A Novel Technique for Forest Height Estimation from SAR Radar Images Using the Omega K Algorithm

The height of forest canopy is a valuable information to estimate or predict of aboveground live biomass in places such as natural reserves or parks. Many of these sites use satellite data based on Lidar or radars to run carbon stock models. The parks in low income countries can not afford to use these expensive solutions. Airborne and small radars tailor made for a given specific area can be deployed as alternative solution. In this paper, we present an algorithm that contains a complete technique for the signal processing needed to estimate canopy height. The algorithm includes a simulator of SAR raw data and SAR images based on impulse response. This algorithm enable testing of focus algorithms and generate three-dimensional images. We use a rigorous and extensive method to implement the SAR image processing part of the algorithm based on the Omega-k algorithm. To confirm the validity of the algorithm we use Quad polarized images of trees from a Polarimetric SAR simulator. Finally, we use SAR raw data acquired by the ERS-2 satellite to test and validate the overall process. The technique is shown to be effective and simple, since even signals with only one and two polarization are sufficient. Height results were compared with Range Doppler based techniques available in the literature.