Comparison of Two Land
Comparison of Two Land Use Change Detection Methods for
Evaluation of their Capability for Environmental Impact Assessment
Jawad Ghasemzadeh
Chabahar Maritime University
Reza Rafiei
Change detection can be used for environment planning and management. This technique is a process used to identify differences of states of objects or phenomena onimages observed at different times. Through a historical change analysis, decision makers and mangers would be able to develop a suitable model between change areas and some most significant factors which lead to detect changes. In this way, several scenarios for future could be implemented. In this regard, the accuracy of change detection technique and its simplicity of implementation are very important. In the present study, two methods of post-classification and principal component analysis were compared with each other in terms of their application and accuracy. In post-classification approach, by selecting some train areas, each data set was separately classified into change and no-change classes. Then, through a pixel-by-pixel comparison, a change map and a detailed matrix of change were produced. We then used the Principal Component Analysis (PCA) method in which two first PCA images belonging to each year were subtracted from each other. After that, the change map was produced by selecting optimum thresholds. The result showed a kappa of 0.82 for post-classification method and 0.62 for principal component analysis, respectively. Selecting the appropriate train areas was the most important downside of post-classification method, but it removes the need to atmospheric correction, while for second method, a detailed matrix wasn't provided and selecting the optimum threshold was very difficult.
Key Words: Change Detection, Evaluation, Post –Classification, PCA
