Robotic Multi-Pick Detection Using Stereoscopic Imaging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing warehouse fulfillment systems face challenges with human-based errors, fatigue, and inefficiencies, and robotic systems struggle with grasping multiple objects or identifying items accurately, especially in varied orientations and backgrounds, leading to incorrect order picking and inventory issues.
Innovation Solution
A robotic multi-pick detection system using software, firmware, and hardware configurations that capture image data from multiple angles, filter and estimate item dimensions, and identify items using stereoscopic cameras, enabling accurate identification and sorting of items, even in complex environments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If robotic systems are used to pick items, then human fatigue and error are reduced, but the difficulty of grasping objects from groups of multiple objects increases
Solution Approach 1:
The patent replaces mechanical detection methods with optical sensing systems. Multiple cameras capture images of items at the scanning area, and image processing algorithms automatically identify and count items based on their visual characteristics and spatial positions, eliminating the need for complex mechanical detection mechanisms
Solution Approach 2:
The patent introduces an intermediary detection system between the robotic arm and the items. This intermediary consists of multiple cameras positioned at different angles that capture item information, and a processing system that analyzes this information to provide accurate identification and quantity detection, serving as a mediator that enables reliable robotic picking
2Productivity
If robots grasp multiple objects, then productivity increases, but incorrect objects may be grasped resulting in incorrect order picking
Solution Approach 1:
The patent segments the detection task into multiple independent components: multiple cameras capture images from different angles, image processing algorithms separately analyze each camera's data, and the system integrates results to identify individual items within a group. This segmentation allows fast processing of multiple items while maintaining high identification accuracy through systematic analysis
Solution Approach 2:
The patent implements feedback mechanisms where the detection system continuously monitors and verifies item identification. The system compares detected item characteristics against expected values and can trigger re-detection or alert operators for verification, ensuring high accuracy while maintaining fast processing speeds
3Loss of time
If existing item identification systems are used, then processing time is reduced, but accuracy is insufficient to distinguish between similar items or multiple items
Solution Approach 1:
The patent transitions from two-dimensional image processing to three-dimensional spatial analysis. Multiple cameras positioned at different angles provide depth information and spatial relationships between items, enabling the system to distinguish between similar items based on their three-dimensional positions and orientations, thereby increasing accuracy without significantly increasing processing time
4Measurement precision
If systems requiring specific item orientation are used, then identification accuracy is improved, but adaptability to varied orientations is reduced
Solution Approach 1:
The patent creates a universal detection system where multiple cameras are positioned to capture items from various angles and orientations. The image processing algorithms are designed to be orientation-invariant, automatically adjusting to accommodate items in any orientation. This multi-functional approach allows the system to maintain high identification accuracy across diverse item orientations without requiring separate processing paths for different orientations
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The system significantly reduces processing time, increases adaptability, and improves accuracy, allowing for fast and reliable item identification and sorting in industrial applications, reducing human intervention and error rates.
Implementation Method 1
capturing, by one or more processors and using one or more cameras, image data of one or more items at a scanning area
Data Source
AI summary
A system may capture image data of one or more items at a scanning area using one or more cameras. The system may determine one or more thresholds for the image data using one or more first dimensions of a set of items and may filter the image data using the one or more thresholds. The system may estimate one or more second dimensions of the one or more items based on the filtered image data, and the system may identify a first item from among the one or more items by comparing the estimated one or more second dimensions of the one or more items with the one or more first dimensions of the set of items.