:Cold image is the important influence factor for the imaging quality of cooled infrared detector.The generation mechanism and frequency characteristics of cold image are analyzed,and an adaptive correction algorithm is proposed.Firstly,the adaptive time-domain low-pass filter was used to obtain an estimated cold image,and then a correlation operation between the estimated image and the pre-stored reference image was carried out.Finally,the noise matrix update was determined and the images were corrected according to the correlation coefficient.Simulated experiments and actual infrared image sequences show that the proposed algorithm can effectively remove the cold image noise.