Improved-SURF-feature-matching-based low-illuminance imaging method

Abstract

The invention, which relates to the computer vision field, particularly discloses an improved-SURF-feature-matching-based low-illuminance rapid imaging method. pre-image signal processing (ISP) is carried out on an original image; improved SURF feature point extraction and matching are carried out on a multi-frame image outputted after processing, thereby obtaining a calibration image, wherein operation of the algorithm is accelerated by using an improved FAST detection feature point method and thus lots of computing time is reduced; multi-frame continuously-shot pictures is carried out; and then contrast adjustment is carried out on the obtained frame accumulation image and an image after processing is outputted. Because the image feature point matching algorithm is improved, the efficiency for obtaining the image and the identifying degree under low illuminance can be improved, so that the real-time requirement is met.

Description

technical field [0001] The invention relates to the field of computer vision, in particular to a low-illuminance imaging method based on improved SURF feature matching. Background technique [0002] Multimedia, video surveillance and other technologies are developing rapidly, and have become common tools for people to communicate and record. However, when shooting in low-light environments such as nights, cloudy days, and conference rooms, the obtained pictures have low contrast, low signal-to-noise ratio, and poor visual effects, and the details in the image cannot be clearly reflected, which makes the imaging system unable to work normally. Work. Therefore, it is of great significance to study how to quickly and effectively process images under low-light conditions, how to increase the brightness of pictures, and reduce the impact of light conditions on imaging systems. [0003] Most of the existing imaging methods under low-illumination conditions only denoise and then ...

AI Technical Summary

Problems solved by technology

[0003] Most of the existing imaging methods under low-illumination conditions only denoise and then enhance the single image acquired under low-illumination conditions. This method cannot accurately reflect the image details to a certain extent. As far as the enhancement algorithm of the image acquired under illumination is concerned, the signal-to-noise ratio of the image under low illumination is close to the detection limit, and at the same time, the output image not only has serious quantum noise added, but also the contrast of the image is close to the sensitivity limit of vision.

The method of frame accumulation improves the signal-to-noise ratio of images acquired under low illumination by increasing the integration time, but in the process of frame accumulation, it is necessary to perform feature point detection, matching and calibration of each frame of multiple images. This process takes a lot of time, so it is very important to improve the real-time performance of the algorithm

Images