A Fast Photometric Stereo Algorithm with Un-calibration Lights for Surface Reconstruction

To solve the problem of complex and time-consuming for light calibration in photometric stereo vision, a three-dimensional(3D) reconstruction algorithm of photometric stereo vision based on isotropic illumination reflection model is studied. For different isotropic materials, the normal direction corresponding to this point is estimated by observing the similarity of the change of light reflection intensity of a point on the object surface under different lighting conditions. Through the dimension reduction method, the original inherent characteristics of the object surface are preserved, and then the normal direction of each pixel on the object is obtained by analyzing the similarity between pixels on the image. Experiments show that this method has good wide applicability, avoids the problem of light source parameter calibration, and has the advantages of fast reconstruction speed.