AUTHOR=Zhou Honghao , Zhou Yang , Long Wei , Wang Bin , Zhou Zhichun , Chen Yue TITLE=A fast phenotype approach of 3D point clouds of Pinus massoniana seedlings JOURNAL=Frontiers in Plant Science VOLUME=Volume 14 - 2023 YEAR=2023 URL=https://www.frontiersin.org/journals/plant-science/articles/10.3389/fpls.2023.1146490 DOI=10.3389/fpls.2023.1146490 ISSN=1664-462X ABSTRACT=Phenotyping of Massoniana seedlings is essential for breeding, vegetation protec-tion, resource investigation, and so on. Few reports regard-ing estimating phenotypic pa-rameters accurately in the seeding stage of Pinus Mas-soniana plants using 3D point clouds exist. In this study, seedlings with heights of ap-proximately 15-30 cm were taken as the research object, and an improved approach was proposed to automatically calculate five key parameters. The key procedure of our pro-posed method includes point cloud preprocessing, stem and leaf segmentation and mor-phological trait extraction steps. In the skeletonization step, the cloud points were sliced in the vertical and hori-zontal directions, gray value clustering is performed, the centroid of the slice is regard-ed as the skeleton point, and the alternative skeleton point of the main stem is deter-mined by the DAG single source shortest path algorithm. Then, the skeleton points of the canopy in the alternative skeleton point are removed, and the skeleton point of the main stem is obtained. Last, the main stem skeleton point after linear interpolation is restored, while stem and leaf segmentation is achieved. Because of the leaf morpho-logical characteristics of Pinus Massoniana, the leaves are large and dense. Even by us-ing a high-precision industrial digital readout, it is impossi-ble to obtain a 3D model of Pinus Massoniana leaves. In this study, an improved algo-rithm based on density and projection is proposed to es-timate the relevant parameters of Pinus Massoniana leaves. Finally, five important pheno-typic parameters, including the plant height, stem diame-ter, main stem length, regional leaf length and total leaf number, are obtained from the skeleton and the point cloud after separation and recon-struction. The experimental results showed that there was a high correlation between the actual value from manual measurement and the predict-ed value from the algorithm output. The accuracies of the main stem diameter, main stem length and leaf length were 93.5%, 95.7%, and 83.8%, respectively, which meet the requirements of real applications.