Enhanced Lane Detection for Autonomous Campus Shuttles Using Hybrid Computer Vision Techniques

Ensuring safe and efficient navigation in smart campus transportation systems is increasingly crucial with the integration of autonomous vehicles, making accurate and robust lane detection a paramount factor. This research presents a combined traditional computer vision algorithm with advanced deep learning techniques to enhance the precision and dependability of lane detection in diverse campus environments in autonomous campus shuttles. The system utilizes a multi-stage process, incorporating color segmentation, edge detection, and Convolutional Neural Network (CNN) for robust lane feature extraction. A dynamic region of interest adjustment mechanism that adapts to varying campus scenarios ensuring adaptability to lighting conditions and road layouts, was introduced. Experimental results using a miniature scaled-down autonomous campus shuttle demonstrated significant improvements in lane detection accuracy compared to conventional methods, validating the effectiveness of our enhanced approach. The proposed system enhances the safety and reliability of autonomous campus shuttles and contributes to the broader field of computer vision applications for intelligent transportation systems.