Object-Text Detection and Recognition System

Abstract

The object recognition system based on deep learning has been applied in different domains, e.g. in the Intelligent transportation system, Autonomous driving system, etc. Along with object detection, in numerous scenes, text detection and recognition have conjointly brought abundant attention and analysis application.Real-time object detection and dimensioning as well as text recognition are important topics for many branches of the industry today.The projected system consists of object – text, detection and recognition module, and a dimension measuring module. This system offers an improved method of classifying objects and calculating their measures in real-time from video sequences. The proposed system uses OpenCV libraries, which comprise erosion algorithms, canny edge detection, dilation, and contour detection. To accomplish the task of the text recognition Tesseract OCR engine is employed.