Document Type : Research Paper
Authors
Abstract
Nowadays, remote sensing images are able to provide the latest information for studying land coverage and land uses, and the value and usability of produced maps depend on their accuracy. Hence, the purpose of this study was to evaluate the classification accuracy of LISS-III sensor's image of IRS-P6 satellite using the Google Earth's database in order to provide a map of land coverage and land uses. Therefore, the QUICKBIRD satellite imagery provided by Google Earth's software was used to determine both educational samples and to evaluate the classification. The studied area is Taleghan city in the Alborz province which is located in the watershed of Taleghan. In this research before determining the educational samples to verify the accuracy of the Google Earth's image, linear digital layers (roads and channels) with terrestrial coordinates were used which obtained an RMSE of 0.77.
In the next step, after determining the educational samples, the mentioned satellite image was classified into 5 categories of garden, agriculture. Pasture, lake and no coverage based on a supervised classification and with maximum probability algorithm using software ENVI 4/2 which obtained a classification KAPPA coefficient of 0.85 and an overall accuracy of 91/4.
The results of this study indicated that Google Earth's software images have a high spatial accuracy in order to evaluate the classification accuracy in some regions and also the use of ecological features such as the slope of the area, hydrologic network and… increase this accuracy. Finally, it is suggested that Google Earth's satellite imagery to be used to evaluate the accuracy of the satellite image classification and even visual interpretation of land coverage and land use.
Keywords