Two-channel Sar Ground Moving Target Indication for Traffic Monitoring in Urban Terrain

This paper introduces and analyses a technique that implements CFAR detection and parameter estimation of moving targets in urban terrain. Firstly, it presents a probability distribution to use with the product model. This model helps to account for the extremely high inhomogeneity encountered in urban terrain. Along with the proposed distribution, the paper then discusses the numerical computation of the CFAR thresholds for an arbitrary GMTI detection metric. For illustration, results have been generated using DPCA as the GMTI metric. Unfortunately, having achieved CFAR for a small probability of false alarm, one finds that the probability of detection has decreased. To counter this, we propose to increase the target SCR by using a SAR processing filterbank where filters in the bank are designed to enhance moving targets since an increase in SCR improves detection [11, 3]. Following the filterbank operation, we construct a master target list of redundant information about the individual targets. This information is then exploited in a quasi-optimal manner to estimate the target velocities and locations while filtering out false alarms that have passed the CFAR test.