Normalized cross-correlation (NCC) measure has often been used for image matching due to its invariance under changes in image bias and gain. We address the problem of using it for pattern matching in practical imaging systems. It presents an empirical relationship between the contrast level in an image and its best-matched NCC. It further derives and confirms this relationship theoretically. A novel, adaptive method of adjusting the best-matched NCC based on the established relationship is subsequently presented to alleviate the problem of pattern matching in practical imaging systems. Experimental results on real scenes of both low and high contrast images are finally presented to demonstrate the usefulness of the proposed method.