Non-destructive Testing Of Machine Vision-based Crop Water And Judge Model

Based on machine vision technology, this paper extractes the surface features which can reveal the water demands of crop(marigold) healthy growth, realizes intactness in the progress and quickness of processing, models crop's water demanding system for healthy growth, provides reliable data support for correct irrigation method, achieves precise, quantitative irrigation, reduces resources, energy and labor consumption.In order to improve the accuracy of the system, the paper put forward preprocessing before processing image to eliminate irrelevant information in the image and recover the useful and real information, so as to enhance detectability of relevant information and simplify data ultimately. Butterworth filter that selected in the paper is not only without the "flap bell" phenomenon, but also can enhance the detail clarity of the image. It is a low-pass filter with the greatest flat range response, and has balanced advantages in linear phase, attenuation slope and loading characteristics. Therefore, in actual use, Butterworth filter has already been listed as the first choice, and used very widely.In the process of extracting eigenvalue of image, the color value and texture are extracted separately. The RGB color model, that means R, G,B is represented a pixel color, is used in the progress of RGB color extracted. RGB means different colors red, green, and blue respectively, the eigenvalue R, G, B is respectively used to judge marigold's moisture condition. There are a lot kinds of parameters which describes the characteristic of image texture, average value, consistency, entropy are used to describe texture of image in the paper.Fuzzy comprehensive evaluation Based on AHP is combined of AHP and fuzzy comprehensive evaluation, the value of each level is calculated through AHP, and the condition of every basic factor, that influents from the upper level to the top, is received through the fuzzy comprehensive evaluation. Finally the project condition of the general object is evaluated and a revaluation system model of the water demand for the crop's healthy growth is established.The design of image processing simulation platform uses GUI in MATLAB to realize the special instruments which can be used repeatedly and simply operated. It provides the function of image input, image enhancement, image analysis and output of water condition.