Project enhances Earth observation and monitors vegetation using space-based data

Imagining a future where we can monitor the health of the planet's forests and crops with millimeter precision is no longer science fiction. This is the objective of INTERSEN, a project included in the 2021 State Research Plan and led by the Visual Engineering (eViS) research group at the Universitat Jaume I in Castelló, which is committed to the intelligent combination of spatial data to improve the way we understand and care for our environment.

The European Copernicus program is a key initiative for Earth observation and includes missions such as Sentinel and FLEX. Sentinel provides detailed information about the terrain and the atmosphere, while FLEX studies how plants use sunlight to grow. The inter-sensor data fusion developed by INTERSEN combines these capabilities to generate more complete and accurate images, maximizing the strengths of each satellite. For example, Sentinel-1 is capable of capturing data in adverse weather conditions, while FLEX offers a unique spectral resolution for analyzing photosynthesis.

To carry out this fusion, the INTERSEN team uses advanced technologies such as machine learning and image processing. These tools allow computers to analyze large amounts of information and produce highly detailed maps of fields and forests. This has direct applications in agricultural planning, water resource management and crop forecasting, as well as contributing to more accurate global climate models.

The project has already made significant progress, such as the development of algorithms that improve the accuracy of data fusion and enable highly reliable identification of inland water areas. For example, techniques have been created that combine data from different sensors to improve the spectral and spatial resolution of the images.

Even so, there are still challenges to be faced, such as the development of semi-supervised classification techniques to identify areas of vegetation and other types of land cover more accurately, as well as the improvement of algorithms for mapping and monitoring vegetation indices. These lines of work are planned for the second phase of the project and could lay the foundations for future research.

The eViS-Visual Engineering research group, part of the University Institute of New Imaging Technologies (INIT), has over 30 years' experience in the field of artificial vision and automatic learning, and has participated in over 40 competitive national and European projects. Its lines of work include shape recognition, color image analysis, visual texture analysis, stereoscopy and automatic classification techniques.

With all the advances of the INTERSEN project, the Universitat Jaume I exemplifies how applied research can offer real solutions to global problems, putting science at the service of society and building a more efficient, responsible and sustainable future for our planet.

Related research is published in Image Analysis and Processing – ICIAP 2023.

More information:
Itza Hernandez-Sequeira et al, Semi-supervised Classification for Remote Sensing Datasets, Image Analysis and Processing—ICIAP 2023 (2023). DOI: 10.1007/978-3-031-43148-7_39

Provided by Universitat Jaume I