Mapping of Land Use with an Object Base Classification Using WorldView-2 High Resolution Satellite Imagery

Land cover maps are important data for environmental assessment. In this study, the accuracy of preparing the land use map of Haft Boram area of Shiraz with the base object algorithm on the World View 2 satellite images of 2015 has been investigated.

In this study, World View 2 images have been used in the forests around Shiraz, which can provide detailed vegetation maps. The method of map preparation analysis is the method of the base object with the decision tree. The work process includes: segmentation, selection and extraction of the desired features including 9 classes with spectral indices, hierarchical classification of the base object and creation of a decision tree to classify the image, and finally classification and evaluating of its accuracy.

The results showed that the classification classes were successful and the overall accuracy was 87.45%. The most accurate classification was obtained in the procedure accuracy for water, forest, crop, and building classes. The highest accuracy is related to water (100%).

Planting trees and natural forests were recognized by the base object method. Additional coastal bands provide the ability to detect Bayer lands from other vegetation. Distance to tree and shadow play an important role in the identification of buildings.