Improving The Accuracy Of 3d Surface Reconstruction By Controlling Height Compensation

The southwell model algorithm based on the least squares method is suitable for a variety of wavefront reconstruction and three-dimensional shape detection through slope fitting. Due to the assumption that the height is linearly related to the gradient in the southwell model, the detailed infoimation of the surface shape needs to be gradually improved by the iterative algorithm. In this paper, we add an attenuation factor in the foul' of a power function controlled by exponential and iterative times to increase the height compensation amount in the early iteration and control the excessive noise data to participate in the iteration, which leads to the deviation of the fitted surface shape from the actual surface shape. Through simulation comparison experiments, the new algorithm has faster iteration rate and higher accuracy than traditional iterative algorithms.