Infrared Image Enhancement Technology Based On Human Visual Characteristics

Infrared thermal imaging technology is widely used in military,public security,medical industry,etc.But the infrared images generally have a series of problems.For example,detail edge is not prominent,contrast is low,gray scale distribution is not intensive and the vision is blurred.Therefore,the enhancement processing of the initial infrared image is an indispensable operation.The human eye plays a major role to evaluate the quality of the image.If the visual characteristics of the human eyes are fully taken into account in the enhancement of the image,better visual effects will be achieved.This paper proposes a multi-scale fractal enhancement algorithm based on frequency characteristic of human eyes.Firstly,by calculating the fractal dimension and the frequency range of the human eyes to choose the image details needed enhancement,and it enhanced image details based on characteristics of the human eye for resolution of gray scale.The main work of this paper is summarized as follows:(1)Based on the frequency sensitivity curves of human eyes,the image details of human eye sensitive area is extracted.Fourier transform the infrared image and filter the human-sensitive area in the frequency domain,and the coordinates of the sensitive region of human eyes in the image are obtained through inverse Fourier transform.(2)Aiming at the extraction of image details based on fractal dimension,an improved fractal dimension calculation method is proposed.Multi-scale space and Euclidean distance are introduced to calculate the fractal dimension of the pyramid image.Compared with the traditional single-scale fractal dimension calculation method,the improved method is more effective.(3)Combining the sensitive area of the human eye with the image details,the information is selectively enhanced according to the characteristics of the human eye for the resolution of gray scale.Firstly,the pixel points are divided into smooth points and minutia points by combining the fractal dimension of the pixels and the coordinates of the human-sensitive pixels,and according to the gray scale resolution characteristics of the human eye,the enhancement value of the corresponding pixel in each gray scale region is obtained,and the enhancement of the smooth point and the minutia point is finally performed.The multi-scale fractal enhancement algorithm based on human eye frequency characteristics proposed in our study is compared with the commonly used Retinex multi-scale enhancement algorithm(MSR),wavelet enhancement algorithm,and fractal wavelet enhancement algorithm.The results of experiments demonstrate that our algorithm is superior on the peak signal-to-noise ratio,average gradient index.The enhanced image highlights the contour of the target,which can solve the problem of blurred and poor visibility of the infrared image.