Dehazing algorithm based on interval estimation and adaptive constraints of transmittance
-
摘要:
针对去雾算法透射率估计不足与结果偏色等问题，提出了一种基于最小通道区间估计与透射率自适应约束模型的图像去雾算法。首先，采用不同尺寸最大值操作得到有雾图像的亮通道，并结合均值处理和频域滤波得到大气光估计；其次，从大气成像理论出发，以有雾图像最小通道为约束，分别以平面模型和自适应映射模型拟合无雾图像最小通道上下边界，并获得无雾图像最小通道和透射率初始估计；最后，对透射率作滤波平滑与自适应边界约束，得到优化透射率，并根据大气散射模型得到复原结果。实验表明：所提算法复原结果颜色自然、亮度适宜、去雾彻底、细节信息丰富且时间复杂度较低，有效解决了透射率估计不足和偏色等问题。
Abstract:In order to solve the problems such as insufficient transmittance estimation and color cast of results of dehazing algorithms, an image restoration algorithm based on minimum channel interval estimation and transmittance adaptive constraint model is proposed. Firstly, bright channel of hazy image is obtained by using maximum operation of different sizes, and average value processing and frequency domain filtering are combined to get the atmospheric light estimation. Secondly, starting from the atmospheric imaging theory, minimum channel of hazy image is regarded as a constraint, then upper and lower boundaries of minimum channel of hazy image are fitted by plane model and adaptive mapping model respectively, and minimum channel of dehazed image and initial transmittance estimation are obtained. Finally, the initial transmittance can be refined by filter smoothing and adaptive boundary constraints to obtain the optimized transmittance, and according to atmospheric scattering model, restoration results are obtained. Experiments show that the restoration results of the proposed algorithm have natural colors, appropriate brightness, thorough degree of dehazing, rich detailed information and low time complexity, which effectively solves the problems of insufficient transmittance estimation and color cast.
-
图 3 大气散射模型空间分布与简化分布
Figure 3. Spatial distribution and simplified distribution of atmospheric scattering model
图 5 最小通道真实值与本文拟合结果对比
Figure 5. Comparison of true value of minimum channel with fitting results of this paper
图 6 含天空区域图像透射率与复原结果对比
Figure 6. Comparison of transmittance and restoration results in sky-images
图 7 不含天空区域图像透射率与复原结果对比
Figure 7. Comparison of transmittance and restoration results in no-sky-images
图 8 本文透射率与标准透射率对比
Figure 8. Comparison of transmittance in this paper and standard transmittance
表 1 远近景区域误差ω
Table 1. Near- and far-field regional error ω
远近景区域 ω 图像1 图像2 图像3 图像4 图像5 近景 0.059 3 0.030 8 0.009 8 0.010 9 0.013 8 中景 0.056 1 0.034 3 0.016 4 0.012 9 0.016 2 远景 0.041 2 0.020 4 0.012 6 0.020 1 0.019 3 
下载: 导出CSV
表 2 MSE指标对比
Table 2. Comparison of MSE indicators
算法 MSE 合成有雾图像1 合成有雾图像2 合成有雾图像3 合成有雾图像4 合成有雾图像5 DCP算法 0.591 9 0.815 7 0.640 7 0.039 8 0.022 4 本文算法 0.028 9 0.069 8 0.049 9 0.038 1 0.004 2
下载: 导出CSV
表 3 真实有雾图像复原结果客观指标对比
Table 3. Comparison of objective indicators of true hazy image restoration results
指标 DCP算法[1] CAP算法[6] Haze Removal算法[5] Dehaze-Net算法[7] AOD-Net算法[9] 多尺度提取算法[8] 本文算法 HCC 0.123 7 0.114 2 0.139 5 0.139 5 0.120 6 0.126 8 0.149 3 E 7.024 2 7.244 8 7.279 8 6.970 4 6.909 4 7.065 0 7.580 8 CG 0.297 4 0.301 8 0.384 9 0.312 7 0.320 6 0.374 7 0.386 5 UQI 0.701 4 0.728 1 0.849 7 0.700 9 0.712 4 0.836 3 0.872 7 VCM 57.98 51.49 60.79 55.74 56.01 57.88 60.62 T/s 2.096 1.180 1.274 2.843 2.874 2.979 1.396 注：黑体数据表示最优值。
下载: 导出CSV
表 4 Middlebury测试集复原结果客观指标对比
Table 4. Comparison of objective indicators of Middlebury test set restoration results
下载: 导出CSV
表 5 RESIDE测试集复原结果客观指标对比
Table 5. Comparison of objective indicators of RESIDE test set restoration results
下载: 导出CSV
-
[1] HE K M, SUN J, TANG X O, et al. Single image haze removal using dark channel prior[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2011, 33(12): 2341-2353. doi: 10.1109/TPAMI.2010.168 [2] TAN R. Visibility in bad weather from a single image[C]//IEEE Conference on Computer Vision and Pattern Recognition(CVPR). Piscataway: IEEE Press, 2008: 1-8. [3] MENG G, WANG Y, DUAN J, et al. Efficient image dehazing with boundary constraint and contextual regularization[C]//IEEE International Conference on Computer Vision (ICCV). Piscataway: IEEE Press, 2014: 617-624. [4] XU Y, GUO X, WANG H, et al. Single image haze removal