基于面积特征算子的猪呼吸频率检测.pdf

中国科技论文在线 基于面积特征算子的猪呼吸频率检测# 朱伟兴，唐 亮，李新城，杨磊* （江苏大学电气信息工程学院，江苏镇江市 212013 ） 5 摘要：为了检测与识别猪的异常呼吸并提供实时预警，本文提出采用机器视觉，利用面积特 征算子非接触地检测猪的呼吸频率的方法。通过相机在猪舍捕获猪站立时的视频，传输到计 算机，在 MATLAB 平台上提取猪脊腹部区域变化的面积特征；利用小波分析自适应和多分 辨率分析的特点，对提取的面积信号进行降噪处理，消除信号的毛刺（局部极值点）；运用 峰值点检测算法检测面积信号的极值点来获得监测时间段内猪的呼吸频次，最后转化成呼吸 10 频率。实验表明结果表明，本方法对猪呼吸的频率检测平均精度约为 92%。本文所提出的 方法对猪的图像轮廓不需要完整性和封闭性，只需要清晰提取出猪的脊腹部轮廓，因此可适 用于实际猪场中光线和环境不是太好的场合。此外，本文虽仅对猪保持站立姿势时进行了检 测，但该方法稍加改进也可适用于猪躺卧姿势的呼吸频率检测。 关键词：农业电气化与自动化；呼吸频率；面积特征算子；猪；机器视觉；小波分析 15 中图分类号：TP391.4 Detection of porcine respiratory frequency based on area feature operator ZHU Weixing, TANG Liang, LI Xincheng, YANG Lei 20 (School of Electrical and Information Engineering, Jiangsu University, Zhenjiang 212013, China) Abstract: In order to detect the abnormal breath of pigs and provide real-time warning, the machine vision technology was used to detect the porcine respiratory frequency based on the feature of area operator. Firstly using the camera, we capture the videos in which pig is standing statically in the piggery and then transmit them to the computer, next extract the changing area 25 signal of pig's abdomen on the MATLAB platform; Secondly the adaptive and multi-resolution characteristics of wavelet analysis were used to denoise the signal; Moreover, the peak point detection algorithm was designed to acquire swine’s respiratory frequency according to the changing area signal during the monitoring time. The experiment results