Text is one of the most important sources of information for humans, and natural scenes are filled with various textual symbols. Optical Character Recognition (OCR) is believed to be familiar to everyone.

The image and text recognition in industrial scenarios is more complex, and OCR appears in many different situations, optically recognizing characters on certain special surfaces and the overall product, facilitating information comparison and error correction of character printing, and providing data support for production. When detecting missing or erroneous characters, the system sends a signal to the removal device for removal or alarm prompts

In this issue, Xiao Ming will share Meiji Sensing's OCR recognition solution for charging heads in the 3C industry

Application scenario

🔸  The detection requirements include contour positioning, screen printing detection, character recognition, and QR code recognition.

🔸  Laser engraving is used for screen printing, and hardware such as cameras cannot be directly installed below. Refractive lenses need to be customized according to the laser wavelength

🔸  After identifying the data of characters and QR codes, it is necessary to compare them with the character data and QR code data in the standard database to ensure the accuracy of laser engraving

solution

1. Using high-precision positioning algorithms to provide accurate position coordinates for subsequent silk screen detection

2. Detecting character defects through differential algorithms to ensure the integrity of character information and appearance defects

3. Using deep learning algorithms for OCR recognition of surface character information, with an accuracy rate of over 99.5%

Solution Process

OCR recognition technology mainly relies on image processing and pattern recognition algorithms. By capturing character features in documents, such as strokes, shapes, sizes, spacing, etc., and comparing them with a preset character library, corresponding text information can be recognized. We need to go through the following processes:

1. Image acquisition

Use high-resolution industrial cameras or smartphone cameras to capture images of the charging head.

It is necessary to ensure that the shooting environment has uniform lighting and that the charging head surface is clean and unobstructed.

2. Image preprocessing

Perform preprocessing operations such as binarization, denoising, and contrast enhancement on the collected images to improve the accuracy of subsequent recognition.

Image preprocessing can be achieved through software algorithms or image processing libraries such as OpenCV.

3. Regional positioning

By using algorithms such as template matching and edge detection, locate the area where the character or identifier to be recognized on the charging head is located.

This step can ensure the accuracy of subsequent OCR recognition and reduce unnecessary computational complexity.

4. OCR recognition

Perform character recognition on the located area.

Characters can be converted into editable and searchable text based on pixel information in the image.

Screen printing detection, character extraction

5. Result output

Output the OCR recognition results in text format, which can be saved to a database, file, or displayed on the user interface.

At the same time, the recognition results can be validated and verified to ensure the accuracy and reliability of the data.