Vehicle Behavior Estimation Using SFM Cycle Adjustment for Repetitive Patterns
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The estimation of ego-motion in optical flow becomes unstable in environments with repetitive patterns, leading to erroneous vehicle behavior estimation, such as when a vehicle travels alongside soundproof walls with the same shape, causing aliasing issues.
Innovation Solution
The system adjusts the Structure From Motion (SFM) cycle by ensuring the vehicle movement is less than half the repetition interval of the pattern, and utilizes specific location information to correct ego-motion estimation, incorporating features like scale factor correction and prediction of feature point positions to prevent erroneous estimations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the vehicle travels alongside soundproof walls with repetitive patterns, then the camera captures continuous images for SFM processing, but the estimation of ego-motion becomes unstable due to aliasing
Solution Approach 1:
The system preemptively detects repetitive patterns in the captured images and identifies potential aliasing conditions before they cause erroneous ego-motion estimation. By detecting the repetition interval of patterns and comparing it with the SFM cycle, the system prevents aliasing errors from occurring in the first place, rather than correcting them after they occur.
Solution Approach 2:
The system continuously monitors the SFM processing results and compares the estimated ego-motion with expected vehicle behavior from other sensors (odometry, GPS). When discrepancies indicate aliasing errors, the system feeds back this information to adjust or discard the problematic SFM results, ensuring reliable ego-motion estimation despite repetitive patterns in the environment.
2Speed
If the SFM cycle is shortened to improve estimation frequency, then real-time performance improves, but aliasing occurs when vehicle movement exceeds half the pattern repetition interval
Solution Approach 1:
The system dynamically adjusts the SFM processing parameters based on detected pattern repetition intervals. When repetitive patterns are detected, the system modifies the SFM cycle or selection criteria to ensure vehicle movement between frames remains less than half the pattern repetition interval, thereby preventing aliasing while maintaining optimal estimation frequency for the given environmental conditions.
Solution Approach 2:
The system changes the SFM processing parameters (such as frame selection interval or motion threshold) based on the detected repetition interval of environmental patterns. By adapting these parameters to the specific environment, the system maintains high estimation frequency where possible while preventing aliasing in environments with repetitive structures.
Data Source
AI summary
In an apparatus for estimating a behavior of a vehicle carrying the apparatus based on images of surroundings of the vehicle captured by an imaging device, an information acquirer acquires beforehand specific location information that is information representing a specific location in which a situation around the vehicle is such that the estimation of the own vehicle behavior based on the images is unstable. In the apparatus, a behavior estimator estimates the own vehicle behavior based on the images captured by the imaging device and the specific location information.