Image forensics has become increasingly important in our daily lives. As a fundamental type of forgeries, Copy-Move Forgery Detection (CMFD) has received significant attention in the academic community. Keypoint-based algorithms, particularly those based on SIFT, have achieved good results in CMFD. However, the most of keypoint detection algorithms often fail to generate sufficient matches when tampered patches are present in smooth areas. To tackle this problem, we introduce entropy images to determine the coordinates and scales of keypoints, resulting significantly increasing the number of keypoints. Furthermore, we develop an entropy level clustering algorithm to avoid increased matching complexity caused by non-ideal distribution of grayscale values in keypoints. Experimental results demonstrate that our algorithm achieves a good balance between performance and time efficiency.