RONELD: Robust Neural Network Output Enhancement for Active Lane Detection
This addresses a critical issue for autonomous vehicle navigation by improving robustness in lane detection across different environments, though it is incremental as it builds on existing CNN-based methods.
The paper tackles the problem of lane detection performance dropping on unseen datasets by proposing RONELD, a method that enhances neural network outputs to identify and optimize active lanes, resulting in up to a two-fold increase in accuracy in cross-dataset tests.
Accurate lane detection is critical for navigation in autonomous vehicles, particularly the active lane which demarcates the single road space that the vehicle is currently traveling on. Recent state-of-the-art lane detection algorithms utilize convolutional neural networks (CNNs) to train deep learning models on popular benchmarks such as TuSimple and CULane. While each of these models works particularly well on train and test inputs obtained from the same dataset, the performance drops significantly on unseen datasets of different environments. In this paper, we present a real-time robust neural network output enhancement for active lane detection (RONELD) method to identify, track, and optimize active lanes from deep learning probability map outputs. We first adaptively extract lane points from the probability map outputs, followed by detecting curved and straight lanes before using weighted least squares linear regression on straight lanes to fix broken lane edges resulting from fragmentation of edge maps in real images. Lastly, we hypothesize true active lanes through tracking preceding frames. Experimental results demonstrate an up to two-fold increase in accuracy using RONELD on cross-dataset validation tests.