关键词: 倾斜摄影, 三维实景模型, 悬空地物, K-means离散点聚类算法, 最小凸包

Abstract: In order to better solve the excessive dependence of the traditional tilting photography real scene model on the three-dimensional editing environment, a real scene model suspended ground object processing technology based on oblique photography triangle network analysis is proposed. The technical method reduces the dimensionality of the three-dimensional triangular network, analyzes the radiation characteristics of the edge on the two-dimensional triangular network node after dimension reduction, and quickly locates the edge point of the feature, and combines the cluster analysis algorithm and the minimum convex edge extraction technology. The rapid positioning and elimination of the boundary of suspended objects is realized. The suspended objects such as vegetation are selected as the research object. It is verified that the technical method has good applicability to dealing with the suspended photographic real scence model, and can greatly reduce the requirements of the three-dimensional environment and the data processing difficulty in the processing of the oblique photography real-world model.

Key words: tilt photography, 3D real scene model, suspended ground objects, K-means discrete point clustering algorithm, minimum convex hull