The morphological structure of the plant is a key screening index for crop cultivation and breeding researches.The reconstruction of the 3D model structure of peanut plants is important for the acquisition of peanut phenotypic parameters.In this paper,we designed a LIDAR-based 3D reconstruction system and method for peanut plant.LIDAR with slide module and high-precision electronically controlled rotary table were used to obtain 3D point cloud data of peanut plant that were processed by direct-pass filtering,statistical filtering,and alignment fusion to remove background information,filter outliers,reconstruct the peanut plant 3D model,and extract phenotypic parameters.Three phenotypic parameters including plant length,plant width and plant height at the seedling and pegging stages were extracted and evaluated for accuracy.The results showed that the average accuracy of the reconstructed peanut plant 3D models at the seedling stage and pegging stage reached 96.47％ and 96.45％ ,respectively.
维普
万方数据