Stereo Camera Egomotion Estimation Using Disparity Maps
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional egomotion estimation methods for moving cameras, such as those on vehicles, are cumbersome, prone to errors due to incorrect feature point extraction and motion vector calculations, especially when landmarks are unclear, and require significant computational resources.
Innovation Solution
An egomotion estimation system employing a stereo camera to generate depth and disparity maps, extract feature points, detect motion vectors, remove errors using a forward-backward algorithm, and determine egomotion with the RANSAC method, optimizing calculations and reducing errors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional egomotion estimation methods are used with single camera and feature point tracking, then egomotion can be estimated, but the system becomes cumbersome and requires large calculation capacity
Solution Approach 1:
The patent introduces disparity maps and depth maps as intermediary data structures between the stereo image input and egomotion estimation output. These intermediate representations encode spatial relationships and enable more efficient motion vector calculation compared to direct feature point tracking, thereby reducing computational complexity while maintaining estimation accuracy
Solution Approach 2:
The patent transitions from 2D single-camera feature tracking to 3D stereo vision by incorporating depth information through disparity maps. This dimensional enhancement allows the system to estimate egomotion more accurately by utilizing spatial relationships in three dimensions, resolving the contradiction between precision and complexity
2Measurement precision
If feature points are extracted from road surface and optical flow is applied, then motion vectors can be estimated, but incorrect feature points increase error likelihood
Solution Approach 1:
The patent applies preliminary filtering and validation steps before final motion vector estimation. By pre-processing feature points using disparity map information and applying consistency checks before optical flow calculation, the system eliminates incorrect feature points early in the pipeline, reducing error propagation and improving reliability
Solution Approach 2:
The patent implements feedback mechanisms where motion vectors are validated against disparity map constraints and depth information. Incorrect motion estimates are detected and corrected by comparing against expected spatial relationships, creating a feedback loop that continuously refines accuracy and reduces error rates
3Ease of manufacture
If landmarks are used for egomotion estimation, then calculations are simplified, but it becomes hard to acquire clearly defined landmarks on actual road surface
Solution Approach 1:
The patent enables the system to automatically generate its own feature points and disparity maps from raw stereo images without requiring external landmarks or pre-defined reference objects. The algorithm extracts necessary information directly from the scene, making the system self-sufficient and adaptable to any environment regardless of landmark availability
Solution Approach 2:
The patent transforms the approach from landmark-based discrete feature matching to continuous disparity map analysis. By changing the fundamental parameter representation from landmark coordinates to pixel-level disparity values, the system achieves both calculation simplicity and reliability without depending on clearly defined landmarks
Data Source
AI summary
An egomotion estimation system may include: a stereo camera suitable for acquiring a stereo image; a map generation unit suitable for generating a depth map and a disparity map using the stereo image; a feature point extraction unit suitable for extracting a feature point from a moving object in the stereo image using the disparity map; a motion vector detection unit suitable for detecting a motion vector of the extracted feature vector point; an error removing unit suitable for removing an error of the detected motion vector; and an egomotion determination unit suitable for calculating and determine an egomotion using the error-removed motion vector.