中文摘要：

目前对饲养于高阶直立式鸡笼内蛋鸡的监控方式主要为人工监控,作业环境恶劣,且对鸡笼内的病、死鸡不能实现实时监控。将机器视觉识别技术应用于高阶直立式鸡笼内蛋鸡的自动化监视过程,通过对笼内蛋鸡进食时的腿部图像进行图像处理,建立以特征区域面积值作为判别鸡笼内是否存在病、死鸡的标准。利用虚拟仪器软件LabWindows/CVI结合RS-485总线,开发了远程、实时及自动判别鸡笼内是否存在病、死鸡的系统。

英文摘要：

At present, artificial monitoring is a main method for monitoring layers in multi-tier vertical cage. Workers can＇t find ill and dead layers in cages in time, and their working environment is worse. With the aid of machine vision recognition technology, firstly raw image of layers＇ leg in the course of tak- ing food was captured and processed by image processing, the value of feature area from image processed was used to act as judgment standard. RS-485 bus and LabWindows/CVI software were used to build a remote, online and automotive supervision system for automatic judgment if ill and dead layers were in cages.