The Use of Tesseract OCR Number Recognition for Food Tracking and Tracing

Abstract

One of the most interesting enabling technologies for digital transformation is computer vision. Object and character recognition has already become very popular and it is used in everyday life. This research focuses on the use of computer vision to read serial numbers from wine labels in order to enable applications based on tracking and tracing of each individual wine bottle. After experimenting with several OCR tools, an open source software called Tesseract OCR engine was selected for the pilot solution. The paper discusses the implementation and image processioning that improved detection accuracy. The coding was done in the Python programming language. The solution code was tested using real-life like images of wine serial numbers. In addition, a custom built web-based evaluation tool was created and used for the interactive evaluation of the system.