2022 Vol.5 March.No1

• Title: Ship re-identification algorithm based on improved metrics and attention fusion for congested waters
• Name: Lv Jinwen1,2,Chen Xianqiao1*
• Company: (1.Department of Computer Science, Wuhan University of Technology,Wuhan 430062,China; 2.Department of Computer Science,Hubei University of Technology,Wuhan 430068,China)
• Abstract:

  The ship re-identification algorithm for congested waters, which has the problem of low recognition accuracy in specific application scenarios, designs a ship re-identification algorithm for congested waters based on improved metrics and attention fusion. The influence factors of ship traffic flow are obtained, the ship heading is modified, a mathematical model of congested water density is constructed, the environmental influence coefficients of the ship domain are derived, and the improved metric and attention fusion techniques are used to design the re-ship identification algorithm. Experimental results: The mean recognition accuracy of the congested water ship re-identification algorithm in the paper is higher than the other two congested water ship re-identification algorithms: 6.336% and 13.127% respectively, proving that the performance of the congested water ship re-identification algorithm is more perfect after making full use of the improved metric and attention fusion techniques.

  
  

• Keyword: improved metrics; attention fusion; congested waters; ship navigation; re-identification; maritime traffic flow;
• DOI: 10.12250/jpciams2021090407
• Citation form: Lv Jinwen,Chen Xianqiao.Ship re-identification algorithm based on improved metrics and attention fusion for congested waters [J]. Computer Informatization and Mechanical System,2021,Vol.4,pp.42-47

References

[1] Li Z ,  Li Y . Ship Identification Characters Detection Method Based on Improved TextBoxes[J]. International Core Journal of Engineering, 2020, 6(5):328-332.

[2] Jung H I ,  Park S H . ISAR Image Quality Analysis and Scattering Point Identification for Maritime Ship Based on Ship Motion[J]. The Journal of Korean Institute of Electromagnetic Engineering and Science, 2020, 31(4):328-337.

[3] Karimi H R ,  Liang B ,  Basin M , et al. An online identification approach for a nonlinear ship motion model based on a receding horizon:[J]. Transactions of the Institute of Measurement and Control, 2021, 43(13):3000-3012.

[4] YAO Hongge, WANG Cheng, YU Jun, et al. Recognition of small-target ships in complex satellite images[J]. Journal of Remote Sensing, 2020, 24(2): 116-125.

[5] Yang J ,  Ma L ,  Liu J . Modeling and application of ship density based on ship scale conversion and grid[J]. Ocean Engineering, 2021, 237(10):109557.

[6]ZHANG Gang-qiang, LI Jun-feng. Research on recognition method of ship water gauge reading based on improved UNet network[J]. Journal of Optoelectronics·Laser, 2020, 31(11): 1182-1196.

[7] CHEN Hai-li, REN Hong-xiang, LI Yuan-hui, et al. Finite Time Control for Marine Surface Vehicle Based on Backstepping Sliding Mode Algorithm [J]. Computer Simulation, 2021, 38(12): 182-187,479.

[8] He X ,  Wang Y ,  Zhao S , et al. 295 Multimodal skin lesion classification in dermoscopy and clinical images using a hierarchical attention fusion network[J]. Journal of Investigative Dermatology, 2021, 141(5):S52.

[9] Shen J ,  Zhang T ,  Wang Y , et al. A Dual-Model Architecture with Grouping-Attention-Fusion for Remote Sensing Scene Classification[J]. Remote Sensing, 2021, 13(3):433.