Research Article Open Access

An Enhanced Framework for Change Detection in Very High Resolution Remote Sensing Images

Abstract

Land-cover (LC) and land-use (LU) change information is important due to its practical uses in various applications. Increasing the geometrical resolution of remote sensing images makes the change detection (CD) process to be complicated. In this paper, An enhanced framework based on the spatial context information of multitemporal adaptive regions homogeneous both in spatial and temporal domain (parcels) is presented to detect the semantic changes in the very high resolution (VHR) remote sensing multitemporal images. The proposed framework uses the absolute difference between {red, green, blue and hue} layers instead of the use of change vector analysis proposed by Lorenzo Bruzzone and Francesca Bovolo in 2013. Two experiments are performed to test the performance of the proposed framework. The first experiment shows that the proposed framework gives 72.8 % CD accuracy while the published framework gives 18.69 %. The second experiment shows that the proposed framework gives overall accuracy, while the published framework gives The results indicate that the proposed framework detects % had been changed from the total area and the published framework gives 10.41%

Eng. Mostafa Mosaad, Dr. Fawzy Eltohamy, Dr.Mahmoud Safwat,Ashraf K. Helmy

To read the full article Download Full Article