Road Boundary Estimation Using Probability-Guided Curve Refinement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies for estimating road markings on vehicle travel paths suffer from errors due to restricted curve fitting and image blurring, leading to instability in vehicle control applications.
Innovation Solution
An image processing device and method that uses a learned model to estimate road boundaries by iteratively updating parameters of a function approximating the road boundary based on probability values and coordinates, rearranging them in descending order, and executing driving control or assistance based on the updated parameters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If road markings are estimated by fitting a curved line to positions extracted from an image, then the road boundary can be approximated, but the estimation error becomes large due to restricted degree of freedom of the curve
Solution Approach 1:
The patent segments the image into multiple scanning lines (horizontal and vertical) and processes coordinates along each line separately. By dividing the road boundary estimation into multiple one-dimensional coordinate sequences along different scanning directions, the system achieves more accurate fitting without requiring complex two-dimensional curve fitting, thus resolving the contradiction between estimation accuracy and computational complexity.
2Speed
If positions are extracted from an image at each time point, then real-time road marking estimation is possible, but blurring occurs in the estimated road markings due to position variation
Solution Approach 1:
The patent employs iterative refinement where the estimated road boundary from previous time points is used as feedback to guide position extraction at current time points. The system continuously adjusts the position extraction based on the evolving boundary model, reducing accumulation of position variations and preventing blurring effects while maintaining real-time processing capability.
Solution Approach 2:
The system performs preliminary road boundary estimation using initial position extractions, then uses this preliminary model to guide subsequent more precise position extractions. By establishing an initial boundary framework before detailed position extraction, the system reduces the impact of position variations and prevents blurring in the final road marking estimation.
3Reliability
If a learned model outputs probability values for each coordinate, then road boundary existence probability can be determined, but computational load increases
Solution Approach 1:
The patent extracts only the necessary probability information from the learned model output - specifically focusing on coordinates with high probability values that indicate road boundary presence. By extracting and processing only the relevant high-probability coordinates rather than all coordinates, the system maintains high detection reliability while significantly reducing computational energy consumption.
Data Source
AI summary
An image processing device acquires a probability value and corresponding coordinates by using a learned model that outputs a probability value indicating an existence probability of a road boundary for each of coordinates in an image in response to an input of the image representing an area in front of a mobile object, captured by a camera mounted on the mobile object, rearranges a plurality of coordinates in descending order of the probability value, iteratively updates parameters of a function that approximates the road boundary in the image using the plurality of coordinates in the rearranged order, and executes driving control or driving assistance of the mobile object on the basis of the road boundary approximated by the function defined by the updated parameters.