Tobacco Stem Recognition and Localization Based on 3D Stereo Vision

With the advancement of industrial automation, accurate recognition and localization of tobacco stems during tobacco processing have become critical for improving processing quality and production efficiency. This study proposes a method for tobacco stem recognition and localization based on 3D stereo vision. By using stereo vision technology to construct depth information and combining image processing and deep learning techniques, we achieve segmentation and 3D localization of tobacco stem regions. Experimental results show that this method achieves a recognition accuracy of over 98% in complex backgrounds, with a positioning error controlled within millimeters, meeting industrial application requirements.