An Ankle Ring Feature Recognition Method Based On Phase Consistency And Direction Filtering

There are a number of loop structures, called coronal loop on a coronal image. Coronal loops appear to loop shape due to hot plasma gas is brightening, they are a form of solar magnetic fields. Moreover, coronal loops have been connected to solar phenomena, such as coronal mass ejections, solar flares, and solar storms. Accurate identification and extraction of coronal loops is crucial to study and analyze solar magnetic field activities. However, the identification based on the intensity and/or gradient threshold has some limitation because coronal loops have a relatively low contrast, blurred boundary, and larger image noise. And there are non coronal loops and corona loops cross overlap in the coronal image. Therefore, we proposed a new method based phase congruency and directional filtering technique for identifying coronal loops.For illustrating the performance of the method, a high-resolution image taken with the Atmospheric Imaging Assembly on aboard the Solar Dynamic Observatory is used to describe the identification process of the procedure. The identification procedures are as follows:firstly, a median filter is used to remove image noise; secondly, the phase congruency technique is used to identify linear shape in an image; thirdly, a binary image is obtained to labeled the identified features; fourthly, a directional filter method is used to remove the redundant features that is nearly perpendicular to the target loops; fifthly, the morphological operators are used to fill holes, remove spur pixels and isolated points for extracting the loop shapes; finally, quadratic fitting is used to obtain the loops.In order to verify the feasibility and accuracy of the procedure, two methods are used in this thesis. First, the recognition result of multiple sets of date are used to compared with the existing algorithm. Furthermore, due to a solar flare eruption and a coronal mass ejection usually can cause coronal loop oscillations nearby. We analyzed two oscillatory loops'phase, oscillation period, and decay coefficient. Moreover, the difference between the two loops was compared and analyzed. We found that the result is similar to the existing literature.The result imply that our proposed method can accurately identify the loop shapes in an image, furthermore can be used to study the physical properties of coronal loops and the relationship between others solar phenomena.