Abstract:

To solve the problem of the typical histogram equalization method used to enhance infrared image, a novel method based on dynamic and generalized histogram equalization is proposed. Firstly, generalized histogram can reflect the different image region feature through the fractional count based on local spatial complexity and prevent overenhancement and noises magnification. Then, after smoothing and filtering the generalized histogram by SG algorithm, the GK clustering algorithm is adopted to realize the histogram’s dynamic partition, which can avoid the dominance of background gray values of high probability on target gray values of low probability. Finally, the output gray value ranges of target and background are allocated dynamically to highlight the target’s details and control the enhancement effect. The experimental results show that the proposed method can overcome the typical histogram equalization method’s shortcomings, such as image details loss, enhancement effect uncontrolled and noises magnification, and is an effective infrared image enhancement method.