Single Remote Sensing Multispectral Image Dehazing Based on a Learning Framework

Table 1

The corresponding values of MSE, SSIM, PSNR,, , and FADE.

Image	SSIM	e		FADE	PSNR	MSE
Original images	0.717	2.517	2.265	0.368	20.15	3.854
	0.695	0.668	2.168	0.706	16.26	4.126
	0.653	1.654	2.463	0.856	18.34	5.021
	0.732	1.786	1.952	0.757	24.32	3.908
MSRCR method	0.754	2.806	1.971	0.269	22.31	3.785
	0.732	0.927	1.932	0.472	19.32	4.103
	0.812	1.781	1.685	0.863	20.08	3.953
	0.795	1.953	1.613	0.617	26.62	3.784
HOT method	0.763	2.763	1.646	0.341	18.31	2.357
	0.742	0.715	2.235	0.620	15.32	2.132
	0.769	1.793	2.361	0.831	16.08	3.416
	0.746	1.851	1.859	0.624	17.62	2.317
DCP method	0.712	2.782	1.836	0.316	21.53	2.365
	0.746	0.757	1.757	0.608	19.56	2.143
	0.812	1.463	2.307	0.741	18.49	3.446
	0.795	1.673	2.135	0.738	25.43	2.512
Qin’s method	0.723	2.402	2.158	0.377	23.13	3.613
	0.734	0.745	2.034	0.712	18.31	3.815
	0.681	1.593	2.207	0.796	19.43	4.236
	0.762	1.810	1.604	0.763	24.69	3.752
Liu’s method	0.743	2.532	2.242	0.330	21.26	2.854
	0.754	0.716	2.093	0.691	17.57	3.867
	0.703	1.803	2.283	0.637	18.96	4.351
	0.764	1.931	1.759	0.743	23.71	3.648
Ours	0.787	2.794	1.825	0.313	25.13	2.244
	0.795	0.783	1.792	0.496	21.41	2.032
	0.823	1.869	1.785	0.756	24.36	3.132
	0.812	2.103	1.654	0.761	27.28	2.137