苹果树冠层三维重建及光照分布计算方法

合理的果树冠层结构有利于获取充足的光照，对提升果实产量及品质具有重要意义。为了智能高效地揭示果树冠层光照分布规律，选择苹果树为研究对象，构建彩色摄像机与ToF摄像机相结合的视觉系统，主要开展：果树冠层快速三维重建方法研究、果树冠层光照分布计算方法研究以及果树冠层光照分布动态模拟系统开发研究。首先分析不同生长阶段果树冠层的三维结构特征，结合前期研究成果-移动平台，动态采集果园多源信息，建立不同生长阶段的果树冠层三维结构模型；研究不同太阳方位角、冠层深度及观测角对彩色图像表观特征的影响规律，建立果树冠层三维空间的颜色特征随时间序列变化的光照分布计算模型；开发果树冠层光照分布动态模拟系统，对模拟剪枝后的果树冠层三维结构与光照分布进行预测；通过在果树冠层三维空间布设无线传感器网络节点，对光照分布计算模型进行验证与优化。项目成果为科学地果树整形修剪与果园管理提供相关理论依据。

The reasonable structure of fruit tree is helpful for canopy to get enough light, and has a great impact on improving fruit quality and productivity. In order to reveal light distribution rule of canopy intelligently and efficiently, apple trees are selected as research objects, a multi-source vision system consisting of color camera and ToF camera is constructed. The key contents of this study are as following：Fast three-dimensional(3D) reconstruction algorithm for canopy, modeling of light distribution in canopy, developing a dynamic simulation system of light distribution for canopy. Firstly, tree canopy images are acquired dynamically using the early achievement which is the mobile platform, and morphologic characteristics of canopy at different growth stages are deeply analyzed, then fast 3D reconstruction methods for each stage are studied. Secondly, the relationship among color images, sun azimuth, canopy layer, and observation angle is researched, and models of light distribution on color characteristics changing with time series are building. Thirdly, based on the above models, a dynamic simulation system of light distribution for canopy is developed, which can be predicted for canopy structure and light distribution after simulation of pruning. Finally, models of light distribution are verified and optimized by establishing wireless sensor network nodes in canopy. The achievements of this study will bring theoretical basis for scientific pruning and orchard management.