IEEE Proceedings of 4 th International Conference on Intelligent Human Computer Interaction, Kharagpur, India, December 27-29, 2012 978-1-4673-4369-5/12/$31.00 ©2012 IEEE An efficient algorithm for automatic fusion of RISAT-1 SAR data and Resourcesat-2 Optical images Indranil Misra, Rajdeep Kaur Gambhir, S. Manthira Moorthi, Debajyoti Dhar and R.Ramakrishnan Data Product Software Group Signal and Image Processing Area Space Applications Centre (ISRO), Ahmedabad-380015 E-Mail:{indranil,rajdeep,smmoorthi,deb,rama}@sac.isro.gov.in Abstract- Satellite Image fusion generates single hybrid image from a collection of input satellite images and helps us to extract maximum information from the remotely sensed datasets to achieve optimal spatial and spectral resolution. The critical steps of image fusion framework are co- registration of Synthetic Aperture Radar(SAR) data with corresponding optical scene, enhance the images for visual clarity and then merge the multi sensor data with a standard fusion technique. The image fusion system should perform all these steps in an automatic manner for providing ease to the user. The primary attention of this work is to examine the improvement that can be obtained by fusion of low resolution multi spectral data obtained from optical Resourcesat-2 platform( LISS-4MX/LISS- III/AWIFS Sensor having 5m/24m/56m spatial resolution) with high resolution RISAT-1 (Fine Resolution STRIPMAP (FRS-1)/Medium Resolution SCANSAR(MRS) mode data having 3m/18m spatial resolution) using SAR-Optical image fusion system discussed above. This integration of optical and SAR images from Indian Remote Sensing satellites facilitates better visual and automatic image interpretation. The Maximum Likelihood algorithm is used for classification of fused image and Resourcesat-2 multispectral data. The quality improvement of the fused product can be observed by comparing the classification accuracies of merged data with original multispectral data of the same region. Keywords: Image Fusion, Resourcesat-2, RISAT-1, SAR, Maximum Likelihood, Image Registration I. INTRODUCTION Remotely sensed satellite images are of great interest in today's world for earth observation that includes resource assessment, vegetation profile mapping and environment monitoring. This is made possible by large amount of datasets acquired by different types of optical and radar sensors data. However for many space applications the information provided by an individual sensor can be incomplete or less in content. The reason behind the same is that a higher spectral resolution is often achieved at the cost of a lower spatial resolution for a given signal to noise ratio. In current high performance computing system, lower resolution multispectral channels are often complemented by a single higher resolution band, which provides a "representation" of the desired spectral and spatial resolution shown in Fig.1 [1]. ISRO’s remote sensing platform constellation offers considerable scope for selection of data with multiple spatial/spectral resolutions. RESOURCESAT-2 was launched in April, 2011 as a follow up mission of RESOURCESAT-1 platform providing LISS-3/LISS-4MX/AWIFS data [2]. RESOURCESAT-2 is now available to many user segments for operational remote sensing based natural resource management projects. RISAT-1 is the latest space born microwave sensing platform in the world and first Indian microwave imaging platform for civilian purpose [3]. This satellite was launched in April, 2012 , provides data in many different modes and the data validation exercises are being carried out by ISRO teams before it is made available to user segments in the country. The image fusion are therefore useful for integrating a high spectral resolution image with high spatial resolution image such as RISAT-1 FRS-1/MRS mode(3m/18m) with Resourcesat-2 LISS-4MX/LISS-3/AWIFS(5m/24m/56m). The aim of the fusion procedure is to produce high quality and accurate data which contains the characteristic of both the multispectral information (object identification) and the spatial detail (object localization and texture) [4]. II. STEPS IN AUTOMATIC IMAGE FUSION Image fusion in automatic mode requires sequential execution of different phases for generating quality data products. Digital Image Processing techniques are used to generate fused image in different stages. The main steps/phases involved in image fusion are: 1. Speckle Suppression: The speckle noise in SAR images is a major obstacle to interpret SAR data for various remote sensing applications. Speckle noise arises