30-06-2023 дата публикации
Номер: CN116363505A
Принадлежит:
The invention relates to a target picking method based on a picking robot vision system. The method comprises the following steps: acquiring an original fruit image by using a camera; the collected original fruit images are preprocessed, the preprocessed fruit images form a data set, and the data set is divided into a training set, a verification set and a test set; a Mask R-CNN model is established, and training is carried out; the training process is optimized, and a Mask R-CNN model with the highest training precision is obtained; the Mask R-CNN model with the highest training precision is loaded on a CPU of the picking robot, and finally target recognition picking is achieved based on a visual system of the picking robot. According to the method, the high-pixel camera is used for acquiring the ripe fruit image, the deep learning technology is applied to intelligent fruit picking, the network structure is adjusted according to the actual use scene, the Mask R-CNN model is trained, finally ...
Подробнее