摘要
传统基于边缘保持滤波的单幅图像快速去雾算法,在雾霾污染大气粒子散射作用下,图像会受到雾化背景的干扰,出现浓雾噪点,图像结构信息复原效果较差.提出一种基于混沌性的加权滤波图像快速去雾算法,通过分析雾天雾化背景干扰下的图像信息,获取图像数据的混沌特性,并对雾化图像进行最小颜色分量估计.在此基础上,结合带雾图像暗原色模型、时域和频域特征分量模型,构建出自适应加权滤波模型,完成基于混沌性的加权滤波图像快速去雾算法的改进设计.实验结果表明,采用该算法能避免雾化图像中间区域的颜色失真,降低雾化背景干扰,减少浓雾噪点,使远景的混沌特征得到合理的保留,在图像质量和运算性能方面都具有优越性.
This paper suggests a quick image defogging algorithm based on the chaotic weighted filtering to im- prove the traditional algorithm, the single image defogging algorithm with edge protection, with which there are the disturbance of foggy background, intense foggy hot pixel and the bad restore of image structure. The chaos of the image data will be acquired and the minimum color components will be estimated with the analysis of those disturbed images by the foggy background in haze. Then on this basis, with the combination of the dark channel model of foggy images and the component model of time domain and frequency domain features, the self - adapted weight filtering model is constructed, and the improved design of the quick image defogging algorithm based on chaotic weight filtering is completed. The results of experiment indicate that the adoption of this algorithm can avoid color distortion in the middle area of foggy images, lower the disturbance of foggy background, reduce the foggy hot pixel, and retain the chaotic characteristics of the distant view, and there- fore, the advantages in image quality and calculation are displayed.
出处
《郑州大学学报（工学版）》
CAS
北大核心
2016年第4期91-96,共6页
Journal of Zhengzhou University（Engineering Science）
基金
河南省科技攻关项目(132102210212)
