Depth-Sensor Picking for Adaptable Box-Like Item Handling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing automated systems for manipulating box-like items on a horizontal support surface face challenges such as positioning difficulties, inflexibility, high implementation costs, inefficiency, and the need for extensive training when new items are introduced, particularly in systems using classical image processing and machine learning.
Innovation Solution
A method and system utilizing a 3-D or depth sensor to generate and rank hypotheses based on surprisals, allowing an autonomous manipulator to handle multiple items efficiently by combining geometric and appearance observations, without requiring extensive training or reconfiguration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If classical image processing and machine learning systems are used to manipulate box-like items, then automation is achieved, but the system requires extensive training and reconfiguration when new items are introduced
Solution Approach 1:
The system performs self-calibration by automatically determining scale factors and geometric properties of new box-like items without requiring external training data or manual reconfiguration. The depth sensor captures images, and the processor automatically calculates scale factors based on known physical dimensions, enabling the system to adapt to new items independently
Solution Approach 2:
The system dynamically adjusts scale factors and geometric parameters based on the specific dimensions of each box-like item. By changing these parameters automatically through depth-based measurement rather than using fixed training data, the system achieves both automation and adaptability to varying item configurations
2Productivity
If automated systems are implemented to manipulate items on a horizontal support surface, then productivity increases, but positioning difficulties and inflexibility arise
Solution Approach 1:
The system replaces traditional mechanical positioning systems with a vision-based depth sensing approach. The depth sensor captures three-dimensional information about item positions, and the processor calculates precise locations and orientations, eliminating mechanical complexity while improving positioning accuracy and flexibility
Solution Approach 2:
The system transitions from two-dimensional image processing to three-dimensional depth-based measurement. By utilizing the depth dimension, the system accurately determines the position, orientation, and geometry of box-like items on the horizontal support surface, resolving positioning difficulties inherent in 2D systems
3Productivity
If traditional automated picking systems are used, then items can be moved from one conveyance to another, but high implementation costs and system complexity occur
Solution Approach 1:
The system uses a single depth sensor and processing unit to perform multiple functions: capturing images, determining geometric properties, calculating scale factors, identifying item positions, and controlling manipulation. This multi-functional approach reduces system complexity while maintaining productivity compared to specialized subsystems for each function
Solution Approach 2:
The depth sensor acts as an intermediary that provides three-dimensional measurement data, eliminating the need for complex mechanical measurement systems. This intermediary device simplifies the overall system architecture while enabling accurate item manipulation and transfer operations
Data Source
AI summary
A method and system for manipulating (i.e. such as by picking) a multitude of target items supported on a substantially horizontal support surface one at a time. The support surface supports a configuration of substantially identical items aligned substantially perpendicular to the support surface. The method includes the steps of providing a plurality of hypotheses which are ranked based on surprisals of the hypotheses. Each of the hypotheses describes an observation of an item in the configuration. The observations include an observation of the appearance of a perimeter of the item and an observation of the geometry of the perimeter of the item. The method also includes generating potential configurations for potential combinations of the multiple items based on the ranked hypotheses. Finally, the potential configurations are ranked. The step of ranking includes the step of combining the surprisals of the hypotheses in each potential configuration.