Online detection system for large-screen TFT-LCD – product brief
Current situation and problems: at present, the final test section is checked manually, the speed of manual inspection is far lower than the rhythm of the production line, and each production line needs to be inspected by several workers in parallel. Workers feel tired while watching the lighted LCD screen for a long time, which leads to missed detection and high subjectivity. But it takes a big cost to train one skilled worker;
The product realizes the screen detection of the large size LCD screen of TV set. According to the actual requirements, different detection rules can be configured to detect various screen display defects, and the purpose of replacing manual inspection can be achieved.
|Manual reinspection and maintenance|
Online detection system for large-screen TFT-LCD – technical scheme
The image is collected by high-definition camera array, the existing manual detection standard is detected by depth learning method, and all kinds of screen defects are identified and judged, and the real-time results are outputted and saved to the database for later reference.
The whole machine is arranged beside the automatic production line, and a signal port connecting with the production line is provided.
Online detection system for large-screen TFT-LCD – detect content
The screen display defects come from all stages of the manufacturing process. From the installation of backlight in every process of LCD panel (CF-TFT-LCD-LCM) to the assembly of the screen of the whole machine, some defects such as uneven imaging, poor circuit, foreign body invasion and so on may be introduced. Finally, it can be detected by visual algorithm on the illuminated screen.
Online detection system for large-screen TFT-LCD – technical feature
- Flexible deployment, line-side fast installation, general linkage control signal, and high efficiency commissioning;
- Universal detection screen supports custom detection screen and detection contents;
- Configurable detection rules and grades are compatible with different levels of screen detection, to avoid false detection and missed detection;
- The method of deep learning is used to identify and judge various defects. According to the requirement of quality, the new defects and overjudgment results can be learned online, and the judgment model can be added. Meanwhile, simple learning methods enable the factory to continuously improve the accuracy of detection;
- Using camera array to capture images, the number of camera arrays can be adjusted flexibly according to the resolution of different LCD modules;
大屏TFT-LCD在线检测系统 – 现场应用