A Review on Multi Sensor Image Fusion Techniques

Most Earth observational satellites are not able to acquire high spatial and spectral resolution data simultaneously because of design or observational constraints. To overcome such limitations, image fusion techniques are use. Image fusion is process combine different satellite images on a pixel by pixel basis to produce fused images of higher value. The value adding is meant in terms of information extraction capability, reliability and increased accuracy. The objective of this paper is to describe basics of image fusion, various pixel level mage fusion techniques for evaluating and assessing the performance of these fusion algorithms. Keywords— -Image Fusion, Pixel Level, Multi-sensor, IHS, PCA, Multiplicative, Brovey, DCT, DWT. INTRODUCTION Image Fusion is process of combine two different images which are acquired by different sensor or single sensor. Output image contain more information than input images and more suitable for human visual perception or for machine perception. Objectives of Image Fusion Schemes are Extract all the useful information from the source images. Figure1.1 Pre-processing of image fusion [1]. Image fusion is applicable at different fields that are: defense systems, remote sensing and geosciences, robotics and industrial engineering, and medical imaging. Goal of image registration is to find a transformation that aligns one image to another. In image registration, one dataset is regarded as the reference data and other as sensed data. Sensed data is matched relative to the reference data, Image registration at a very basic level. Image re-sampling [2] is the process to produce new image with eight in different size. Re-sampling can change the size of the image. Increasing the size is called up-sampling; decreasing the size is called down-sampling. Note that the spatial resolution would not change after the RS procedure, either up-sampling or down-sampling. In multi-sensor image fusion, the images of the same scene come from different sensors of different resolution. In multi-focus image fusion, the images of the same scene come from the same sensor are combined to produce an image in which all the objects are in focus. Pohl & Genderen 1998 presents three types of image fusion levels: pixel, feature, and decision levels, In this paper, we are only concerned about pixel level fusion. International Journal of Engineering Research and General Science Volume 2, Issue 3, April-May 2014 ISSN 2091-2730 343 www.ijergs.org Figure 1.2 Level of Image Fusion [3]. Image Fusion Techniques Image fusion techniques are classified into several techniques, which are described below Figure1.3 The categorization of pixel level image fusion techniques [4]. IHS (Intensity, Hue, Saturation) IHS is a color space, intensity relates to the total amount of light that reaches the eye, hue is defined as the predominant wavelength of a color, and saturation is defined as total amount of white light of a color. Steps 1. first converts a RGB image into intensity (I), I v1 v2 = 1/3 1/3 1/3 − 2 6 − 2 6 2− 2 6 1 2 −1 2 0 R G B Hue (H) and Saturation (S) components. 2. Replacement of I by high resolution image, F(R) F(G) F(B) = 1 −1 2 1 2 1 −1 2 −1 2 1 2 0 PAN v1 v2 3. Reverse IHS, converting IHS components into RGB colors. Pixel level Image Fusion Techniques Component Substitution IHS PCA Mathematical Combination multiplicative Brovey Filtering