Abstract
We propose a new feature-detection technique based on phase-congruency (PC) measurements to automatically recognize or enhance faint features in solar observations, such as off-limb coronal loops and umbral dots. Compared with other feature-detection methods that are based on gradient illuminance and imaging filtering, PC-based measurements are particular efficient for recognizing faint features, which generally have a low-intensity contrast to their surroundings. In the present article, we carry out a PC-based measurement of the local weighted mean phase angle (LWMPA) at each point in an image to indicate or highlight low-contrast features. We first used artificial images to check the detection accuracy and sensitivity to the noise of this approach. Subsequently, we applied this approach to an EUV observation obtained by the Solar Dynamics Observatory/Atmospheric Imaging Assembly to highlight off-limb coronal loops, and a photospheric observation obtained by the Hinode/Solar Optical Telescope to recognize faint dots within the cores of sunspots and pores. The results illustrate that this PC-based measurement of the LWMPA is a robust detection method for faint structures in solar observations.
Author
Feng, Song; Xu, Zhi; Wang, Feng; Deng, Hui; Yang, Yunfei; Ji, Kaifan
Journal
Solar Physics
Paper Publication Date
October 2014
Paper Type
Astrostatistics