Wavelet-based Infrared Image Enhancement Algorithms

Along with the development and application of Modern infrared technology, civilian infrared detection systems and military infrared remote monitoring systems have been widely applied. However, because of the impact in infrared detection devices and environmental, there are no satisfactory effects in infrared images. In practical applications, It's necessary to enhance infrared images in order to get better target recognition and visual observation.Wavelet analysis, also called multi-resolution analysis, a developmental discipline on the foundation of the Fourier analysis, has the deep theories meaning and extensive applications. It has generated strong impulse on the new technique application subjects and the ancient natural subjects. At present, wavelet analysis is international deeply concerned leading field of science and technology.Wavelet theory is introduced to enhanced infrared images in this research. The wavelet transform decomposes an image into a finite number of resolution scales that is very suitable for image analysis. And after the wavelet reconstruction, the visual quality of the processed images can be improved effectively.The research mainly focused on the infrared images enhancing, including two aspects: Infrared images de-noising and enhancing from the visual. the new results are as follow:at first, introduced some key technology of wavelet transform; Focus on analysis of multi-scale wavelet coefficients approach and image reconstruction, proposed a scheme of infrared image de-noising based wavelet threshold processing. According to the distribution of noise, both containing additive noise also contains a large number of multiplicative noise in infrared images, In this research, infrared images are decomposed and reconstructed twice, the additive noise is weakened at the first wavelet threshold treatment; Second wavelet decomposition and reconstruction, the multiplicative noise is weakened . After two wavelet threshold treatment, the noises in infrared images are greatly inhibited. Comparing with traditional de-noising methods, it is more significant. Wavelet theory and retinex algorithm are both used in infrared image enhancement processing. The wavelet transform decomposes an image into the wavelet coefficients, Low-frequency coefficients are treated with retinex algorithm to adjust image brightness. There are greater dynamic brightness dynamic range and more uniform illumination in the processed images. High frequency coefficients are treated with wavelet threshold, then weakens noise and enhances the details of the images. With the two methods, good enhanced results have been achieved in image contrast,Information Entropy and Relative SNR.