In scenarios such as inland rivers, automatic navigation for marine vessels is becoming more and more important. To improve the effect of automatic navigation, a visual navigation method based on machine learning for vessels is designed, and simulation validation is conducted. Firstly, vessel images are captured by using visual sensing technology. A semantic segmentation model is employed to separate the shoreline from the water region and identify the shoreline. Secondly, based on the identified shoreline, a reference anchor point is set for the vessel’s navigation. The vessel’s automatic navigation is achieved through a control algorithm. Finally, the proposed visual navigation method is validated through a simulation platform. Simulation experiments have shown that in simple inland waterway scenarios, the proposed method enables autonomous navigation of the vessel.
Visual Navigation Method and Simulation Validation
2023-08-04
616127 byte
Conference paper
Electronic Resource
English
Navigation environment visualization representation method for ship visual navigation
European Patent Office | 2015
|Human visual sampling processes : a simulation validation study
TIBKAT | 1969
|