电信科学 ›› 2019, Vol. 35 ›› Issue (3): 99-106.doi: 10.11959/j.issn.1000-0801.2019042

• 研究与开发 • 上一篇    下一篇

一种基于人眼视觉特性的Curvelet域紫外图像增强算法

张斌,吴鹏,袁杰,孙凌卿,傅启明,张永泽,陆宇   

  1. 江苏电力信息技术有限公司,江苏 南京 210009
  • 修回日期:2019-02-19 出版日期:2019-03-01 发布日期:2019-03-23
  • 作者简介:张斌(1973- ),男,江苏电力信息技术有限公司教授级高级工程师,主要研究方向为模式识别。|吴鹏(1979- ),男,江苏电力信息技术有限公司高级工程师,主要研究方向为知识发现、模式识别。|袁杰(1981-),男,江苏电力信息技术有限公司高级工程师,主要研究方向为机器学习、模式识别。|孙凌卿(1982- ),男,现就职于江苏电力信息技术有限公司,主要研究方向为图像处理。|傅启明(1980- ),男,江苏电力信息技术有限公司助理工程师,主要研究方向为图像处理。|张永泽(1991- ),男,现就职于江苏电力信息技术有限公司,主要研究方向为图像处理。|陆宇(1990- ),男,现就职于江苏电力信息技术有限公司,主要研究方向为图像处理。

An ultraviolet image enhancement algorithm in Curvelet domain based on human visual characteristics

Bin ZHANG,Peng,YUANJie WU,Lingqing SUN,Qiming FU,Yongze ZHANG,Yu LU   

  1. Jiangsu Electric Power Information Technology Co.,Ltd.,Nanjing 210009,China
  • Revised:2019-02-19 Online:2019-03-01 Published:2019-03-23

摘要:

针对紫外图像本身存在的对比度低、成像不清晰、整体成像过暗或过亮且受噪声干扰等特点,从人眼视觉的掩蔽特性出发,提出了一种基于人眼视觉特性的 Curvelet 域紫外图像增强算法。通过 Curvelet 变换将紫外图像信息分解到不同尺度、不同方向子带,再依据各子带建立HVS模型,利用其遮蔽性对高频分量进行非线性增益来提高图像对比度,同时非线性灰度拉伸低频分量可以提高图像整体亮度;最终将处理过的高低频分量通过 Curvelet 逆向重建,得到增强后的紫外图像。实验结果表明:增强后的图像符合人眼视觉对图像信息的提取,突出了边缘和目标的轮廓信息,在主观视觉和定量指标方面都具有明显的效果。

关键词: 紫外图像, 掩蔽特性, Curvelet变换, 非线性增益, 增强

Abstract:

Aiming at the characteristics such as low contrast,unclear imaging,over-dark or over-bright overall imaging and noise interference in ultraviolet image itself,a Curvelet domain ultraviolet image enhancement algorithm was proposed based on the masking characteristics of human vision.Firstly,the ultraviolet image information was decomposed to different scales and direction sub-bands by Curvelet transform.Secondly,the HVS model was established according to each sub-band,and with its masking characteristics,the nonlinear gain of the high frequency component was used to improve the image contrast.Meanwhile,the low frequency component of nonlinear grayscale stretching could improve the overall brightness of the image.Finally,the processed high and low frequency components were reconstructed by Curvelet backward to obtain enhanced ultraviolet images.The experimental results show that the enhanced image accords with the extraction of image information from human vision and highlights the contour information of edge and target.In addition,it has obvious effect in subjective vision and quantitative index.

Key words: ultraviolet image, masking characteristic, Curvelet transform, nonlinear gain, enhancement

中图分类号: 

No Suggested Reading articles found!