Determination of opium poppy (Papaver somniferum) plantation using by high-resolution satellite imagery
2017
Demir, S.
The aim of thesis was to determine the identifying capability of poppy parcels with high-resolution satellite images. In the study, Quickbird-2 satellite image was used and suitable processes of ERDAS Imagine, eCognition Developer and ArcGIS software were run. Accuracy assessment was done with the ground references and the best classification method for poppy parcels was reveal.For this purpose, a test area of 4 km2 was selected. Multi-threshold segmentation, spectral difference segmentation, multiresolution segmentation, quadtree-based segmentation and pixel-based classification algorithms of object-based classification method were used. Unsupervised classification, supervised classification, fuzzy classification and fuzzy convolve algorithms of pixel based classification method were used. Vegetation indices were also determined using RVI, NDVI, TNDVI, SAVI, MSAVI and MTVI classifications. The classification maps produced with vegetation index were compared with the field survey data.As a result, the object-based classification was found best accuracy method. The producer-user accuracy were estimated as 97.99% using spectral difference segmentation process of multiresolution segmentation algorithm in object-based classifications method. Using this algorithm, opium poppy can be determined with the highest accuracy.
Показать больше [+] Меньше [-]