OCR recognition technology

OCR recognition technology plays a crucial role in industrial production. It utilizes advanced image processing technology to automatically and efficiently recognize and convert textual information in images into editable text data.

The application of this technology ensures the accuracy and real-time nature of production data, providing strong support for optimizing production processes, improving product quality, and making business decisions.

In the plastic processing industry, the plastic mixing station is a key link that involves mixing different types of plastic raw materials according to specific formula ratios to ensure the quality and performance of the final product. At this workstation, it is crucial for workers to check the part number on the outer packaging of plastic pellets, as it directly affects the accuracy of raw materials and the continuity of production.

In traditional work methods, staff will carefully check the part number information on the outer packaging before loading. These part numbers usually appear in the form of numbers, letters, or combinations thereof, used to uniquely identify key information such as the type, specification, manufacturer, and production date of the batch of plastic pellets.

Nowadays, it is only necessary to use more efficient visual sensors to quickly confirm the correctness of the part number.

Scene requirements - OCR recognition

At the plastic mixing station, sensors need to be installed here to detect the part number of the plastic particle outer packaging, so that the staff can confirm the correct part number before feeding and prevent the wrong material from being loaded

Testing difficulties:

1. The part number printing on the outer packaging may vary in size and may not be standard OCR characters

2. The packaging may have wrinkles, which can affect camera recognition to some extent

3. The material is manually lifted in front of the camera for inspection, and the inspection position is not fixed, requiring the camera to have a sufficiently large field of view

solution

The Akusense visual sensor CDS-X112RD series is selected. When the material reaches the designated area, the visual sensor quickly and accurately obtains the part number and related information, and outputs a signal of OK/NG by comparing the specified part number in the pre-set production plan or process requirements.

1. The CDS series uses sensors and optical components to obtain images of the tested object, and implements functions such as detection, classification, OCR recognition through the device's built-in deep learning algorithm. It can also output detection results through various communication methods

2. Integrating deep learning algorithms, with only a small number of samples and no need for configuration or parameter tuning, can meet the requirements of visual detection

3. Easy installation and debugging, one click automatic parameter tuning, non professionals can also quickly get started

4. 2 million pixels paired with a liquid lens for clearer imaging

5. Adopting an ultra flexible light source system, red/white light can be customized according to the scene, and all four sets of light sources can be independently controlled, with high flexibility

6. IP67 protection, fearless of harsh industrial application environments