We introduce a novel neural network architecture-Spectral ENcoder for SEnsor Independence (SEnSeI)-by which several multispectral instruments, each with different combinations of spectral bands, can be used to train a generalised deep learning model. We focus on the problem of cloud masking, using several pre-existing datasets, and a new, freely available dataset for Sentinel-2. Our model is shown to achieve state-of-the-art performance on the satellites it was trained on (Sentinel-2 and Landsat 8), and is able to extrapolate to sensors it has not seen during training such as Landsat 7, Per úSat-1, and Sentinel-3 SLSTR. Model performance is shown to improve when multiple satellites are used in training, approaching or surpassing the performance of specialised, single-sensor models. This work is motivated by the fact that the remote sensing community has access to data taken with a hugely variety of sensors. This has inevitably led to labelling efforts being undertaken separately for different sensors, which limits the performance of deep learning models, given their need for huge training sets to perform optimally. Sensor independence can enable deep learning models to utilise multiple datasets for training simultaneously, boosting performance and making them much more widely applicable. This may lead to deep learning approaches being used more frequently for on-board applications and in ground segment data processing, which generally require models to be ready at launch or soon afterwards.
Opportunistic constant target matching is a new method for satellite intercalibration. It solves a long-standing issue with the traditional simultaneous nadir overpass (SNO) method, namely, that it typically provides only data points with cold brightness temperatures for humidity sounding instruments on sun-synchronous satellites. In the new method, a geostationary infrared sensor (SEVIRI) is used to select constant target matches for two different microwave sensors (MHS on NOAA 18 and Metop A). We discuss the main assumptions and limitations of the method and explore its statistical properties with a simple Monte Carlo simulation. The method was tested in a simple case study with real observations for this combination of satellites for MHS Channel 3 at 183 ± 1 GHz, the upper tropospheric humidity channel. For the studied 3-month test period, real observations are found to behave consistently with the simulations, increasing our confidence that the method can be a valuable tool for intercalibration efforts. For the selected case study, the new method confirms that the bias between NOAA 18 and Metop A MHS Channel 3 is very small, with absolute value below 0.05 K.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.