Display Panel Inspection Autofocus via Fourier Magnification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing autofocus mechanisms for display panel inspection, such as contrast-detection AF, suffer from slow operation due to uncertainty in determining the direction of out-of-focus camera position, while laser AF requires expensive dedicated hardware and necessitates calibration.
Innovation Solution
An inspection method that estimates the current focus position of an imaging device using Fourier transform on a one-dimensional luminance profile of a display panel image, calculating imaging magnification without requiring dedicated hardware, enabling fast autofocus.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If contrast-detection AF is used for autofocus, then high focus accuracy is achieved, but operation speed becomes slow
Solution Approach 1:
The patent replaces the mechanical search-based contrast-detection AF system with a computational optics system using Fourier transform. Instead of mechanically moving the lens and evaluating contrast at each position, the system uses optical field analysis through Fourier transform to directly calculate focus position and imaging magnification from the captured image data, achieving both high accuracy and fast operation speed.
Solution Approach 2:
The patent changes the fundamental parameter used for focus detection from image contrast (which requires iterative search) to optical field distribution characteristics (which can be directly analyzed). By transforming the image data into frequency domain through Fourier transform, the system extracts focus information from the optical field parameters themselves, enabling direct calculation without iterative searching.
2Speed
If laser AF is used for fast autofocus, then operation speed is improved, but dedicated hardware cost increases
Solution Approach 1:
The patent substitutes the laser-based optical measurement system with a computational image processing system. Instead of using laser illumination and dedicated distance sensors, the system uses standard imaging device data processed through Fourier transform to obtain focus position and imaging magnification, eliminating the need for expensive dedicated AF hardware while maintaining fast operation speed.
Solution Approach 2:
The patent creates a computational model of the optical system behavior through Fourier transform analysis. By analyzing the frequency domain characteristics of the captured image, the system reconstructs focus position and imaging magnification information that would traditionally require dedicated optical measurement hardware, effectively creating a virtual copy of the distance measurement function through software-based optical field analysis.
3Speed
If laser AF is used for distance measurement, then fast operation is achieved, but calibration complexity increases
Solution Approach 1:
The patent replaces the laser distance measurement system with a self-calibrating computational optics system. The Fourier transform analysis inherently provides imaging magnification information that directly correlates with focus position, eliminating the need for separate calibration procedures between distance measurement and focus control systems. The system automatically determines the relationship between optical field characteristics and focus position through the mathematical transformation.
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
Enables highly accurate and fast autofocus without the need for dedicated hardware, improving inspection efficiency and reducing operational costs.
Implementation Method 1
an inspection image of a pixel region of the display panel captured by an imaging device
Implementation Method 2
imaging device with an optical system
Implementation Method 3
generating a frequency spectrum intensity profile by performing Fourier transform on the one-dimensional luminance profile; calculating, based on a fact that the plurality of pixels are arranged in a periodic structure in the pixel region, an imaging magnification of the imaging device from one or more peak positions included in the frequency spectrum intensity profile
Data Source
AI summary
An inspection method includes: obtaining a one-dimensional luminance profile by performing image processing of compressing a luminance of an inspection image of a pixel region of a display panel captured by an imaging device, into a one-dimensional luminance; generating a frequency spectrum intensity profile by performing Fourier transform on the one-dimensional luminance profile; calculating, based on a periodic structure of pixels shown in the pixel region, an imaging magnification of the imaging device from one or more peak positions included in the frequency spectrum intensity profile; and estimating a current focus position from the imaging magnification calculated and magnification information indicating a relationship between a focus position and an imaging magnification of the imaging device, the current focus position being a focus position of the imaging device in the inspection image.