论文部分内容阅读
本文利用目标提取技术，根据近似点位进行房屋角点的自动定位，自动搜索同名点，然后建立一个尽可能多地包含房顶面的目标窗口，取代传统的以房角为中心的窗口进行房顶角点影像匹配，提高了房顶角点匹配的可靠性，实现了航空影像房屋三维数据的半自动量测。进而在地面与建筑物表面二维半不规则三角网和原始数字影像的基础上，实现了建筑物可见表面纹理的恢复，重建了城市真实三维景观。
In this paper, we use the target extraction technique to automatically locate house corners based on approximate points and automatically search for points of the same name, and then create a target window that contains as many roofs as possible to replace the traditional corner-centered window The matching of vertex images improves the reliability of corner matching of roofs and realizes the semi-automatic measurement of three-dimensional data of aerial image houses. Based on the two-dimensional semi-irregular triangulation of the ground and the building surface and the original digital image, the visible surface texture of the building is restored and the urban real three-dimensional landscape is reconstructed.