Research on Automatic Dodging Method for Road Surface Line Array Images

Table 1

Quantitative evaluation parameters of the four images before and after dodging by different methods.

Image	Evaluation indicators	Original image	Mask dodging	Retinex enhancement	Homomorphic filtering	Wallis transform	My method
a	M^1	74.42	74.00	183.50	162.74	86.67	130.21
	SD^2	13.50	10.26	12.11	15.61	10.94	25.96
	H^3	5.79	5.37	5.58	5.98	5.48	6.71
	G^4	8.41	8.41	9.91	15.59	8.21	21.70
	LDV^5	5.74	0.23	0.36	2.63	0.95	0.30
b	M	70.91	70.00	216.81	156.93	113.13	154.33
	SD	21.96	13.46	12.08	7.90	19.59	25.12
	H	6.43	5.78	5.42	4.91	6.32	6.64
	G	10.66	10.65	8.65	6.17	14.19	20.48
	LDV	11.51	0.41	0.56	2.59	1.49	0.61
c	M	62.34	62.00	182.53	150.41	97.81	163.61
	SD	26.15	6.56	17.36	23.81	7.77	30.53
	H	6.36	4.58	4.69	4.97	4.96	6.85
	G	4.80	4.82	5.52	4.93	4.35	23.07
	LDV	22.28	0.60	6.14	15.28	1.10	10.58
d	M	67.49	67.00	191.26	152.25	102.67	177.90
	SD	25.49	11.19	17.49	13.49	12.59	39.99
	H	6.66	5.47	5.27	5.09	5.68	7.25
	G	10.39	10.37	9.90	6.82	10.44	35.34
	LDV	19.74	0.45	2.30	6.80	0.95	8.86

¹M: mean of image; ²SD: standard deviation; ³H: information entropy; ⁴G: mean gradient; ⁵LDV: luminance distribution value.