Crane-Mounted Vision Positioning for Moving Container Bays
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The challenge of automating crane operations over container vessels is hindered by the need for manual interaction, varying container types, and the movement of vessels during berthing, which requires real-time updates and reliable cargo information for precise positioning.
Innovation Solution
A method using machine learning algorithms to process image data from cameras mounted on cranes, combined with lidar data, to determine the position of containers on a container bay, enabling autonomous operation and efficient crane management.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If manual interaction is used for crane operations, then operators can handle various container types, but automation cannot be achieved and productivity is limited
Solution Approach 1:
The crane system performs self-positioning and self-adjustment by automatically determining container positions using camera images and machine learning algorithms. The system serves itself by autonomously updating its understanding of the container bay layout without manual intervention, enabling adaptive autonomous operation across different container configurations
Solution Approach 2:
The system adapts to different container types by dynamically changing operational parameters based on image analysis. The machine learning algorithm identifies container characteristics from images and adjusts positioning parameters accordingly, allowing the crane to handle various container types while maintaining autonomous operation
2Measurement precision
If real-time updates are implemented to track vessel movement, then positioning accuracy is maintained, but data processing complexity increases
Solution Approach 1:
The system continuously captures images of the container bay and feeds this data back to the machine learning algorithm for real-time position determination. This feedback loop automatically updates container positions as the vessel moves, maintaining measurement precision without requiring complex manual data processing
Solution Approach 2:
The patent replaces complex mechanical measurement systems with optical sensing (cameras) and computational methods (machine learning). Instead of using sophisticated physical instruments to track position, the system uses image processing and algorithmic analysis, simplifying the overall system while maintaining or improving precision
3Productivity
If detailed cargo information is collected for autonomous operation, then crane efficiency increases, but information reliability becomes critical
Solution Approach 1:
The system creates a digital copy or representation of the container bay by capturing images and processing them through machine learning algorithms. This digital model serves as a reliable information source for autonomous operation, replacing the need for potentially unreliable manual cargo manifests or paper-based information systems
Solution Approach 2:
The patent replaces manual information collection methods with automated optical sensing and computational processing. Cameras and machine learning algorithms automatically extract cargo information from images, eliminating human error and improving both the speed and reliability of information acquisition for crane operations
Data Source
Figure 1~2
Figure 3~4
Figure 5~6
AI summary
A method for determining a position of a container (24) on a container bay (22) of a container vessel (20) is described. The method comprises: receiving image data from at least one camera (40, 44) mounted on a structure of a crane (30), wherein the structure of the crane (30) at least partly extends over the container bay (22) and wherein the image data are representative of a first image and at least a second image each showing at least an area of the container bay (22) in which the container (24) is arranged; and determining the position of the container (24) depending on the image data by a machine learning algorithm.