Kitchen garbage classifying and sorting method and system based on computer vision

The invention belongs to the field of garbage classification, and particularly relates to a kitchen garbage classification and sorting method and system based on computer vision, and the method comprises the steps of target detection and segmentation, track planning, grabbing posture estimation and garbage recovery. According to the scheme, a garbage sorting robot system based on an ROS framework is adopted, a YOLACT target detection model with ResNet-50 as a backbone network is constructed and trained to detect different types of recoverable garbage, the 3D position of the garbage is collected and extracted, trajectory planning is automatically conducted, a mechanical arm grabs and recovers the kitchen garbage, and health risks caused by direct contact of workers are avoided; a GPD grabbing control algorithm is adopted to process non-rigid objects, grabbing gestures are directly generated from point clouds, an accurate geometric model is not needed, the garbage sorting robot system is more suitable for proce