Abstract
In this study, to the development of the visual system of tomato cultivation robot by using a Web camera is an inexpensive image input device and OpenCV is an open source. Was the recognition in the binary process by the color data as the target petals, fruit, a auxiliary bud. In addition, I examined the noise removal method that has occurred. At that time, recognition narrowly accurate recognition range is difficult to auxiliary bud. So I examined the object detection by machine learning. As a result, I was considered to be the detection of possible side Buds by machine learning.