UTAR Institutional Repository

Optical character recognition (OCR) solution for capturing data from legacy manufacturing machines

Lai, Suk Ling (2019) Optical character recognition (OCR) solution for capturing data from legacy manufacturing machines. Final Year Project, UTAR.

[img]
Preview
PDF
Download (3365Kb) | Preview

    Abstract

    This project is named Optical Character Recognition (OCR) Solution for Capturing Data from Legacy Manufacturing Machines. Literally, it is a system that can read data from screen of the legacy machines for the operator for monitoring purpose. In fact, there are still exist some factories and manufacturers applying old machines rather than upgrading the infrastructure to Industrial 4.0 which makes everything connected. This trouble the productivity and monitoring process. However, upgrading the environment to the Industrial 4.0 is not easy as the cost is expensive and the machines are proprietary solutions therefore no customization is allowed. Some machines have to run 24 hours therefore it is unable to off for customization and modification as it will cause loss to the company. This project would help in solving the problems in the lowest cost without interrupting the operation. This project is to find the most suitable OCR engine, to develop an automated OCR solution with minimal human calibration and to transmit the data to the cloud for data storage. This system will be hanging in front of the machine’s screen and collects the data so that operator can have a clear view of the status of the machine’s operation. This system able to capture the screen of the machines using the Raspberry Pi camera, recognize the character/data on the screen using existing OCR engine and upload the data to the cloud for storage purpose Google cloud. The OCR engine chosen for the system is Tesseract after doing the researches. Researching, analyzing and evaluating are a must for a good project. The review on the existing solutions based on the criteria. The criteria are 1) the solution is online or offline, 2) need of cropping, 3) detect the character automatically, 4) font and language limitation, 5) needs of human intervention. OCR engines has been reviewed to find the most suitable engine to the project.

    Item Type: Final Year Project / Dissertation / Thesis (Final Year Project)
    Subjects: Q Science > Q Science (General)
    Divisions: Faculty of Information and Communication Technology > Bachelor of Computer Science (Hons)
    Depositing User: ML Main Library
    Date Deposited: 20 Aug 2019 12:20
    Last Modified: 20 Aug 2019 12:20
    URI: http://eprints.utar.edu.my/id/eprint/3486

    Actions (login required)

    View Item