CN-121982077-A - Image registration method, device and equipment

CN121982077ACN 121982077 ACN121982077 ACN 121982077ACN-121982077-A

Abstract

The embodiment of the application provides an image registration method, device and equipment. The method comprises the steps of obtaining a plurality of images to be registered, classifying the images to be registered to obtain feature rich images and feature sparse images, determining feature rich images adjacent to the feature sparse images based on the position relation of the feature sparse images, determining homography matrixes of the adjacent feature rich images, determining homography matrixes of the feature sparse images according to the homography matrixes of the adjacent feature rich images, and registering the images to be registered according to the homography matrixes to obtain registered images. The method is used for achieving the effect of improving the reliability of image registration.

Inventors

ZHOU WEN
XU CHEN
WANG JIANXUN
ZHAO SHANMIN

Assignees

转转一零二四(北京)科技有限公司

Dates

Publication Date: 20260505
Application Date: 20251217

Claims (10)

1. A method of image registration, comprising: acquiring a plurality of images to be registered, wherein the images to be registered have a position relationship; Classifying the images to be registered to obtain feature rich images and feature sparse images; Determining a feature rich image adjacent to the feature sparse image based on the position relationship of the feature sparse image; Determining homography matrices of the adjacent feature-rich images; Determining a homography matrix of the feature sparse image according to homography matrices of the adjacent feature rich images; and carrying out registration processing on the image to be registered according to the homography matrix to obtain a registered image.
2. The method according to claim 1, wherein the classifying the image to be registered to obtain a feature-rich image and a feature-sparse image includes: Extracting feature points of the image to be registered to obtain feature points of the image to be registered; determining feature sparsity of the image to be registered according to the feature points of the image to be registered, wherein the feature sparsity represents the degree of density of the feature points of the image to be registered; If the feature sparsity of the image to be registered is smaller than or equal to the preset threshold, the image to be registered is determined to be a feature sparse image.
3. The method according to claim 2, wherein the determining the feature sparsity of the image to be registered according to the feature points of the image to be registered includes: acquiring the number of the characteristic points of the image to be registered and the size information of the image to be registered; determining area information of the image to be registered according to the size information of the image to be registered; And taking the ratio of the number of the characteristic points of the image to be registered and the area information of the image to be registered as the characteristic sparsity of the image to be registered.
4. The method of claim 1, wherein the determining the homography matrix of the feature sparse image from the homography matrices of the neighboring feature rich images comprises: If the adjacent feature rich images are determined to be one, taking a homography matrix of the adjacent feature rich images as a homography matrix of the feature sparse images; and if the number of the adjacent feature-rich images is determined to be multiple, carrying out weighted average processing on the homography matrix of the adjacent feature-rich images to obtain the homography matrix of the feature sparse image.
5. The method according to claim 4, wherein the weighted average processing is performed on the homography matrices of the adjacent feature-rich images to obtain homography matrices of the feature-sparse images, including: Determining the weight of the adjacent feature rich images according to the feature sparsity of the adjacent feature rich images; according to the weight of the adjacent feature-rich images, carrying out weighted average processing on homography matrixes of the adjacent feature-rich images to obtain processed homography matrixes; and taking the processed homography matrix as the homography matrix of the characteristic sparse image.
6. The method of claim 1, wherein the image to be registered has pose information of a segmented acquisition system, wherein the determining the homography matrix of the feature sparse image from the homography matrices of the neighboring feature rich images comprises: determining an initial matrix according to the homography matrix of the adjacent feature-rich images and pose information of the block acquisition system corresponding to the adjacent feature-rich images; and carrying out weighted average processing on the initial matrix to obtain the homography matrix of the characteristic sparse image.
7. The method of any of claims 1-6, wherein the determining a homography matrix of the neighboring feature rich images comprises: Extracting feature points of the adjacent feature-rich images to obtain feature points of the adjacent feature-rich images; and carrying out matching processing on the feature points of the adjacent feature-rich images and a preset reference image to obtain homography matrixes of the adjacent feature-rich images.
8. An image registration apparatus, comprising: The acquisition module is used for acquiring a plurality of images to be registered, wherein the images to be registered have a position relationship; the classification module is used for classifying the images to be registered to obtain feature rich images and feature sparse images; The first determining module is used for determining a feature rich image adjacent to the feature sparse image based on the position relation of the feature sparse image; A second determining module, configured to determine a homography matrix of the adjacent feature-rich images; The third determining module is used for determining the homography matrix of the feature sparse image according to the homography matrix of the adjacent feature rich image; And the registration module is used for carrying out registration processing on the image to be registered according to the homography matrix to obtain a registered image.
9. An electronic device is characterized by comprising a memory and a processor; the memory stores computer-executable instructions; the processor executing computer-executable instructions stored in the memory, causing the processor to perform the method of any one of claims 1-7.
10. A computer readable storage medium having stored therein computer executable instructions which when executed by a processor are adapted to carry out the method of any one of claims 1-7.

