A method and system for reducing clutter in a number of images. Sub-image pixel values are identified for a sub-image at a location in an image in the number of images. A first portion of the sub-image pixel values corresponds to a number of suspected target pixels in the sub-image. A second portion of the sub-image pixel values corresponds to clutter pixels in the sub-image. Modeled pixel values are generated using a clutter model fitted to the second portion of the sub-image pixel values. The modeled pixel values are subtracted from the sub-image pixel values to form new pixel values.