Detection method of coronal mass ejection based on multi-feature fusion

Abstract

The invention discloses a coronal mass ejection detection method based on multi-feature fusion. By segmenting the CME differential image, the CME detection is modeled as a classification problem of the brightest block in the differential image; first, the original image is converted into polar coordinates secondly, since the typical performance of CME is a bright enhanced structure with complex texture, therefore, find the brightest block of CME, use the brightest block as the representative of the image, and extract the grayscale, texture and HOG features of the brightest block ; Finally, based on the extracted grayscale features, texture and HOG features, the decision tree is used as the base classifier, and the integrated decision tree is used to complete the detection of CME. Experimental results show that the proposed multi-feature fusion CME detection algorithm can achieve better CME detection results.

Description

technical field [0001] The invention relates to a detection method for coronal mass ejection based on multi-feature fusion. Background technique [0002] Coronal mass ejections (Coronal Mass Ejections, CMEs) are a frequent eruption phenomenon in the solar atmosphere, which is characterized by obvious changes in the structure of the corona on a time scale of minutes to hours, accompanied by observable mass ejections, usually in the form of A bright, complexly textured enhanced structure followed by a dark region of insufficient brightness. CME is not only a transient phenomenon, it may also play an important role in the long-term evolution of the corona; at the same time, it has a close relationship with many interplanetary disturbances, and can cause drastic changes in the Earth's space environment; and the intensity of CME, Angles can have a significant impact on space weather, so the quantitative study of CME is of great significance to the disciplines of solar physics an...

AI Technical Summary

Problems solved by technology

However, it uses the number of bright spots as the grayscale statistical feature, ignoring the texture information of the CME structure; on the other hand, it is difficult to describe the radial propagation mode of CME by using the rectangular block cutting method

[0006] In short, the existing research in the field of CME detection often relies on the detection methods of single features such as grayscale and texture, and most of them use traditional digital image processing technology for detection, so it is difficult to obtain good detection results

Images