Abstract
Small-flowered chrysanthemums are a major type of cut flower used in Japan for which we are developing flowering stage inspection equipment. This equipment can correctly determine whether it is appropriate for shipping as same as the expert in the case of parameters for the classification is set properly. However, classification parameters depend on breed, season, distance to consumers, and days allowed for shipping, and the expert must find the proper parameters in practical use. We studied the automatic acquisition method about the classification parameters using several appropriate and inappropriate learning samples. We introduce this method and the results of evaluation experiments.
