Comprehensive Survey On Computer Aided SAR Image Segmentation and Classifi cation Methods

Synthetic Aperture Radar (SAR)is a satellite imaging technology which uses radio waves for capturing the images. Here a RADAR(Radio Detection And Ranging) is mounted on a moving platform which emits radio waves and the refl ected echoes from earth’s surface is collected back. Signal processing of these backscattered echoes results in SAR images. The major advantages of SAR imaging is that it is unaffected by weather conditions and can penetrate through cloud and soil. Hence it has got wide range of applications in homeland security, environmental protection, traffi c monitoring, 3D map generation, land resource management etc.SAR images are affected by Speckle noise which is multiplicative in nature and is diffi cult to remove. Therefore despeckling is carried out as a pre-processing step to SAR image segmentation and classifi cation. . Here some papers related to SAR image segmentation and classifi cation for last one decade is being reviewed. Image segmentation and classifi cation are the crucial step for SAR image analysis. SAR image segmentation can be carried out by using either contextual or non-contextual method and from the analysis it is found that contextual segmentation technique is better than other techniques.SAR image classifi cation can be carried out under supervised and unsupervised techniques. And from the analysis it is found that supervised classifi cation methods for SAR image produces better result as compared to unsupervised methods.