Detecting dim moving point targets using transformation of pixel statistics

A new algorithm based on track-before-detect (TBD) technology is proposed for the detection of dim moving point targets of unknown position and velocity in a sequence of digital images collected from a staring sensor. The algorithm, using the transformation of the pixel statistics on a composite image produced by integrating a few differential images, then performs the detection process followed by the TBD target enhancement procedure. This algorithm is applied to a variety of image sequences with moving clutter backgrounds that are usually supposed to be non-Gaussian processes. Compared with other algorithms, the new algorithm is not only computationally simple, but also has high performance. Theoretical and experimental results are also given.