Most important scientific results of the project group

  • Kanjir, Urška, Đurić, Nataša, Veljanovski, Tatjana (2018): Sentinel-2 based temporal detection of agricultural land use anomalies in support of common agricultural policy monitoring. ISPRS international journal of geo-information, ISSN 2220-9964, 13. Oct. 2018, vol. 7, iss. 10, 1-24,  https://www.mdpi.com/2220-9964/7/10/405, DOI: 10.3390/ijgi7100405 [COBISS.SI-ID  43706157]

The European Common Agricultural Policy (CAP) post-2020 timeframe reform will reshape the agriculture land use control procedures from a selected risk fields-based approach into an all-inclusive one. The reform fosters the use of Sentinel data with the objective of enabling greater transparency and comparability of CAP results in different Member States. In this paper, we investigate the analysis of a time series approach using Sentinel-2 images and the suitability of the BFAST (Breaks for Additive Season and Trend) Monitor method to detect changes that correspond to land use anomaly observations in the assessment of agricultural parcel management activities. We focus on identifying certain signs of ineligible (inconsistent) use in permanent meadows and crop fields in one growing season, and in particular those that can be associated with time-defined greenness (vegetation vigor). Depending on the requirements of the BFAST Monitor method and currently time-limited Sentinel-2 dataset for the reliable anomaly study, we introduce customized procedures to support and verify the BFAST Monitor anomaly detection results using the analysis of NDVI (Normalized Difference Vegetation Index) object-based temporal profiles and time-series standard deviation output, where geographical objects of interest are parcels of particular land use. The validation of land use candidate anomalies in view of land use ineligibilities was performed with the information on declared land annual use and field controls, as obtained in the framework of subsidy granting in Slovenia. The results confirm that the proposed combined approach proves efficient to deal with short time series and yields high accuracy rates in monitoring agricultural parcel greenness. As such, it can already be introduced to help the process of agricultural land use control within certain CAP activities in the preparation and adaptation phase.

  • Mesner, Nika, Tič, Katja, Mangafić, Alen (avtor, kartograf), Triglav Čekada, Mihaela (2018): Automatic recognition of overgrowth of agricultural land in the Karst region. V: CIGLIČ, Rok (ur.), et al. Pokrajina v visoki ločljivosti, (GIS v Sloveniji, ISSN 1855-4954, 14). 1. izd. Ljubljana: Založba ZRC, str. 103-113 [COBISS.SI-ID 43527213]

Overgrowth of agricultural land and consequently its abandonment is becoming a serious problem in Slovenia. Identification of overgrowth is very important for establishment of permanent monitoring of the overgrowth in the first place and for evaluation of agricultural policy with the aim of reducing the rate of abandonment land. Traditionally monitoring of overgrowth is based on agricultural and forest land use data, which has two main disadvantages. Firstly, data based on different versions of methodology are inappropriate for permanent monitoring and secondly overgrowth photointerpretation of orthophotos time series is very complex and time demanding. Therefore, we propose a time independent methodology based on automatic object-based image analysis of orthophotos and height data.

  • Đurić, Nataša, Kanjir, Urška, Veljanovski, Tatjana (2018): Analysis of Sentinel-2 time series for detection of ineligible use in permanent meadows. V: CIGLIČ, Rok (ur.), et al. Pokrajina v visoki ločljivosti, (GIS v Sloveniji, ISSN 1855-4954, 14). 1. izd. Ljubljana: Založba ZRC, str. 93-102 [COBISS.SI-ID 43644973]

In this paper, we are dealing with the ability to interpret mass satellite data for identification of un-justified uses of permanent meadows. We want to identify ploughing or other permanent changes or anomalies in the grassland land cover. For the analysis, we focused on the three-year time series of Sentinel-2 satellite images. These were analysed on a sample of selected permanent meadows using three different approaches: BFAST Monitor time series method, the standard deviation of time series and the analysis of time profiles. With the latter two methods in addition to test their -ability to detect irregularities, we also validated the first method since it can be limited while dealing with short time series. The anomalies while using BFAST Monitor method are seen as a deviation from the historical trend, in time profiles as a drop of the NDVI values and as an increase of standard deviation. The results indicate usefulness of Sentinel-2 data and suitability of the time series analysis methodology to detect unjustified use of permanent meadows.

