Fig. 1. Different categories of image dehazing methods

Fig. 2. Comparison of other algorithms with bilateral weighted histogram^[20]

Fig. 3. Comparison of the other algorithms with the CLAHE + Guided filtering^[24]

Fig. 4. Graphical representation of the GC and GCP processes^[26]

Fig. 5. Comparison of multi-scale Retinex dehazing^[29]

Fig. 6. Comparison of the effect of other dehazing algorithms and MWTO dehazing algorithm^[32]

Fig. 7. Comparison of the effect of other dehazing algorithms and saturation line prior dehazing algorithm^[34]

Fig. 8. Comparison of noise removal methods for night images^[38]

Fig. 9. Dehaze imaging model^[10]

Fig. 10. Flowchart of DCP dehazing algorithm

Fig. 11. Comparison of the effect of other dehazing algorithms and LU Z dark channel dehazing algorithm^[45]

Fig. 12. Comparison of various other algorithms with NANDAL S anisotropic diffusion algorithm^[48]

Fig. 13. Comparison of various other algorithms with NLTV algorithm^[50]

Fig. 14. Schechner 's restoration effect^[12]

Fig. 15. Results of in different dehazed scenes^[56]

Fig. 16. The dehazing result of LIU^[57]. (a) Raw image; (b) Dehazed result using the proposed method; (c) Magnified region on the green rectangle A in Fig.(a); (d) Magnified region on the green rectangle B in Fig.(a); (e) Magnified region on the green rectangle A in Fig.(b); (f) Magnified region on the green rectangle B in Fig.(b); (g) Histograms of Fig.(a); (h) Histograms of Fig.(b)

Fig. 17. LIANG's dehazing results^[61]. (a) Dehazed color image; (b) Dehazed infrared image; (c) The fusion image of visible light and near-infrared haze images; (d) The fusion image of visible light and near-infrared dehazing images is the final dehazing result

Fig. 18. TIAN’s dehazing results^[64]. (a) Raw intensity; (b) Imaging results by M-PDI; (c) Intensity profiles by raw intensity, traditional PI, and M-PDI; (d) Contrast distributions of the object obtained by raw intensity, traditional PI, and M-PDI versus optical thickness

Fig. 19. Images in the corresponding target^[65]. (a) Clear water and (b) Turbid condition; (c) Stokes vector elements for the direct measurement (D), separated scatter (B) and target (T); (d) Generation of numerical plots for the direct measurement and separated target

Fig. 20. Real-time optical sensing and detection system^[66]

Fig. 21. Comparison of experimental results^[66]. (a)-(d) Polarization images corresponding to polarization angles of 0°, 45°, 90°, and 135°; (e) Polarization image with minimum airlight; (f) Polarization image with maximum airlight; (g) Total intensity image; (h) Detection results of Schechner method; (i) Detection results of WANG and LIU’s method

Fig. 22. The restoration result of LIANG^[67]. (a) Original hazy image; (b) The dehazing image of Fig.(a); (c) Original hazy image with specular surface scene; (d) The dehazing image of Fig.(c)

Fig. 23. Comparison of various other algorithms with the LIANG Z polarization algorithm^[69]

Fig. 24. Comparison of various other algorithms with AOD-Net algorithm^[71]

Fig. 25. Comparison of various other algorithms with variational and deep CNN-based dehazing methods^[77]

Fig. 26. General framework of unsupervised image dehazing^[81]

Fig. 27. Comparison of various other algorithms with unsupervised dark channel dehazing methods^[85]

Fig. 28. General framework of semi-supervised image dehazing^[89]

Fig. 29. Comparison of various other algorithms with semi-supervised dehazing methods^[89]