The country-wide monitoring service generates CBK-Floods product, which provides the current surface water cover over Poland. The automatic detection algorithm has been developed. It uses Copernicus Sentinel-1 radar satellite images as well as proposed improved model of valleys derived from fusion of various data sources (e.g., Copernicus Riparian Zones, LIDAR, flood hazard zones). The overall accuracy of the algorithm is around 86%. The map is updated after each pass of the satellite and shows different stages of inundation: new water extent, areas with long-lasting water and those from which water has receded in the last days. Two kinds of information are generated: (1) flood water extent; and (2) hydroperiod regime. Information about flood water extent is of critical importance for rescue and crisis management activities. Availability of recent water cover maps can support rapid situational assessment and influence decision processes taken in regional and local crisis management centers during flood. Information about hydroperiod regime allows the proper management of water needed for agriculture and can be an indicator of the state of ecosystems present in the valley. In 2022 service worked in pre-operational mode and produced a series of surface water maps for the entire Poland. In 2023 service will go into operational mode. The water extent maps will be available to visualize in the Sat4Envi Crisis Management Portal and downloadable from its repository. In this paper, we aim to present the data processing chain applied in the flood monitoring system, including the surface water detection method and the way of visualizing the final product. We present the limitations of the service based on satellite radar data and give examples of the use of flood products.
Post-processing is the last and often optional stage of land cover (LC) classification from satellite images. In the traditional approach, it is usually applied to remove the effect of “salt and pepper” from the classified image and also to standardize the image details according to the defined minimum mapping unit (MMU). The proposed post-processing method presented in this paper, has been used in the Sentinel-2 Global Land Cover (S2GLC) project. Its main goal is to remove or minimize typical classification errors that can appear in the classification output. Therefore, a set of functions that are able to improve the result of LC classifications has been developed. These include relatively simply defined rules that operate based on predefined threshold values of selected spectral channels, spectral indexes or auxiliary data. Additionally, logical relations between certain LC classes have been implemented. The proposed post-processing has been applied to the classification results of the S2GLC project and helped to improve LC classification in all test sites representing different parts of the globe.
Supervised classification of satellite images is performed based on utilization of reference training data. Therefore, the availability and quality of reference data highly influences the results and the course of the entire classification process. In the Sentinel-2 Global Land Cover (S2GLC) project Sentinel-2 images are classified using Random Forest (RF) algorithm powered by training points selected from existing low resolution land cover databases. This approach allows to perform the classification process in a highly automatic manner without much intervention of an operator. An alternative method for creating training dataset has been developed in order to ensure the implementation of the S2GLC classification in case of limited access to the required land cover databases or their low quality. The proposed method is a semi-automatic process initiated by an operator, who by a visual interpretation, indicates only several starting samples for the classes of interest. Afterwards, utilizing this limited set of initial training samples, hundreds or thousands of training samples with similar spectral characteristics are automatically selected from the image. Such a set of data, can be further used as an alternative source of training data for land cover classification on much greater scale. Comparing to the traditional approach, in which all samples or training areas are manually indicated, the developed method is very effective and also allows for processing data more rapidly. The semi-automatic training can be used as an alternative or supplement the training dataset applied in the S2GLC classification approach.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.