Ship Information Measurement with Machine-Learned Feature Points
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for measuring ship information, such as vessel position, dimension, and azimuth, face challenges in accuracy and ease of use, particularly when relying on automatic identification systems (AIS) due to potential failures or human errors, and other measurement techniques like three-dimensional laser scanning and deep learning of segmentation are costly or require manual operations.
Innovation Solution
A ship information measurement device installed on an unmanned aircraft uses a camera, distance meter, and inclinometer, combined with a machine learning model to extract feature points from vessel images, calculating distances, inclination angles, and identifying ship information accurately and automatically.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If AIS (Automatic Identification System) is used to obtain ship information, then the information acquisition is automatic and convenient, but the accuracy is compromised due to potential failures and human errors
Solution Approach 1:
The patent introduces an intermediary measurement system consisting of an image acquisition device, distance measurement device, and inclination measurement device that acts as a mediator between the ship and the information gathering process. This intermediary system captures images and measurements directly from the ship's physical attributes, bypassing the need to rely on AIS data which may contain errors or failures.
Solution Approach 2:
The patent replaces the electronic/AIS-based information acquisition system with a physical measurement system using cameras, distance meters, and inclination meters. This substitution transitions from relying on electronic data transmission (prone to errors) to direct physical measurement of the ship's attributes through imaging and sensor data.
2Measurement precision
If three-dimensional laser scanning is used to measure ship information, then the measurement accuracy is improved, but the cost and device complexity increase significantly
Solution Approach 1:
The patent combines multiple simple measurement devices (camera, distance meter, inclination meter) into an integrated measurement system. By merging these individual components that are relatively simple and inexpensive, the system achieves accurate ship information measurement without requiring complex three-dimensional laser scanning equipment.
Solution Approach 2:
The patent uses two-dimensional image capture as a simplified copy or representation of the three-dimensional ship structure. Instead of directly scanning the full 3D geometry with complex laser equipment, the system captures 2D images and derives ship information through image processing and geometric calculations, reducing device complexity while maintaining measurement accuracy.
3Extent of automation
If deep learning segmentation is used to extract ship features, then the automation level is improved, but the computational complexity and processing time increase
Solution Approach 1:
The patent extracts only the essential feature points needed for ship information measurement from the captured images, rather than performing comprehensive deep learning segmentation of the entire ship structure. By taking out only the critical features (such as key structural points for dimension calculation), the system achieves automation without the heavy computational burden of full segmentation.
Solution Approach 2:
The patent applies partial action by focusing on extracting specific feature points rather than performing complete segmentation of the ship. This partial extraction approach provides sufficient information for measuring ship position, dimension, and azimuth without the excessive computational requirements of full deep learning segmentation.
Data Source
AI summary
A non-transitory computer-readable recording medium stores a program for causing a computer to execute a process for measuring ship information which includes: imaging a predetermined vessel with an image acquisition device to obtain a captured image of the predetermined vessel; extracting a feature point of the predetermined vessel in the captured image based on an output obtained by inputting the obtained captured image to a machine learning model that extracts a feature point of a vessel from an image; obtaining a distance from the image acquisition device to a horizontal plane on which the predetermined vessel is positioned; obtaining an inclination angle of the image acquisition device with respect to the horizontal plane; and identifying the ship information of the predetermined vessel based on a position of the feature point in the captured image, the distance, and the inclination angle.