using light and dark channel prior[C]//IEEE/CIC International Conference on Communication in China (ICCC). Piscataway: IEEE Press, 2016: 1-6. [5] YANG Y, WANG Z. Haze Removal: Push DCP at the edge[J]. IEEE Signal Processing Letters, 2020, 27: 1405-1409. doi: 10.1109/LSP.2020.3013741 [6] ZHU Q, MAI J, SHAO L. A fast single image haze removal algorithm using color attenuation prior[J]. IEEE Transactions on Image Processing, 2015, 24(11): 3522-3533. doi: 10.1109/TIP.2015.2446191 [7] CAI B, XU X, JIA K, et al. Dehaze-Net: An end-to-end system for single image haze removal[J]. IEEE Transactions on Image Processing, 2016, 25(11): 5187-5198. doi: 10.1109/TIP.2016.2598681 [8] REN W Q, LIU S, ZHANG H, et al. Single image dehazingvia multi-scale convolutional neural-networks[C]//ECCV 2016: Computer Vision, 2016: 154-169. [9] LI B Y, PENG X, WANG Z, et al. AOD-Net: All-in-one de-hazing network[C]//IEEE International Conference on Computer Vision (ICCV). Piscataway: IEEE Press, 2017: 4780-4788. [10] LIU R, FAN X, HOU M, et al. Learning aggregated transmission propagation networks for haze removal and beyond[J]. IEEE Transactions on Neural Networks and Learning Systems, 2019, 30(10): 2973-2986. doi: 10.1109/TNNLS.2018.2862631 [11] QIAN W, ZHOU C, ZHANG D. CIASM-Net: A novel convolutional neural network for dehazing image[C]//IEEE International Conference on Computer and Communication Systems (ICCCS). Piscataway: IEEE Press, 2020: 329-333. [12] LI C, GUO C, GUO J P. PDR-Net: Perception-inspired single image dehazing network with refinement[J]. IEEE Transactions on Multimedia, 2020, 22(3): 704-716. doi: 10.1109/TMM.2019.2933334 [13] ZHANG J, TAO D. FAMED-Net: A fast and accurate multi-scale end-to-end dehazing network[J]. IEEE Transactions on Image Processing, 2020, 29: 72-84. doi: 10.1109/TIP.2019.2922837 [14] WANG A N, WANG W H, LIU J L, et al. AIP-Net: Image-to-image single image dehazing with atmospheric illumination prior[J]. IEEE Transactions on Image Processing, 2019, 28(1): 381-393. doi: 10.1109/TIP.2018.2868567 [15] LI R, PAN J, HE M, et al. Task-oriented network for image dehazing[J]. IEEE Transactions on Image Processing, 2020, 29: 6523-6534. doi: 10.1109/TIP.2020.2991509 [16] QIN X, WANG Z, BAI Y, et al. FFA-Net: Feature fusion attention network for single image dehazing[C]//Association for the Advance of Artificial Intelligence(AAAI), 2020: 11908-11915. [17] SULAMI M, GLATZER I, FATTAL R, et al. Automatic recovery of the atmospheric light in hazy images[C]//IEEE International Conference on Computational Photography. Piscataway: IEEE Press, 2014: 1-11. [18] SUN W, WANG H, SUN C, et al. Fast single image haze removal via local atmospheric light veil estimation[J]. Computers & Electrical Engineering, 2015, 46: 371-383. [19] HE K, JIAN S, TANG X. Guided image filtering[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2013, 35(6): 1397-1409. doi: 10.1109/TPAMI.2012.213 [20] BI G, REN J, FU T, et al. Image dehazing based on accurate estimation of transmission in the atmospheric scattering model[J]. IEEE Photonics Journal, 2017, 9(4): 1-18. [21] MIM X, ZHAI G, GU K, et al. Quality evaluation of image dehazing methods using synthetic hazy images[J]. IEEE Transactions on Multimedia, 2019, 21(9): 2319-2333. doi: 10.1109/TMM.2019.2902097 [22] 郭璠, 蔡自兴. 图像去雾算法清晰化效果客观评价方法[J]. 自动化学报, 2012, 38(9): 1410-1419. https://www.cnki.com.cn/Article/CJFDTOTAL-MOTO201209004.htmGUO F, CAI Z X. Objective evaluation method of image dehazing algorithm sharpening effect[J]. Acta Automatica Sinica, 2012, 38(9): 1410-1919(in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-MOTO201209004.htm [23] 杨燕, 张得欣, 岳辉. 基于最小值通道与对数衰减的图像融合去雾算法[J]. 北京航空航天大学学报, 2020, 46(10): 1844-1852. doi: 10.13700/j.bh.1001-5965.2019.0552YANG Y, ZHANG D X, YUE H. Image fusion dehazing algorithm based on minimum channel and logarithmic attenuation[J]. Journal of Beijing University of Aeronautics and Astronautics, 2020, 46(10): 1844-1852(in Chinese). doi: 10.13700/j.bh.1001-5965.2019.0552 -
下载:
点击查看大图
计量
- 文章访问数: 278
- HTML全文浏览量: 41
- PDF下载量: 48
- 被引次数: 0
