Remote sensing image registration method and system based on SURF algorithm

Abstract

The invention discloses a remote sensing image registration method and system based on an SUFR algorithm. The method comprises the following steps: uploading a reference remote sensing image and a remote sensing image to be registered to a GPU video memory; respectively carrying out feature extraction on the reference remote sensing image and the remote sensing image to be registered in a GPU in aparallel calculation mode by adopting an SURF algorithm to obtain a first feature descriptor and a second feature descriptor, the first feature descriptor is a feature descriptor of the reference remote sensing image, and the second feature descriptor is a feature descriptor of the remote sensing image to be registered; carrying out feature matching on the first feature descriptor and the secondfeature descriptor in the GPU by adopting a fast approximate nearest neighbor search algorithm to obtain a feature matching image; and downloading the feature matching image into a CPU memory to obtain a registered image. According to the method, image registration is realized in the GPU in a parallel calculation mode by adopting an SURF algorithm, so that the operation speed of remote sensing image registration can be improved, and the registration efficiency is further improved.

Description

technical field [0001] The invention relates to the technical field of image registration, in particular to a remote sensing image registration method and system based on SUFR algorithm. Background technique [0002] Registration of remote sensing images is a key step in applications such as remote sensing image matching, change detection, and multispectral fusion. UAV remote sensing images show a trend of high resolution and large amount of data. Improving the utilization efficiency of remote sensing images has always been the goal pursued by people. In the process of image registration, the selection of feature extraction determines the overall registration efficiency. Although the SURF feature descriptor is much faster than SIFT in terms of computational efficiency, it still cannot meet the real-time requirements of remote sensing image registration speed in practical applications. Require. [0003] In the field of UAV remote sensing image registration, at present, the ...

AI Technical Summary

Problems solved by technology

Although the SURF algorithm improves the feature extraction and description method compared with SIFT, and completes the feature extraction and description in a more efficient way, the operation speed is higher than that of the SIFT algorithm, but there are still problems due to feature point description, monitoring, and extraction. Slow speed leads to low image registration efficiency, low feature point matching accuracy, and poor image registration effect, which cannot meet the requirements of mass remote sensing data registration within milliseconds.

Images