Computer Vision Approach for Traffic Sign Recognition Using Raspberry Pi

• Published in: 2024 International Conference on Power, Energy, Control and Transmission Systems (ICPECTS) pp. 1 - 4
• Main Authors: Azhagumurugan, R., A, Hariharan, A, Logeshraaj, R, Sathish G
• Format: Conference Proceeding
• Language: English
• Published: IEEE 08.10.2024
• Subjects: Advanced Driver Assistance System(ADAS); Computer vision; Control systems; Location awareness; Neural Network; Open CV; Optical character recognition; Programming; Python; Raspberry pi; Representation learning; Software algorithms; Subtraction techniques; Text recognition; Traffic sign recognition
• DOI: 10.1109/ICPECTS62210.2024.10780157

The Optical Character Recognition system translates scanned images of printed or handwritten text into audio output, using a Raspberry Pi, hence lies in the broader field of computer vision and a subfield of Artificial Intelligence. OCR systems for many global languages are already being used successfully. A Gaussian-based background subtraction technique is used to separate object regions. The system performs recognition and text localization in order to acquire text information. A combination of text localization techniques and the Tesseract algorithm, which utilizes edge pixel distribution, stroke orientation, and gradient feature learning in an AdaBoost model, enables automatic localization of text regions within an object. Once the localized text characters are binarized , they can be identified by common OCR software, particularly Tesseract. For visually impaired users, the text that has been identified is subsequently translated to speech. Noted is the effectiveness of the suggested text localization algorithm. After the recognition process is over on the Raspberry Pi, character codes from text files are further processed using Python programming and a Tesseract algorithm to give audio output.