张俊林, 石冬阳, 杨慧敏, 聂玲, 刘天光, 武正萍. 基于受限光值与透射率修正的图像去雾算法[J]. 红外技术, 2023, 45(6): 613-621.
引用本文: 张俊林, 石冬阳, 杨慧敏, 聂玲, 刘天光, 武正萍. 基于受限光值与透射率修正的图像去雾算法[J]. 红外技术, 2023, 45(6): 613-621.
ZNANG Junlin, SHI Dongyang, YANG Huimin, NIE Ling, LIU Tianguang, WU Zhengping. Image Defogging Algorithm Based on Limited Light Value and Transmittance Correction[J]. Infrared Technology , 2023, 45(6): 613-621.
Citation: ZNANG Junlin, SHI Dongyang, YANG Huimin, NIE Ling, LIU Tianguang, WU Zhengping. Image Defogging Algorithm Based on Limited Light Value and Transmittance Correction[J]. Infrared Technology , 2023, 45(6): 613-621.

基于受限光值与透射率修正的图像去雾算法

Image Defogging Algorithm Based on Limited Light Value and Transmittance Correction

  • 摘要: 针对暗通道先验去雾算法在滤波窗口较小时得到的去雾后图像存在颜色失真、对引入因子的选择及明亮区域透射率的计算存在误差、去雾后图像的抗噪性能较弱等问题,提出基于受限光值与透射率修正的图像去雾算法。首先对大气光值A设定阈值上限;其次通过建立引入因子与结构相似度的对应关系以获得最佳引入因子;并在引入容差机制的基础上进一步提出透射率优化方法;最后在所提去雾算法基础上融入高斯滤波算法,并调整去雾图像亮度以提升可视化效果。仿真结果表明,运用所提算法得到的图像PSNR值、SSIM值、Entropy值相对于改进前分别平均提升9.9964 dB、8.57%、0.3732,验证了所提算法的有效性与优越性。

     

    Abstract: In this study, an image demist algorithm based on limited light value and transmittance correction is proposed. The aim of the study was to address the issues of color distortion in the demist image obtained by dark channel prior demist algorithm when the filtering window is small, error in the selection of introduction factor and calculation of the transmittance of the bright area, and weak anti-noise performance of the demist image. First, the upper threshold of atmospheric light value A was set. Second, the best introduction factor was obtained by establishing the corresponding relationship between the introduction factor and structural similarity. On the basis of introducing the tolerance mechanism, the transmission optimization method was further proposed. Finally, based on the proposed defogging algorithm, a Gaussian filtering algorithm was incorporated, and the brightness of the defogging image was adjusted to improve the visualization effect. The simulation results showed that the PSNR and SSIM values and entropy value of the image obtained by the proposed algorithm were 9.9964 dB, 8.57%, and 0.3732 higher than those before the improvement, respectively; thus, the effectiveness and superiority of the proposed algorithm were verified.

     

/

返回文章
返回