基于正弦变换的红外图像均衡算法

An Algorithm of Infrared Image Equalization Based on Sine Transform

针对红外图像的特点,提出一种基于正弦变换的红外图像均衡算法。通过设定合适的阈值,将图像灰度直方图中像素分布为零的灰度级进行完全压缩;将小于阈值的灰度级映射到最近的大于阈值的灰度级,然后将有效灰度级做基于正弦函数的拉伸变换后进行直方图均衡;最后将均衡后的图像灰度级等间距排列。该算法在压缩灰度冗余的同时进行基于正弦变换的直方图均衡,保留了图像细节部分,避免了常规直方图均衡所带来的图像细节丢失和灰度断层现象,提高了图像的细节和清晰度。实验结果表明该方法解决了热像仪信号处理中输入信号大动态范围和显示输出的小动态范围的矛盾,且算法简单、运算量小,适合在实时图像处理系统中采用。

Considering the characteristics of infrared image, a new algorithm of infrared image equalization based on sine transform is proposed. The appropriate threshold is calculated out. Then, the gray levels occupied no pixel are compressed completely through statistic of image gray histogram; the gray levels which pixel sum is less than the threshold are mapped to the nearest gray level greater than the threshold. Then, the gray levels are equalized after sine transform. Finally, the gray levels are arranged in equal distance. The gray-level redundancy can be compressed through this algorithm, which can also overcome the drawbacks of gray-level truncation and detail losing in histogram equalization. The details and definition of image can be improved. The simulation results show that ： 1 ） with the proposed algorithm, the conflict of the wide dynamic range of input and the narrow dynamic range of output in the thermal imaging systems is solved; and 2） the algorithm is simple, easy to realize, and is suitable for use in real-time image processing system.