CME是空间天气研究的基础内容之一，其检测易受各种伴随的或相关的物理过程影响，地基观测还受到大气散射、抖动和设备跟踪误差等影响，使得准确的CME自动检测依然是一个挑战。针对这一挑战，本课题旨在研究日冕观测序列图像中CME自动准确检测的关键问题。研究内容包括日冕图像预处理、观测场景建模、CME区域提取、CME形态学动力学特征提取。拟解决的关键问题包括：基于日冕图像具有多尺度、非刚体运动、有挡盘等特征，通过互信息、局部滤波等技术达到对日冕序列图像配准、滤噪；统计观测场景的静态、动态特征实现对观测场景的自适应描述；通过贝叶斯条件后验概率、形态学、特征匹配及机器学习方法实现CME检测。本课题将对LASCO等空间日冕仪观测的日冕图像序列为实验数据建立CME目录，并与CDAW等人工目录和CACTus等自动检测目录进行对比；同时将研究的日冕图像处理方法应用到我国近期投入使用的羊八井日冕仪观测的图像数据。

CME is the basic content of the study of space weather and its detection is easily affected by various accompanying or related physical processes, ground based observations also affected by atmospheric scattering, jitter and equipment tracking error. All above reasons make the accurately and automatically detecting CME to be still a challenge. In response to this challenge, the main purpose of this project is researching the key issues of how to accurately and automatically detect CME from observed images of coronal sequence. The research content includes preprocessing about observed images, scene model establishing, extract the CME candidate regions, extracting morphological features and evolution features from CME candidate regions. The key issues to be resolved by this project are: base on multi-scale, non-rigid motion, a baffle plate and other characteristics of coronal images, completing registration and filtering process of images by mutual information and local filtering technology, according to the statistics of static and dynamic characteristics to complete adaptive description about the observation scene, through the conditions of Bayesian posterior probability, morphology, feature matching and machine learning methods for detecting the CME. This project will use image data from LASCO to establish CME catalogs and compare it to manual catalogs such as the CDAW catalogs and automated CACTus catalogs, the ultimate goal is to provide new method which can bring high accuracy CME detection; at the same time, the study of coronal image processing method will be applied to the image data of Yangbajing coronagraph observations which will be put into used recently.