节点文献

基于对比度保持下的图像去色方法研究

Contrast-Reserved-Based Decolorization of Color Images and Its Applications

【作者】 张欣

【导师】 孙佳宁;

【作者基本信息】 东北师范大学 , 计算数学, 2016, 硕士

【摘要】 对彩色图像进行去色(Decolorization of Color Image),是将彩色图像灰度化的一种操作。然而在传统的"RGB2GRAY"的去色操作中,彩色图像的色彩对比度无法在转化得到的灰度图像的灰度对比度中得到很好的保持。于是,就如何在彩色图像去色过程中,更快捷高效的把彩色图像的色度、亮度、色彩饱和度在去色后的灰度图像中通过灰度对比度保持下来,则成为当今彩色图像去色问题研究的重点。本文通过结合拉普拉斯分布(Laplace distribution)和双峰分布(Bimodal distribution),提出了一种全新的基于对比度的彩色图像去色模型。并通过引入弱光滑函数(Weak smooth function)对新的去色模型的求解形式进行了重建,从而给出了快速有效的求解算法。为了说明本文成果的有效性和应用性,本文的数值实验具体分为色彩标准图像的测试、普通数码图像的应用测试、雾霾图像和色盲图像的应用测试这三个部分,并对具体的去色测试效果进行了评估。相关成果为未来的图像去色方法研究提供了良好的依据。

【Abstract】 A color image to desaturate(Decolorization of Color Images), is an operation of color image gray processing. However, in the traditional "RGB2GRAY" operation of decolorization, color contrast of a color image can not be well maintained by grayscale contrast in the converted to gray one. So, how to maintain the chrominance, brightness, color saturation in the color image by holding down the gray contrast of the gray im-age more quickly and efficiently becomes the research focus on decolorization of color images.In this paper, by combining Laplace distribution and a bimodal distribution, we present a new contrast-based model of color images. And by introducing a weakly smooth function. The solving form of the new model has been restructured, then we give a fast and efficient algorithm.To illustrate the effectiveness and applicability of the results of this paper, Numer-ical experiments have been divided into three parts including standard color image test, a conventional digital image application testing and haze image and color-blind image application testing. And the specific test results have been evaluated. Related outcomes for the future research methods of decolorization provide a good basis.

  • 【分类号】TP391.41
  • 【被引频次】2
  • 【下载频次】85
节点文献中: 

本文链接的文献网络图示:

本文的引文网络