【摘要】 三维地籍产权体包含空间产权信息和物理信息，但不同视点下表达的权籍信息量不同。因此，选定符合人眼视觉感知特征且能传递尽可能多信息量的视点，对三维地籍制图具有重要意义。该文提出顾及人眼视觉感知特征的三维地籍产权体最优视点选择方法，将包围三维地籍产权体的正二十面体的12个顶点和每个三角形面的重心作为候选视点集合，选取三维界址点可见性、可见面积比和视点舒适性作为模型特征，利用多层感知器进行训练，得到三维地籍产权体最优视点选择模型。实验结果表明：利用多层感知器可有效避免求解过程陷入局部最优解的情况，提高了算法的稳定性；输出的最优视点充分考虑了人眼视觉感知特征，且能包含较大权籍信息量，并传递三维宗地图、三维产权图的更大视觉信息，提升三维地籍制图可读性，可用于三维地籍系统社会化应用制图。更多还原

【Abstract】 Spatial and physical information are both contained in the 3D cadastral unit, although the amount of cadastral data expressed from various perspectives varies.Therefore, choosing a viewpoint that complies with the features of human visual perception and can convey as much information as possible is crucial for 3D cadastral mapping.The strategy for selecting the best view for a 3D cadastral property unit is suggested in this study while taking into account aspects of human visual perception.The center of gravity of each triangle and the 12 icosahedron vertices encircling the 3D cadastral property unit are utilized as candidate viewpoint sets, and the visibility of 3D boundary points, visible area ratio, and viewpoint comfort are chosen as model features.Finally, based on the training using multilayer perceptron, the optimal viewpoint of 3D cadastral property unit is then obtained.The experimental results show that the usage of multilayer perceptron can effectively avoid falling into the local optimal solution, which can increase the stability of the algorithm.The output optimal viewpoint fully considers the human visual perception characteristics, and can contain an amount of cadastral information.It can transfer more visual information between the 3D parcel map and the 3D property map, and enhance the readability of 3D cadastral mapping.更多还原