Author Affiliations
1School of Electronic and Information Engineering, Liaoning Technical University, Huludao, Liaoning 125105, China2State Key Laboratory of Robotics, Shenyang Institute of Automation, Chinese Academy of Sciences, Shenyang, Liaoning 110016, China3Institutes for Robotics and Intelligent Manufacturing, Chinese Academy of Sciences, Shenyang, Liaoning 110016, Chinashow less
Fig. 1. Underwater optical imaging system
Fig. 2. Comparison of algorithm results. (a) Non-red channel compensation image; (b) red channel compensation image; (c) figure (a) histogram; (d) figure (b) histogram
Fig. 3. Background light point extraction regions. (a) Highlight region; (b) flat region; (c) intersection region
Fig. 4. Transmittance images. (a) Original image; (b) backscatter component transmittance; (c) R channel direct component transmittance; (d) G channel direct component transmittance; (e) B channel direct component transmittance
Fig. 5. Algorithm flow diagram
Fig. 6. Color recovery experiment results. (a) Original image; (b) UDCP algorithm; (c) UGAN algorithm; (d) DUIENet algorithm; (e) fusion algorithm; (f) algorithm in Ref. [7]; (g) proposed algorithm
Fig. 7. Comparison of different color deviation experiments results. (a) Original image; (b) UDCP algorithm; (c) UGAN algorithm; (d) DUIENet algorithm; (e) fusion algorithm; (f) algorithm in Ref. [7]; (g) proposed algorithm
Fig. 8. Comparison of different turbidity experiment results. (a) Original image; (b) UDCP algorithm; (c) UGAN algorithm; (d) DUIENet algorithm; (e) fusion algorithm; (f) algorithm in Ref. [7]; (g) proposed algorithm
Fig. 9. Feature matching test results. (a) 1st group; (b) 2nd group; (c) 3rd group; (d) 4th group
No. | Original image | UDCP | UGAN | DUIENet | Fusion | Ref. [7] | Proposed |
---|
1 | 0.6656 | 0.6159 | 0.6826 | 0.6533 | 0.6530 | 0.6696 | 0.7102 | 2 | 0.5113 | 0.6301 | 0.6440 | 0.5216 | 0.5906 | 0.5375 | 0.6589 | 3 | 0.4893 | 0.5906 | 0.5913 | 0.4878 | 0.5713 | 0.5899 | 0.6473 | 4 | 0.4198 | 0.5585 | 0.5034 | 0.5100 | 0.5826 | 0.5687 | 0.6332 | 5 | 0.4859 | 0.6028 | 0.5946 | 0.4804 | 0.5985 | 0.5318 | 0.6110 | 6 | 0.5194 | 0.5755 | 0.5885 | 0.5092 | 0.6193 | 0.5959 | 0.6748 | Average | 0.5152 | 0.5956 | 0.6007 | 0.5271 | 0.6026 | 0.5822 | 0.6559 |
|
Table 1. Evaluation values of UCIQE index
No. | Original image | UDCP | UGAN | DUIENet | Fusion | Ref. [7] | Proposed |
---|
1 | 3.4290 | 2.8489 | 5.4006 | 5.0863 | 3.2882 | 2.7955 | 2.6430 | 2 | 2.2981 | 2.3943 | 2.1456 | 2.1435 | 13.3725 | 2.1435 | 2.1417 | 3 | 3.9414 | 3.4532 | 4.9135 | 4.2319 | 3.3507 | 3.2522 | 3.2152 | 4 | 3.6825 | 3.2579 | 7.5442 | 18.8819 | 3.1639 | 2.9914 | 2.9401 | 5 | 3.4250 | 3.2047 | 5.7049 | 4.8825 | 3.1791 | 3.0694 | 2.9555 | 6 | 3.5470 | 3.4551 | 6.4132 | 4.8828 | 3.3517 | 3.0824 | 2.7889 | Average | 3.3872 | 3.1024 | 5.3537 | 6.6848 | 4.9510 | 2.8891 | 2.7807 |
|
Table 2. Evaluation values of NIQE index