Scholarlink

In response to the pressing need for improved accessibility in public transportation, this paper introduces an Automated Multilingual Bus Route Announcement System designed specifically for KSRTC buses. Leveraging advanced technologies including YOLOv8 for license plate detection, EasyOCR for number recognition, and GTTS for multilingual announcements, the system efficiently identifies buses upon entry to bus stands and disseminates route information in multiple languages. Despite challenges posed by the variable image quality and camera positions, our system demonstrates exceptional performance metrics: an accuracy of 93.89%, an F1 score of 96.85%, a precision of 96%, and a recall of 97.6%. By offering real-time, multilingual route information, our system aims to significantly enhance the accessibility and usability of public transportation for diverse user demographics. This automated approach not only reduces the dependency on manual announcements made by people but also ensures consistent, timely, and accurate information delivery, thus improving the overall efficiency and reliability of the service.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Next-Generation Cross-Language Visual Transit Navigation System Leveraging Deep Learning and NLP

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • K. J. Ashik,
  • Edwin Martin,
  • A. Gauri,
  • P. V. Rajaraman,
  • Asha Rose Thomas,
  • M. G. Sabitha

摘要

In response to the pressing need for improved accessibility in public transportation, this paper introduces an Automated Multilingual Bus Route Announcement System designed specifically for KSRTC buses. Leveraging advanced technologies including YOLOv8 for license plate detection, EasyOCR for number recognition, and GTTS for multilingual announcements, the system efficiently identifies buses upon entry to bus stands and disseminates route information in multiple languages. Despite challenges posed by the variable image quality and camera positions, our system demonstrates exceptional performance metrics: an accuracy of 93.89%, an F1 score of 96.85%, a precision of 96%, and a recall of 97.6%. By offering real-time, multilingual route information, our system aims to significantly enhance the accessibility and usability of public transportation for diverse user demographics. This automated approach not only reduces the dependency on manual announcements made by people but also ensures consistent, timely, and accurate information delivery, thus improving the overall efficiency and reliability of the service.