Binocular camera image depth information acquisition method and device

Abstract

The invention provides a binocular camera image depth information acquisition method and device. The method includes the first step of carrying out image correction on input data of two target images according to binocular camera calibration information obtained in advance, so that corresponding pixel points on the target images are located on the same horizontal line, the second step of calculating pixel matching energy according to the image correction information, constructing a matching pair directed connection graph according to the pixel point adjacent relationship and setting the weight of each side in the connection graph, the third step of utilizing a graph theory maximum flow algorithm to solve the maximum flow of the connection graph through the adoption of a line matching method, converting the maximum flow into a matching result and caching the matching result, the fourth step of carrying out postprocessing on the cached matching result and obtaining a postprocessing result, and the fifth step of calculating the pixel depth of the target images according to the binocular camera calibration information and the postprocessing result and obtaining the depth information. According to the method, the depth of a binocular camera can be regained and the depth information can be obtained while reasonable hardware expenditure, low power consumption and performance requirements are met.

Description

technical field [0001] The invention relates to the technical field of information collection and processing, in particular to a binocular camera image depth information acquisition method and device that utilizes a binocular camera to restore depth information and meets reasonable hardware overhead, low power consumption and performance requirements. Background technique [0002] Simulating human binocular vision to perceive the depth of the scene is to recover the depth information of the scene by using the image pair captured by a pair of cameras whose relative positions are similar to those of the eyes. It is an important part of computer vision and artificial intelligence, and it lays the foundation for computers / robots to use visible light information to automatically perceive and understand the real world, and based on this, to exhibit certain intelligent behaviors. In recent years, more and more researchers are committed to applying this technology to many applicatio...

AI Technical Summary

Problems solved by technology

At present, due to the large amount of calculation to obtain depth information, the hardware implementation of this complex process is difficult and the cost is very high, which cannot meet reasonable hardware overhead, low power consumption and performance requirements.

Images