In recent years, frequent poultry diseases have seriously impacted the poultry industry and threatened human health, which has become the focus of the industry and the country. The traditional manual inspection method is time-consuming and easy to miss, and the poor breeding environment will damage the health of the staff. In view of the above problems, the team takes the laminated cage broiler as the research object, and plans to use embedded development technology and convolutional neural network technology to develop an automatic identification system for the health status of broilers, so as to solve the identification problem of sick chickens and provide intelligent monitoring scheme for farms. According to the above technology, the team plans to design and implement an inspection robot, which has automatic inspection, identifies sick chickens and sends out alarms through multiple information sources (such as chicken behavior state, feces color and shape), collects and visualizes sensor information at the robot end, sets up a control panel at the robot end to adjust the robot route and other parameters in real time, and monitors poultry farm information in real time through the website background. Comprehensive utilization of a variety of advanced technologies: This project comprehensively uses embedded development technology and convolutional neural network technology. Through the organic combination of these technologies, the comprehensive monitoring and identification of broiler health status is realized, thereby improving the accuracy of disease detection And efficiency. Comprehensive analysis of multiple information sources: Compared with the traditional single information source, the project adopts the method of comprehensive analysis of multiple information sources, including the behavior state of broilers, the color and shape of feces, etc., through comprehensive analysis of these information, can be more accurate Identification of sick chickens, improve the accuracy of diagnosis. Intelligent inspection and monitoring system: through the design and implementation of an intelligent inspection robot and website background monitoring system, the project realizes the automatic monitoring and management of poultry farms, which not only improves the timeliness and comprehensiveness of monitoring, but also reduces the cost and risk of manual inspection. Real-time early warning and response mechanism: through the automatic identification and alarm function of the robot end, the project can detect the epidemic disease in time, and realize real-time monitoring and response through the background system, which provides timely and effective early warning information for farm managers, which helps to take timely measures to curb the spread of the epidemic. Promote the intelligent improvement of the breeding industry: The successful research and development of this project will promote the improvement of the intelligent level of the poultry breeding industry, promote the transformation of the traditional breeding mode to intelligence and automation, and provide a safer and healthier production environment for the farms. Make greater contributions to poultry health and human health. In this study, computer vision technology was used to capture images by camera, identify the behavior state of chickens, feces color and other characteristics, and build a health state model combined with daily behavior. At the same time, using the Internet of Things and web visualization technology, combined with the environmental factor data monitored by the robot end sensor and the recognition results of the recognition module, the integration of monitoring the health status of broilers and the environment of the chicken farm is realized.