张赛赛,田益民,王飞,杜云飞.基于对比度和饱和度新的先验图像去雾算法[J].包装工程,2024,45(11):191-197.
ZHANG Saisai,TIAN Yimin,WANG Fei,DU Yunfei.A New Prior Image Defogging Algorithm Based on Contrast and Saturation[J].Packaging Engineering,2024,45(11):191-197. |
| 基于对比度和饱和度新的先验图像去雾算法 |
| A New Prior Image Defogging Algorithm Based on Contrast and Saturation |
| 投稿时间：2023-09-05 |
| DOI：10.19554/j.cnki.1001-3563.2024.11.022 |
| 中文关键词: 图像去雾 导向滤波 非线性逆策略 对比度 饱和度 |
| 英文关键词: image defogging guided filtering nonlinear inverse strategy contrast saturation |
| 基金项目:北京印刷学院博士启动金项目(27170122004);北京印刷学院校级一般项目(Ec202205);北京印刷学院校级创新团队项目(Ea202304);北京市教育委员会项目(KM 202110015001) |
|
| 摘要点击次数: |
| 全文下载次数: |
| 中文摘要: |
| 目的 针对目前图像印刷中清晰度和分辨率不高，为了提高图像的质量和观感，使雾图像细节更清晰，处理效率更高，文中设计一种基于对比度和饱和度的新的图像去雾先验算法。方法 首先对采集的雾图像利用对比度、饱和度和亮度先验来初步估计透射图，然后借助伽马校正的非线性逆策略增强图像的对比度和饱和度来防止传输估计过高或过低，并利用导向滤波对非线性逆策略的粗透射估计进行优化处理，最后使用非线性对比度拉伸方法对复原图像进行亮度进一步的增强。结果 通过实验测试得出，对一个450´450的图像，本文提出的算法相较目前最新算法在PNSR值和SSIM值分别提高了9.220 4和0.332 7，处理时间减少了0.079 38s。结论 图像复原效果较好，细节更加丰富，清晰度更高，效率更高。 |
| 英文摘要: |
| The work aims to design a new prior image defogging algorithm based on contrast and saturation tosolve the low clarity and resolution in current image printing, and improve the quality and perception of images, so that the fog image details can be clearer and more efficient in processing. The initial estimation of the transmission map was first performed on the captured fog image with contrast, saturation and luminance as prior. Then the contrast and saturation of the image were enhanced with the help of a gamma-corrected nonlinear inverse strategy to prevent the transmission estimation from being too high or too low, the coarse transmission estimation of the nonlinear inverse strategy was applied for optimization by guided filtering. And finally, the recovered image was further enhanced with luminance using a nonlinear contrast stretching method. Through experimental tests, it is concluded that for a 450×450 image, the algorithm proposed in this paper improved the PNSR and SSIM values by 9.220 4 and 0.332 7, respectively, and reduced the processing time by 0.079 38 s compared with the current state-of the-art algorithm. In conclusion, the image restoration effect is better. It has more detail, and higher clarity and efficiency. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
|
|
|
渝公网安备 50010702501716号