECH
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
CHENGDU UNIV OF INFORMATION TECH
Filing Date
2026-05-13
Publication Date
2026-07-24

AI Technical Summary

Technical Problem

Traditional models for detecting abnormal behavior in musk deer are susceptible to background interference, resulting in large detection biases and poor accuracy in limb and joint positioning, making it difficult to support refined behavioral analysis. Furthermore, traditional bioelectric monitoring methods cause stress to musk deer and are complex to operate.

Method used

An improved YOLOv8-Pose model was adopted, embedding a lightweight CBAM attention module and a CoordConv layer, combined with a LiteACmix module and a FocalLoss loss function. The musk deer's scent expulsion behavior was monitored through a pressure distribution sensor array. The sensor was discarded, and the behavior was distinguished by the pressure center of gravity and the pressure fluctuation characteristics of the anal region.

Benefits of technology

It enables accurate identification and non-invasive monitoring of abnormal behaviors in musk deer, improves the accuracy of limb joint positioning, reduces stress on musk deer, lowers operational complexity, and meets the needs of real-time monitoring.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122454640A_ABST
    Figure CN122454640A_ABST
Patent Text Reader

Abstract

The application discloses a kind of abnormal behavior identification methods of forest musk deer, it is related to animal behavior monitoring technical field, method includes: the core key point of historical forest musk deer image is labeled;Key point detection model is obtained based on historical forest musk deer image, key point detection model is identified to obtain target data to the forest musk deer image to be detected, and abnormal behavior is obtained based on target data;Key point detection model includes successively connected backbone module, neck module and head module, backbone module includes successively connected convolution module and first module, and convolution module includes several CoordConv layers, and CoordConv layer is used to carry out convolution operation and extract feature;Neck module includes successively connected second module, CBAM module and third module, and second module and third module are used to enhance and fuse feature, and CBAM module is used to focus key feature, solve the problem that the detection model of traditional forest musk deer abnormal behavior is susceptible to background interference and leads to large detection deviation, and limb joint positioning precision is poor.
Need to check novelty before this filing date? Find Prior Art