Feasibility study of diagnosis of physiological diseases of pistachio trees with image processing

In today’s modern world, in order to accelerate the work, it has been attempted to replace the human force with machines so that the jobs would not be dependent on certain people. Agriculture is no exception and today, image processing is used for mechanizing the methods and for replacing human beings with intelligent machines. One of the cases in which image processing is used is for identifying types of plants, their defects and the type of their diseases. The deficiencies of nutrients are one of a variety of physiological diseases of trees that its identification is very important. The common method for this is to review the leaf through human eyes, soil decomposition and laboratory methods, which needs experts and so much time and costs which makes it uneconomical. Thus, we have considered a method for solving this problem. In this study, few types of defect have been investigated. For this purpose, at first, the leaves were collected and their images were taken and after processing the image, leaf features were initially extracted and then, by using the chromatic features of each pixel of the leaf, their defects were identified. This method identifies the defects with an accuracy of 95%.