  • Foški, Mojca, Đurić, Nataša, Tič, Katja, Triglav Čekada, Mihaela (2019): Comparative analysis of land cover and land use models in selected countries. Geografski vestnik: časopis za geografijo in sorodne vede, 90-1, str. 97-114, https://ojs.zrc-sazu.si/gv/article/view/7298 ,  DOI: 10.3986/GV90106 [COBISS.SI-ID 8709217]

This paper presents land cover and land use models in Austria, Germany, The Netherlands, and Great Britain and the activities under the European projects EAGLE, HELM and LUCAS. These models have been set up for, and depending on, national requirements and also depending on international spatial data harmonisation. The data on land use of agricultural and forest land, set up for the purpose of implementing the Common (European) Agricultural Policy, are also applicable for setting up the national data base on land cover and land use. In Slovenia, the land cover data set up by the Statistical Office of the Republic of Slovenia are no longer updated, while the data by the ministry responsible for agriculture do not cover all land uses with the same level of detail. The comparative analysis of the models can serve as a reflection and starting point for the establishment of a land cover and land use model in Slovenia as well.

The Slovenian national orthophoto in the scale 1 : 5000 is made from the Cyclical Aerophotogrammetric Survey (CAS) images and the digital terrain model (DTM). It is the most widely used spatial source for a variety of GIS analyses and visual photointerpretation in Slovenia. It is used also as a main source for acquisition of changes in the actual use of agricultural land. Different vector or raster spatial layers can be shown on it, which may be more precise than the orthophoto. The users of the orthophoto are often not aware, that orthophoto may include positional errors. Using the geometrical connection between the original image and the orthophoto we derive the maximal radial displacement expected on orthophoto for objects of different height above or below the DTM. We show that radial displacement of higher objects, like higher rocky cliffs (height 50 m), tall buildings (like church towers) and forest edge, may exceed the permitted positional error of the orthophoto, at least in the extreme cases when these objects are located near the seam lines of the orthophoto. We also mention orthophotos made from images of unmanned aerial vehicles (drones), where the problem of its positional accuracy is even broader, as for their production the DTM with nonuniform vertical accuracy is used.

Overgrowth of agricultural land and consequently its abandonment is becoming a serious problem in some European countries. Monitoring and identification of overgrowth and abandonment of agricultural land is significant for the process of evaluation of agricultural policy with the aim of reducing the rate of abandonment of agricultural land. Traditionally monitoring of overgrowth and abandonment of agricultural land in Slovenia is done in the scope of maintenance of agricultural and forest land use/land cover data. Data is interpreted directly, using comparison of old and new ortho-photo images to detect changes. The photointerpretation of overgrowth is very complex and time demanding, therefore we propose two novelties in the existing method:

– use of normalized digital surface model data (derived from aero photos) as auxiliary data in the photointerpretation process and

– automatic identification of overgrowth and cut areas based on a series of orthophotos and normalized digital surface model data.

Object based image analysis has been used for extraction of cut and overgrowth areas in different stages (low, middle and high overgrowth). Completeness, correctness and potential use of the automatic identification results in the future maintenance process has been assessed by expert photointerpretators. The major advantage of proposed automated change detection in the spatial data maintenance process is, that automatically identified changes higliht the areas of potential overgrowth. This way the work of photointerpretator is focused on limited areas instead of entire working unit and consequently overgrowth areas, that often get failed to notice on ortho-photo, are easier to interpret.

Most important socio-economic results of the project group

  • Mesner, Nika, Mangafić, Alen, Tič, Katja, Veljanovski, Tatjana, Kanjir, Urška, Đurić, Nataša, Foški, Mojca, Triglav Čekada, Mihaela (2018): Project results : Automatic procedures for the identification of changes in the actual use of agricultural land. Geodetski vestnik : glasilo Zveze geodetov Slovenije, ISSN 0351-0271, 62-3, str. 501-509, http://geodetski-vestnik.com/62/3/gv62-3_mesner.pdf  [COBISS.SI-ID 8568673]

This professional paper briefly presents all the essential results of the project. An analysis of alternative data of remote sensing is described, which can be used to study changes in the actual land use of agricultural and forest land. The methodology for automatic identification of changes has been described, where we limited ourselves to recognizing the overgrowing of permanent meadows, spreading of built-up areas to agricultural land, deforestation of forests, vineyards and orchards, and changes of permanent meadows into the field or vice versa. We tested three approaches to automatic identification of changes: the analysis of classified orthophoto images and digital surface models, the analysis of Sentinel-2 time series, and an automatic analysis of a direct comparison of two or more images (satellite or aerial).

