朝鲜族建筑虚拟场景重建中配准质量优化仿真

Simulation of Registration Quality Optimization in Virtual Scene Reconstruction of Korean Architecture

对建筑虚拟场景重建中参数配准的研究,能够有效的提升建筑质量和建筑特色,达到提高参数配准质量的目的。对建筑重建中的参数进行配准,需要先获取基于像素迁移的相似性匹配准则,构建建筑图像特征矩阵,完成对场景重建中的参数配准。传统方法先对图像特征点进行检测,生成图像特征描述符,构建建筑特征数据点集,但忽略了构架特征矩阵,导致参数配准精度低。提出基于聚类的朝鲜族建筑虚拟场景重建中参数配准方法。融合于几何约束的能量方程得到图中各个顶点结构稳定的三维面片,并赋值给对应顶点在三维空间中的坐标值,获取基于像素迁移的相似性匹配准则,并构建Hessian图像特征矩阵,给出图像特征点的主方向,完成对朝鲜族建筑虚拟场景重建中参数配准。实验结果表明,所提方法匹配精度较高,可以为提升朝鲜族建筑的建筑质量和建筑特色奠定基础。

A parameter registration method of in virtual scene reconstruction of Korean architecture based on clus- ter is presented. A three-dimensional （3D） facet of each vertex with stable structure in graph is obtained integrated with energy function of geometric constraint, and coordinate value of corresponding vertex in 3D space is assigned, then matching criterion of similarity is acquired based on pixel migration. The Hessian characteristic matrix of image is also built, and main direction of feature point of image is provided. Thus, the parameter registration is completed. Experimental results show that the method has high matching precision. It can lay the foundation of improving quality of Korean architecture and architectural feature.