Self-position estimation correcting feature values by illuminant direction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing self-position estimation systems for vehicles face challenges in accurately estimating position due to changes in brightness levels between daytime and nighttime, as feature values extracted from images change significantly, leading to unsuccessful correlation with registered map features.
Innovation Solution
A self-position estimation apparatus that captures driving images and reference images along a route, detects feature points, and corrects feature values based on illuminant direction estimation, allowing for accurate correlation and position estimation by selecting similar images and adjusting feature values to match the environment map's conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If feature values are extracted directly from images without correction, then the extraction process is simple and fast, but the correlation accuracy deteriorates due to brightness changes between daytime and nighttime
Solution Approach 1:
The patent applies parameter changes by correcting feature values based on illuminant direction. Specifically, the system estimates the illuminant direction from the image and corrects the feature values (such as ORB features) according to this direction, thereby compensating for brightness changes caused by different lighting conditions. This allows accurate correlation between daytime and nighttime images despite significant brightness variations.
2Measurement precision
If feature values are corrected based on illuminant direction, then correlation accuracy improves, but the processing time and computational complexity increase
Solution Approach 1:
The patent applies preliminary action by pre-calculating and storing correction information (illuminant direction and correction amounts) for different regions of the image. When processing images, the system retrieves these pre-computed values and applies the corresponding corrections, rather than performing complex calculations in real-time. This significantly reduces processing time while maintaining high correlation accuracy.
3Measurement precision
If the system accounts for illuminant direction in feature value correction, then position estimation accuracy improves, but the system complexity increases
Solution Approach 1:
The patent introduces an intermediary approach by using a predetermined correction amount stored in advance for different image regions. Instead of directly computing complex illuminant direction effects, the system uses these pre-stored correction values as intermediaries to adjust feature values. This simplifies the overall system complexity while maintaining the ability to compensate for lighting variations and improve position estimation accuracy.
Data Source
AI summary
A self-position estimation apparatus includes an image capturing unit capturing driving images and reference images at a plurality of positions along a predetermined driving route, a detection unit detecting feature points on each of the driving images and feature points on each of the reference images, a storage unit storing map information which includes the feature points on each of the reference images and a position and a posture of the image capturing unit at a time each of the reference images is captured by the image capturing unit, and an estimation unit selecting a similar image similar to one of the driving images from the reference images to correlate the feature points on the one of the driving images and feature points on the similar image, the estimation unit estimating a position and a posture of an own vehicle a predetermined driving route based on a correlation result.