Remote Sensing Vegetation Indices: Evaluation on a Mandarin
Orchard - Abstract
Dimitrios Koulouris 1, Ioannis L. Tsirogiannis 2 and Nikolaos Malamos 1
1
    University of Patras, Department of Agriculture, Theodoropoulou Terma, Amaliada, 27200, Greece
2
    University of Ioannina, Department of Agriculture, Kostaki, Arta, 47100, Greece


                Summary
                Precision agriculture involves modern crop surveillance methods, such as spectral vegetation
                indices via satellite remote sensing. This study examines the performance of Normalized
                Difference Vegetation Index (NDVI) and Normal Difference Moisture Index (NDMI) spectral
                indices as irrigation management tools. The IRMA_SYS (https://arta.irmasys.com) intelligent
                irrigation decision support system for the plain of Arta, which has an area of 45000 hectares,
                provides means for downloading the above-mentioned indices for each field. These indices are
                obtained from the available Sentinel-2 satellite images i.e., every 2 or 3 days. These images
                have pixel dimension of 20 m that attains the spatial accuracy of each spectral index.
                The study was carried out for two mandarin orchards. The fields were coded in IRMA_SYS
                as: a) field 728, with an area of 5000 m2 and b) field 729, with an area of 30000 m2. Also, a
                complete meteorological dataset is available for each field through IRMA_SYS, by spatial
                interpolation of data from seven agro-meteorological stations in the area. Daily values of
                temperature (°C), solar energy (W m–2) and rainfall (mm) were obtained from 15/3/2021 to
                28/10/2021. Also, NDVI and NDMI values were obtained for a characteristic point for each of
                the two fields. The total effective rainfall for field 728 was 232 mm while for field 729 was
                252 mm, correspondingly. The two farmers followed different irrigation strategies: The owner
                of field 728 applied 2 irrigations totaling 59 mm with an average of 29.5 mm, while the owner
                of field 729 applied 7 irrigations totaling 78 mm with an average of 11.2 mm.
                Based on the analysis performed, the aforementioned indices are strongly related to rainfall or
                irrigation, since in both fields’ irrigation was followed by a clear upward trend in the values of
                the two vegetation indices. In particular, the NDVI index seems to be more responsive than the
                NDMI index in cultivation practices such as weed removal. It should be noted that after rainfall
                or irrigation, the variation of both indices is not immediately noticeable, but a period of 2 or 3
                days is required since the trees do not correspond instantly to the applied water. Therefore, the
                changes of the vegetation indices should be evaluated not as individual values but as
                continuous time series. This evaluation should consider information concerning cultivation
                practices carried out during this period.
                Conclusively, irrigation is reflected through the NDVI and NDMI vegetation indices, to a
                satisfactory degree so that they can be used under specific conditions as a tool for monitoring
                crops within the context of precision agriculture practices.

                Keywords 1
                Precision agriculture, remote sensing, IRMA_SYS, NDVI, NDMI, irrigation




Proceedings of HAICTA 2022, September 22–25, 2022, Athens, Greece
EMAIL: dkoulouris@upatras.gr (A. 1); itsirog@uoi.gr (A. 2); nmalamos@upatras.gr (A. 3)
ORCID: 0000-0002-3814-0813 (A. 1); 0000-0001-9102-8372 (A. 2); 0000-0002-5292-2870 (A. 3)
             ©️ 2022 Copyright for this paper by its authors.
             Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).
             CEUR Workshop Proceedings (CEUR-WS.org)




                                                                                  426