Research On The Technology Of Image Fusion

Image fusion is an important branch of information fusion. It will fuse multiple images of the same scene or target the image to generate a fusion image that is more clarity and accurate description of the scene or the target. And, the fused image is more suitable for human vision and image processing. Image fusion technology is an emerging discipline,which involves the sensor technology, signal processing, image processing. Image fusion has been widely applied to military and civilian fields, such as medical treatment, monitoring,traffic and so on.Visual image and infrared image fusion, and multi-focus images fusion as an important part of image fusion, are the main research object of this article. In this article, the CSF of HVS is as the definition model of the visual image and infrared image fusion to extract the required information from the source images to be saved into the fusion image; gradient-based structure salience tensor serves as focus measure of multi-focus images, and we put forward an adaptive weighted filter function to generate an all-in-focus image. The main research work of this paper is as follows:(1) In view of the current definition model lack of HVS in visual image and infrared image fusion, a new definition model based on CSF in HVS is proposed. And, this model is as the fusion rules of the high frequency band in the Non-subsampled Contourlet Transform(NSCT).(2) The structure salience tensor based on gradient serves as the focus measure of the multi-focus images that can effectively extract the structure information of source images, which is to overcome the effects of blur in the images.(3) Using the Gaussian model in the shape from focus to calculate the PSNR of multi-focus images generating a selection matrix, which can eliminate noises in the resource images.(4) Analysisof behavior characteristics of the focus measure, we find that the behavior characteristics is in line with a high-pass filter. Based on adaptive hyperbolic tangent filter is proposed to calculate the weights of the source images. Finally, an all-in-focus image is generated.Finally, the validity of the fusion algorithm proposed in this paper is verifiedthrough contrast test. And, the experimental data also suggests that this algorithm is superior to other algorithms.