2nd International Conference on Agro-Geoinformatics (Agro-Geoinformatics), Virginia, United States Of America, 12 - 16 August 2013, pp.93-96
This study focused on crop type identification and area estimation of the cultivated parcels using images with 1.5m spatial resolution and 60km x 60km coverage from newly launched SPOT 6 satellite. A test site and a district in Sanliurfa Province, Southeastern Turkey were selected as study region where ground truth data was also available. After the geometric correction of SPOT 6 images; further image processing methods were applied. In the first step, a comparative analysis was performed at a test site to find out optimum parameters to be applied for this research. For this analysis, object based image classification (OBC) was applied to SPOT 6 image and feature extraction was performed from classification results to identify different parcels within the study area. Extracted features were compared to parcel vector database and accuracy metrics of parcel identification and area estimation was produced. In the second step, analysis was extended to district scale and classification OBC classification was performed for Suruc district of Sanliurfa. Accuracy assessment was performed with stratified random point selection for district scale classification results.