基于偏振双目视觉的无纹理高反光目标三维重构

摘要

For high reflective targets with smooth surface,single or even none texture,a large area of data void may appear on the reconstruction surface since the traditional three-dimensional reconstruction method depending on texture and reflective characteristic.While three-dimensional reconstruction method based on polarization vision does not rely on the texture information on the surface of objects,the polarization imaging can restrain highlight to a certain extent,which can effectively solve the problems existing in the traditional three-dimensional reconstruction method.However,the depth information obtained by three-dimensional reconstruction method based on polarization vision is under pixel coordinates.Therefore,polarization and binocular three-dimensional reconstruction method was put forward.Taking absolute three-dimensional coordinate of a few feature points obtained by binocular stereo vision as the "bridge",combined with the camera parameters obtained by binocular calibration,the point cloud data acquired by polarization in image pixel coordinates can be transformed into absolute data in world coordinates.%对于表面光滑、纹理单一甚至缺失、高反光的目标,传统的三维重构方法由于对纹理以及反光特性的依赖会使得重构表面出现大面积的的数据空洞,而基于偏振的三维重构方法不依赖于物体表面的纹理信息,同时偏振成像能够在一定程度上抑制耀光,能够有效解决传统三维重构方法存在的问题,但是,基于偏振视觉的三维重构方法得到的是像素坐标系下的深度信息,因此提出了偏振双目视觉三维重构方法,以由双目立体视觉获得的少量特征点的世界三维坐标为“桥梁”,利用双目立体标定得到的相机参数,将偏振得到的图像像素坐标系下的点云数据转化为世界坐标系下的绝对数据.