Marine Vessel Vision Docking Detection for Open Water Availability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Identifying available docking positions for marine vessels is challenging due to the difficulty in distinguishing docks from other long, straight structures and ensuring adjacent water is unoccupied, leading to potential selection of unsuitable docking locations.
Innovation Solution
A system using a combination of image segmentation and object detection models to identify docks and adjacent water availability, enabling automatic maneuvering to available docking positions without user intervention.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional docking identification methods are used, then the system is simple to operate, but the accuracy of identifying available docking positions is low
Solution Approach 1:
The patent divides the docking identification task into multiple specialized models: an object detection model to identify docks and other structures, and an image segmentation model to analyze water occupancy. This segmentation allows each model to specialize in specific aspects of the problem, improving overall accuracy while managing complexity through modular architecture.
Solution Approach 2:
The patent introduces trained machine learning models as intermediaries between the raw image data and the docking decision. These models act as mediators that process and interpret visual information, enabling accurate identification of docking positions and water occupancy without requiring complex manual analysis systems.
2Productivity
If automated docking identification is implemented, then the productivity of docking operations is improved, but the reliability of docking position selection may deteriorate due to false identification
Solution Approach 1:
The system employs multiple verification mechanisms including object detection results, image segmentation analysis of water occupancy, and overlap detection between dock bounding boxes and water regions. This multi-layered feedback approach ensures reliable identification of available docking positions by cross-validating information from different analysis streams before confirming a docking location.
3Measurement precision
If manual verification of docking positions is required, then the accuracy of docking position selection is improved, but the time required for docking operations increases
Solution Approach 1:
The system performs self-verification through automated image segmentation and overlap detection algorithms that independently analyze water occupancy and dock boundaries. This self-service capability eliminates the need for manual verification while maintaining high accuracy, as the system automatically cross-checks detection results against segmentation data to confirm valid docking positions.
Data Source
Figure 1
Figure 2
Figure 3A1
AI summary
A system for automatically identifying available docking positions using a vision system of a marine vessel (10) is provided, the system comprising: an imaging device (72, 73, 74, 75, 76, 78), wherein the imaging device (72, 73, 74, 75, 76, 78) is configured to be mounted to the marine vessel (10) with an associated field of view (92, 93, 94, 95, 96, 98) of an environment (300, 320) of the marine vessel (10); and one or more hardware processors configured to: identify, based on image data captured using the imaging device (72, 73, 74, 75, 76, 78), a dock instance; determine, based on the image data, whether the dock instance is available for docking; indicate that the dock instance is an available docking position; and cause the marine vessel (10) to maneuver to the available docking position.