Ego-Motion Estimation Using Affine Projected Optical Flow
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current driver assistance systems face challenges in accurately estimating vehicle ego-motion, particularly in complex environments, due to issues with optical flow estimation and outlier removal, which affects the precision of vehicle positioning and motion tracking.
Innovation Solution
The system employs a multi-camera setup with an affine projection to generate a surround view, using the Horn-Schunck method for optical flow estimation and RANSAC for outlier filtering, combined with an Ackermann steering model to integrate multiple motion sources, including GPS and vehicle sensors, for precise ego-motion calculation and kinematic state determination.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If optical flow estimation is used to determine vehicle ego-motion, then vehicle positioning information can be obtained, but measurement errors and outliers affect the precision of motion estimation
Solution Approach 1:
The patent extracts and removes outlier measurements from the optical flow data using statistical methods. By identifying and eliminating erroneous optical flow vectors that deviate significantly from the majority of measurements, the system improves the reliability of ego-motion estimation without sacrificing the precision gained from optical flow analysis.
Solution Approach 2:
The patent implements a feedback mechanism where the estimated ego-motion is continuously refined by comparing multiple optical flow measurements and adjusting the estimation based on the consistency of results. This iterative process allows the system to correct for measurement errors and improve both precision and reliability over time.
2Measurement precision
If multiple motion sources (GPS, vehicle sensors, optical flow) are combined for ego-motion estimation, then positioning accuracy is improved, but system complexity increases
Solution Approach 1:
The patent merges multiple motion sources including GPS data, vehicle sensor readings, and optical flow measurements into a unified ego-motion estimation. By combining these diverse data sources through a consistent mathematical framework, the system achieves improved positioning accuracy while managing complexity through integrated processing.
Solution Approach 2:
The patent creates a universal estimation framework that can process multiple types of motion data (GPS, sensors, optical flow) through a single coherent algorithmic approach. This multi-functional system handles different data sources uniformly, improving accuracy without proportionally increasing system complexity.
3Loss of information
If feature extraction and correlation methods are used to construct optical flow field, then vehicle motion information can be derived, but computation time increases
Solution Approach 1:
The patent applies partial action by focusing computational resources on extracting and correlating only the most relevant features for motion estimation. Rather than processing all image features equally, the system identifies and processes key features that provide sufficient motion information, reducing computation time while maintaining information completeness.
Data Source
AI summary
A method and device for determining an ego-motion of a vehicle are disclosed. Respective sequences of consecutive images are obtained from a front view camera, a left side view camera, a right side view camera and a rear view camera and merged. A virtual projection of the images to a ground plane is provided using an affine projection. An optical flow is determined from the sequence of projected images, an ego-motion of the vehicle is determined from the optical flow and the ego-motion is used to predict a kinematic state of the car.