Target Detection and Recognition Method of Farming Machine Based on Machine Vision

摘要

In operation system of farming machine, the moving target detection is a key part of the monitoring system. Aimed at the shortcoming of poor detection effect of Gaussian background modeling, this paper proposes a detection method based on non-reference background model. The model utilizes the previous series of sampling values of current pixel to estimate the probability model of observed pixel; then based on the probability model to carry on binary detection of moving object. In terms of moving target recognition, this paper puts forward several characteristics, and trains identification through BP neural network. The experimental results show that the background model can effectively detect the moving target, and also achieve the satisfactory effect of recognition rate of farming machine based on machine vision.