一种基于人眼视觉最小探测概率最大化的图像增强方法 | |
朱枫; 嵇冠群; 郝颖明; 吴清潇; 范晓鹏; 刘勋; 吴锦 | |
2014-06-04 | |
专利国别 | 中国 |
专利号 | CN103839231A |
专利类型 | 发明 |
产权排序 | 1 |
权利人 | 中国科学院沈阳自动化研究所 |
其他题名 | Image enhancement method based on maximization of human vision minimum detection probability |
中文摘要 | 本发明是一种基于人眼视觉最小探测概率最大化的图像增强方法,针对当前几种典型算法的不足提出了以目标函数为中心的图像质量增强算法,并且该目标函数充分考虑人眼视觉特性,旨在尽量提高原图像中不被人眼所感知的细节信息的对比度。算法的实施上包括三个主要环节:1、根据原始图像建立图像灰度值最小相邻关系表;2、根据图像灰度值最小相邻关系表对图像灰度进行合并;3、依据不同灰度背景下人眼灰度差探测概率函数,将合并后的灰度值进行灰度值重新分配,实现目标函数最大化。本发明可用于增强逆光、雾天或者照明昏暗等条件下拍摄得到的图像。 |
是否PCT专利 | 否 |
英文摘要 | The invention discloses an image enhancement method based on maximization of human vision minimum detection probability. An image quality enhancement algorithm with an objective function as a center is put forward aiming at defects of current some kinds of typical algorithms, and the objective function takes full consideration of human vision characteristics and aims to improve the contrast ratio of detail information which cannot be sensed by the human eyes. The implement of the algorithm comprises three main links of (1) establishing an image gray value minimum adjacent relation table according to an original image, (2) carrying out combination on image gray levels according to the minimum adjacent relation table of image gray values, and (3) carrying out gray value reallocation on the combined gray values according to a human eye gray level difference detection probability function under different gray level backgrounds to achieve the maximization of the objective function. The image enhancement method based on the maximization of the human vision minimum detection probability can be applied to images shot and obtained in enhanced backlight conditions or a foggy condition or a dark illumination condition or the like. |
申请日期 | 2012-11-27 |
语种 | 中文 |
专利申请号 | CN201210494757.1 |
专利代理 | 沈阳科苑专利商标代理有限公司 21002 |
内容类型 | 专利 |
源URL | [http://ir.sia.cn/handle/173321/14883] |
专题 | 沈阳自动化研究所_光电信息技术研究室 |
推荐引用方式 GB/T 7714 | 朱枫,嵇冠群,郝颖明,等. 一种基于人眼视觉最小探测概率最大化的图像增强方法. CN103839231A. 2014-06-04. |
个性服务 |
查看访问统计 |
相关权益政策 |
暂无数据 |
收藏/分享 |
除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。
修改评论