Robust cabbage recognition and automatic harvesting under environmental changes

In Japanese agriculture, labor shortages are becoming increasingly severe due to the lack of farmers and aging. Therefore, extensive research has been conducted on automation cabbage harvesting. automatic harvesting, detection is performed using deep learning. evening hours, if backlight enters camera, not possible. Moreover, cabbages in back row that be harvested detected targeted for harvest. To solve these problems, we have proposed new recognition methods this paper. We lower half selection an RGB-D camera. sliding-mode control was incorporated enable harvesting soft soil. The experimental results demonstrate effectiveness methods.