A novel progressive estimation scheme is proposed to detect moving targets in infrared video sequences in this work. The proposed algorithm does not place constraints on camera motions nor requires rigid backgrounds. The complexity of the algorithm is low so that it can be performed in real time. Computational results of the algorithm applied to a real-world infrared video sequence are given to demonstrate its superior detection performance.
Shahryar Karimi-Ashtiani, C. C. Jay Kuo