Hans Hellsten, Saab AB, Electronic Defence Systems
When using radar with meter wavelengths for imaging the ground surface, the images produced have several salient features. Above all they highlight the presence of targets and suppress irrelevant clutter. However, image resolution is modest—of meter order. The imaging method is very useful in the context of change detection, comparing images of the same scene collected at different instances. By this comparison, targets that have appeared or disappeared are detected. An efficient scheme for change detection exploits Bayesian methods to establish relevant statistics from image data, in order to assess the probability for any change to represent a genuine appearance, or disappearance, of a target. Although performance of the scheme is high, it contains several numerically intense processing stages. Using the vector methods embedded in MATLAB®, the processing routines turn out to be surprisingly fast, even in the case of large data sets. The development indicates that MATLAB could be advantageously used for a wider set of cases for radar processing. The advantages of using MATLAB are obvious, as code can be accessed and modified during radar development and experimentation.
Recorded: 21 Apr 2016