Improved Image Fusion Algorithm Based on Wavelet Packet Analysis in Intelligent Transport System

Based on wavelet packet analysis, this paper introduces a new image fusion algorithm to combine infrared image and visible image to a single image in intelligent transport system. The Gaussian smoothing operator is utilized to filter out the noise component of original infrared image and visible image before wavelet packet transform. The fusion method is determined by wavelet modulus: if wavelet modulus of visible light image and infrared image are discrepant widely, the larger coefficients will be adopted; otherwise weighted average arithmetic will be applied. Experimental results have shown that most of the final composite images have more integrated information content and better perceptual quality than either of the original images. The results indicate that application of this algorithm improves performance of Intelligent Transport System.