Projects / Programmes
Automatic procedures for the identification of changes in the actual use of agricultural land
Code |
Science |
Field |
Subfield |
2.17.00 |
Engineering sciences and technologies |
Geodesy |
|
Code |
Science |
Field |
T181 |
Technological sciences |
Remote sensing |
Code |
Science |
Field |
2.07 |
Engineering and Technology |
Environmental engineering
|
actual land use, farmland, remote sensing, automatic identification of changes, methodology
Researchers (13)
Organisations (3)
Abstract
The common agricultural policy is one of the most important policies of the European Community. Common agricultural policy mechanisms include the direct payments, which represent a steady income for farmers and have a positive impact on the preservation of agricultural production and preservation of cultural landscape in Slovenia.
Currently the data layer of the actual land use of the agricultural and the forest land represents the basic source for determining eligibility for obtaining direct payments and is updated on the basis of national orthophotos in three-year cycles. In some areas the information about actual land use is more than three years old, which causes problems in the process of determining eligibility for direct payments. It is therefore necessary to introduce an alternative source of information about the actual situation in the area and optimization of process data maintenance.
Automatic procedures for identification of land cover changes using various remote sensing data have unexploited potential to improve the quality of data on the actual land use of the agricultural and the forest land. The usage of automated processes for classification of land use changes can also contribute to the efficiency of the maintenance procedures. The proposed project will contribute to up-to-date status and quality of the actual land use to reduce irregularities in the agricultural beneficiaries reports for individual measures of the common agricultural policy.
Up-to-date status of actual land use of agricultural and forest land is closely linked to the availability of state orthophotos, which are produced in the national projects of the Cyclical aerial photographing of Slovenia (CAS). Actual land use of the agricultural and the forest land is updated every year only for the part the national territory for which aerial images and ortophoto was produced in this year (CAS was carried out). Aerial photography is carried out in a period of two to four years, in exceptional circumstances this period may be extended, due to for example adverse weather conditions at the time of recording. Thus, the state orthophotos as a primary source of information for updating the actual land cover use of agricultural and forest land, sometimes do not meet the requirements with regard to up-to-date status. Possible solution represents the use of alternative sources for actual land use updating by means of remote sensing data such as optical satellite imagery, radar imagery and lidar data.
Updating actual land use of agricultural and forest land is based on photointerpretation of state orthophotos and a numerous auxiliary data. Updating is conducted in two stages; in the first stage each polygon is review and, if necessary, updated by operator – photointerpreter. In a second stage, the updated data are confirm and, if necessary, correct by the operator - controller. The two-stage updating process provides quality data, but it needs a lot of resources in time or staff to be conducted. Updating procedures may be improved by the introduction of at least partially automated methods in the production line. This means the introduction of partly automated classification of land use changes, when the operator would check only the areas identified as potential areas with land use change by manual photointerpretation. In this procedure the operator would make targeted updating, he will check and correct only the areas where this method automatically detect changes. In this way we can optimize the updating of the actual land use of agricultural and forest land.
This proposal for a research project is based on the use of alternative data sources and partially automated maintenance procedures to ensure the effective updating of the actual land cover use of agricultural and forest land. We expect to focus on automatic identification of key changes in actual land use of agricultural and forest land: the reforestation of agricultural land, the urban a
Significance for science
This project is designed entirely as an applied expertise, which will be directly applicable for the client. This is also the intention of this target research project call. It will establish an improved methodology and the process of determining the changes of actual land use, which will be supported by practical instructions for the client.
From the standpoint of science and profession, the key innovative aspect is the use of advanced remote sensing methods and a combination of different techniques to obtain the source data (satellite, aerial). The process of data acquisition with the use of existing and selected new source data will not be more expensive than in the classical procedure. From a scientific point of view, a substantial shift from manual to automated methods of land use change identification will be made in a way to maximize the probability of correct recognition for the type of land use change, which is an unsolved problem on a global scale. In this area, the project team has internationally recognized results, which were, inter alia, also acquired in the Centre of Excellence Space.si, which is preparing to launch Slovenian microsatellite for remote sensing.
