Research on 3D Phenotypic Reconstruction and Micro-Defect Detection of Green Plum Based on Multi-View Images

Rain spots on green plum are superficial micro-defects. Defect detection based a two-dimensional image is easily influenced by factors such as placement position and light prone to misjudgment omission, which the main problems affecting accuracy of defect screening plum. In this paper, using computer vision technology, an improved structure from motion (SFM) patch-based multi-view stereo (PMVS) algorithm similar graph clustering matching proposed perform three-dimensional sparse dense reconstruction plums. The results show that, compared with traditional algorithm, running time lower, at only 26.55 s, mean values camera optical center error pose 0.019 0.631, respectively. This method obtains higher meet subsequent micro-defect requirements. Aiming point cloud model plums, through preprocessing, adaptive segmentation Lab color space realizes effective experimental that average 2.56 showing faster speed better effect than K-means K-means++ algorithms. After cloud, information plums was extracted basis random sample consensus (RANSAC) plane fitting, provides theoretical for further improving sorting appearance quality