Depth map super-resolution optimization method, device, processing equipment and storage medium

Abstract

The present invention provides a depth image super-resolution optimization method, device, processing equipment and storage medium; wherein, through the boundary fusion detection of the color camera image and the depth image, the accurate boundary of the depth image can be obtained, and noise and other factors can be avoided. The boundary of the map is not clear, which affects the performance of the depth map; through dense optical flow tracking on the depth map, combined with weighted mean filtering, the problem of flickering between adjacent frame images is alleviated; through the judgment of the connected domain, the same type of pixels is completed. The smooth filter ensures soft image display and further avoids flickering.

Description

technical field [0001] The present invention relates to the field of image processing, in particular to a depth map super-resolution optimization method, device, processing equipment and storage medium. Background technique [0002] With the development of science and technology, the rendering effect of images is also constantly improving, from traditional 2D images to 3D depth images. Compared with 2D images, 3D depth images have added depth information, which can better and more accurately reconstruct the environment in the images, and have brought qualitative improvements to many intelligent fields, including robotics, AR, etc. [0003] The current 3D depth images are mainly obtained by 3D cameras / Lidar, however, the 3D cameras / Lidar are expensive, and the resolution is much lower than that of 2D cameras; on the other hand, due to the inaccuracy of depth measurement, the obtained 3D depth images generally There is a lot of noise or part of the data is missing, which will...

AI Technical Summary

Problems solved by technology

However, in the process of smoothing and filtering, the boundary of the object displayed in the image will be blurred; on the other hand, the process of solving the sparse equation will greatly increase the amount of calculation, which will cause the image displayed by the device to lag behind; finally, due to the single-frame image The filtering process can suppress the flickering problem when the image is played continuously to a certain extent, but it is difficult to solve the flickering problem

Images