摘要
图像拼接是石窟建筑三维形貌测量中不可或缺的环节,拼接质量直接影响石窟建筑三维形貌重构的可行性和精度。以往三维图像拼接方法由于受到视场的限制,难以通过一次测量获得曲率较大的石窟建筑物体的整体形貌,存在测量速度慢以及拼接精度低的问题。为此,提出一种基于Canny算法的石窟建筑三维数字图像形貌拼接优化方法。首先对于平面圆靶标在摄像机成像平面所构成的石窟建筑三维数字图像,依据石窟建筑三维数字图像中像元恢复值的突变特性,自适应的改变石窟建筑三维数字图像滤波器的权值,运用Canny边缘检测提取石窟建筑三维数字图像候选边缘线,并对边缘位置沿梯度方向做亚像素级的校正,采用最小二乘方法结合几何相交的圆特征三维重建方法求解石窟建筑物体前后测量位置的坐标系转换矩阵,由此实现三维数字图像形貌拼接优化。实验结果表明,优化的三维拼接方法对石窟建筑三维图像具有较好的检测精度和拼接精度。
This paper proposes an optimization method for morphology mosaic of three -dimensional (3 D) digital image of grotto architectural based on Canny algorithm. The coplanar circle is used to constitute the 3D digital image in imaging plane of camera, and the weight of filter of the digital image is adaptively changed according to mutation feature of pixel recovery value in the digital image, then the Canny edge detection is used to extract candidate edge line of the digital image. The sub - pixel level correction for border position along gradient direction is carried out, and the coordinate transformation matrix of position before and after measurement of the grotto architecture integrating least square method with circle feature of geometrical intersection is solved. Thus, the morphology mosaic is achieved. Simulation results show that the optimized 3D mosaic method has good detection precision and mosaic precision.
出处
《计算机仿真》
北大核心
2017年第12期250-253,330,共5页
Computer Simulation
关键词
石窟建筑
三维数字图像
形貌拼接方法
Grotto architecture
3 D digital image
morphological mosaic method
Canny algorithm