Research On Point Cloud Information Processing Of Plant Leaf Based On VTK

Virtual plant growth model has a very important significance to crop yield forecasting, land’ productivity evaluation, resource and environmental analysis, crop cultivation guidance, crop growth mechanism and regulation of crop growth and its response to environment studies. Among them, the precise three-dimensional morphological structure of the plant model can be used to study the spatial structure associated with the nature of plants, is an important aspect of virtual plants, plant modeling and other related studies. Study of plant three-dimensional information can obtain crop growth parameters in the process of dynamic data, change the traditional agriculture difficult situation of quantitative research, and provide the basis for the precision agriculture.Morphological visualization technology is a key technology of virtual plant research. In this paper, we use plant leaves as the research object. First introduced the virtual plant and plant three-dimensional visualization research status, discussed the feasibility and necessity of leaf three-dimensional visualization. Combined visualization library VTK implements plant point cloud3D visualization, and make the point cloud coordinate transformation, color mapping and other operations, has a strong sense of reality, can simulate three-dimensional shape of plant leaves well. Subsequently, the paper describes the iterative closest point ICP algorithm, using multi-field scanning and stitching methods combine with the iterative closest point ICP algorithm to registration, obtain a good practice effect. Additionally, this article describes the Delaunay triangulation and its related algorithm, and using VTK visualization library to achieve an efficient and stable Delaunay triangulation effect. Finally, light irradiation on plant leaves was simulated.This paper includes:1) The use of structured light scanner to obtain plant morphological features point cloud, which further on plant physiological structure and morphology analysis and processing. The author use plant leaves as the research object, through the leaf point cloud processing, access point cloud containing plants coordinate and color information.After de-noising in the reverse engineering software, the measured object was multi-field scanned and stitched through the rotary table. Thus obtaining the plants point cloud meet the requirements of subsequent experiments.2) Self-invented a reference body to detect the accuracy of plant three-dimensional point cloud information retrieval system, the system accuracy in the length, angle measurement were tested with the reference body, the test results show that the system for3D point cloud data in terms of length measurement accuracy of98%, while the accuracy of measurements of the angle reached96%. The three-dimensional point cloud data obtained by this system has high precision, can be performed subsequent data processing.3) Using Microsoft Visual Studio C++2010integrated programming environment, combined with visualization library VTK (Visualization Toolkit), we achieved leaf3D reconstruction and visualization on computer. Carried out on plant point cloud coordinate conversion, color mapping and other operations, it has a sense of reality and to facilitate subsequent operations. Using multi-field scanning and stitching methods combine with the iterative closest point ICP algorithm to registration, obtained a good practice effect. We use VTK visualization library to achieve an efficient and stable Delaunay triangulation effect, in order to provide a feasible method for further study.4) Established a leaf-based3D visualization system based on all the previous algorithms. Combined with the use of VTK and MFC programming in Microsoft Visual Studio C++2010platform enables reconstruction of three-dimensional model of plant leaves. This software’s main features include leaf three-dimensional point cloud display, rotation, scaling. ICP registration algorithm, Delaunay triangular mesh, and leaves under illumination effects simulation. So it can be concluded that this system can realize non-destructiveobservation and3D reconstruction of plant roosin situ on normal PCs with outstanding visualization quality.