Adaptive Retinex Image Defogging Algorithm Based on Depth-of-Field Information

Xiangping Wu; Qingqing Gao; Shaowei Huang; Ke Wang

doi:10.3788/LOP221632

Laser & Optoelectronics Progress, Volume. 60, Issue 12, 1210013(2023)

Adaptive Retinex Image Defogging Algorithm Based on Depth-of-Field Information

Xiangping Wu^1,2、*, Qingqing Gao¹, Shaowei Huang¹, and Ke Wang¹

Author Affiliations

¹College of Information Engineering, China Jiliang University, Hangzhou 310018, Zhejiang, China

²Key Laboratory of Electromagnetic Wave Information Technology and Metrology of Zhejiang Province Hangzhou, China Jiliang University, 310018, Zhejiang, China

show less

Abstract Get PDF(in Chinese)

Figures & Tables(15)

Fig. 1. Flow chart of dehazing algorithm

Download full size

Fig. 2. Network structure diagram of LPG

Download full size

Fig. 3. Comparison of depth-of-field estimation. (a) Original drawings; (b) depth-of-field estimation map of color attenuation priori; (c) depth-of-field estimation map of BTS depth learning model

Download full size

Fig. 4. Linear fitting diagram

Download full size

Fig. 5. Color unrecovered pseudo defog map

Download full size

Fig. 6. Bilinear interpolation

Download full size

Fig. 7. Effect drawing of bilinear interpolation. (a) Before bilinear interpolation; (b) after bilinear interpolation

Download full size

Fig. 8. Comparison of dehazing effects of each algorithm. (a) Original drawings; (b) algorithm of reference [5]; (c) algorithm of reference [2]; (d) algorithm of reference [6]; (e) algorithm of reference [10]; (f) algorithm of reference [8]; (g) proposed algorithm

Download full size

Fig. 9. Partial enlarged contrast diagrams. (a) Algorithm of reference [5]; (b) proposed algorithm

Download full size

Fig. 10. Partial enlarged contrast diagrams. (a) Algorithm of reference [6]; (b) proposed algorithm

Download full size

Table 1. Comparison of standard deviation and mean brightness of each algorithm's renderings

View table

Table 1. Comparison of standard deviation and mean brightness of each algorithm's renderings

Image	Standard deviation						Brightness mean
Image	Original	MSRCR	Algorithm of reference［6］	Algorithm of reference［10］	Algorithm of reference［8］	Proposed algorithm	Original	MSRCR	Algorithm of reference［6］	Algorithm of reference［10］	Algorithm of reference［8］	Proposed algorithm
Village	20.934	23.299	26.797	24.394	24.258	27.901	112.079	139.913	31.638	50.048	67.810	106.871
City	51.793	74.811	60.859	58.074	55.774	70.897	125.364	128.969	99.567	82.007	115.867	136.971
Building	57.373	72.715	51.101	54.444	60.606	77.745	108.797	156.699	70.252	59.746	87.254	120.942
Town	53.519	53.562	44.580	57.859	61.846	54.350	125.194	132.757	73.041	72.792	94.869	134.442
Peak	55.249	56.647	59.680	60.224	61.790	60.249	129.364	128.857	85.156	90.586	95.957	96.980

Table 2. Comparison of standard deviation and mean brightness of each algorithm's renderings

View table

Table 2. Comparison of standard deviation and mean brightness of each algorithm's renderings

Image	Information entropy						Average gradient
Image	Original	MSRCR	Algorithm of reference［6］	Algorithm of reference［10］	Algorithm of reference［8］	Proposed algorithm	Original	MSRCR	Algorithm of reference［6］	Algorithm of reference［10］	Algorithm of reference［8］	Proposed algorithm
Village	4.247	5.196	4.944	4.202	4.210	5.281	214.274	447.548	517.583	549.089	337.272	491.281
City	5.138	5.248	5.371	5.211	5.213	5.450	2413.29	4124.39	4252.078	4928.586	3413.171	4533.830
Building	4.947	5.021	5.128	4.692	4.819	5.184	454.610	678.696	681.916	755.753	690.425	1738.238
Town	4.822	4.852	4.947	4.914	4.868	5.173	471.522	871.047	709.575	936.947	836.029	1834.301
Peak	4.464	4.697	4.956	5.126	4.932	5.187	527.537	571.859	809.645	909.864	810.238	1283.670

Table 3. Comparison of structural similarity of renderings of various algorithms

View table

Table 3. Comparison of structural similarity of renderings of various algorithms

Image	Structural similarity
Image	MSRCR	Algorithm of reference［6］	Algorithm of reference［10］	Algorithm of reference［8］	Proposed algorithm
106_Hazy	0.779	0.877	0.853	0.873	0.881
287_Hazy	0.646	0.879	0.734	0.777	0.892

Table 4. Influence of different modules on algorithm performance
View table
Table 4. Influence of different modules on algorithm performance
SSR √ √ √
PSNR 19.24 23.75 29.26 35.40
Color correction √ √
Linear model √

Table 5. Comparison of demisting time of each algorithm
View table
Table 5. Comparison of demisting time of each algorithm
Algorithm MSRCR Algorithm of reference［6］ Algorithm of reference［10］ Algorithm of reference［8］ Proposed algorithm
Time /s 8.25 1.53 0.58 2.83 3.26

Tools

Get Citation

Copy Citation Text

Xiangping Wu, Qingqing Gao, Shaowei Huang, Ke Wang. Adaptive Retinex Image Defogging Algorithm Based on Depth-of-Field Information[J]. Laser & Optoelectronics Progress, 2023, 60(12): 1210013

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites

Paper Information

Category: Image Processing

Received: May. 17, 2022

Accepted: Jun. 16, 2022

Published Online: Jun. 5, 2023

The Author Email: Xiangping Wu (xiangpingwu@cjlu.edu.cn)

DOI:10.3788/LOP221632

Topics

laser devices and laser physics

Lasers and Laser Optics

Laser physics

laser manufacturing

Instrumentation, Measurement and Metrology

Table 1. Comparison of standard deviation and mean brightness of each algorithm's renderings

Table 1. Comparison of standard deviation and mean brightness of each algorithm's renderings

Table 2. Comparison of standard deviation and mean brightness of each algorithm's renderings

Table 2. Comparison of standard deviation and mean brightness of each algorithm's renderings

Table 3. Comparison of structural similarity of renderings of various algorithms

Table 3. Comparison of structural similarity of renderings of various algorithms

Table 4. Influence of different modules on algorithm performance

Table 4. Influence of different modules on algorithm performance

Table 5. Comparison of demisting time of each algorithm

Table 5. Comparison of demisting time of each algorithm