基于偏微分法的光度立体三维重建

Photometric stereo 3D reconstruction based on partial differential method

光度立体法通过物体表面光照变化来恢复物体的三维信息.目前许多算法都是基于灰度图像,同时深度需要通过法向量的求解来得到,这在一定程度上影响了重建的精度和效率.采用了一种基于偏微分法的光度立体三维重建技术,通过原始RGB图像间的光度比构建线性偏微分方程组来对深度直接进行求解,从而避免了由彩色图向灰度图转化造成的信息丢失和由于变量的多步求解造成的累积误差,并且消除了多反射率对重建结果的影响.通过人工数据集的仿真和真实环境中物体的重建实验验证了该方法的有效性.

The photometric stereo recovers ured through the change of illumination the three dimensional information of object meas direction. Many algorithms are generally based on grayscale images,and the depth is obtained by solving the normal beforehand. Therefore,the reconstruction efficiency and accuracy are not ideal. This article uses a photometric stereo re construction technology which based on partial differential method to get the depth directly. This method avoids the information loss caused by the conversion from the color image to the grayscale image and the accumulated error caused by the multi step solution of the varia ble,and eliminates the influence of multiple reflectivity on the reconstruction result. The ef fectiveness of the proposed method is verified by the simulation under artificial data and the reconstruction experiment in real environment.