Camera Vision Parts Matching With OCR and Depth Sensing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current parts matching processes in manufacturing operations, such as pairing power supply units (PSUs) with customer premises equipment (CPE), rely heavily on manual methods like barcode reading and suffer from long processing times and high incompatibility rates.
Innovation Solution
A camera vision system using optical character recognition (OCR) and a depth sensor to automatically read CPE barcodes and PSU model numbers, with a dichroic mirror to separate depth sensor laser light from camera vision, enabling accurate and fast matching of PSUs to CPEs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If manual barcode reading and parts matching processes are used, then device complexity is reduced, but processing time increases significantly (1-5 minutes per order)
Solution Approach 1:
The patent replaces manual barcode reading and parts matching processes with an automated vision system that uses cameras, depth sensors, and pattern recognition algorithms. The system captures images of power supply units and customer premises equipment, extracts visual features and barcode information, automatically matches compatible parts, and validates compatibility rules, thereby eliminating manual intervention and significantly reducing processing time from 1-5 minutes to under 12 seconds per order
Solution Approach 2:
The system creates digital copies of physical parts through image capture and stores their visual features, barcode data, and compatibility information in a database. When matching parts, the system retrieves and compares digital copies rather than physically handling each component, enabling rapid automated matching and validation without manual inspection
2Reliability
If manual parts matching is used, then measurement precision requirements are reduced, but compatibility accuracy decreases (5-10% incompatibility occurrence)
Solution Approach 1:
The system replaces manual visual inspection and barcode scanning with automated computer vision technology. Cameras capture high-resolution images of power supply units and customer premises equipment, depth sensors provide precise distance measurement, and pattern recognition algorithms automatically extract features, read barcodes, and match compatible parts with high accuracy, eliminating human error in compatibility determination
Solution Approach 2:
The system incorporates multiple validation feedback loops that continuously verify compatibility by comparing extracted visual features, barcode data, and compatibility rules against stored reference information. The depth sensor provides feedback on part positioning and orientation to ensure accurate image capture, while the pattern recognition system validates matches against multiple compatibility criteria before finalizing the pairing
3Device complexity
If a single vision system camera is used for both imaging and depth sensing, then device complexity is reduced, but optical alignment becomes critical
Solution Approach 1:
The patent combines a single vision system camera with a depth sensor into an integrated optical assembly where both imaging and depth measurement functions share common optical components. A dichroic mirror directs laser light from the depth sensor to the device under test while allowing reflected light to reach the camera, merging two functional systems into one compact unit that reduces overall complexity
Solution Approach 2:
The dichroic mirror serves as an intermediary optical element that manages the interaction between the depth sensor laser and the camera. It reflects the laser light toward the device while transmitting the reflected light from the device to the camera, enabling both functions to operate simultaneously through a single optical path without direct interference
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
Achieves 100% accurate PSU-CPE pairing in 12 seconds per unit, processing 300 pairs in 1 hour and 2400 pairs in 8 hours, significantly reducing processing times and eliminating incompatibility.
Implementation Method 1
a dichroic mirror that passes light of an image scene of a device under test to the camera and reflects laser light emitted from the depth sensor to the device under test and reflects the laser light reflected from the device under test to the depth sensor
Implementation Method 2
a variable focus liquid lens, and an optical component developed for this application. With the described system, PSUs may be accurately matched to routers (or other CPEs)
Implementation Method 3
a depth sensor; a dichroic mirror that passes light of an image scene of a device under test to the camera and reflects laser light emitted from the depth sensor to the device under test and reflects the laser light reflected from the device under test to the depth sensor
Data Source
AI summary
A parts matching system includes a camera: a depth sensor: a dichroic mirror that passes light of an image scene of a device under test to the camera and reflects laser light emitted from the depth sensor to the device under test and reflects the laser light reflected from the device under test to the depth sensor; and a computer that processes (i) data received from the depth sensor to determine a distance of the device under test to the camera and (ii) image data of the device under test received from the camera.