A change detection statistic for repeat pass interferometric SAR

In repeat pass synthetic aperture radar interferometry (In-SAR) the scene coherency can be used to detect subtle scene changes that may not be evident in the image magnitude data alone. The sensitivity of the coherency estimate for identifying man-made disturbances is dependent on the size and nature of these disturbances compared to the other sources of disturbance such as wind, system noise and processing aberrations that affect the scene images. In this paper the detection problem is formulated in a hypothesis testing framework and a new change statistic is proposed. Analytic expressions for the probability of detection and false alarm are derived which show a significant improvement over the commonly used coherence based change detection. Application of the new change statistic to experimental data obtained using the DSTO Ingara SAR demonstrates the improved detection performance.