Study of monitoring maize leaf nutrition based on image processing and spectral analysis

The nutrition monitoring of maize leaf is studied with computer vision technology in this research. Planting schemes are implemented on the experimental farm. Images of the fresh leaves are grabbed and the nutrition content of leaf is acquired in the laboratory. And then, the relationship between the color features and the nutrition content is studied to find a quick and reliable method by which the maize nutrition can be estimated in a non-destructive way. The color components are extracted from the preprocessed images, include R, G, B components of the RGB model and H, I components of the HSI model. Based on the studying of the spectral characteristics of maize leaf, the absorption and reflection characteristics of chlorophyll at green and red band are analyzed. Furthermore, some indexes of color features, such as G/R, G/B, G-R, 2G-R, 2G-R-B are analyzed and compared. And regression analysis is made between chlorophyll content and the above color components and features. The G component, I component and 2G-R color features showed good correlation with leaf chlorophyll content of maize. Multi-regress analysis is made between color features and chlorophyll content on the basis of the correlation between leaf color features and components, and a related multi-linearregress model is established. It indicates that the image processing technology can be used to analyze and predict the nutrition of maize leaf.