Land cover classification using object-based image analysis (OBIA) in Delta Tumpat, Kelantan

Abstract

Delta Tumpat is one of Kelantan's fastest-growing districts, with a population that is growing over the years. This may result in land cover changes and rapid development for human activities. The land cover changes must be monitored to ensure that it is in line with the regional development planning and does not cause problems in the future. Therefore, continuous monitoring of land use and land cover of this area is very essential. By utilizing remotely sensed data from Landsat 8-OLI, the land cover can be classified and detected efficiently. In this study, the object-based image analysis (OBIA) was used to classify the land cover in Delta Tumpat with three different levels of set parameters were examined. The results of the study presented a segmentation scale parameter value of 50, shape and compactness parameters at 0.2 and 0.8, respectively is the best segmentation rule set for the land cover detection in the study area. The classification results produced four types of land cover classes with vegetation class is the major land cover type which dominated 45.2% of the total area using the OBIA technique. Hence, the selected ruleset had achieved 92.32 % accuracy using this method.