International Open Access Journal Platform

多尺度Retinex算法在不同彩色空间图像去雾效果评估

Multi-scale Retinex Algorithm to Evaluate the Fogging Effect in Different Color Space Images

Modern Computer Technology and Application / 2019,1(1): 1-10 / 2019-10-29 5834 9028

View Download PDF

• Authors: 崔倩倩*      李瑞洋     
• Information:

  西京学院信息工程学院，西安

• Keywords: Retinex；去雾；YCbCr
• Retinex; Defogging; YCbCr
• Abstract: 由于近年来雾霾频发，对于室外图像的采集造成一定困扰。图像去雾成为新的热点话题。本文针对图像去雾，首先研究了Retinex算法，并介绍了图像处理中常用的RGB、HSV、RGB、YCbCr这4种颜色空间。对HSV、RGB、YCbCr这3种颜色图像只进行亮度分量去雾，加快了速度。同时给出了信息嫡（Entropy）、对比度（Contrast）、平均亮度（Luminance）、平均梯度（Gradient）4个指标对去雾后的图像进行评价。最后得出结论，Retinex算法在YCbCr颜色空间中去雾的效果最佳。
• Due to the frequent occurrence of haze in recent years, the collection of outdoor images has caused some problems. Image defogging has become a new hot topic. First, Retinex algorithm is studied for fog removal, and four color Spaces commonly used in image processing, including RGB, HSV, RGB and YCbCr, are introduced. HSV, RGB, YCbCr color images are only used for brightness componentdefogging, which picks up the speed. Then, Information Entropy, Contrast, average Luminance and average Gradient that the four indicators are given to evaluate image after fog removal. Finally, Retinex algorithm has the best effect to remove fog in YCbCr color space.
• DOI: https://doi.org/10.35534/mcta.0101001
• Cite: 崔倩倩，李瑞洋．多尺度 Retinex 算法在不同彩色空间图像去雾效果评估［J］．现代计算机技术与应用，2019，1（1）：1-10．