SAR imaging: From Intensity to Interferometry and Tomography

The scope of this thesis is to report my research and perspective about the methodology and potential applications of the remote sensing in urban and natural scenarios through synthetic aperture radar (SAR) data for four years in the TETIS laboratory. It provides a picture of my main activities from SAR intensity to interferometry and tomography. Mapping agricultural land cover is a challenging problem in SAR. This is due to the speckle noise nature of radar, leading to a less intensive use of radar rather than optical images. We exploited temporal filtering to reduce noise while retaining as much as possible the fine structures present in the images. We proposed recurrent neural network (RNN) models to intelligently exploit temporal dependency among data. We demonstrated the suitability of C-band Sentinel-1 time series intensity data for agriculture land cover applications and revealed that the results of the proposed RNN models clearly outperformed classical machine learning approaches (e.g., Support Vector Machine and Random Forest).Recent advances in multi-temporal spaceborne SAR interferometry, especially with a persistent scatters (PS) interferometry approach, has made this a power technique for measuring ground subsidence. Here we exploited an advanced technique, which jointly manages the estimation of both PS and DS (distributed scatters) targets. We demonstrated this robust technique using L-band ALOS images for providing an unprecedented spatial extent and continuous temporal coverage of the subsidence with millimetric accuracy in Ho Chi Minh City, Vietnam. This confirms the promising PS/DS approach to monitor ground subsidence.New technologies are urgently needed to observe forest biomass stocks globally, providing data for efforts to mitigate greenhouse gas emissions by reducing deforestation and forest degradation. To meet this challenge, the next Earth Explorer BIOMASS will collect P-band SAR data to produce three-dimensional maps of most of the world’s forests by using SAR tomography. We demonstrate that the high sensitivity of SAR tomography to forest vertical structure enables to measure the full range of biomass values. Interestingly, tomographic data are significantly correlated with biomass in the upper vegetation layer (i.e., 20–40 m), where the influence of large trees predominates. These findings increase our confidence that BIOMASS can provide accurate mapping of global forest biomass.