Research On Obtaining The Information Of Growing Rapeseed Based On Computer Vision

Computer is widely used in agriculture, but it is a new field to use computer vision to discriminate leaf color and growth vigor. Based on two-factor of breed and fertilization experiment with randomized block design, the growing images, biochemical and biophysical indexes of the rapeseed leaves are collected. On the base of those collected information, the individual plant can be distinguished from the images via multi-media image processing. Besides, the measuring models can be established to measure the contents of chlorophyll, full nitrogen, soluble sugar and ratio of carbon to nitrogen.The major contents and conclusions are summarized as:1. The computer vision system of obtaining and analyzing images is built and used to research the methods of obtaining images.2. Some usual color models of digital image and their transfer are studied to identify the characters and application of them. The new color system is also modeled based on normalized RGB color system.3. The analyzing on histogram of 18 color characters and statistical color parameters of representative objects shows that 2G-R-B, 2g-r-b, G-R, g-r, red and normalized red which are the maximum of RGB or normalized RGB can be used to recognize image, so six methods are applied to split and recognize the images to obtain the experiential values of six methods and filter the best suitable method to the research.4. After splitting the images, the measuring models are established by regression analysis on the value of SPAD (the relative content of chlorophyll), contents of soluble sugar, contents of full nitrogen, ratio of carbon to nitrogen and 18 means of color characters.The results indicate that computer vision application in capturing the information of growth vigor is feasible. It is an accurate, convenient, clipping method to measure the feature parameters of growing crop. The model and theory is developed to monitor the growth of crop by using computer vision.