Projects / Programmes
Using Sentinel satellite imagery and selected other remote sensing sources for controlling direct payments in agriculture
| Code |
Science |
Field |
Subfield |
| 4.03.06 |
Biotechnical sciences |
Plant production |
Agricultural technics |
| Code |
Science |
Field |
| T181 |
Technological sciences |
Remote sensing |
| Code |
Science |
Field |
| 4.01 |
Agricultural and Veterinary Sciences |
Agriculture, Forestry and Fisheries |
remote sensing, agriculture, common agricultural policies, monitoring, satelite imagery, Sentinel, time series analysis
Researchers (11)
Organisations (2)
Abstract
Several remote sensing methods enable the identification of changes in use of agricultural surfaces. Slovenian agriculture has several specific traits, for example a highly fragmented parcel structure, which require an adequate system for monitoring justification of agricultural policy measures. Satelite imaging is an appropriate method for this purpose, as it can cover large areas, including difficult to access parts. Furthermore, a comparatively high imaging frequency furthers the detection of any changes in the field.
The main aim of this project is to implement an automated system for detecting change in agricultural fields, by using remote sensing data from different sources. The system will detect not only large changes (e.g. harvest, plowing, change of use, buiding), but also changes in crop type in the same growing season on the same field (e.g. main crop, non-overwintering and overwintering crops).
By carrying out a review of published research on the use of spatial data and ways of implementing monitoring to verify the viability of agricultural policy measures, we will achieve a state of the art implementation of the system for automatic change detection. Remote sensing analysis result will be available on the on-line platform established under the H2020 Project PerceptiveSentinel, coordinated by colleagues from Sinergise. The platform will combine all the acquired knowledge with other objectives of this project and enable users to quickly and accurately inspect current status and changes in agricultural land. In addition, we will test the products and services that will be created within this project as well as the H2020 PerceptiveSentinel project. We will provide testing of suitable products and services in Slovenia's specially fragmented agricultural area, which will be developed within these projects.
The final results of the project include the establishment of an on-line platform for automatic monitoring of changes in agricultural land using satellite and other remote sensing data. In order to achieve this goal, we will carry out expert literature reviews on the use of remote sensing methods in landscape control (agricultural and other areas), and establish spectral libraries of multi- and hyperspectral data, which will serve as learning patterns for the training of machine learning algorithms. New satellite recordings will be available every 5 days, due to the high imaging frequency of Sentinel satellites. The usefulness of multispectral satellite imagery depends on weather conditions, so the platform will also include other sensors to cover data holes and provide more accurate status estimates. In addition to spectral databases (multispectral satellite and UAV, and hyperspectral airborne), ground truthing databases will also be established. Both will be aimed at the development of algorithms and the validation of the developed methods.
Significance for science
The use of remote sensing methods is nowadays a constant in many scientific disciplines. However, it is mainly based on individual recordings, and the temporal sequences are mostly treated as individual recordings with longer time intervals between them. As part of this project, time series with high temporal resolution will be processed and will provide direct added value, in the form of detecting changes, identifying crops and calculating spectral indices. With this, the scientific community will receive a tool for processing large amounts of data in a time sequence, which will enable more reliable identification of objects, processes and changes.
Significance for the country
The system for automatic identification of changes in agricultural land is primarily intended for the Ministry of Agriculture or its bodies in the composition (AKTRP), which will be able to use it for direct control over the implementation of agricultural policy and for granting financial support to growers. In further development, it will be possible to add other functionalities to the system, for example, control of flood areas and fires.
Access, use and interpretation of satellite data have so far been difficult because they require extensive expertise and time. Also, many products that are designed to solve specific problems are not available. To increase the reliability of remote sensing results, it is also necessary to combine different data sources and integrate them into existing databases. Our system for automatic identification of changes on agricultural land will be accessible and understandable to the non-expert public, which will enable the development of new and optimization of existing products.
Most important scientific results
Annual report
2019,
final report
Most important socioeconomically and culturally relevant results
Annual report
2018,
2019,
final report
