Change Detection of Surface Elevation by Alos/prism for Disaster Monitoring

In October 2006, remotely sensed data of PRISM (Panchromatic Remote-Sensing Instrument for Stereo Mapping) sensor that mounted on Japanese satellite ALOS (Advanced Land Observation Satellite) was began to provide for general users. ALOS has some missions such as mapping, regional observation, disaster situation observation and survey of natural resources. In the mission of disaster situation observation, it is important to do change detection of the affected area with high accuracy. ALOS has three sensors namely PRISM, AVNIR-II, PALSAR. The PRISM sensor is three lines scanner with 2.5 meters sampling distance. Triplet imagery can be acquired to, the same location at the same time from the same orbit. Therefore, acquired data of PRISM sensor is expecting to generate 3D data. The 3D data can be used for natural disaster monitoring and many others. Authors report the change detection of surface elevation by ALOS/PRISM. The DSM (Digital Surface Model) generation technique that used stereo PRISM images was constructed. The proposed method needs some GCPs (Ground Control Point) and corresponding points by stereo matching. GCPs are measured by VRS-GPS (Virtual Reference Station Global Positioning System) measurement. And the basic geometric model is established using GCPs. Corresponding points of stereo matching are computed by template matching and least squares matching. Surface elevation is calculated by least squares method using basic geometric model and corresponding points of stereo images. The accuracy of generated DSM showed about 2~10 meters. The amount of the changes can be detected by computing the generated DSM before and after the changes. Target area was selected around Kochi Prefecture. For change detection of surface elevation, accuracy of DSM must be improved. Recently, method of stereo matching is revised. In future, change detection of surface elevation will be carried out. 1. BACKGROUND In October 2006, remotely sensed data of PRISM sensor that mounted on Japanese land observation satellite ALOS was began to provide for general users. PRISM sensor has three lines scanner (forward, nadir, backward looks) with the ground resolution of 2.5 meters and it can be observed at the same time from the same orbit. Therefore, acquired data of PRISM sensor is expecting to generate accurate 3D data by processing triplet images. 3D data can be used for natural disaster monitoring, etc. Generally, 3D data can be generated using several geometric models. PRISM’s RPC (Rational Polynomial Coefficient) is usually use, which is produced from RESTEC (Remote Sensing Technology Center of Japan). On the other hand, basic geometric model with GCP data can also make 3D data. Then each geometric model should be evaluated for accurate change detection. After that change detection with the most accurate transformation will be carried out.