FPGA Image Acceleration for Real-Time LCD Optical Inspection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current automated optical inspection systems for TFT-LCD production face challenges in processing large amounts of image data in real time, leading to inefficiencies and visual fatigue in manual inspections, as traditional PCs struggle to handle the data processing demands of mass production.
Innovation Solution
An image acceleration processing device utilizing a field programmable gate array (FPGA) processing platform, combined with a personal computer (PC) and optical fiber interfaces, enables comprehensive control of screen lighting and data processing, supporting distributed processing and parallel computing to enhance image processing capacity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional automated optical inspection uses multiple industrial cameras and CCDs to improve inspection capability, then measurement precision and reliability are improved, but the amount of image calculation data increases dramatically and traditional PCs cannot handle the processing demand
Solution Approach 1:
The patent replaces the traditional PC-based processing system with an FPGA-based parallel processing system. The FPGA architecture provides massive parallel computing capability to handle the large volume of image data from multiple cameras and CCDs, substituting the insufficient sequential processing of traditional PCs with high-speed parallel processing that can keep up with the inspection requirements.
Solution Approach 2:
The patent divides the image processing task into multiple parallel processing channels, each handling data from specific cameras or CCDs. The FPGA is configured with multiple processing units that simultaneously process different portions of the image data, enabling the system to handle the increased data volume from multiple sensors without overwhelming a single processing unit.
2Device complexity
If traditional PC is used for image data processing, then device complexity is kept simple, but processing time increases and cannot meet real-time inspection requirements
Solution Approach 1:
The patent replaces the PC-based processing architecture with an FPGA-based system that provides deterministic real-time processing. The FPGA's hardware-level parallel processing eliminates the software overhead and sequential execution limitations of PC-based systems, achieving consistent real-time performance for inspection applications.
3Device complexity
If manual inspection is used to maintain simplicity, then device complexity is low, but workers experience visual fatigue and inspection accuracy decreases due to repetitive detection
Solution Approach 1:
The patent implements an automated optical inspection system that performs detection tasks autonomously without human intervention. The system uses multiple cameras and CCDs combined with automated image processing algorithms to detect defects, eliminating the need for manual visual inspection and the associated visual fatigue and human error.
Data Source
AI summary
An image acceleration processing device including a field programmable gate array (FPGA) processing platform; and a personal computer (PC). The FPGA processing platform includes: a first fiber interface configured to receive configuration parameters and test commands, and to output test results; a second fiber interface configured to exchange data with the PC; a third fiber interface configured to receive image data and output the configuration parameters and the test commands; a fourth fiber interface configured to control the generation of a screen lighting signal; and a fifth fiber interface configured to control an input/output (IO) light source.