Vanishing Point Correction for ADAS Lane Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing ADAS systems, such as FCWS and LDWS, face errors in lane recognition due to variations in the vanishing point caused by vehicle pitch motion or road bends, leading to false warnings or missed collision alerts.
Innovation Solution
A vanishing point correction method that extracts contour points of a forward vehicle, tracks their movement, calculates the variation in width between contour points, and uses this information to correct the vanishing point position, preventing false warnings by accurately determining the vanishing point variation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If the vanishing point position is fixed based on initial camera calibration, then the system complexity is low, but measurement precision of lane recognition and object distance deteriorates when vehicle pitch motion occurs
Solution Approach 1:
The patent implements dynamic vanishing point correction by continuously tracking the vanishing point position based on detected vehicle pitch angle and road gradient, transforming the static calibration approach into a dynamic adaptation system that maintains measurement precision under varying vehicle conditions
Solution Approach 2:
The system changes the vanishing point position parameters in real-time according to detected pitch motion and road gradient variations, allowing the measurement system to adapt to changing conditions without increasing overall system complexity
2Ease of operation
If the vanishing point variation is not detected, then the system operation is simple, but false warnings are generated in ADAS systems
Solution Approach 1:
The patent introduces feedback mechanisms where the corrected vanishing point information is continuously fed back to the ADAS system, enabling reliable collision and lane departure warnings while maintaining simple operation through automated correction processes
Solution Approach 2:
The vanishing point correction module acts as an intermediary between the image processing system and ADAS warning systems, filtering out false alarms caused by pitch motion before warnings are generated, thus improving reliability without complicating user operation
3Measurement precision
If contour points are extracted and tracked to correct the vanishing point, then measurement precision of vanishing point position is improved, but device complexity increases
Solution Approach 1:
The patent segments the image processing into distinct modules: contour point extraction, vanishing point calculation, pitch motion detection, and correction application. This modular segmentation improves measurement precision while managing complexity through organized processing stages
Data Source
AI summary
A vanishing point correction method may include the steps of: (a) extracting a contour candidate point of an object considered as the forward vehicle from the image taken by the camera; (b) tracking a movement of the contour candidate point over time; (c) determining whether the contour candidate point belongs to the vehicle; (d) setting the contour candidate point to a vehicle contour point of the forward vehicle, when it is determined at the step (c) that the contour candidate point belongs to the vehicle; (e) calculating a variation in width between vehicle contour points at the same height among the vehicle contour points; and (f) calculating a variation of the vanishing point.