【作者】 刘博；

【导师】 胡正平； 王成儒；

【作者基本信息】 燕山大学 ， 电路与系统， 2009， 硕士

【摘要】 在图像处理中,图像增强技术对于提高图像的质量起着重要的作用。它通过有选择地强调图像中某些信息而抑制掉另一些信息,以改善图像的视觉效果,将原图像转换成一种更适合于人眼观察和计算机进行分析处理的形式。但由于图像增强与感兴趣的物体特性、观察者的习惯和处理目的密切相关,带有很强的针对性,因此,图像增强算法的应用也是有针对性的。本文针对图像增强过程中存在的问题,在借鉴国内外最新思路的基础上,提出了一些新的图像增强思路。首先,为解决多层次图像增强过程缺乏可控性和可选择性的不足,提出了基于模糊松弛迭代的分层图像增强算法。通过设置模糊松弛参数和选择合适的迭代次数,实现了对不同层次图像内容的可控式增强。其次,为解决直方图均衡过程中过增强高频数灰度级,压缩低频数灰度级和传统图像增强算法无法较好突出原图直方图特征的不足,同时能够使增强对象更具目的性,提出了极大灰度频数抑制结合动态直方图均衡图像增强算法和基于最优自适应直方图规定化函数引导的动态分层图像增强算法。两算法利用极大灰度频数抑制策略较好解决了过度增强高频数灰度的问题,利用构建最优自适应规定化目的直方图思路较好解决了无法突出原直方图特征的问题,利用动态映射思路解决了增强对象缺乏目的性的问题。最后,为有效实现灰度图像彩色化,使彩色化后的图像更自然、逼真,提出了基于相似性竞争选择的多彩色图像自适应混合颜色迁移算法,该算法利用自动图像检索技术和多图像自适应混合颜色迁移技术较好地达到了这个目的。更多还原

【Abstract】 Image enhancement technology plays a very important role to improve image quality in image processing. By enhancing some information and restraining other information selectively it can improve image visual effect and transform an image to another form which adapts to human observation or computer analysis and processing better. Because image enhancement is closely related to the property of the interested target, the habit of observers and the specific processing object. So, the image enhancement algorithm is only aimed at the given processing object, too. According to some new thoughts, this paper puts forward some new image enhancement thoughts to solve some problems during the process of image enhancement.Firstly, in order to solve the drawback of the lack of control and option in the multi-level image enhancement process, multi-level image enhancement algorithm based on fuzzy relaxation iterative procedure is proposed. The different level image content can be enhanced by choosing fuzzy relaxation parameters and suitable iterative processing times.Secondly, in order to solve the problem that histogram equalization technology leads to excessive enhancement of high frequency gray level and compresses low frequency gray level, and the problem that traditional image enhancement algorithm could not emphasize the characteristics of original histogram. At the same time, in order to make the enhance process possess purpose, image enhancement algorithm combines maximum gray frequency restrict with dynamic histogram equalization and dynamic partition image enhancement algorithm based on optimal adaptive histogram specification function induct are proposed. The problem of excessive enhancement is solved by maximum gray frequency restrict, the problem that could not emphasize the characteristics of original histogram is solved by constructing a optimal desired histogram adaptively, and the problem of lacking purpose is solved by dynamic mapping.Finally, in order to color gray image effectively, and make the color image more natural and realistic, adaptive coloring grayscale image based on multiple color images competing selection model is proposed. The purpose can be achieved by the automatic image retrieval technology and multi-image adaptive mixed color transfer technology.更多还原