Homography-Based Camera Shake Correction for Roadside 3D Obstacle Positioning

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

In Vehicle-to-Everything (V2X) roadside sensing scenarios, camera shake due to wind or heavy vehicles introduces shake errors in determining the 3D position of obstacles, leading to inaccurate results, and existing stabilization methods like optical and mechanical image stabilization are costly and inaccurate for large-scale implementation.

Innovation Solution

The method determines 3D coordinates of obstacles by establishing a homography relationship between the ground surface in the processed image and a template image, correcting for camera shake without the need for optical or mechanical stabilization, using a 3*3 homography matrix and image registration algorithms to transform pixel coordinates, thereby eliminating shake effects.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Area of stationary object

If the roadside camera is installed at a higher position to increase field of view, then more road traffic information can be acquired, but camera shake occurs due to wind or heavy vehicles causing inaccurate 3D position determination

Engineering Contradiction:
Improvefield of viewVSAvoid3D position determination accuracy
Core Design Contradiction:
Area of stationary objectVSMeasurement precision

Solution Approach 1:

The patent uses a template image (a stable reference image captured when the camera is not shaking) to create a virtual copy of the ground surface. By establishing homography between the current to-be-processed image and this template image, the system can correct camera shake effects without physically stabilizing the camera at the high installation position

Inventive Principle:
Principle #26Copying

Solution Approach 2:

The patent replaces mechanical image stabilization systems with a computational approach using homography transformation. Instead of using mechanical gimbals or stabilization mounts to physically counteract camera shake, the system uses image processing algorithms to correct the visual effects of shaking through coordinate transformation based on the homography matrix

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

2Measurement precision

If optical or mechanical image stabilization methods are used to correct camera shake, then 3D position accuracy can be improved, but implementation costs increase significantly

Engineering Contradiction:
Improve3D position determination accuracyVSAvoidimplementation cost
Core Design Contradiction:
Measurement precisionVSEase of manufacture

Solution Approach 1:

The patent replaces expensive optical or mechanical stabilization hardware with a software-based homography transformation method. The solution uses standard image processing algorithms and coordinate transformations that can be implemented on existing roadside computing devices without additional stabilization hardware, significantly reducing implementation costs while maintaining accuracy

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The patent uses a template image (a digital reference) instead of expensive physical stabilization equipment. This digital reference can be captured once and reused multiple times for correcting subsequent images, providing a low-cost, reusable solution that eliminates the need for continuous mechanical stabilization

Inventive Principle:
Principle #27Cheap short-living objects (Disposable)

3Ease of manufacture

If electronic image stabilization is used to reduce costs, then implementation cost decreases, but stabilization accuracy becomes insufficient leading to large errors

Engineering Contradiction:
Improveimplementation costVSAvoidstabilization accuracy
Core Design Contradiction:
Ease of manufactureVSMeasurement precision

Solution Approach 1:

The patent creates a precise digital copy of the ground surface through homography transformation by establishing geometric correspondence between the template image and to-be-processed image. This mathematical copying approach preserves spatial relationships and provides accurate correction without the errors associated with simpler electronic stabilization methods

Inventive Principle:
Principle #26Copying

Solution Approach 2:

The patent replaces inaccurate electronic image stabilization with a mathematically rigorous homography transformation approach. By using perspective transformation based on ground plane geometry and coordinate mapping, the system achieves high accuracy in correcting camera shake effects without relying on less accurate electronic stabilization algorithms

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentEP3968266B1Obstacle three-dimensional position acquisition method and apparatus for roadside computing device
Publication Date: 2023.11.15 APOLLO INTELLIGENT CONNECTIVITY (BEIJING) TECH CO LTD
  • EP3968266B1 patent drawingFigure 1
  • EP3968266B1 patent drawingFigure 2
  • EP3968266B1 patent drawingFigure 3

AI summary

The present application discloses a method and an apparatus of obstacle three-dimensional position acquisition for a roadside computing device, and relates to the fields of intelligent transportation, cooperative vehicle infrastructure, and autonomous driving. The method may include: acquiring pixel coordinates of an obstacle in a to-be-processed image; determining coordinates of a bottom surface center point of the obstacle according to the pixel coordinates; acquiring a homography relationship between a ground surface corresponding to the to-be-processed image and a ground surface corresponding to a template image; transforming the coordinates of the bottom surface center point of the obstacle into coordinates on the template image according to the homography relationship; and determining three-dimensional coordinates of the bottom surface center point of the obstacle according to the coordinates obtained by transformation and a ground equation corresponding to the template image. By use of the solution of the present application, implementation costs can be saved, and the accuracy is better.