Técnicas de procesamiento
Calibración, recuperación y validación
Técnicas para los primeros niveles de datos producidos en teledetección, incluyendo métodos de calibración y estimación de errores
Image formation and calibration in SAR
The development of the Ground Segment for PAZ mission by INTA has given capabilities to process diverse SAR signals
- Development of an adapted filter to improve object discrimination
- Internal and external calibrations are applied
- Different processing modes
Calibration of image sensors on board drones using low-cost effective panels
Calibration of hyperspectral data must be done frequently in order to ensure the quality of the data. In situ data for calibration require the deployment of panels, that must be light weighted and affordable, in order to repeat the operation as frequently as possible. SpecLab has produced a new generation of calibration panels with good spectral signature and low cost that can be used to confirm the radiometric quality of multi or hyperspectral cameras onboard drones.
Correction of RFI errors in interferometry
The presence of sources of Radio Frequency Interferences can damage very badly interferometric images, as its effect spread all over the image. By the method of Nodal Sampling, this impact can be greatly reduced: the original image is oversampled and it is resampled at the locations where the nodes of the oscillation are found, reducing the distortion.
Differential Syntetic Aperture Radar
Advanced Differential Synthetic Aperture Radar Interferometry (DInSAR) is a microwave remote sensing technique that allows us to investigate surface deformation phenomena with a centimeter to millimeter precision and with a large spatial coverage capacity.
Copernicus Sentinel-1 mission allow us to study almost any geophysical event (earthquakes, volcanoes, anthropogenic activities, etc..) in the world with a low revisit time (between 6 and 12 days) and an extensive historic archive.
Inverse problem models of the deformation sources
Solving the inverse problem by modelling process allow us to obtain the sources of the observed deformation without making a priori assumptions like the geometry or the size of the sources. Combining this powerful models with high resolution and high precision techniques like DInSAR, LIDAR or microgravimetry can lead to results with unprecedented level of detail which can help in the understanding of the physics behind different natural events like volcanos or earthquakes.
Estimation of measurement errors using Triple Collocation
It is possible to infer the error associated to three different sensors/systems measuring the same variable by intercomparing them. The Triple Collocation method allows to estimate the standard deviations of the errors of three collocated datasets by means of an appropriate algebraic combination of its statistical order-2 moments. The method accounts for the different spatial and temporal representation of the data sources (r2), a key parameter for an accurate estimation of the errors.
Estimation of error of maps using Triple Collocation
When comparing satellite-derived and/or model fields of the same variable, error maps can obtained using Triple Collocation, by either estimating r2 or by using products of similar spatial and temporal resolution even if two of them are correlated.