伪随机阵列和 Harris角测量在视觉检测中的应用

在三维视觉场景重构过程中，一个难题就是场景图像上的坐标点的匹配问题。在工业检测中，通常解决该问题的方法是采用结构光主动视觉技术，如点结构光、线结构光扫描法以及伪随机编码结构光法等。但是要想用一幅图象在三维欧氏空间中重构三维场景，有效的方法是采用伪随机编码结构光照明主动视觉技术。该方法利用编码结构光照明被测场景，利用编码的窗口特性，使场景被测表面每一个特征点都具有唯一的代码而唯一地被辨识。而图像的特征点提取可以采用简单而稳定的Harris角测量。

Application of multicolor pseudo- random encoded array for active visual inspection

In 3D reconstruction with machine vision method, correspondence between the surface to be recovered and its image are very difficult to register. In order to tackle this problem, the practical methods in industry inspections are to use active machine techniques with structured light illumination, such as dot structured light, line structured light and encoded structured light pattern. The last one usually is more effective, since the 3D reconstruction can be implemented only with one single image, especially for dynamic 3D scenes. In reconstruction process, the scene surface to be recovered is illuminated with a pseudo-random encoded array pattern. Then correspondence in the grabbed image is identified uniquely based on the window characteristics of the array pattern. The interest point in the image can be obtained by Harris corner detector.