Multisensor multiframe change detection via optimal likelihood ratio test

Based on the likelihood ratio test with the maximum likelihood estimation, optimal multi-sensor data fusion is addressed for the detection of changing scene in consecutive multiple image frames. It is shown that the information necessary for the sensors to transmit to the fusion center is the estimation of the correlation coefficients. The optimal threshold of the testing is obtained from the criterion of a constant false alarm. The experimental results illustrate that the performance of the detection scheme is improved when the number of the sensors and/or the frames increased. Furthermore, the changing areas of interest can be extracted accurately from the change detection algorithm combined with preprocessing and mask techniques.<>