Vehicle Vision Free-Space Detection Using Horizontal Optical Flow
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing mono camera systems for vehicles face challenges in reliably detecting drivable space in front of a car, as they either require stationary objects for 3D reconstruction or rely on unreliable comparisons with theoretical flat surfaces for optical flow analysis.
Innovation Solution
The system determines transitions between regions of constant and non-constant horizontal optical flow, using the fact that horizontal optical flow is constant along a column on the road regardless of object movement, without needing a prior flat surface, and calculates drivable space by analyzing these transitions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If structure-from-motion is used for 3D reconstruction, then depth information can be obtained, but the method requires target objects to be stationary which limits applicability
Solution Approach 1:
The patent changes the parameter being analyzed from spatial structure (3D reconstruction) to temporal motion patterns (optical flow characteristics). By analyzing how horizontal optical flow varies vertically in the image plane, the system can detect drivable space without requiring objects to be stationary, thus resolving the contradiction between measurement precision and adaptability.
2Ease of operation
If optical flow comparison with theoretical flat surface is used, then drivable space detection can be performed, but the method is not always reliable
Solution Approach 1:
Instead of comparing actual optical flow with a theoretical flat surface model, the patent inverts the approach by identifying regions where horizontal optical flow is constant (indicating raised structures) versus regions where it varies (indicating drivable road surface). This inversion eliminates the need for theoretical models and improves reliability.
3Device complexity
If prior flat surface assumption is made, then distance determination can be simplified, but accuracy decreases when road is not flat
Solution Approach 1:
The patent introduces a dynamic approach by allowing the system to adapt to different road conditions. While the basic method works with simple flat road assumptions, the system can incorporate road profile information from satellite navigation when needed, dynamically adjusting the complexity based on environmental conditions to maintain both simplicity and accuracy.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach provides a simple and reliable method to determine drivable space in front of a vehicle, allowing for accurate distance estimation to structures and free space detection without relying on stationary objects or flat surface assumptions, and can use road profile information for enhanced accuracy.
Implementation Method 1
an imaging apparatus (11) for capturing images of a region surrounding a motor vehicle
Implementation Method 2
transitions between regions of essentially constant horizontal optical flow and regions of essentially non-constant horizontal optical flow are determined
Data Source
Figure 1
Figure 2
AI summary
A vision system for detecting free space in front of a motor vehicle comprises a mono imaging apparatus (11) adapted to capture images (30) from the surrounding of a motor vehicle, and an electronic processing device (14) adapted to perform image processing of images (30) captured by said mono imaging apparatus (11) in order to detect objects in the surrounding of a motor vehicle. The electronic processing device (14) is adapted to calculate a horizontal component of the optical flow (31), and to determine transitions (33, 35) between regions of essentially constant horizontal optical flow and regions of essentially non-constant horizontal optical flow.