Mapping the poplar plantations using Landsat-8 data (Case Study: Talesh and Sumehsara region, Guilan province)

This research aims at investigating the feasibility of Landsat8 images for poplar plantation mapping. Three areas in Talesh and Sumehsara regions, Guilan province, with a total 1025.35 ha with different vegetation mixtures were chosen as study area. Five image-sets were selected from growth and autumn seasons. Studying the images showed no radiometric and geometric distortions. The image enhancements like rationing, PCA and Tasseled-Cap transformation were performed. By preparing training samples in each region and choosing the suitable band-sets, the images were classified using maximum likelihood algorithm. The ground truth was prepared for Hefzabad and Toularoud in Talesh using GPS. It was extracted for Haft-Deghnan and Sheikhneshin from available update-maps. To ensure suitability of the training samples, assessing the accuracy of classification were performed in the training samples. After ensuring this, the accuracy of classification was assessed in whole areas of all images. The results in all areas and dates were almost close to each other. However, the best results were always obtained in growth season, especially in June 2013. Using calculated enhanced bands along with the original bands in classifications in Talesh region showed slight improvement in accuracies, with overall accuracy and Kappa about 95% and 71%, respectively. In Sumehsara region, the best results were obtained in June with 83% overall accuracy and the relatively low Kappa 0.43. While, the results of accuracy assessment in training samples in all cases were about 0.89 to 0.99. It could be stated that the Landsat 8 images had medium capability for poplar plantation mapping. More investigations are needed to create better poplar plantation mapping.