提出一种采用投影仪投射彩色伪随机编码图像实现物体高精度三维重构和测量的方法。首先设计了彩色伪随机编码投影模板,在Harris角点检测的基础上,提出了彩色角点提取算法,利用二次曲面拟合实现了角点亚像素定位,使左、右摄像机摄取的图像反生成伪随机阵列,从而实现特征点的快速匹配,并最终得以重构。
A 3-D reconstruction method was introduced to apply color pseudo-random coded image from a projector.A color projecting template was designed based on the pseudo-random principle.Then,a modified color corner extraction method based on Harris corner testing was introduced,which applying quadric fitting to realize sub-pixel location of the corners.Therefore,the left and right images were decoded to pseudo-random array on the basis of the corner testing results.According to the window property,the matching between the left and right images was realized and the reconstruction of the scene could be obtained combined with the vision calibration.The surface of a cylinder was reconstructed.The result proved that the method introduced here had the advantage of high matching speed and high matching accuracy.