Click on the "Xiaobai Learns Vision" above, select to add "Star" or "Top"
Heavy content delivered first

Source | Machine Vision Salon

With the rapid development of the new generation of information communication technology, the market demand for terminal devices such as smartphones, tablets, and smart home products continues to grow, leading to a sharp increase in the production scale of LCD panels, making them one of the pillars of the information industry.

The rise of the LCD panel industry has driven a surge in demand for panel quality inspection.

Traditional defect detection mainly relies on human eyes, which brings significant difficulties for enterprises to control production quality and improve production efficiency. Therefore, adopting a new type of panel quality detection method to replace manual inspection is an urgent technical problem that needs to be solved.

Surface Defects and Detection Methods of LCD Panels

01 Origins of LCD Panel Defects

Thin Film Transistor Liquid Crystal Display (TFT-LCD) has advantages such as high resolution and low power consumption, and is therefore widely used in the display industry.

However, the complexity of the production process and environmental factors make it difficult to avoid defective displays, resulting in a high product defect rate. The manufacturing process of TFT-LCD displays includes coating, etching, developing, panel assembly, sealing, and installing driver chips. The complexity of these processes leads to defects, with common defects including point defects, line defects, and Mura defects.

The term “Mura” comes from Japanese, meaning spots or stains, and is also known as “cloud spots”. It is one of the most difficult defects to detect in displays. The traditional method for detecting Mura is through manual visual inspection, primarily using naked-eye identification. This method is inefficient and prone to visual fatigue, leading to a decrease in result accuracy.

Machine vision-based LCD panel inspection can detect defects caused by various production processes of LCD panels, including Array process, CF (Color Film) process, CELL (Cell Box) process, Module process, effectively distinguishing defects such as cracks, breakages, chips, scratches, burrs, and drops.

02 Challenges in LCD Panel Detection

Large product area and high production capacity make it impossible for manual inspection of every panel, resulting in missed detections.

Before cutting, LCD panels are usually very large. For example, a 10.5 generation line panel can have an area of 3370mm*2940mm, making manual inspection extremely difficult and prone to fatigue, thus manual inspection cannot guarantee production capacity and accuracy.

Panel defects can easily lead to product breakage.

During the production of LCD panels, defects such as chipping, breakage, cracks, and scratches may occur probabilistically. If not detected and addressed in time, these issues may lead to glass breakage within the main production equipment during subsequent processes, necessitating a halt for maintenance and cleaning, resulting in economic losses due to downtime.

The location and type of defects cannot be statistically analyzed or summarized, making digitization impossible.

Manual inspection lacks image documentation, resulting in poor consistency in inspection standards and making product traceability impossible.

High difficulty in recruitment and high labor costs.

As wage levels continue to rise, labor costs also increase year by year, and recruitment difficulties and high personnel turnover are significant issues. Additionally, manual inspection poses considerable strain on the eyes, leading to high turnover rates and recruitment challenges.

03 LCD Panel Detection Methods

In recent years, many computer vision-based panel defect detection systems have emerged, but the defect detection algorithms used by various manufacturers are generally traditional visual methods. However, as production processes continue to advance, detection requirements have also increased, with demands for detection accuracy and real-time performance significantly heightened.

With the rise of deep learning, we utilize Convolutional Neural Networks (CNN) to detect difficult-to-detect defects.

For detection issues, there are many existing object detection algorithms capable of achieving high precision and real-time detection, effectively identifying and classifying defects in LCD panels—detecting defects such as chipping, cracks, stains, water droplets, and scratches, and integrating with main production equipment to automatically feedback judgments based on defect types: stop or alarm, eliminating the need for real-time monitoring by staff on-site.

Increase Efficiency

Deep learning algorithms can more accurately identify production hazards such as chipping, eliminating downtime caused by glass fragments, and preventing false alarms due to misidentification, thereby improving production efficiency.

Reduce Costs

Avoid equipment damage caused by glass breakage, reduce costs associated with downtime, and lower maintenance costs.

Improve Quality

Reduce product quality issues caused by surface scratches and stains affecting circuit production and liquid crystal imaging.

Enhance Processes

By digitizing quality inspection, we enhance data traceability and analysis capabilities, providing strong support for improving production processes.

LCD Screen Appearance Defect Detection Scheme

Using line array cameras and lenses, select and match backlight sources, coaxial light, or high-brightness linear light sources according to detection requirements to capture images of LCD panels. During the scanning process, intelligent algorithms simultaneously process images, displaying defect information in real-time, determining whether the surface has foreign objects, scratches, color anomalies, bulges, dents, pinholes, burrs, and bright spots. For defective glass, real-time linkage with main equipment is performed to carry out relevant operations such as stopping production to prevent downtime due to glass breakage during the manufacturing process.

Machine Vision Detection Process Design:

1. Open the product to 180°, placing the LCD screen horizontally on the stage;

2. Start detection: the stage moves, and the camera begins to capture images;

3. After image acquisition is complete, the stage resets;

4. Detection software analyzes the quality of the LCD screen;

5. Based on the analysis results, an OK or NG judgment is given;

6. Operator: performs sorting and re-inspection based on the results;

A) Visual Detection Anomaly: Black Spots

B) Visual Detection Anomaly: Foreign Objects

C) Visual Detection Anomaly: Scratches

D) Visual Detection Anomaly: Bubbles

E) Visual Detection Anomaly: Foreign Objects

This article is for academic sharing only. If there is any infringement, please contact to delete the article.

Download 1: OpenCV-Contrib Extension Module Chinese Version Tutorial

Reply "Extension Module Chinese Tutorial" in the background of "Xiaobai Learns Vision" public account to download the first OpenCV extension module tutorial in Chinese available online, covering installation of extension modules, SFM algorithms, stereo vision, target tracking, biological vision, super-resolution processing, and more than twenty chapters of content.

Download 2: 52 Lectures on Python Vision Practical Projects

Reply "Python Vision Practical Projects" in the background of "Xiaobai Learns Vision" public account to download 31 practical projects including image segmentation, mask detection, lane detection, vehicle counting, eyeliner addition, license plate recognition, character recognition, emotion detection, text content extraction, face recognition, etc., to assist in quickly learning computer vision.

Download 3: 20 Lectures on OpenCV Practical Projects

Reply "OpenCV Practical Projects 20 Lectures" in the background of "Xiaobai Learns Vision" public account to download 20 practical projects based on OpenCV to advance learning in OpenCV.

Discussion Group

Welcome to join the reader group of the public account to exchange ideas with peers. Currently, there are WeChat groups for SLAM, 3D vision, sensors, autonomous driving, computational photography, detection, segmentation, recognition, medical imaging, GAN, algorithm competitions, etc. (these will gradually be subdivided). Please scan the WeChat ID below to join the group, and note: "Nickname + School/Company + Research Direction", for example: "Zhang San + Shanghai Jiao Tong University + Vision SLAM". Please follow the format for notes, otherwise, you will not be approved. After successfully adding, you will be invited to related WeChat groups based on research direction. Please do not send advertisements in the group, otherwise, you will be removed. Thank you for your understanding~

LCD Screen Appearance Defect Detection Scheme

Related posts

Leave a Comment Cancel reply