Despite the significant progress that was achieved throughout the recent years, to this day, automatic change detection and classification from synthetic aperture radar (SAR) images remains a difficult task. This is, in large part, due to (a) the high level of speckle noise that is inherent to SAR data; (b) the complex scattering response of SAR even for rather homogeneous targets; (c) the low temporal sampling that is often achieved with SAR systems, since sequential images do not always have the same radar geometry (incident angle, orbit path, etc.); and (d) the typically limited performance of SAR in delineating the exact boundary of changed regions. With this paper we present a promising change detection method that utilizes SAR images and provides solutions for these previously mentioned difficulties. We will show that the presented approach enables automatic and high-performance change detection across a wide range of spatial scales (resolution levels). The developed method follows a three-step approach of (i) initial pre-processing; (ii) data enhancement/filtering; and (iii) wavelet-based, multi-scale change detection. The stand-alone property of our approach is the high flexibility in applying the change detection approach to a wide range of change detection problems. The performance of the developed approach is demonstrated using synthetic data as well as a real-data application to wildfire progression near Fairbanks, Alaska.

This research adapts an effective change detection approach originally applied to mapping fire scar from a stationary synthetic aperture radar (SAR) scene to the problem of oil spills from SAR data. The method presented here combines several advanced image processing techniques to mitigate some of the common performance limitations of SAR-based oil spill detection. Principally among these limitations are the following. First, the radar cross section of the areas affected by an oil spill strongly depends on wind and wave effects, and is, therefore, highly variable. Second, the radar cross section of oil covered water is often indistinguishable from other dark ocean features such as low wind areas, which leads to errors and uncertainties in oil spill detection. In this paper, we introduce a multi-image analysis, the Lipschitz regularity (LR), and wavelet transforms as a combined approach to mitigate these performance limitations. We show that the LR parameter is much less sensitive to variations of wind and waves in an oil spill SAR imagery, lending itself well for normalizing and suppressing ocean background using image ratio processing. We combine LR processing with a multiscale technique based on the wavelet transform to additionally achieve high-quality noise suppression. To describe the performance of this approach under controlled conditions, we applied our method to simulated SAR data of wind-driven oceans with oil spills and low wind areas. We also applied our method to several real-world oil spill data acquired during the 2010 Deep Water Horizon spill in the Gulf of Mexico.

The traveling public judges the quality of a road mostly by its roughness and/or ride quality. Hence, mapping, monitoring, and maintaining adequate pavement smoothness is of high importance to State Departments of Transportation in the US. Current methods rely mostly on in situ measurements and are, therefore, time consuming and costly when applied at the network scale. This paper studies the applicability of satellite radar remote sensing data, specifically, high-resolution Synthetic Aperture Radar (SAR) data acquired at X-band, to the network-wide mapping of pavement roughness of roads in the US. Based on a comparison of high-resolution X-band Cosmo-SkyMed images with road roughness data in the form of International Roughness Index (IRI) measurements, we found that X-band radar brightness generally increases when pavement roughness worsens. Based on these findings, we developed and inverted a model to distinguish well maintained road segments from segments in need of repair. Over test sites in Augusta County, VA, we found that our classification scheme reaches an overall accuracy of 92.6%. This study illustrates the capacity of X-band SAR for pavement roughness mapping and suggests that incorporating SAR into DOT operations could be beneficial.