Research on Plant State Detection Based on Image Processing Technology

With the rapid development of plant science and technology, computer vision and intelligent analysis become more and more important on the study of plant growth state detection. In the field of intelligent agriculture, the main research of direction in China is the establishment of facility agriculture, and there is no more in-depth research on the application of image processing technology in it. With the example of Epipremnum aureum leaves, we use images to record the changes of the leaves which had been stopped watering for some days, and extract some parameters of leaves in different sate from three fields of texture, geometry and color in this paper. First, extract blade images by Closed-Form algorithm; second, analyze textural features by gray scale difference statistics, self-correlation function, gray scale level co-occurrence matrix and frequency domain analysis; and at last, analyze color features by moments of color images and color histograms. The experiment results show that the differences in parameters of texture, geometry, and color characteristic for plant’s different growth states do indeed exist and those parameter values have some relationship with the different plating state.