Study On Methods And Applications Of Layer Chicken Information Monitoring Based On Visual Technology

In recent years,precision livestock husbandry with welfare,precision,and automation as a breeding goal has attracted more and more attention at domestic and foreign.This article adopts machine vision technology and IoT(Internet of Things)technology to build an intelligent monitoring system for poultry breeding,which can monitor the behavior of chickens and welfare status of the henhouse and realizes the automatic control of breeding equipment in the henhouse.Focusing on solving chicken body segmentation and multi-moving body tracking problems in real chicken house environment,the chicken body behavior identification method and the relationship between flock behavior and welfare status was studied.The application of the equipment not only increases the degree of precision and automation of the cultivation,but also ensures the welfare level in the breeding process.The main works and conclusions of this paper are as follows:(1)The intelligent monitoring system for poultry production has been designed.It adopts a multi-camera scheme for capturing images synchronously from a top view and take a scheme which is consisted of wired(RS485)and wireless(ZigBee)communication to achieve remote control.The selection of the camera and the hardware and software design of the coordinator and terminal control node in the ZigBee network have been completed.(2)A chicken body segmentation method based on the maximum inter class variance method in complex background and a method of individual tracking was proposed based on ellipse fitting and shortest distance matching are proposed.In order to solve the effect of individual adhesion on the motion tracking,an algorithm for the separation of adhesions based on convex defects was designed.Through experimental verification,the correct segmentation rate of the above image segmentation method is 0.9493,and the mean square error is 0.007 at different time periods.The experiment confirmed that the image acquisition frequency is 2 frames/s.Under this sampling frequency,the maximum tracking time of the tracking algorithm is 9S.(3)According to the position of the chicken body,four simple individual behaviors discrimination methods for feeding,drinking,moving,and rest are proposed;four group behavior indicators-distribution index,average speed,occupation index,and activity level-that can directly reflect the welfare status of the flock are proposed.By comparing the change of chicken behavior before and after feeding or water supply、under the different ambient temperature、different activity spaces and different health conditions,the method flow discriminating the overall welfare state was obtained through the index of group behavior.(4)Use the Microsoft Foundation Class Library MFC and computer vision library OpenCV to write poultry information monitoring system software.The application of the poultry information monitoring system monitors the welfare status of layers raised under floor raising mode and automatically controls the farming equipment to verify the feasibility and stability of the system.The test results show that the system can accurately determine the welfare status of the chickens.