Lane Marker Recognition Using Front and Side Camera Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems struggle to accurately estimate lane marker information on the far side of a lane after a lane change due to visual limitations and obstruction by other vehicles, which is crucial for safe lane changes in autonomous driving systems.
Innovation Solution
A road information recognition system that combines data from front and side cameras to detect lane markers on both sides of the vehicle, utilizing stereo cameras for accurate positioning and radar for distance measurement, allowing estimation of lane markers on the far side by incorporating lane width information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If only a front camera is used to photograph the front of the vehicle, then the system structure is simple, but the lane marker information on the far side of the lane after change of lane cannot be obtained accurately
Solution Approach 1:
The patent divides the detection task into multiple segments by using different cameras for different purposes: front camera for current lane detection, side cameras for adjacent lane detection. This segmentation allows each camera to focus on specific areas, achieving comprehensive lane marker information without requiring a single complex camera system
Solution Approach 2:
The patent transitions from a single-front viewing dimension to multiple dimensions by adding side cameras. This dimensional expansion enables the system to capture lane markers that are invisible from the front view alone, particularly the far side lane markers needed for safe lane change decisions
2Loss of information
If a side camera is mounted to photograph the side of the vehicle, then lane marker information on the side can be obtained, but the lane marker on the far side of the lane after change of lane still cannot be detected
Solution Approach 1:
The patent merges information from multiple sources: front camera data, side camera data, and vehicle state information (steering angle, speed). By combining these diverse information sources, the system can infer far side lane marker positions even when they are not directly visible, overcoming the limitations of individual sensors
3Reliability
If another vehicle is traveling on the lane after change of lane, then the lane marker may be shielded, but the system cannot obtain the lane marker information on the far side of the lane after change of lane
Solution Approach 1:
The patent performs preliminary detection and recording of lane marker information before the lane change occurs. By capturing and storing this information in advance, the system has reference data available even when the far side lane markers become obscured by other vehicles during the actual lane change maneuver
Solution Approach 2:
The system creates a copy or model of the lane marker information based on detected parameters (position, orientation, curvature). This copied information can be used to infer the presence and position of far side lane markers even when direct visual detection is blocked by other vehicles
Data Source
Figure 1
Figure 2
Figure 3A~3B
AI summary
The purpose of the present invention is to provide a novel road information recognition system capable of highly accurately estimating the positions of lane markers (LM4) to be located on the far side of a lane (L3) after change of lane. The present invention is equipped with: a front lane marker detection means for detecting, by using a front imaging means, lane markers (LM3) located ahead (FF1) of a host vehicle; a side lane marker detection means for detecting, by using a side imaging means, the positions of lane marker (LM3, LM4) located on both sides (RR1) of the vehicle and a lane width (W); and a front lane marker estimation means for estimating the positions of the lane markers (LM4) to be located on the far side of the lane (L3) after change of lane on the basis of the information regarding the lane markers (LM3) located ahead (FF1) of the vehicle, which is obtained by the front lane marker detection means, and the positions of the lane markers (LM3, LM4) located on the sides (RR1) of the vehicle and the lane width (W), which are obtained by the side lane marker detection means .