  • Veljanovski, Tatjana, Marsetič, Aleš, Čotar, Klemen, Oštir, Krištof (2016): Observations of disturbances in forests with satellite imagery of low spatial resolution. V: CIGLIČ, Rok (ur.), et al. Digitalni podatki, (GIS v Sloveniji, ISSN 1855-4954, 13). 1. izd. Ljubljana: Založba ZRC. [COBISS.SI-ID 40181037]

This paper presents the possibility of using satellite images of low spatial resolution for the observation of disturbances in forests such as fires, ice sleet damage, wind breakages, illegal logging and other events. In the study we focused on the SPOT-VEGETATION and PROBA-V satellite data, which are used for vegetation monitoring. For them we acquired long-term time series of observations with span off more than a decade. We developed new method for determination of the extent and the degree of disturbances by the vegetation indicators that are able to detect intensly expressed disturbances of various origins, despite of its relatively low spatial, but high temporal resolution. The capability to detect various events was investigated at the level of individual forest management units, but in the case of a disturbance of a smaller extent it is more reasonable to observe only the affected area.

  • Kanjir, Urška, Pehani, Peter, Veljanovski, Tatjana, Đurić, Nataša, Kokalj, Žiga (2017): Program Kopernik bo vključeval več kot 30 satelitov =. Delo, ISSN 0350-7521, 2. mar. 2017, leto 59, št. 50, str. 14. [COBISS.SI-ID 41267501]

In the daily newspaper Delo, the authors presented the Copernicus program. With this, the knowledge of the potential of using satellite technologies and about freely available satellite imagery, which we also used in the framework of this CRP project (e.g. Sentinel-2), was presented to the general public.

  • Lectures titled »Processing and the use of Lidar data« intended for the professional public, 5. 12. 2017, Nova Gorica [COBISS.SI-ID 8295777, 8296289]

In December 2017, we have conducted (Mihaela Triglav Čekada and Alen Mangafić, Geodetic institute of Slovenia) training for the professional public. We presented the Laser Scanning of Slovenia 2014-2015, the Cyclical Aerial Surveying of Slovenia and the possibility how to use such data in the QGIS Open Source Program.

  • Co-tutor in the master thesis: Orthophoto and lidar data object classification for application in spatial planning [COBISS.SI-ID 8202337]

Mojca Foški (UL FGG) was a co-tutor to the master’s thesis dealing with the issues we dealt with also within this CRP project.

In presented master thesis we are dealing with problematics of delineating vacant building lands. Precise data about location and area of such plots of land are a major problem of many municipalities in Slovenia. Manually updating such spatial data demands extensive financial budget and is often time consuming. However, automatic updating is only rarely used. Presented methodology of delineating vacant building lands uses modern concepts of object-based classification to produce land cover layer and built up areas. Using GIS and spatial overlays we create potentially vacant building lands layer, which is further investigated and enhanced. As a result, we present approximated layer of vacant building lands. Presented methodology was tested on three study areas in municipality Trebnje and results were compared to vacant building lands layer created with visual interpretation and vectorisation. As results show, automated delineation of vacant building land combined with visual interpretation can achieve comparable quality as visual interpretation only.

Other results

  • Kanjir, Urška, Đurić, Nataša, Veljanovski, T. (2018): Sentinel-2 time series based automatic detection of agriculture land use anomalies. V: GEOBIA in a changing world : from pixels to ecosystems and global sustainability : program. [S. l.: s. n. 2018], str. 37. [COBISS.SI-ID 43398189]
  • Triglav Čekada, Mihaela, Mesner, Nika, Veljanovski, Tatjana, Tisu, Miran (2018). Automatic procedures for the identification of changes in the actual use of agricultural land. V: Workshop on checks and management of agricultural land in IACS, IACS workshop, Vilnius, Lithuania. https://ec.europa.eu/jrc/sites/jrcsh/files/11_automation_si.pdf  [COBISS.SI-ID 8472161]