Visible Light Image And Infrared Image Fusion Technology Research

With the development and improvement of science and technology, the quality of images is demanded much higher than ever before. A visual image describes the shape of all the objects in the scene and has good ability to represent objects’ outline which reflects same information as naked eyes. On the other hand, infrared images cannot display the shape clearly but present the temperature distribution of objects in the scene which is particular important for object detection and so on. To fuse them into a single image is very useful in reality and that is the main point of this dissertation.This dissertation is aimed at fusing the visual image and infrared image about a same scene into a single image to display both images’ information. According to wavelet transform (Mallat algorithm), the visual image and infrared image are both transformed into several child-band images which carry different information among. The low frequency part mainly lays out the whole outline and high frequency parts describe the details of the source image. By fusing these child-band images with different fusion methods and reverse Mallat transform, a single fused image is achieved at last. It not only remains the shape information of the visual image, but also displays the temperature distribution of objects from the infrared image.The characteristic of this dissertation is improving the traditional fusion algorithm, giving the objective assessment of the fusion quality and making best use of Qt platform to combine the algorithm of image processing and application development based on C++language. It successfully completed a software system to fuse visual and infrared images and the software puts the both images’ information into a fused image through the improved algorithm. The successful image fusion processing truly means significant importance for human observing, target recognition and other further image processing.