Depth map structure restoration method based on two-layer full-connection conditional random field model

Abstract

The invention discloses a depth map structure restoration method based on a two-layer full-connection conditional random field model, and the method comprises the steps: inputting a color map and a depth map into a first-layer full-connection conditional random field model, and obtaining a roughly restored depth map through the minimization of an energy function of the first-layer full-connectionconditional random field model; and inputting the roughly recovered depth map into a second-layer full-connection conditional random field model, and obtaining an accurately recovered depth map by minimizing an energy function of the second-layer full-connection conditional random field model. According to the method, the depth map containing serious structural distortion can be accurately recovered, and the problem of color map texture mapping is effectively solved.

Description

technical field [0001] The invention belongs to the technical field of image processing, and in particular relates to a depth map structure restoration method based on a two-layer fully connected conditional random field model. Background technique [0002] With the development of depth data acquisition technology, depth maps are widely used in various 3D vision tasks, such as: 3D reconstruction, robot navigation, virtual reality, etc. Currently, there are mainly two methods for obtaining the mainstream depth map. The first is to calculate a depth map based on one or more color images, such as stereo matching or estimation methods based on deep neural networks. The second is to obtain depth maps through physical sensors, such as TOF or structured light sensors. However, due to the immaturity of the current technology, there are many defects in these two types of methods. For example: the depth map calculated by the stereo matching method is prone to content loss in the we...

AI Technical Summary

Problems solved by technology

The neural network-based repair method usually uses a trained neural network to repair the depth map, but the depth map repaired by this method is not accurate enough, the geometric structure is often not accurate enough, and there are phenomena such as blurred boundaries.

Images