Color image depth up-sampling method and system based on adaptive stability model

Abstract

The invention discloses a color image depth up-sampling method and system based on an adaptive stability model, and the method comprises the steps: dividing a depth map into a flat region and an edgeregion, carrying out the clustering of pixels of the edge region, and dividing the edge region into a region I and a region II according to a clustering result; mapping the clustering depth map to a high-resolution grid with the same resolution as the color image to obtain a to-be-filled matrix, and obtaining an observation matrix according to the to-be-filled matrix; obtaining a prediction matrixaccording to the adaptive stability model, and obtaining a depth matrix according to the prediction matrix and the observation matrix; and performing bicubic interpolation on the clustering depth mapto obtain an initial depth map, filling the target points of the to-be-filled matrix area I with the initial depth map, and filling the target points of the to-be-filled matrix area II with the depthmatrix to complete up-sampling of the depth map. The problem that depth images such as depth edge blurring and depth mixing in an up-sampling depth image are not smooth enough is solved, and the up-sampling depth image with clear edges is generated.

Description

technical field [0001] The present disclosure relates to the technical field of image processing, in particular to a color image depth upsampling method and system based on an adaptive stabilization model. Background technique [0002] The statements in this section merely provide background information related to the present disclosure and do not necessarily constitute prior art. [0003] With the development of unmanned driving, 3D TV and 3D movies, 3D content is closely related to our life. Among them, the reconstruction of relevant 3D information of dynamic and static objects and scenes is the core issue of computer vision. In the early days, people could perceive dynamic 3D scenes based on their vision, so they believed that computers could solve problems with the help of visual information, but even if an accurate 3D model was built, it was not necessarily possible to obtain accurate 3D information. Depth sensors have become an important tool for generating 3D depth ...

AI Technical Summary

Problems solved by technology

However, since color images have color or structural discontinuities, this method may still lead to erroneous depth information output, so how to suppress texture replication artifacts is still a challenging problem.

[0007] Ferstl treats depth map upsampling as a convex optimization problem with high-order regularization. Depth upsampling will be done according to the anisotropic diffusion tensor of the HR intensity image as a guide, where the high-order regularization term is enforced as a piecewise Refined solution that preserves sharp edges from texture while compensating for acquisition noise; but ignores structural inconsistencies that exist between depth and color images

Images