基于暗原色先验与小波变换的图像去雾方法.PDFVIP

收稿日期:2015-05-22 修回日期:2015-08-25 网络出版时间:2016-01-26 基金项目:安徽省自然科学基金重大项目(KY2014ZD004) 作者简介:许 杰(1989-),男,硕士研究生,研究方向为图像处理与模式识别;杨会成,教授,硕士生导师,研究方向为图像处理与模式识别。 网络出版地址:http: / / www. cnki. net / kcms / detail / 61. 1450. TP. 1521. 058. html 基于暗原色先验与小波变换的图像去雾方法 许 杰,杨会成 (安徽工程大学 电气工程学院,安徽 芜湖 241000) 摘 要:针对雾天采集图像清晰度差、对比度低的问题,提出一种基于暗原色先验与小波变换的图像去雾方法。 先利用暗 原色先验规律直接从一幅有雾的图像中估计出景物透射率,恢复出比较清晰化的图像;再将图像由 RGB空间转换到 YUV 空间,对亮度 Y图像采用正交 Haar小波变换进行处理,得到小波各层的分解系数,利用阈值滤波的细节系数增强法对雾天 图像的高频子带小波系数进行处理,经小波重构得到增强后的亮度 Y图像,最后与颜色分量 U 、 V合并得到清晰化的雾天 图像。 实验结果表明:该方法对雾天模糊图像处理后,能有效提高图像的清晰度、细节和层次感,一定程度上改善了图像 的视觉效果,是一种有效的雾天图像清晰化方法。 关键词:小波变换;图像清晰化;暗原色先验法;小波分解;阈值去噪 中图分类号:TP391. 41 文献标识码:A 文章编号:1673-629X(2016)02-0066-03 doi:10. 3969 / j. issn. 1673-629X. 2016. 02. 015 A Defogging Method for Image Based on Binding Priori Dark Colors and Wavelet Transform XU Jie,YANG Hui-cheng (School of Electric Engineering,Anhui Polytechnic University,Wuhu 241000,China) Abstract:For poor visibility and low contrast in fog image,a defogging method was proposed based on wavelet transform and dark chan- nel prior colors. First use of dark colors rule estimates the scene transmittance rate directly from a foggy image,recovering relatively clear picture. Then the image space is converted from RGB to YUV,the bright Y image was processed by using orthogonal Haar wavelet,obtai- ning coefficients of each layer,and use of detail coefficients image enhancement method to process high frequency sub-band wavelet co- efficients,luminance Y images were enhanced after wavelet reconstruction. Finally,the color components U , V are merged to get a clear picture with the fog color. The results show that the method,which is effective,is applied to image after fog,the clearance,detail and laye- ring has been improved significantly,to some extent improved visual effects. Key words:wavelet transform;image clarity;dark colors priori method;wavelet decomposition;threshold denoising 0 引 言 雾天拍摄图像时,图像因受雾气的干扰,所获得的 图像质量存在退化现象,变得模糊。 因