Abstract:

Leaf blade is one of the most important plant organs. A geometric modeling method based on point cloud of plant leaf blades was proposed to build a plant leaf blade model with high precision. According to the morphological characteristics of plant leaves, a suitable 3D scanner was chosen to obtain point cloud data of leaf blades. Then operations including registration, simplification and denoising were conducted, and high quality point cloud data of leaf blades was obtained. On this basis, leaf blades mesh generation and mesh optimization were carried out to obtain mesh leaf model with high precision. Geometric models of cucumber, maize and 2 different varieties of grape leaves were modeled using this method. The results showed that this model could better maintain the morphology characteristics of leaf blades. And this method had greatly improved model accuracy and realistic sense compared with the former ones. This study has great significance for promoting geometric modeling and further visual computing in digital plants.

Key words: plant, leaf blade, three\, dimensional reconstruction, point clouds

摘要：

叶片是植物最重要的器官之一,为构建植物叶片的高精度几何模型，提出了一种基于三维点云数据的植物叶片几何建模方法。针对植物叶片形态特征，选定最适三维扫描仪进行叶片点云数据获取，通过点云的配准、简化及去噪等操作得到高质量叶片点云数据，在此基础上进行叶片网格生成与网格优化，最终得到高精度植物叶片网格模型。利用该方法分别对黄瓜、玉米和两个品种的葡萄叶片进行几何建模，结果表明，所构建的叶片模型能够较好地保持叶片形态特征，且较以前的方法在精确度和真实感方面有了较大的提高。该研究对于推动数字植物几何建模及进一步基于几何模型的可视化计算具有重要意义。

关键词: 植物；叶片；三维重建；点云