Robust depth map structure reconstruction and denoising method based on guided filter

Abstract

The invention discloses a robust depth map structure reconstruction and denoising method based on a guided filter. A structure error region is detected; detecting a place where the input depth map isgreatly different through the guided filtering of the large window and the guided filtering of the small window; an eclosion effect can be achieved due to guided filtering under a large window; the guide filtering of the small window only plays a smoothing role. Therefore,. Therefore, areas with large differences can be regarded as structure error areas. Marking as a potential structure error region, then constructing a weight based on an iterative reweighted least square algorithm, after the weight construction is completed, carrying out overall solution and updating the depth map, judging whether a set iteration frequency is reached or not according to a result, if so, outputting the depth map to end calculation, and otherwise, carrying out structure error region detection again. According to the method, strong noise can be suppressed, structure error regions of the depth map and the color map can be repaired, the consistency of the depth map and the color map is improved, a correctdepth map boundary is recovered, and the method has important guiding significance for improving the quality of the synthesized view.

Description

technical field [0001] The invention belongs to the technical field of image processing, and in particular relates to a method for reconstructing and denoising a robust depth image structure based on a guiding filter. Background technique [0002] With the advent of depth sensors and the rapid development of stereoscopic display technology, depth maps have become a research hotspot in recent years. There are two ways to obtain the depth map: active and passive. The active method mainly uses the visible light data of a single viewpoint for depth estimation, or performs stereo matching for the visible light data of two (or more) viewpoints to calculate the disparity of the corresponding position, and then converts it into a depth map according to the geometric relationship. With the successful application of deep learning in the field of computer vision, the accuracy of actively obtained depth maps has been greatly improved. However, this type of method has high requirements...

AI Technical Summary

Problems solved by technology

However, this method is not suitable

Images