Research On Wheat Leaf Water Content Based On Machine Vision

With the vigorous development of machine vision technology,many experts are more inclined to application of machine vision technology to diagnosis information research.In this paper,using machine vision technology,combined with the local wheat moisture,the moisture content of wheat leaf prediction research,the results show that the method is used to forecast leaf research is practical and reliable,provide important basis for subsequent research.In this paper,using machine vision technology to extract leaf characteristics correlation with water content from the 100 images,realized the feature extraction of condition and rapidness,water content of evaluation model is set up,using BP neural network,with this model can predict leaf water content.In order to improve the accuracy of the prediction system,first of all,the choice of algorithm for theoretical study of image preprocessing,image segmentation and feature extraction of image.In the process of the image preprocessing,image enhancement part chose median filtering algorithm,and then directly enhance contrast,to achieve the effect of target edge is clear.In the part of image segmentation,due to the grayscale difference is obvious,so choose the image binarization operation of the Ostu between-cluster variance method.Then use Matlab program to feature extraction of image,the evaluation index of the present study was to leaf image color,texture and shape,reduce the influence of a single parameter to determine,improve the accuracy of the comprehensive judgement.Finally establish the BP neural network,the network training and output simulation,and forecast the network authentication,the predicted results of the accuracy reached 96%,reaches the expected goal.The results show that,wheat leaf water content detection based on machine vision is feasible,and can be applied to leaf water content in practical forecast.