Construction of the Automatic Diagnosis System of Plant Disease Using Genetic Programming Which Paid Its Attention to Variety

The purpose of this study is to establish the automatic system of diagnosis for plant diseases, which can be a core of the environmentally-friendly production technology, thereby aiming for the realization of a sustained symbiosis of human beings and nature. In the past, I prepared parameters to differentiate healthy leaves from diseased leaves by using 650nm optical filtered images. In this study, I investigated the methods to prepare distinction parameters, utilizing genetic programming (GP), as a tool to identify the type of diseases and improved the GP itself as well. I will chiefly describe the followings: 1. I established the identification technology for the type of diseases using the shape feature amounts of visible lesions. 2. I developed parameters with higher distinction rate by introducing a method to combine multiple parameters by GP. In particular, I improved the GP by focusing attention to the variety of groups to be searched and tried to offer technical advantages for the disease-type discriminator.