Signal processing is a powerful tool in environmental sciences and physics. In my PhD project I am taking techniques developed for processing Lidar and Laser Scanning data and applying it to full-waveform data obtained using radar. The processing steps are likely to be analogous, but the application of the algorithms will differ significantly. In essence, the techniques are optimised to look for targets along a line of sight (i.e. along a waveform) and to obtain a measurement of distance. Detection methods include, but are not limited to:
- Gaussian decomposition
- Deconvolution of the radar antenna pattern with the signal echo
- Geospatial waveform stacking
- 3D segmentation for enhanced data visualisation
I am also particularly interested in understanding the impact of the Radar (or Lidar) beam divergence on the backscattered signal and surface elevation retrieval. This of major relevance in determining the accuracy of a ranging system and can potentially be used to derive additional measurements such as local slope values, improved Normalised Radar Cross Section (NRCS) and additional range measurements. Also of interest is the derivation of classification algorithms using the signal echo and the backscattered power, also denoted by the NRCS. While I am currently building algorithms to distinguish between snow, ice and rock, I am also looking to apply Machine Learning to this area of research.