一种融合深度学习与几何建模的鲁棒性车道线检测方法
By integrating deep learning and geometric modeling, this method solves several bottleneck problems in lane detection technology under complex environments, achieving accurate lane detection in changing road conditions and complex driving scenarios, thus ensuring driving safety and stability.
CN118230273BActive Publication Date: 2026-07-17UNIV OF JINAN
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- UNIV OF JINAN
- Filing Date
- 2024-04-28
- Publication Date
- 2026-07-17
Smart Images
Figure CN118230273B_ABST
Abstract
本发明提供一种融合深度学习与几何建模的鲁棒性车道线检测方法，首先利用基于空间变换的卷积神经网络网络对预处理后的车载图像进行关键特征提取，将生成后的特征图进行水平翻转处理，再将原始特征图和翻转后的特征图进行有效融合得到聚合特征图，采用数据驱动的策略拟合并重建三次样条曲线，形成一种可自适应学习的车道线几何模型，最后，通过分类分支和回归分支输出预测结果，精确估计出三次样条曲线的控制点和车道线存在的概率。本发明在确保曲线检测所特有的低计算复杂度优势前提下，实现了鲁棒性和准确性的统一，为自动驾驶及其他相关应用领域提供了坚实可靠的技术。
Need to check novelty before this filing date? Find Prior Art