Optical Character Recognition

Client Overview

A US based heavy machinery manufacturer.

Business Challenge

Client wants to develop a mobile based solution that could take pictures of metallic plates on their machinery, uses OCR to automatically detect and categorize information and sends them to cloud. The main challenge was the plates were metallic and not in very good condition for any regular OCR to run on them.

Acclivis Contribution

  • Complete mobile application development
  • Pre-processing image module development
  • Customized image processing algorithm development for OCR on metallic plates
  • Database storage on cloud server (Microsoft Azure)

Product Features Developed/Supported by Acclivis

  • The app is able to capture images of metallic tags and convert selected area
    of image into text.
  • Selection on image can be made using touch functionality of the screen.
    Moreover image can be zoomed via pinch-in-out method for more precise ROI
    selection. Alternatively, option of reselection is also available in case user
    selected wrong ROI
  • The app also has functionality to upload the recognized text along with few
    other info related data (example: GPS location, image thumbnail, date/
    time stamp etc) to cloud over Wi-Fi or
    mobile data.
  • The app has inbuilt image pre-processing module which enhances the quality of
    image before passing it to the OCR engine. The app can also recognize whether image is captured
    properly or not.

Tools and Technology

  • Cloud | Microsoft Azure
  • Database | SQL server with .Net
  • Image Processing | Visual studio, OpenCV
  • iOS Application Development | Xcode
  • Project Management Tools | Gitlab, Bitbucket

Client Benefit

  • Because of mobile application and highly customized OCR for metallic plates by Acclivis, client had very ease in cataloging machinery and collecting data of the machines that are going in field for work..