Lane Detection for Self-Driving Cars

Project Overview

Implemented a sophisticated lane detection system for self-driving vehicles using deep learning and computer vision techniques. The system processes road imagery to accurately identify and track lane markings, providing essential information for autonomous navigation.

Technical Approach

  • Developed a VGG-16-based Convolutional Neural Network (CNN) for road segmentation
  • Trained and validated on the industry-standard KITTI Road/Lane Detection Evaluation dataset
  • Implemented advanced preprocessing techniques:
    • Perspective transformation to obtain bird’s-eye view of the road
    • Sobel edge detection for identifying gradient changes
    • Color thresholding in multiple color spaces to isolate lane markings
  • Enhanced detection accuracy through:
    • Sliding window lane detection approach
    • Region-of-interest masking to focus processing on relevant road areas
  • Leveraged transfer learning to improve model performance and reduce training time

Technologies Used

  • Python
  • TensorFlow
  • Keras
  • Pandas
  • NumPy
  • OpenCV

Results

The system achieved 98.58% pixel-wise accuracy in lane segmentation, demonstrating excellent performance across various road conditions and scenarios present in the KITTI dataset.

See Repo Here