In this paper, an efficient algorithm is developed to automatically detect and extract coronal loops. First of all, in the algorithm, three characteristics associated with coronal loops are used to construct a match filter able to enhance the loops. Secondly, the method combining a high-pass filter (unsharp-mask enhancement) with a global threshold is used to further enhance and segment the loops. Thirdly, to extract every individual coronal loop and obtain their parameters (the 2D projected space coordinates and lengths) from the segmented loops, a clustering method of the pixels with approximate local direction and connected domain is further used. Fourthly, to evaluate the performance of the developed algorithm, images observed by the Transition Region and Coronal Explorer (TRACE), the Atmospheric Imaging Assembly (AIA) of the Solar Dynamics Observatory (SDO) and the High-Resolution Coronal Imager (Hi-C) are used, and comparison experiments between the existing algorithms and the developed algorithm are performed. Finally, it is found that the developed algorithm is commensurate with the two most promising algorithms, oriented coronal curved loop tracing (OCCULT) and its improved version, OCCULT-2, in performance. Therefore, for scientific applications associated with coronal loops, the developed algorithm will be a powerful tool.