Radiometric calibration and correction converts the sensor’s digital numbers (DNs) to radiance values and subsequently reflectance values. Additionally, the term “correction” points to the fact that radiometric measurements with satellite sensors contain error. Therefore, radiometric correction is concerned with improving the accuracy of surface spectral reflectance, emittance, or back-scattered measurements obtained using a remote sensing system. The Earth’s atmosphere, land and water are complex and can never be captured perfectly because of the limitations of remote sensing devices that lie in their spatial, spectral temporal and radiometric resolution. Therefore, error occurs in the data acquisition process and degrades the quality of remotely sensed data. The most common errors in remote sensing are radiometric and geometric. This concept is focused on the correction of remote sensing data to account for radiometric error that is to some degree systematic. Systematic errors in radiometric measurements come from the interaction of the sensed radiance with the atmosphere, the acquisition geometry in relation to the radiance source (the sun) and the Earth surface geometry (terrain). There are several levels of radiometric calibration and correction. The first is sensor calibration that converts the DNs to top-of-atmosphere (TOA) reflectance. It converts to radiance values and further to reflectance values by accounting for the viewing angle and sun angle during acquisition. The second is atmospheric correction that converts TOA reflectance to bottom-of-atmosphere (BOA) reflectance. The third is topographic correction that converts BOA reflectance to surface reflectance. Radiometric calibration is necessary to ensure radiometric comparability of the measurements. There is a need for calibration when comparing different spectral bands within one image, e.g. for the calculation of geo-biophysical parameters with band math operations. Results from uncalibrated image data would differ from results achieved with calibrated data because the unaccounted cal_gain and cal_offset of the used spectral bands would lead to distortions. In addition, radiometric calibration complements the geospatial comparability that is achieved with geo-referencing an image to geographic coordinates. Geo-referencing enables comparison of an image pixel to the geospatially matching pixel in another image acquired with a different sensor but with comparable resolution. Radiometric calibration enables a radiometric comparison between these two pixels’ radiance values. In case the two images are from different acquisition dates, a calculated radiometric difference would indicate change. This example shows the relevance of radiometric calibration for inter-sensor comparisons. Radiometric comparability is particularly relevant in studies that require inter-sensor comparisons, comparisons of surface features over time, or comparisons to laboratory or field reflectance data. Then the radiometric correction should cover atmospheric, solar and topographic effects. A full radiometric correction that also includes topographic correction can benefit the accuracy of image classifications by reducing the internal variability of vegetation types, since the corrected reflectance relates better to the geometrical or biological properties of the plant than to the original reflectance.
Transform imagery into radiometrically/atmospherically corrected state
In progress