Polarization image smoke removal based on precheck mechanism

YAN Qing; YE Mengmeng; ZHANG Jingjing; LIU Xiao; NIAN Fudong; LI Teng

doi:10.3969/j.issn.1673-6141.2023.02.003

Journal of Atmospheric and Environmental Optics, Volume. 18, Issue 2, 108(2023)

Polarization image smoke removal based on precheck mechanism

YAN Qing¹, YE Mengmeng^1,2、*, ZHANG Jingjing^1,2,3, LIU Xiao³, NIAN Fudong^1,4, and LI Teng^1,2

¹Key Laboratory of Computational Intelligence and Signal Processing (Anhui University), Ministry of Education, Hefei, Anhui 230601, China

²Anhui Key Laboratory of Polarized Light Imaging Detection Technology, Hefei 230031, China

³Key Laboratory of Optical Calibration and Characterization, Chinese Academy of Sciences, Hefei 230031, China

⁴School of Advanced Manufacturing Engineering, Hefei University, Hefei 230031, China

show less

Abstract Get PDF(in Chinese)

Figures & Tables(11)

Fig. 1. Flow chart of local smoke removal

Download full size

View in Article

Fig. 2. Multi-scale attentional adversarial network

Download full size

View in Article

Fig. 3. Polarization image. (a) Stokes I image; (b) Stokes Q image; (c) Stokes U image; (d) image of polarization degree P

Download full size

View in Article

Fig. 4. PSNR iterative graph of multi-scale attentional adversarial network with single-channel I image input and four-channel I, Q, U, P image input

Download full size

View in Article

Fig. 5. Comparison of smoke removal effect. (a) Input image; (b) true image; (c) input I image de-smoking effect;(d) input I, Q, U, P image de-smoking effect

Download full size

View in Article

Fig. 6. Smoke detection results of Yolov3

Download full size

View in Article

Fig. 7. Smoke removal effect of different smoke removal networks based on precheck mechanism. (a) True image; (b) smoke removal effect of GCANet; (c) smoke removal effect of FFANet; (d) smoke removal effect of multi-scale attention adversarial network

Download full size

View in Article

Table 1. Location index of crater smoke area
View table
View in Article
Table 1. Location index of crater smoke area
Method mAP Precision/% Time/s
Yolov3 85.09 93.5 0.112

Table 2. Quantitative comparison of global PSNR and SSIM
View table
View in Article
Table 2. Quantitative comparison of global PSNR and SSIM
Method PSNR SSIM Time/s
GCANet 26.69 0.792 0.741
FFANet 26.91 0.804 2.540
Ours 28.51 0.866 1.694
Yolov3 + GCANet 27.13 0.832 0.352
Yolov3 + FFANet 27.42 0.875 0.972
Yolov3 + Ours 29.22 0.915 0.719

Table 3. Quantitative comparison of local PSNR and SSIM
View table
View in Article
Table 3. Quantitative comparison of local PSNR and SSIM
Method PSNR SSIM
Yolov3 + GCANet 24.96 0.733
Yolov3 + FFANet 26.42 0.819
Yolov3 + Ours 27.78 0.848

Table 4. Ablative analysis
View table
View in Article
Table 4. Ablative analysis
Method PSNR-I PSNR-IQUP
BaseNet 24.48 24.83
BaseNet + Multi-scale attention 26.29 27.32
BaseNet + Multi-scale attention + Gan 27.27 28.51

Tools

Get Citation

Copy Citation Text

Qing YAN, Mengmeng YE, Jingjing ZHANG, Xiao LIU, Fudong NIAN, Teng LI. Polarization image smoke removal based on precheck mechanism[J]. Journal of Atmospheric and Environmental Optics, 2023, 18(2): 108

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites