Quick semi-global dense matching method and device

Abstract

The invention provides a quick semi-global dense matching method and device. The quick semi-global dense matching method includes the steps that a base image and a matching image which are to be matched and have a epipolar line relation are obtained; matching calculation is carried out on the base image and the matching image according to a dense matching algorithm and the epipolar line relation; an obtained base image disparity map and an obtained matching image disparity map are denoised and refined; consistency detection is conducted on the base image disparity map and the matching image disparity map obtained through denoising and refining processing to obtain a matching base image disparity map. According to the method, the base image and the matching image are calculated according to the dense matching algorithm and the epipolar line relation, the two obtained disparity maps are denoised and refined, noise in the disparity maps is filtered out, consistency between the object boundary in the disparity maps and the object boundary in the original images is kept, left-right consistency detection is realized fast, and it is unnecessary to conduct two individual left-right image matching; the matching time is shortened on the premise of considering matching quality and matching speed at the same time.

Description

technical field [0001] The invention relates to the field of digital photogrammetry, in particular to a fast semi-global dense matching method and device. Background technique [0002] The process of generating image disparity maps based on image dense matching algorithms is an important step in digital photogrammetry, and it is also the basis for subsequent photogrammetry products. After the image disparity map is generated, with the help of the disparity value of each pixel and the outer orientation of the two images, the dense 3D point cloud can be obtained through the space forward intersection, and then the object can be 3D modeled and digitally modeled according to the obtained dense 3D point cloud. Surface model generation and orthophoto correction, etc. [0003] Specifically, each image for dense image matching (including the reference image and the matching image, corresponding to the left image and the right image respectively) includes a lot of pixels, because it...

AI Technical Summary

Problems solved by technology

[0003] Specifically, each image for dense image matching (including the reference image and the matching image, corresponding to the left image and the right image respectively) includes a lot of pixels, because it is necessary to find a corresponding point with the same name for each pixel in the above images, Therefore, image intensive matching is the step that occupies the most computing resources and consumes the longest time in the entire production process.

Specifically, in actual production, in addition to considering the time factor, the matching quality must also be guaranteed, and the following reasons can cause wrong matching, including: first, due to the difference in shooting angle between the left and right images, there will be different degrees of The object occlusion phenomenon, so it will cause the boundary occluded by the object to be unclear during matching; second, if there are different regions in the images captured in the left and right images, but the different regions correspond to similar or repeated textures, it will make the left and right The matching of images is ambiguous; third, some areas (such as grasslands, deserts) do not have significant textures, which will make local areas unable to match

[0004] At present, the method to solve the above problem is to use a semi-global matching algorithm. Based on the above algorithm, related technologies provide a semi-global dense matching method, which includes: first matching the left image (ie, the reference image) to the right image through the above semi-global matching algorithm (i.e. matching images) for the first independent matching to obtain the disparity map of the left image; however, this method will cause more errors in the matching of the edge of the captured object; in order to solve this problem, the related technology also provides another semi- The global dense matching method includes: through the above semi-global matching algorithm, first match from the left image (i.e. the reference image) to the right image (i.e. the matching image) for an independent first match, and then in turn match from the right image to the left image. An independent second matching, and then compare the two matching results, remove the wrong matching points in the left image, and obtain an accurate disparity map of the left image; however, performing two independent image matching requires more computer memory and will make double the match time

[0005] The inventor found in the research that in the semi-global dense matching method provided in the prior art, on the basis of both speed and quality, it still needs to consume more computing time and occupy more computer memory. To solve this problem, No effective solution has yet been proposed

Images