A fast corner detection method based on dynamic frequency conversion

Abstract

The present invention proposes a fast corner detection method based on dynamic frequency conversion, including: dividing the image into multiple blocks, calculating the mean square error of each block, and judging that the image is a flat area or rich in corner points according to the mean square error area; when it is judged that the image is an area rich in corner points, then calculate the gradient kernel according to the mean square error; calculate the absolute gradient of the current pixel according to the calculated gradient kernel, and judge whether the current pixel is a corner or not according to the absolute gradient point; integrate the corner points of all blocks, and the integrated corner points are all corner points of the current image. The invention realizes fast corner point detection and can effectively improve the performance of image corner point detection.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to a fast corner detection method based on dynamic frequency conversion. Background technique [0002] The corner feature of an image is an important feature in image processing. There is a generally accepted point of view: the point where the brightness of the two-dimensional image changes sharply or the point with the maximum curvature on the edge curve of the image can be called the corner point. The corner points retain the important features in the image graphics, which can effectively reduce the redundancy of information and make the information content high, which can effectively improve the speed of image calculation, which is beneficial to the analysis and processing of images, and makes real-time processing possible. . Corner detection plays a very important role in computer vision fields such as 3D scene reconstruction, motion estimation, target tracking, target...

AI Technical Summary

Problems solved by technology

[0004] (1) The Plessey operator of scholar Harris. The advantage of this algorithm is that it is relatively simple to operate and implement. It is currently widely used in corner detection applications. The main disadvantage is that the accuracy of detection and positioning is not particularly ideal, especially in some special corners. It is easy to lose gradient information or corner points with large obtuse angle blur in detection, and the calculation time is not very ideal.

[0005] (2) Later, Smith proposed the concept of USAN for the first time, and based on this, designed the corner detection method of SUSAN. Its biggest advantage is that the method is very simple, has integral characteristics, good anti-noise performance, and the positioning accuracy is not affected by the type of corner. Influence; the disadvantage is that due to the fact that there are a lot of fuzzy edges in the actual image, it is easy to generate false responses or lose the real corners, so the detection rate of the actual image is generally lower than that of the Harris algorithm. In addition, the integration process also takes more time.

[0007] To sum up, although these classic algorithms have many advantages, their common disadvantages are that the execution time is long, the accuracy of corner point positioning is not high, and the execution performance of detection is poor, so It is very necessary to design a method that can accurately locate corner information and has a short execution speed and time, especially in computer vision applications where corner time information and spatial information are of high practical value.

Images