Research On Color Fusion Methods For Multi-sensor Images

With the development of imaging sensors, multi-sensor image fusion technology is generated and applied in a wide variety of fields such as military surveillance, medical diagnosis, remote sensing and intelligent robot. Image fusion technique has accomplished pretty good fusion effects; especially in the field of fusion of grayscale images, but not much study in the color image fusion. The request for color image fusion techniques is becoming more and more crucial in the application of the relevant fields due to the excellent visualization and more information compared with the gray image.The focus of the research is to develop the multi-sensor image color fusion algorithms.Firstly, the thesis introduces and implements several typical image fusion methods, whose fusion principles, advantages and disadvantages are carefully analyzed. Secondly a novel color fusion method for visible and infrared (IR) images is presented. In the method the visible image is fused with the original infrared image and the reverse polarity of infrared image by a grayscale fusion scheme based on Curvelet transformation, respectively. The fusion rules are consistent with human's vision associations of warm red and cool blue. The experimental results show that the color fused image maintains high resolution details and improves efficiently the discrimination between targets and the environment.Next, considering the fact that the color of the fused image using false color method is not natural, a novel region-based color mapping method is proposed to render the image fused multi-sensor images. The method involves image segmentation, object recognition, image fusion and color transfer, which can color the image segment-by-segment. Experiments show that the image colorized by the proposed method resembles the natural color appearance and helps the observer make scene interpretation more intuitive compared with that produced by other coloring method.After introducing some indexes for color fused image quality assessment, we analyzed the performance of several color image fusion methods from subjective and objective aspects.Finally the conclusions summarized the work of this paper and points out the work needing to deal with in the further research.