A space-time adaptive detector for moving targets based on the analysis of variance

Abstract

We present a new methodology for detecting moving objects based on the analysis of variance. This new detector is a generalization of feature based and optical-flow based techniques. Our algorithm is effective in enhancing the detection of range-spread/Doppler-spread targets and in suppressing background interference, with application to synthetic aperture radar and high resolution millimeter wave imaging sensors, in addition to optical and infrared cameras. The statistics necessary for implementation are estimated under both the null hypothesis and its alternative. The performance is demonstrated via analysis of measured data.