3D visual detection method for position and pose discrimination of stacked IC devices

In response to the difficulties in identification and positioning caused by stacking IC devices in the feeding tray during the sorting and testing process, the difficulty in identifying depth information caused by stacking placement, and the need to set different depths for different IC devices due to the diverse types of IC devices. Propose a multi-objective stacked electronic component 3D pose recognition method, introduce 3D machine vision technology to perceive the environment, and elevate traditional sequential loading to intelligent active search and suction loading. This method improves the traditional Census stereo matching method and enhances the quality of the obtained disparity map. A disparity based region growing image segmentation algorithm was used to effectively segment the target IC device in response to the problem of IC aliasing stacking. Finally, the target device was reconstructed in 3D using the obtained disparity values and camera parameters, effectively obtaining the 3D pose of the target device in the camera coordinate system. The recognition and detection effect is good in real samples and can be effectively applied to the detection of IC devices.