Ground Surface-Object Monitoring Using Multi-Temporal High Resolution SAR Data

Today’s space based remote sensing is the era of constellation missions. Synthetic Aperture Radar (SAR) sensor constellation of small satellites provide information at frequent time intervals suitable for monitoring in day-night and all weather conditions with large volume of data. Automatic Change Detection is desired to continuously track and monitor the changes on Ground surface. In this paper, an in-house developed Automatic Change Detection method is presented to track changes from high resolution of SAR data. The process is divided into four major parts namely, Preprocessing, Change detection, Post-processing to remove false alarms and Classification. To track actual changes and reduce false predictions due to change in imaging geometry, it is necessary to select the image pair for extracting change information. Pre-processing step involves filtering for selection of image pair and with reduction of speckle noise of the selected image pairs to increase the detection sensitivity of algorithm of detecting the small changes. Algorithm uses amplitude information to detect the changes and generate the Change detected Image and Masked output image of the region. The Post-processing stage then reduces the false alarm of smaller erroneous regions and generates the final change map and then classification information is generated, which is used for identify the area of interest in SAR image and shows the individual classes of changed pixels from change detected output images. The results and analysis are presented in the paper on ICEYE’s constellation consists of X band with 0.5m azimuth and range spacing, VV polarization SAR. High resolution SAR datasets indicate that the proposed method provides accurate results which is very useful for identifying significant changes and monitoring objects on ground surface, and are evaluated.