Series acquisition secondary positioning method based on machine vision

Fecha de publicación: 18/06/2021
Fuente: Wipo tomato
The invention relates to a string picking secondary positioning method based on machine vision, and the method comprises the steps: inputting a to-be-processed tomato fruit string image into a mature tomato fruit string recognition model based on a YOLOv4 network, so as to judge whether there is a mature tomato fruit string in the to-be-processed tomato fruit string image; if yes, extracting an image of a region of interest from the image of the tomato fruit string to be processed by the recognition model and outputting the image; and judging whether the region-of-interest image exceeds a corresponding image size range, if not, inputting the region-of-interest image into a Mask R-CNN network-based mature tomato stem and fruit string segmentation model, positioning a picking point of a tomato fruit string image to be processed according to an output result, and completing picking by a picking robot according to the picking point. According to the method, mature tomato fruit strings can be effectively identified, picking points on the tomato fruit strings and fruit stems can be accurately positioned, and high-precision lossless picking can be realized.