Deep image intensification method fused with RGB image information

Abstract

The invention discloses a deep image intensification method fused with RGB image information. The deep image intensification method comprises the steps of S1, obtaining a mapping relation of coordinates of a deep image and an RGB image; S2, preprocessing the deep image and the RGB image separately; extracting an invalid area in the deep image and labeling the pixel position of the invalid area; conducting edge detection on the RGB image, extracting edge information, and comparing the edge information with the preprocessed deep image to determine an effective supporting edge used for repairingthe deep image; S3, estimating the deep information of the invalid area; along a boundary of the invalid area in the deep image, using information of the effective supporting edge to conduct depth calculation layer by layer from outside to inside; S4, conducting filtering and noise reduction optimizing on a miniature isolation area from which the effective supporting edge cannot be extracted in the deep image to improve the precision of the deep image. By means of the deep image intensification method fused with the RGB image information, the invalid area in the deep image can be repaired on the premise of guaranteeing that the edge is clear, and the sharpening degrees of edges of all deep missing areas can be improved.

Description

technical field [0001] The invention belongs to the field of image processing, in particular to a depth map enhancement method for fusing RGB image information. Background technique [0002] In recent years, the application scenarios of mobile robots have become more and more extensive, and have penetrated into many fields such as industry and agriculture, home services, medical care, and aerospace. Simultaneous Localization and Mapping (SLAM) refers to carrying specific sensors on some mobile platforms such as robots, and collecting sensors at various moments during the movement without prior information of the environment. Information, estimate its own trajectory, and build an environment map that fits the application scenario. SLAM technology provides accurate positioning and maps so that the robot can move autonomously in the room. It is not only used in the field of robotics, but also the core technology in related fields such as virtual reality and augmented reality....

AI Technical Summary

Problems solved by technology

However, this method is mainly limited by its calculation speed, and the real-time performance of the algorithm is not high

Images