液晶与显示, 2020, 35 (10): 1079, 网络出版: 2021-01-22
结合双阈值定位与透射率约束的航拍图像去雾
Aerial image defogging based on dual-threshold position and transmittance constraint
图像去雾 航拍图像 透射率 大气散射模型 图像复原 image defogging aerial image transmittance atmospheric scattering model image restoration
摘要
为了解决航拍图像受雾气影响而导致视见度降低的问题,提出一种结合双阈值定位大气光与约束优化透射率的去雾算法。针对航拍图像基本不含天空且边缘细节丰富的特点,首先通过梯度阈值和亮度阈值的双重定位限制,提高大气光强度的估计准确性;其次构造约束条件求解透射率,并利用双指数滤波获取精确值;最后根据人眼视觉对信号波动的感知定律,提出针对去雾图像的色调映射方法。仿真实验表明,恢复图像中的雾气被有效去除,颜色自然真实,图像的平均信息熵为7.659,平均梯度为16.631,均比现有算法有了一定程度的提升。研究表明,该算法对不同地貌场景的重现和细节特征的恢复非常有效,能够满足航拍图像去雾的应用需求。
Abstract
In order to improve the visibility of aerial image in the foggy weather, an image defogging algorithm based on dual-threshold position and transmittance constraint is proposed. Firstly, according to the features of non-sky and rich details from aerial images, both the gradient threshold and brightness threshold are used to keep positional limitation of the atmospheric light and therefore the accuracy has been improved effectively. Secondly, the transmittance is estimated by creating constraint, then the precise value is obtained by performing bi-exponential edge preserving filtering. Finally, a tone mapping method for restored image is proposed based on the perception law of human vision to the signal fluctuation. Simulation results show that the restored images not only remove fog effectively, but also keep color natural and real. The average entropy and gradient from test images are 7.659 and 16.631, which achieve a degree of improvement compared with other algorithms. The proposed algorithm gets satisfactory restoration for different landforms and detail features, thereby it can meet the application requirement of aerial image defogging.
王伟鹏, 项文杰, 戴声奎. 结合双阈值定位与透射率约束的航拍图像去雾[J]. 液晶与显示, 2020, 35(10): 1079. WANG Wei-peng, XIANG Wen-jie, DAI Sheng-kui. Aerial image defogging based on dual-threshold position and transmittance constraint[J]. Chinese Journal of Liquid Crystals and Displays, 2020, 35(10): 1079.