| CHANGE INDEX BASED ALGORITHMS FOR CHANGE DETECTION IN FULL POLARIMETRIC UAVSAR IMAGES |
| Paper ID : 1318-SMPR |
| Authors: |
|
Leila Yousefizadeh1, Reza Shahhoseini *2, Saeid Homayouni3 1School of Surveying and Geospatial Engineering, College of Engineering, University of Tehran 2Photogrametry and Remote Sensing Department, School of Surveying and Geospatial Engineering, College of Engineering, University of Tehran, Tehran, Iran 3Department of Geography, University of Ottawa |
| Abstract: |
| Change detection is one of the most important applications of polarimetric SAR images in monitoring urban development and supporting urban planning. In many change detection methods, only one difference image is produced; however, when the area has complex land cover, detecting changed the area in different classes based on single difference image is not feasible. In this study, quad-polarized Uninhabited Aerial Vehicle Synthetic Aperture Radar (UAVSAR) data over an urban area in San Anderson was used to make a change map. Input features play an important role in the accuracy assessment of a change map. In many studies intensity images were used to extract changed area; however, a few studies examined the potential of polarimetric decomposition in detecting changed area. Therefore, we analyzed different decomposition feature based on the targeted nature. After producing a feature change vector by ratio operator, a kernel-based clustering method was used to discriminate changed and unchanged areas. A comparison of results with ground truth data showed a good level of agreement. |
| Keywords: |
| Change Index, UAVSAR, Polarimetric, Kernel-KMeans, Decomposition |
| Status : Conditional Accept (Poster) |
