Rapeseed 3D reconstruction and morphological parameter measurement based on laser point cloud

Abstract

Rapeseed's morphological parameters provide important pattern information by corresponding to structure in the physical, optical and geometrical properties of plants. As one of the most important oil crops, the high-accuracy rapeseed geometric model contributes to agronomic horticulture, crop-breeding and plant genetics research. This paper proposes a new method to measure the morphological parameters by using the three-dimensional (3D) model of rapeseed plants, which is different from the traditional measurement method, for its higher data accuracy and efficiency of parameter acquisition. In this research, we use a portable 3D laser scanner to obtain and process rapeseed point clouds, and then we provide data preprocessing, process point cloud through building Delaunay triangulation mesh, and generate 3D model. Based on the rapeseed plant leaves high-precision model, we obtain the boundaries of rapeseed plant leaf model and characteristic points of rapeseed plant vein by the utility curvature extreme value method, and we fit the contour of rapeseed plant leaves and veins form by minimum spanning tree algorithm and the B-spline function. During the experiment, rapeseed plant 3D models are accurately reconstructed, the leaf areas are measured accurately based on the reconstruction models. We also verify the validity of the approaches and the accuracy of the models, and achieve a non-destructive measurement of rapeseed plants leaf area, acquire perimeter by blade boundary features line of blade models. According to the results, it can be concluded that the proposed methodology can not only reconstruct real rapeseed plant model, but also provides an accurate morphological parameter measurement in three dimensional way.