红外图像的多尺度分形增强算法

本文提出了一种红外图像增强算法,结合多尺度分形维数和人眼频率特性对图像进行增强.该算法首先在多尺度条件下计算图像的分形维数,然后采用巴特沃斯滤波器,从频域抽取人眼敏感区的图像信号,最后将滤波图像结合分形维数值进行处理.结果表明,本文算法使图像中边缘及细节部位的亮度和对比度有提升,相较于单一尺度,图像的细节和边缘更加丰富,相比于传统的分形算法图像更适合人眼观看.

Multi-scale Fractal Enhancement Algorithm for Infrared Image

An infrared image enhancement algorithm is proposed in this paper. To enhance the infrared image, this algorithm combined multi-scale fractal dimension with the frequency characteristic of human eyes. First, the fractal dimension of image was computed in multi-scale space, and then the Butterworth filter was used to extract the sensitive image region of the human eyes from the frequency domain. Finally, the part that conformed to the theoretical value of the fractal dimension in the sensitive image region of human eyes was enhanced. The experimental result showed that the algorithm in this study improved the brightness and contrast of the edges and the details in infrared image. And the details and edges of the image were more prominent than the algorithm in single-scale space. Compared with the traditional algorithm of fractal value, the image is more suitable for human eyes.