Infrared Imaging Enhancement Algorithms Based On Retinex Theory

Infrared imaging has been applied to a lot of areas, such as military, industry, agriculture and so on, on account of its large coverage, strong anti-jamming ability, no radiation, and effective invisibility. The radiation wave length of infrared band are shorter than that of radio waves, but longer than that of visible radiation. Thus the spatial resolution of infrared images is higher than that of radar, but lower than that of visible radiation. Therefore, in order to obtain the images which are easy for men and machines to distinguish them, the infrared images require to be enhanced. In that way, the goals can be identified and detected correctly.The algorithms of traditional image enhancement are diverse. They are mainly divided into two categories, including Spatial domain and Frequency domain. These algorithms are the basic principle of image enhancement. With the support of the traditional algorithm theory, started with the analysis of the features of target Infrared imaging. To facilitate men’s observation and machines’identification, we analyze the human visual characteristic and color immutability, and finally chose a TV-Retinex on the basis of Retinex Theory. The algorithm based on the variational theory establishes an energy model, and achieves better enhancement effect By solving the global optimal solution of the model.After the optimization and manipulation of the iterative algorithm of the original Variational Framework, the computing cost becomes lower. A series of simulation experiments proves that the refined algorithm can not only eliminate random noise and halo effects, also improve the images’ contrast and definition.So using this algorithm,the experimental result shows the effect of image enhancement is better.