在禽群病死个体检测中的应用机器视觉技术的探讨  

Machine Vision Technology in Application to Detecting Ill Individuals of Flocks

作　　者：魏长宝[1] 李平[1] 

机构地区：[1]黄淮学院信息工程学院,河南驻马店463000

出　　处：《电子器件》2015年第4期826-830,共5页Chinese Journal of Electron Devices

基　　金：国家自然科学基金青年基金项目(61304172)

摘　　要：目前的农场中,环境温度、风量、禽蛋、禽粪、食物和水的供给都已完成自动控制,但是对于农场中的死禽,并无达到自动监测和控制的目的,造成农场环境污染。针对上述死禽巡查检测中的各种缺陷,以养鸡场为研究对象,采用逻辑与操作排除活鸡存在的部分,提取出鸡的各个重要特征,以这些特征的变化量作为特征向量,判断是否为死鸡。通过实验结果可得,本文的图像识别方法能有效提升检测死鸡的精度。The current farm,ambient temperature,air flow,eggs,poultry manure,food and water supplies have been completely automatic controlled. However,the dead poultry in the farm has not automatic monitor and it can cause the environmental pollution. For the above mentioned various defects detected in dead poultry inspections to farm,we used the method of logic and operations to exclude the presence of part of live chickens,and by means of extracting the important features of each chicken. Based on the change of such features,It can be judged whether the chickens died or not. The experimental results can be obtained that the image recognition method can effectively enhance the detection accuracy of dead chickens.

关 键 词：机器视觉 自动检测 特征提取 检测精度 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]