Image Segmentation For Sun Activity In Sun Images

With the expansion of the astronomical observatory and the rapid development of solar technology, the number of the solar image also has the rapid growth, in order to carry out the solar aspects research, we need the full-disk solar images and solar activity images segmentation results for the subsequent related work. Because the image segmentation results will directly affect the quality of the subsequent work, and at present, there is no systematic analysis and comparison to confirm the most suitable image segmentation for the solar images, so in this thesis, the core problem is how to deal with the increasing solar images effectively and how to identify and segment the solar activities in the solar images accurately. To solve above problems, this thesis mainly includes three aspects of work.First of all, in this thesis, the solar image processing including the preprocessing and the solar activity image segmentation. In the pretreatment stage, after comparing various filtering denoising methods and according to the characteristics of solar activity image noise, the de-noising experiments are made with median filtering and mean filtering respectively to the image sample database, and MSE and PSNR are used for de-noising results analysis, the conclusion is that the median filtering algorithm can achieve better effect for solar image de-noising.Secondly, in the solar activity image segmentation, this thesis put forward an improved algorithm based on the seed point selection of the region growing, this improved algorithm combines the region growing and the edge detection, and it can effectively reduce the time consumption during image segmentation to improve processing efficiency. In addition, this thesis systematically compares the results of eight segmentation algorithms, which comprises the improved algorithm, and establishes a perfect performance evaluation system from two aspects of subjective and objective, the conclusion is that the region growing algorithm can achieve better effect for solar activity image segmentation processing work.Finally, according to the optimization algorithm for each stage of the solar image processing, this thesis propose the best process of the solar image processing, and it greatly increase the effectiveness and precision of the solar image processing.