Laser & Optoelectronics Progress, Volume. 58, Issue 22, 2210012(2021)
Image Dehazing Algorithm Based on Haze Distribution and Adaptive Attenuation
Figures & Tables(14)
Fig. 1. Hazy image and its characteristic distribution. (a) Hazy image; (b) brightness component; (c) saturation component
Fig. 2. Hazy image and haze distribution. (a1)--(a5) Hazy image; (b1)--(b5) haze distribution
Fig. 4. Analysis of restoration results of synthetic hazy images obtained by different methods. (a1)--(a4) Hazy image; (b1)--(b4) clear image; (c1)--(c4) transmittance obtained by model in Ref.[5]; (d1)--(d4) restoration result obtained by model in Ref.[5]; (e1)--(e4) transmittance obtained by proposed method; (f1)--(f4) restoration result obtained by proposed method
Fig. 5. Comparison of proposed atmospheric light and local atmospheric light in Ref. [6]. (a1) (a2) Hazy image; (b1) (b2) local atmospheric light; (c1) (c2) restoration result obtained by local atmospheric light; (d1) (d2) proposed atmospheric light; (e1) (e2) restoration result obtained by proposed atmospheric light
Fig. 7. Schematic diagram of dehazing by proposed method. (a) Hazy image; (b) haze distribution; (c) transmittance; (d) atmospheric light; (e) dehazed image
Fig. 8. Comparison of restoration results of natural hazy images. (a1)--(a5) Original image; (b1)--(b5) restoration result obtained by algorithm in Ref.[1]; (c1)--(c5) restoration result obtained by algorithm in Ref.[3]; (d1)--(d5) restoration result obtained by algorithm in Ref.[5]; (e1)--(e5) restoration result obtained by algorithm in Ref.[13]; (f1)--(f5) restoration result obtained by algorithm in Ref. [14]; (g1)--(g5) restoration result obtained by proposed algorithm
Fig. 9. Comparison of restoration results of dense hazy images. (a1)(a2) Original image;(b1)(b2) restoration result obtained by algorithm in Ref.[1]; (c1)(c2) restoration result obtained by algorithm in Ref.[3]; (d1)(d2) restoration result obtained by algorithm in Ref.[5]; (e1)(e2) restoration result obtained by algorithm in Ref.[13]; (f1)(f2) restoration result obtained by algorithm in Ref. [14]; (g1)(g2) restoration result obtained by proposed algorithm
Fig. 10. Comparison of restoration results of synthetic hazy images. (a1)--(a4) Original image; (b1)--(b4) restoration result obtained by algorithm in Ref.[1]; (c1)--(c4) restoration result obtained by algorithm in Ref.[3]; (d1)--(d4) restoration result obtained by algorithm in Ref.[5]; (e1)--(e4) restoration result obtained by algorithm in Ref.[13]; (f1)--(f4) restoration result obtained by algorithm in Ref. [14]; (g1)--(g4) restoration result obtained by proposed algorithm
Table 1. Texture information values of different images
View tableTable 1. Texture information values of different images
Image MEne MEnt MCon image1 0.0502 0.0045 0.1354 image2 0.0582 0.1103 0.2177 image3 0.0487 0.0049 0.1792 image4 0.0368 0.0993 0.1013 image5 0.0924 0.0921 0.1669
Table 2. Comparison of quantitative indicators
View tableTable 2. Comparison of quantitative indicators
Image PSNR of model in Ref. [5] PSNR of proposed model MSE of model in Ref. [5] MSE of proposed model Image6 2.7082 3.6905 0.0004 0.0002 Image7 3.7785 4.8033 0.0002 0.0002 Image8 2.7594 4.8073 0.0003 0.0002 Image9 9.3287 1.0116 0.0007 0.0005
Table 3. Comparison of indicators of each algorithm in case of natural hazy image
View tableTable 3. Comparison of indicators of each algorithm in case of natural hazy image
Indicator Ref. [1] Ref. [3] Ref. [5] Ref. [13] Ref. [14] Proposed e 0.1607 0.1635 0.1436 0.0915 0.0955 0.1661 r 1.3034 1.3686 1.2446 1.3138 1.3127 1.3763 SHCC 0.1207 0.1397 0.1226 0.1198 0.1179 0.1472 t /s 2.1026 2.0798 2.1248 2.4366 2.3927 1.9049
Table 4. Comparison of indicators of each algorithm in case of test set image
View tableTable 4. Comparison of indicators of each algorithm in case of test set image
Indicator Ref. [1] Ref. [3] Ref. [5] Ref. [13] Ref. [14] Proposed SSIM 0.8954 0.9678 0.8856 0.9589 0.9666 0.9713 PSNR 18.9677 20.6146 18.7633 20.4653 19.1398 20.6179
Tools
Get Citation
Copy Citation Text
Yan Yang, Jinlong Zhang. Image Dehazing Algorithm Based on Haze Distribution and Adaptive Attenuation[J]. Laser & Optoelectronics Progress, 2021, 58(22): 2210012
Paper Information
Category: Image Processing
Received: Feb. 1, 2021
Accepted: Feb. 12, 2021
Published Online: Nov. 10, 2021
The Author Email: Yan Yang (yangyantd@mail.lzjtu.cn)
© Copyright 2018-2021 | Chinese Laser Press.
All Rights Reserved 沪ICP备15018463号-20