Comparison of Pixel-Based and Object-Oriented Classification Methods for Extracting Built-Up Areas in Coastal Zone

The monitoring of the earth surface and the atmosphere on a global, regional, and even local scale has become very accessible, thanks to the new and powerful techniques of remote sensing. Thus, it has become easy to access important coverage, mapping, and classification of the land covering features, including soil, vegetation, as well as water. Monitoring the coastal environment using remote sensing and GIS techniques has been undertaken in this study, with a particular focus on the comparison between the classical and object-oriented image classifications of remote sensing imagery in coastal areas. In fact, the investigation was based on the testing of a coastal zone image classification, pixel-based image classifiers such as SVM classifier and an object-oriented image classifier. The method was later compared using a Pleiades image. The use of reference data sets that were taken from high-resolution satellite images, aerial photographs, and field investigation was considered as an effective way to assess the accuracy of this method. Overall accuracy of 88% with a kappa coefficient of 0.74, compared with 79% (0.71) that was concluded from the conventional pixel-based method, was the result of this object-oriented method.