After investigating the technique for amplitude ranking correlation，this paper presents an algorithm for quantizing and ranking grey levels of pixels of a sensor image. The optimum threshold for quantization has been found; after ranking a sensor image can be decomposed into several 0-1 coded images. Then the coded images will be correlated with a reference image without multiplication operations. For a give quantization function the means and variance of metrics for correlation can be calculated theoretically. Detection threshold sequences {T_k} for different signal to noise ratios can be obtained for the image of Gaussian distribution，Computer simulation shows that the detection threshold sequence can detect each match point effectively for the images of Gaussian distribution with different correlation lengths and with signal to noise ratios from 5 to 1. It is also shown that， for actual terrain maps with signal to noise ratios from 3 to 1，correct match can be reached.