DETECTION OF LINE SEGMENTS OF TRANSPORTATION MONUMENTS FOR AUTONOMOUS AND SEMI-AUTOMATIC SYSTEMS AND APPLICATIONS

Machine learning models for detecting road marking line segments enhance navigation precision and safety in autonomous vehicles by accurately determining vehicle locations and navigation rules based on road markings, addressing the limitations of conventional systems in environments with sparse traffic features.

DE102025138906A1Pending Publication Date: 2026-03-26NVIDIA CORP
1 Cites 0 Cited by

Patent Information

Application Number
DE102025138906
Authority / Receiving Office
DE · DE
Patent Type
Applications
Current Assignee / Owner
Priority Date
2024-09-26
Filing Date
2025-09-25
Publication Date
2026-03-26

AI Technical Summary

Technical Problem

Conventional systems fail to accurately determine the locations of road markings for precise longitudinal localization of vehicles, especially in environments with few traffic signs or signals, limiting the navigation capabilities of autonomous and semi-autonomous systems.

Method used

The use of machine learning models to process sensor data for detecting and determining the locations of line segments associated with road markings, enabling accurate localization and navigation by aligning these segments with map data, thereby enhancing the safety and precision of vehicle maneuvers.

Benefits of technology

Enables precise determination of vehicle locations along roads without relying on traffic signs, improving navigation accuracy and safety by accurately identifying permitted and restricted navigation areas based on road markings.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

This document describes the detection of line segments of traffic features for autonomous and / or semi-autonomous systems and applications in various examples. The systems and methods described herein can determine the locations of line segments (e.g., line markings) associated with road markings within environments. For example, one or more machine learning models can process sensor data (e.g., image data, etc.) to determine points associated with the line segments as represented by the sensor data, and / or directional information (e.g., direction vectors) associated with the points. As described herein, the points can be associated with the edges of the line segments, the midpoints of the line segments, and / or other locations of the line segments.Systems and procedures are then described that perform operations based on the locations of these line segments, such as updating a localization map, performing localization and / or determining trajectories to be navigated.
Need to check novelty before this filing date? Find Prior Art