Syllabus
1. What can be seen from space ?
- Physical principles of acquisition (active / passive sensors, light spectra, atmospheric absorption…)
- Geometry of satellite acquisition (orbits, geo-referencing,…)
- New missions of CNES, ESA, NASA, DLR
- How to download a video of your house from the sky
2. Geometric modeling of optical satellites and application to 3D reconstruction
- Geometric modeling of pushbroom sensors (example of Pléiades)
- Approximation of the geometric model (RPC, errors, precision, …)
- Application to stereo rectification for 3D reconstruction (stereovision, epipolar curve)
3. Modeling a Synthetic Aperture Radar instrument
- Synthetic aperture and chirp
- Geometric distorsions and geo-referencing
- Radiometry (backscattering properties, speckle phenomenon,…)
4. How to recover 3D information with optical sensors ?
- Stereovision overview
- Matching algorithms classification
- Local methods (matching costs, cost volume, aggregation, patchmatch)
- Match filtering
- Global methods (dynamic programming, semi-global matching)
5. Sub-pixel accuracy in stereo matching with low baseline
- Error sources in stereo matching: matching errors, fattening effect, noise, aliasing, interpolation errors.
- How to obtain images without aliasing ?
- Estimating and attaining subpixel accuracy in stereo matching
6. Generation and exploitation of 3D data
- DEM, DSM, DTM, orthoimage
7.How to recover 3D information with SAR sensors ?
- SAR complex signal information
- interferometry principles
- beyond 3D with differential interferometry
8. Processing and exploitation of SAR data
- despeckling approaches for complex vectorial data
- 3D reconstruction with non local or markovian methods
9. Time series analysis
- Optical time series analysis
- registration, normalization, change detection
- SAR time series analysis
- multi-temporal denoising
- change detection methods
- multi-temporal change analysis