The Color Model of Rice Leaf Based on SVM and BP Neural Network

Abstract

In order to construct a color model of rice leaf based on physiology and ecology, a modeling method based on SVM and BP neural network was proposed for the relationship between the chlorophyll, carotenoid of rice leaf and its RGB value. The chlorophyll a, chlorophyll b and carotenoid were used as model input parameters, the R, G and B values of the rice leaf image were used as the model output parameters respectively and the corresponding RGB component values of leaf image were predicted by using SVM and BP neural network. The results show that the prediction accuracy of BP neural network is significantly higher than that of SVM. The research can meet the needs of agriculture research and provide a theoretical basis for rice leaf color simulation modeling. It also provides a theoretical basis for the digitization and visualization of plant growth. The research method has good universality and generalization.