Research On Fattening Pig Growth Information Monitoring Based On Machine Vision

Body size and weight of fattening pigs are important indicators for monitoring their growth and health.Timely and accurate collection of body size and weight information is very important for the scientific breeding of fattening pigs.At present,the body size of fattening pigs is mainly measured manually using tools such as tape measures.The body weight is mainly weighed by using weighbridge after driving.This method not only brings additional stress and stimulation to pigs,but also causes more large measurement error.Therefore,it is necessary to use non-contact methods to monitor pig body size and weight data.This paper analyzes the current domestic and foreign research data on Machine Vision and body weight monitoring of pigs,and develops a software and hardware system for fattening pig growth information monitoring in view of the existing problems.The system can automatically collect the top view,side view,weight information and ear tag number of fattening pigs.For the collected images of fattening pigs,pre-processing analysis is first performed.The method of histogram equalization and median filtering is adopted to reduce the influence of lighting and noise in the shooting environment.Then research and analyze the image segmentation and edge detection methods such as basic global threshold segmentation method,Otsu segmentation method,iterative method,Sobel edge detection method,Canny edge detection method,LOG edge detection method and morphological edge detection method,and propose a suitable monitoring system.Better image segmentation and edge detection algorithm.It provides relevant theoretical basis for accurately detecting the morphological characteristics of fattening pigs.Using MATLAB to compile a fattening pig growth information detection program,a fattening pig body size detection system based on multi-feature point matching is realized.Use binocular vision technology to extract the three-dimensional body size information of fattening pigs.Based on SPSS and MATLAB,a multiple statistical regression method,a MLP neural network,a RBF neural network and a feed forward neural network were used to study the relationship between body size and weight parameters,and an estimation model was established.By comparing the fitting correlation and average relative error of different methods,an optimal estimation model based on feed forward neural network is established.