Abstract: A small-target detection approach was proposed, which is based on the modified local reverse entropy operator and the improved region growing technology. The approach adopted the fact that the emergence of small target could lead to the large change of local reverse entropy. The proposed approach made use of the local reverse entropy map to suppress the small-target image background, and then adopted the improved region growing technique to effectively detect small targets. Experimental results show that the proposed approach can detect small targets accurately and quickly.