Description

Image registration method, device and equipment Technical Field The present application relates to the field of image processing technologies, and in particular, to an image registration method, apparatus, and device. Background Because of the large size of 3C products such as computers (computers), communication (Communication), consumer electronics (Consumer Electronics), etc., if a single shot is adopted, the resolution requirement is difficult to meet, therefore, the product needs to be divided into a plurality of areas for block image acquisition, and then each block image is fused into a complete image by an image registration technology, so as to realize the functions of defect detection, size measurement, etc. In the related art, the registration is achieved by performing feature point extraction on each segmented image and then performing matching processing on each segmented image and a reference image, but in a sparse feature area such as a battery shell, a solid-color coating or a texture-free area, the registration fails due to the lack of effective feature points. Disclosure of Invention The image registration method, the device and the equipment provided by the embodiment of the application are used for achieving the effect of improving the reliability of image registration. In a first aspect, an embodiment of the present application provides an image registration method, including: acquiring a plurality of images to be registered, wherein the images to be registered have a position relationship; Classifying the images to be registered to obtain feature rich images and feature sparse images; Determining a feature rich image adjacent to the feature sparse image based on the position relationship of the feature sparse image; Determining homography matrices of the adjacent feature-rich images; Determining a homography matrix of the feature sparse image according to homography matrices of the adjacent feature rich images; and carrying out registration processing on the image to be registered according to the homography matrix to obtain a registered image. In a possible implementation manner, the classifying the image to be registered to obtain a feature rich image and a feature sparse image includes: Extracting feature points of the image to be registered to obtain feature points of the image to be registered; determining feature sparsity of the image to be registered according to the feature points of the image to be registered, wherein the feature sparsity represents the degree of density of the feature points of the image to be registered; If the feature sparsity of the image to be registered is smaller than or equal to the preset threshold, the image to be registered is determined to be a feature sparse image. In a possible implementation manner, the determining the feature sparsity of the image to be registered according to the feature points of the image to be registered includes: acquiring the number of the characteristic points of the image to be registered and the size information of the image to be registered; determining area information of the image to be registered according to the size information of the image to be registered; And taking the ratio of the number of the characteristic points of the image to be registered and the area information of the image to be registered as the characteristic sparsity of the image to be registered. In a possible implementation manner, the determining the homography matrix of the feature sparse image according to the homography matrix of the adjacent feature rich image includes: If the adjacent feature rich images are determined to be one, taking a homography matrix of the adjacent feature rich images as a homography matrix of the feature sparse images; and if the number of the adjacent feature-rich images is determined to be multiple, carrying out weighted average processing on the homography matrix of the adjacent feature-rich images to obtain the homography matrix of the feature sparse image. In one possible implementation manner, the weighted average processing is performed on the homography matrix of the adjacent feature-rich images to obtain a homography matrix of the feature sparse image, including: Determining the weight of the adjacent feature rich images according to the feature sparsity of the adjacent feature rich images; according to the weight of the adjacent feature-rich images, carrying out weighted average processing on homography matrixes of the adjacent feature-rich images to obtain processed homography matrixes; and taking the processed homography matrix as the homography matrix of the characteristic sparse image. In one possible implementation, the image to be registered has pose information of a segmented acquisition system, and the determining the homography matrix of the feature sparse image according to the homography matrix of the adjacent feature rich image comprises: determining an initial matrix according to the homogr