LED Panel Pixel Measurement Using Camera and AI Colorimetry
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for measuring luminance and chromaticity of LED elements in LED panels are time-consuming and costly, requiring frequent calibration and using expensive equipment like image colorimeters.
Innovation Solution
An LED panel measuring apparatus that uses a camera to capture images of the panel while displaying preset colors, classifies pixels into regions based on brightness, and employs a trained artificial intelligence model to obtain tristimulus values, luminance, and chromaticity, reducing measurement time and cost.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If an image colorimeter is used to measure luminance and chromaticity of LED elements, then measurement accuracy is improved, but measurement time increases and equipment cost increases
Solution Approach 1:
The system performs preliminary calibration by capturing images of a reference chart with known luminance and chromaticity values. This calibration data is stored and used to establish conversion relationships, eliminating the need for time-consuming calibration before each measurement. The AI model is pre-trained with calibration data to enable direct measurement without repeated calibration processes.
Solution Approach 2:
Instead of using expensive specialized measurement equipment (image colorimeter), the system uses a standard camera to capture images of the LED panel. The camera captures visual information that is then processed through AI algorithms to derive luminance and chromaticity measurements, creating a functional copy of the measurement capability using cheaper, more accessible equipment.
2Measurement precision
If an image colorimeter is used to measure luminance and chromaticity of LED elements, then measurement accuracy is improved, but equipment cost increases
Solution Approach 1:
The system replaces expensive specialized measurement equipment (image colorimeter with CCD sensor and tricolor filters) with a standard camera. The camera captures images that are then processed using AI algorithms to extract luminance and chromaticity information, achieving comparable measurement accuracy with significantly lower equipment cost.
Solution Approach 2:
The system replaces the mechanical/optical measurement system (image colorimeter with physical filters and sensors) with a computational approach using standard camera imaging combined with AI image processing. This substitution eliminates the need for specialized optical components while maintaining measurement capability through algorithmic analysis.
3Measurement precision
If calibration process is performed every time measurement is made using image colorimeter, then measurement accuracy is maintained, but measurement complexity increases
Solution Approach 1:
The system performs calibration once during system initialization or setup, capturing images of a reference chart and storing the calibration data. This preliminary calibration establishes the relationship between camera responses and actual luminance/chromaticity values. Subsequent measurements use this pre-established calibration without requiring repeated calibration processes, significantly reducing operational complexity.
Solution Approach 2:
The AI model automatically performs measurements using the pre-stored calibration data without requiring manual calibration operations. The system self-calibrates by referencing the stored calibration information, eliminating the need for operators to perform complex calibration procedures before each measurement session.
Data Source
AI summary
Disclosed are an apparatus, a system, and a control method for measuring an LED panel. The method for measuring an LED panel comprises the steps of: capturing an image of an LED panel displayed in a preset color; classifying each pixel of the LED panel from the captured image of the LED panel into pixel regions having different sizes according to a predetermined threshold brightness value; obtaining a tristimulus value of each pixel by using the data of the classified pixel regions and a trained artificial intelligence model; and obtaining the luminance and chrominance of each pixel on the basis of the obtained tristimulus value.