红外图像自适应逆直方图增强技术

Infrared image adaptive inverse histogram enhancement technology

在红外图像中,传统直方图均衡图像时细节像素容易被大量的背景像素淹没,导致图像产生过亮、过暗等现象。基于这样的状况,提出一种自适应逆直方图均衡化细节增强算法。该算法通过逆向统计、自适应选取阈值以及分段映射来增强图像细节。相比于传统直方图均衡化算法,逆直方图均衡化算法明显改善了图像在不同灰度层分布的视觉效果,使图像的不同区域亮度得到不同程度的增强。而且该算法在能够达到更好的图像处理效果的前提下仍然能够通过优化计算方法保证实时性,高效性,并且适合在FPGA硬件移植中采用。

In infrared images, when the traditional histogram equalizes the image, the detail pixels are easily immerged by the background pixels, resulting in the image being too bright and too dark. Based on this situation,an adaptive inverse histogram equalization algorithm was proposed in this paper. The algorithm enhanced image details by inverse statistics, adaptive selection threshold and segmentation mapping. Compared with the traditional histogram equalization algorithm, the inverse histogram equalization algorithm significantly improve the image visual effect in different gray level distributions and enhance the details of different areas of the image to different degrees. Moreover, under the premise of achieving better image processing effects, this algorithm can still guarantee real-time performance and high efficiency by optimizing calculation methods, and is suitable for FPGA hardware transplantation.