Vessel Draft Depth Detection Using Multi-Task Learning Network
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting vessel draft depth rely heavily on manual labor, resulting in low reading efficiency, especially in visual image-based systems which require accurate water gauge scale recognition and are affected by vessel fouling.
Innovation Solution
A method and device utilizing a multi-task learning network model with a multi-scale convolutional neural network, target detection sub-network, and water surface and hull segmentation sub-network for automatic detection of vessel draft depth, extracting local area image blocks and determining draft depth through image processing and feature extraction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If manual reading methods are used for vessel draft depth detection, then the system is simple to implement, but the reading efficiency is low
Solution Approach 1:
The system enables automatic detection of vessel draft depth through self-service mechanisms. The multi-task learning network model automatically processes hull images to extract waterline positions and draft depth values without requiring manual intervention. The system performs autonomous feature extraction, scale recognition, and measurement calculation, transforming a manual reading process into an automated self-service detection system that significantly improves reading efficiency
Solution Approach 2:
The patent replaces manual mechanical reading operations with an automated image processing system. Instead of human operators manually reading draft markings from vessel hulls, the system uses computer vision technology with multi-scale convolutional neural networks to automatically detect and measure draft depth from hull images, substituting mechanical human labor with automated computational processing
2Extent of automation
If existing automatic reading methods based on visual images are used, then manual labor is reduced, but the methods require high accuracy in water gauge scale recognition and are affected by vessel fouling
Solution Approach 1:
The patent applies segmentation by dividing the complex draft depth detection task into multiple specialized sub-tasks handled by different network modules. The multi-task learning network model separates waterline detection, draft scale recognition, and measurement calculation into distinct functional components. This segmentation allows each module to specialize in specific aspects, improving overall detection accuracy and robustness against vessel fouling by focusing computational resources on relevant features
Solution Approach 2:
The system performs preliminary action by pre-processing hull images to enhance relevant features before main detection. The multi-scale convolutional neural network performs preliminary feature extraction and waterline localization before draft depth measurement. This preliminary processing prepares the data by removing irrelevant information and enhancing critical features, making the subsequent detection more accurate and less susceptible to vessel fouling interference
3Extent of automation
If acoustic signal based detection methods are used, then automatic detection is achieved, but the equipment deployment cost is high
Solution Approach 1:
The patent uses copying by creating a digital replica of the vessel hull through image capture instead of using physical acoustic sensing equipment. The system captures visual images of the hull and creates a digital representation that can be processed automatically. This copying approach replaces expensive acoustic hardware with affordable imaging devices and computational processing, achieving automatic detection at lower deployment cost
Solution Approach 2:
The patent substitutes acoustic mechanical systems with optical imaging systems. Instead of deploying acoustic signal transmitters and receivers that require complex hardware infrastructure, the system uses visual image capture and digital processing. This substitution replaces expensive acoustic equipment with more affordable imaging technology, reducing deployment costs while maintaining automatic detection capability
Data Source
AI summary
Disclosed is a method and device for automatic detection of vessel draft depth, which processes the image of a vessel's hull and extracts local area image blocks with vessel's water gauge scale separately to improve the pertinence of data processing and reduce the complexity of data processing; and based on a multi-task learning network model, performing data processing on local area image blocks to extract scale characters and waterline position, reducing the computational complexity of the model; finally, based on the relative positions of the scale and waterline, determining the vessel's draft depth, thus achieving automatic acquisition of the vessel's draft depth, this method greatly improves the efficiency of reading the vessel's draft depth.