| Differentiation of vineyards using Sentinel-2 and Sentinel-1 satellite images (A case study in Hamedan province) |
| Paper ID : 1074-NICAME1402 |
| Authors: |
|
Hamidreza Maleki *1, Hosna Mohamadi Monavar2, Mehraneh Khodamoradpour3 1Department Of Biosystem, Agriculture , Bu Ali Sina, Hamedan, Iran 2Department Of Biosystem, Agriculture, Bu Ali Sina, Hamedan, Iran 3Department of Water Engineering, Agriculture, Bu Ali Sina, Hamedan, Iran |
| Abstract: |
| The changes that occur over time in the cultivation of grape vineyards make it essential to create updated maps for vineyards in regional agricultural planning and management. Grapes are one of the most valuable horticultural products, with Hamedan province producing over 12% of the country’s grapes. Remote sensing technology and the utilization of satellite data are effective tools in the field of agriculture. The objective of this research was to differentiate vineyards in Hamedan province using remote sensing vegetation cover indices extracted from combined Sentinel-2 and Sentinel-1 satellite images. In this study, the SVM machine learning algorithm was utilized to identify and distinguish vineyards from other crops in Hamedan province. The classification process was carried out using the Google Earth Engine system. The maximum monthly NDVI (Normalized Difference Vegetation Index) was extracted from Sentinel-2 satellite images. To enhance the classification accuracy, these images were combined with radar images from Sentinel-1 satellites. After classifying the images, the classification accuracy was evaluated, resulting in an overall accuracy of 70% and a kappa coefficient of 0.61. The results demonstrated that by combining Sentinel-2 and Sentinel-1 satellite images with machine learning algorithms, vineyards can be distinguished from other crops. |
| Keywords: |
| Vineyard, NDVI, Support Vector Machine, Sentinel-2 |
| Status : Paper Accepted (Poster Presentation) |
