Research on wheat leaf water content based on machine vision

This paper is based on Matlab software to predict the water content of wheat leaves. The object of study are 100 wheat leaves which collected in the field, the moisture content of the blade was measured by drying, preprocess the image with Matlab so as to denoise the image, segmentation of blade images by image two valued operation of Otsu method then, image features are extracted. By correlation analysis, the H feature and the area of the shape feature of the color feature which are related to the water content are extracted, Through correlation analysis, we extracted the five components of the color feature, which are related to the water content, the area of the shape feature, the average value of the texture features, consistency and entropy, and so on, and the H features are extracted, it can reduce the influence of single parameter on decision and improve the precision of comprehensive decision. Finally, the BP neural network was used to train 80 samples and meet the requirements, and then to predict the 20 new samples. The results show that the prediction accuracy can reach above 96%.