基于数码相机的果树枝干三维重建关键技术研究

Resrarch on key technologies for 3D reconstruction of fruit tree stems based on digital camera

基于图像的三维重建是计算机视觉的一个重要分支,是虚拟现实和科学可视化领域的基础。本文针对果树枝干三维重建的关键技术进行了研究,改进了枝干提取、特征点匹配、三维点云计算等算法。真实图像实验表明,本文所提出基于分型与颜色特征相结合的图像分割算法、改进的SIFT特征匹配算法在果树枝干三维重建中都优于文中提到的传统算法;基于最小二乘法的三维点云数据的计算方法匹配准确率达到85.16%,验证该方法的可行性。

3D reconstruction based on image is an important branch of computer vision, and the foundation of virtual reality and scientific visualization. In this paper, the key technology for 3D reconstruction of fruit trees was studied; stems image extraction.feature matching and 3D point cloud computing algorithm were improved. that image segmentation algorithm based on fractal Experiment with real image demonstrated and color theory and SIFT feature matching algorithm is superior to the traditional algorithm in 3D reconstruction of fruit tree stems; it reached 85.16% that the matching accuracy of 3D point cloud computing algorithm based on least square method. The proposed methods were proved to be applicability and feasibility.