Photometric stereo light calibration optimization and high-reflective distance fitness compensation method to improve reconstruction

Abstract

This paper proposes an interpolated light calibration optimization and high-reflective area compensation method to solve the accuracy loss caused by idealization of the photometric stereo (PS) model and high-reflective area. The spatial distribution model of light intensity is defined as a cubic interpolation function, which is used to obtain an intensity coefficient matrix to optimize the PS model. A light source adaptation model is adopted to select the appropriate position of the light source based on the characteristics of the object. Finally, a distance fitness compensation model is established to repair the high-reflective area, which selects the optimal pixel values of the non-high-reflective area. The experiment result shows that the proposed method can obtain the light source spatial characteristics, and high-reflective areas could be repaired.