Lane Line Estimation Using Orthogonal Projection of Keypoints
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Accurately identifying lane lines from sensor data is complex due to detection errors and unordered keypoint detections, leading to erratic representations of lanes that can degrade the accuracy of downstream processes.
Innovation Solution
A heuristic-based approach is employed to estimate lane lines by acquiring sensor data, transferring headings to key point detections, and using these detections to project lines onto an orthogonal line, averaging projection points to define an estimated line point, and iteratively connecting these points to define the lane line.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If lane lines are directly connected from raw sensor data, then the process is simple, but the accuracy of lane line representation deteriorates due to detection errors and unordered keypoint detections
Solution Approach 1:
The patent applies preliminary action by sorting keypoint detections in chronological order before connecting them to form lane lines. This preprocessing step ensures that detections are arranged in the correct temporal sequence, preventing erratic lane line representations while maintaining a relatively simple generation process.
Solution Approach 2:
The patent introduces an intermediary filtering mechanism that processes raw sensor detections before final lane line generation. This intermediary step validates and orders keypoint detections, resolving the contradiction by adding a processing layer that improves accuracy without significantly complicating the overall system.
2Ease of operation
If keypoint detections are connected in chronological order, then the process is straightforward, but the lane line exhibits erratic characteristics due to unordered detections and overlaps
Solution Approach 1:
The system performs preliminary sorting of keypoint detections in chronological order before connection. This ensures that overlapping and unordered detections are properly sequenced, eliminating erratic lane line characteristics while maintaining straightforward operation.
Solution Approach 2:
The patent implements feedback mechanisms that monitor the quality of lane line representations and adjust the ordering and selection of keypoint detections accordingly. This feedback loop ensures smooth, stable lane lines by correcting erratic patterns detected during the representation process.
Data Source
AI summary
Systems, methods, and other embodiments described herein relate to improving the estimation of lane lines along a roadway. In one embodiment, a method includes acquiring sensor data about a roadway. The sensor data includes at least frames and associated point detections about a lane line associated with the roadway. The method includes identifying a selected point and nearby points of the point detections. The method includes projecting lines from the nearby points to an orthogonal line extending from the selected point to define projection points. The method includes estimating a line point of the lane line according to the projection points. The method includes generating an estimated line for the lane line using the line point and other line points.