Abstract: On the basic of computer vision principle, a free surface measurement system is proposed. The paper uses the more rational model of two plane to calibrate the camera to acquire the sophisticated parameter matrix, a distorted stripe is obtained on surface. The authors finds the mating points at sub-pixel level by using epipolar geometry constrain. In coordinate computing, multiple variables non-linear optimized problems with constraint are changed into least squares problems of linear equation group utilizing Screw theory, which simplifies the computing complexity and advances stability in solving course.

摘要： 利用计算机视觉原理,建立了自由曲面的测量系统。采用更合理更有效的双平面模型标定射像机,获得高精度的参数矩阵,以激光亮线为特征,利用内极几何约束辅助确定匹配点,同时结合摄像机参数模型实现亚像素匹配。在多视角测量数据融合中采用Screw理论,简化了计算的复杂性,提高了求解过程的稳定性。

关键词: 双平面模型, 基础矩阵, Screw理论

CLC Number: 

• TP391.41