Method for positioning fruit stem shear point of bunched tomato picking robot

The invention discloses a fruit stem shear point positioning method of a bunched tomato picking robot. A string tomato is used as a research object, a string tomato data set is established by using a data enhancement technology, and the maturity of a fruit string is judged by combining a YOLOv5 deep learning algorithm. The method comprises the following steps of: obtaining a cluster of fruit stems of the fruit bunches by using an European clustering method, extracting a cluster of a central region, and fitting a straight line segment of a fruit stem space by using an RANSAC straight line fitting method so as to output position information of shear points of the mature fruit bunches. The cluster tomato picking robot recognizes cluster tomatoes, mature fruits, color changing period fruits and immature fruits at first, a visual fruit stem area point cloud picture is generated through fruit stem depth images, after point cloud data is subjected to European clustering, fruit stem clusters are extracted and subject