Spaceborne multibaseline SAR tomography for retrieving forest heights
Abstract/Contents
- Abstract
- Climate change is perhaps the biggest challenge of our time, impacting all life on Earth. During the last 50-100 years, changes in global climate have been much bigger and happened much faster than any such changes the planet has ever seen in recorded human history. Forests play a key role in the global carbon cycle, and hence in the global climate. The yearly carbon flux from forests to atmosphere is still an order of magnitude greater than that from human activities. However, the role of forests remains poorly quantitatively characterized as compared to other ecosystems, due to the practical difficulties in measuring forest biomass stocks globally. Without a quantitative model of the global carbon cycle, sensible policies and actions can be hard to make. Synthetic aperture radar (SAR) is a promising candidate for measuring forest biomass. Since low frequency microwaves penetrate the forest canopy, SAR systems operating at low frequencies can provide essential information in the elevation direction for assessing the forest biomass. In addition, a significant number of satellites equipped with SAR sensors will be scheduled and launched in the next five to ten years. Vast amounts of image data will be generated by current and future spaceborne SAR systems. This calls for new algorithms to exploit the diverse SAR images for forest mapping. In this thesis, we present a new technology, called SAR tomography, to retrieve forest heights. SAR tomography utilizes multiple images acquired from slightly different orbital locations to form a synthetic aperture in the elevation direction. Hence, it can achieve a full three dimentional (3-D) reconstruction of the imaged scenes. We first introduce the basic concepts of the SAR imaging system, and then establish the signal model for SAR tomography. Based on the signal model, we propose the Capon beamformer as the reconstruction algorithm and test its performance via simulation. In the second part, we investigate particular problems when using spaceborne repeat-pass SAR systems on forest areas, including phase errors, forest distributed scatterers and temporal decorrelation. For each problem, we propose a mitigation method and evaluate the performance via simulation. In the last part, we show two case studies using the proposed SAR tomography to reconstruct the elevation profiles in urban and forest areas. The SAR images used in this thesis are from the Advanced Land and Ocean Sensor (ALOS).
Description
| Type of resource | text |
|---|---|
| Form | electronic; electronic resource; remote |
| Extent | 1 online resource. |
| Publication date | 2017 |
| Issuance | monographic |
| Language | English |
Creators/Contributors
| Associated with | Lin, Qiuhua |
|---|---|
| Associated with | Stanford University, Department of Electrical Engineering. |
| Primary advisor | Zebker, Howard A |
| Thesis advisor | Zebker, Howard A |
| Thesis advisor | Linscott, Ivan |
| Thesis advisor | Pauly, John (John M.) |
| Advisor | Linscott, Ivan |
| Advisor | Pauly, John (John M.) |
Subjects
| Genre | Theses |
|---|
Bibliographic information
| Statement of responsibility | Qiuhua Lin. |
|---|---|
| Note | Submitted to the Department of Electrical Engineering. |
| Thesis | Thesis (Ph.D.)--Stanford University, 2017. |
| Location | electronic resource |
Access conditions
- Copyright
- © 2017 by Qiuhua Lin
Also listed in
Loading usage metrics...