Road Marking Detection Using 2D Images and 3D Point Clouds

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current methods for detecting road traffic markings in car navigation and vehicle automatic driving systems are affected by sparse, occluded, or missing three-dimensional point clouds, and manual detection is inefficient and prone to errors.

Innovation Solution

A data processing method that collects a two-dimensional streetscape image, a three-dimensional point cloud, and inertial navigation data, performs region segmentation, and calculates three-dimensional coordinates of road traffic markings using image recognition techniques to improve accuracy and practicality.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Extent of automation

If road traffic markings are detected directly from three-dimensional point cloud data, then the detection process is automated, but the detection accuracy deteriorates when point cloud data is sparse, occluded, or missing

Engineering Contradiction:
Improveautomation of detection processVSAvoiddetection accuracy
Core Design Contradiction:
Extent of automationVSMeasurement precision

Solution Approach 1:

The patent introduces a two-dimensional streetscape image as an intermediary medium between the three-dimensional point cloud data and the final detection result. The image serves as a mediator that fills in gaps where point cloud data is sparse or occluded, allowing automated detection to maintain high accuracy by cross-referencing the image information with the point cloud data

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent transitions from exclusively using three-dimensional point cloud data to incorporating two-dimensional image data. This dimensional shift allows the system to detect road traffic markings from a different data perspective, compensating for deficiencies in the three-dimensional point cloud representation

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Measurement precision

If manual detection method is used to extract road traffic markings from colorful point cloud, then detection accuracy can be maintained, but processing efficiency deteriorates and human errors increase

Engineering Contradiction:
Improvedetection accuracyVSAvoidprocessing efficiency
Core Design Contradiction:
Measurement precisionVSProductivity

Solution Approach 1:

The patent enables the system to automatically perform detection tasks that were previously requiring manual intervention. By training the automated detection algorithm to recognize road traffic markings in the two-dimensional image and map them to the three-dimensional point cloud, the system serves itself without human involvement, achieving both high accuracy and high efficiency

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The patent replaces the manual mechanical detection process with an automated computational system. The manual extraction of markings from colorful point clouds is substituted by an algorithm that automatically processes the two-dimensional image and three-dimensional point cloud data, eliminating human errors and significantly improving processing speed

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentEP3407294B1Information processing method, device, and terminal
Publication Date: 2021.03.31 TENCENT TECHNOLOGY (SHENZHEN) CO LTD
  • EP3407294B1 patent drawingFigure 1
  • EP3407294B1 patent drawingFigure 2a~2c
  • EP3407294B1 patent drawingFigure 3a~3b

AI summary

An information processing method, device, and terminal. The method comprises: collecting road data, the road data comprising a two-dimensional street view image, a three-dimensional point cloud and inertial data (SI01); dividing the two-dimensional street view image into regions on the basis of a spatial position relationship of the two-dimensional street view image, the three-dimensional point cloud and the inertial data, and extracting a ground regional image (S102); detecting at least one target road traffic marking in the ground regional image, the road traffic marking comprising lane lines and/or road signs (S103); calculating a three-dimensional coordinate of the at least one target road traffic marking on the basis of the spatial position relationship of the two-dimensional street view image, the three-dimensional point cloud and the inertial data (SI04). By means of detecting the road traffic marking from a two-dimensional street view image, the present invention is able to improve the accuracy of detection results and improve the practicality of data processing.