中文摘要：

尝试将相位一致性方法增强的图像用于冕环结构的识别．针对相位一致性方法增强图像，提出一个新的冕环识别方法（简称PCB），该方法依据冕环形态变化比较平缓的形态学特性，将冕环识别方向限制在一个较小的范围之内，并且首次同时将识别过程中推进方向的变化情况和识别点的相位一致性大小作为终止识别的判据．对多组数据的识别和分析则表明相位一致性方法增强的图像的确适合用于冕环识别，而PCB方法所识别的冕环结构则同时具有很好的完整性和准确性，是一套现实可行的冕环自动识别方法．

英文摘要：

We test a phase congruency method for automatically extracting coronal loops from the coronal observation data, and propose a new coronal loop detective method. On account of smooth changes of the direction along coronal loops, the possible direction of loops is restricted in a small range for improving the detective result. Beyond that, inspired by the structural characteristics of coronal loops, we firstly suggest that both the change of result directions and the flux of result points can be used to terminate the loop detection. Finally, several coronal images are detected by our loop extracting method, and the result indicates that the congruency enhanced image is exactly suitable for coronal detection and extraction, and our new loop detective method can also extract most of loops in the test images completely and accurately. Besides that, our research also suggests that the characteristics of coronal loop should be a significant reference for improving other loop detective methods.