一种倾斜摄影模型精细融合方法

现有三维模型融合方法多采用对原有三角面片顶点重新构网的方式，容易导致原有模型几何特征丢失及纹理失真等。本文针对倾斜摄影模型自身特点，提出了一种模型精细融合方法。首先，以一个模型作为中心模型，并以另一模型与中心模型外轮廓的空间关系计算融合区域；然后，以中心模型外轮廓作为约束，对原始模型三角面片进行精细剖分；最后，基于精细剖分的三角面片进行纹理坐标的插值计算。以实际倾斜摄影数据进行试验的结果表明，本文方法融合后的网格模型拓扑结构正确，融合边界不存在漏洞、裂缝和错位等情况，而且纹理过渡自然，颜色无明显差异。

An accurate oblique photography models merging method

Existing 3D models merging methods are mostly implemented by re-triangulation vertices of original triangle face, which causes unpleasant results, like geometric feature losing and texture distortion of original 3D models. In this paper, an accurate oblique photography models merging method is proposed based on the characteristics of the oblique photography models. Fistly, a model is selected as center model, and the fusion area is calculated by the spatial relationship between the center model outline boundary and another model. Then the outline boundary of center model is set as constraint and triangles of raw mesh are subdivided accurately. Finally, Texture coordinates is calculated by interpolation of accurate subdivided trianglular face. At the end of this paper, we verify proposed merge method with actual oblique photography data, and the result shows that the merged mesh model of proposed method is topologically correct, and there have no holes, cracks and misalignments on the fusion boundary, and the texture is transitioned naturally and no obvious differences of color exist.