Innovativeness of the proposal lies in a partial replacement of the classic data source with the freely available satellite and aerial imagery in the process of determining the actual changes in land use. Classical data source (digital orthophoto) has a fixed and therefore too long period of updating, which may lead to wrong decisions in the allocation of direct payments to farmers. Semantically more correct way of definitions and classifications of the types of changes of actual land use will be utilized, because interfering objects and shadows are incorrectly and quite frequently detected as a change in actual land use. The examples of good practice abroad will be considered. They will increase the professional relevance and international comparability of the project. The recommended and prescribed information to control the allocation of direct payments, requested by the European Commission, will be upgraded with alternative data of various new platforms (aerial laser scanning, Sentinel-2 satellite, radar imagery). Scientific findings show that the optimal results in the automatic detection and identification of objects and their changes are achieved through a combination of data sources and not just by using classical data sources like digital orthophotos.
Furthermore, the definition of the layer of land use changes will be focused only on objects and areas that are relevant to the client and for agricultural production. This will allow the use of narrowly targeted methods that may be more effective than general methods. The focus will be made on at least the following changes: overgrowing permanent grassland, expansion of urban sprawl on agricultural lands, and deforestation. The novelty will be also a semantic and numeric definition of the characteristics of changes of actual land use such as the type of changes, size and time component. This will increase the objectivity and usefulness of the results.
We will experiment with both free as well as commercial satellite imagery, and, additionally with airborne laser scanning data, which in Slovenia are free of charge. A scientific contribution will be in evaluation of the free satellite imagery of lower resolution ("aggregated data") in relation to commercial high resolution images. Finding a compromise solution is likely to give the optimum result. To identify changes in actual land use dense clouds of points from laser scanning will also be used. Because the free (open) data from the Sentinel-2 satellite and national laser scanning will be used, the research will also become interesting internationally as laser scanning data for the whole country do not exist everywhere. Even if they do exist, they are generally not available free of charge.
Several approaches to identify land use changes w
Significance for the country
The layer of actual agricultural and forest land use covers the area of ??entire Slovenia. It is updated periodically. Because of these two features, the data layer is directly or indirectly used in other processes, even if it has certain shortcomings for other purposes. Based on the Real Estate Recording Act (ZEN, 2006) the land use data are connected by graphical cross-section with land parcel data in the land cadastre. The actual land use data are one of the most important data for mass real estate evaluation and consequently for the determination of a land property tax. In the context of the compensation for the change of agricultural land use under the Agricultural Land Act (ZKZ-D, 2012) it is necessary to indicate the the type and area of the actual land use pertaining to the floor area of ??the building, while the actual land use has to be taken from the records of actual land use (Article 3f). The type of actual land use influences the value of the cadastral income on the basis of the Act of determination of the cadastral income (UKD-1, Official Gazette of RS, no. 9/2011, 47/2012). Data about the type of property, which include information on the type of actual land use, is also used as an input in social transfers (Article 61 ZSVarPre, 2011, and revisions). On the basis of the Spatial Planning Act (ZPNačrt, 2007) the topographic data for spatial planning have to contain the actual land use and the calculation of differences between the actual and planned land use, where the most important feature is the areal difference between the category of built-up and related land versus the planned building space.
At the same time the actual land use and land cover are fundamental data to obtain basic knowledge about the area. Changes in the volume of a particular category of land use and transitions between categories are often a reflection of the socio-economic changes (Antrop, 2005). Due to this fact the status of land use remains one of the starting points of spatial planning and a subject of technical and scientific surveys. Details of land cover and land use are also important for statistical analysis, and are used in various indicators (eg. Floor and surface TP03, environmental indicators, EARS). Country-level data are important in an international context, for example at the European level (comparison of the condition and land use changes in different countries). Also, in order to standardize the environmental monitoring of land cover the European institutions merged under the umbrella of Project EAGLE (Internet 1) in order to enable harmonization of data structures and to compare data. Actual land use data is in line with the European INSPIRE directive as a part of national spatial data infrastructure.
Recording of land use and land cover, and monitoring of their changes is not only important in terms of the effectiveness of agricultural policy, but has a broader social significance. It is related to virtually all government departments as well as to private sector, particularly in the design of new information services. If you actually utilize land use data as a time series from several consecutive periods, they become interesting for the development of IT and for the geoinformation companies. Spatio-temporal applications of past landscapes have enormous potential for use in the services of local, national, transnational and international character. Spatio-temporal development of actual land use and land cover in the near and distant history had an influence on virtually all economic and social activities: urbanization, transport, agriculture, forestry, hydrology, tourism, cultural heritage, quality of life, ecology, climate and demographic change.
From the actual land use one can carry out a number of spatial indicators and evaluations, which are a central issue in several European tenders of cross-border relevance. When data on the actual land use become open public sector information (in accordance with Pub
Most important scientific results
Annual report
2016,
final report
Most important socioeconomically and culturally relevant results
Annual report
2016,
final report