Depth image enhancement processing method and device

Abstract

The invention discloses a depth image enhancement processing method and device in the technical field of depth image enhancement processing device. An image conversion module is in electrical output connection with a plane image separation module, and the plane image separation module is in electrical output connection with an edge extraction module; the edge extraction module is in electrical output connection with a mask image generation module, the mask image generation module is in electrical output connection with a filling and repairing module, the filling and repairing module is in electrical output connection with a dual filtering module, and the image selection processing module is in electrical output connection with a deep learning module; the deep learning module is in electrical output connection with a pixel coordinate and depth value extraction module, the pixel coordinate and depth value extraction module is in electrical output connection with a three-dimensional coordinate calculation module, and the three-dimensional coordinate calculation module is in electrical output connection with an image enhancement module. According to the invention, the resolution of the depth map can be effectively improved, and the marginal definition after target enhancement in different scenes can be ensured.

Description

technical field [0001] The present invention relates to the technical field of depth image enhancement processing devices, in particular to a depth image enhancement processing method and device. Background technique [0002] Depth information plays an important role in many computer vision applications, such as AR, scene reconstruction, 3D TV-assisted sensors, etc. AR technology can make the real-time world richer, more interesting and more efficient by superimposing rich text and multimedia information. AR The scene has three-dimensional information. By superimposing virtual objects in real-time three-dimensional space, the seamless fit with the real scene is realized. However, the detection efficiency and accuracy of the image recognition technology that the existing AR technology relies on remain the same. The failure to meet the requirements of most application scenarios directly leads to the fact that AR technology can only be applied in some specific scenarios, which ...

AI Technical Summary

Problems solved by technology

[0002] Depth information plays an important role in many computer vision applications, such as AR, scene reconstruction, 3D TV-assisted sensors, etc. AR technology can make the real-time world richer, more interesting and more efficient by superimposing rich text and multimedia information. AR The scene has three-dimensional information. By superimposing virtual objects in real-time three-dimensional space, the seamless fit with the real scene is realized. However, the detection efficiency and accuracy of the image recognition technology that the existing AR technology relies on remain the same. The failure to meet the requirements of most application scenarios directly leads to the fact that AR technology can only be applied in some specific scenarios, which greatly limits the limitations of the use of AR technology, and the data needs to be preprocessed in the actual visual image application. Filling empty pixels and smoothing image noise, most of the existing technical solutions are based on RGB images and their corresponding edge information as a guide for bilateral filtering, rarely considering the scene where depth discontinuity is invisible in RGB, and cannot effectively solve this scene At the same time, over-enhancement of the image is likely to cause blocky discontinuity in the image, and the edge of the target is blurred after the depth image is enhanced. Therefore, we propose a depth image enhancement processing method and device.

Images