Monocular Camera Optical Flow Discontinuity Detection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current driver assistance systems using monocular image sensors struggle to accurately detect three-dimensional structures and differentiate between moving objects in real-time, especially in complex and changing vehicle environments, due to limitations in extracting spatial information from two-dimensional images.

Innovation Solution

The method employs optical flow analysis to segment images based on discontinuities, enabling the detection of static and moving obstacles, and distinguishing between them, which supports lane guidance and hazard warnings, even in poor visibility conditions, by using monocular image sequences and accounting for terrain characteristics.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If monocular image sensors are used to reduce cost, then device complexity is reduced, but the ability to detect three-dimensional structures and differentiate moving objects deteriorates

Engineering Contradiction:
Improvesensor system complexityVSAvoidspatial information extraction accuracy
Core Design Contradiction:
Device complexityVSMeasurement precision

Solution Approach 1:

The patent segments the image sequence into different motion components by analyzing optical flow discontinuities. By dividing the image into regions with different motion characteristics (static background vs. moving objects), the system can extract spatial information without requiring complex stereo sensors. This segmentation approach enables 3D structure detection from monocular images by identifying boundaries where motion patterns change.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent transitions from analyzing single static images to analyzing temporal sequences of images. By adding the time dimension and examining how pixel intensities change across multiple frames, the system extracts optical flow information that reveals three-dimensional structure and motion. This temporal dimension compensates for the lack of depth information in monocular imaging.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Measurement precision

If model analysis is used to detect three-dimensional objects from monocular images, then measurement precision improves, but device complexity and computational requirements increase

Engineering Contradiction:
Improvethree-dimensional object detection accuracyVSAvoidcomputational processing complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent uses the motion information inherently present in the image sequence itself to solve the 3D detection problem. Instead of requiring external models or assumptions about scene geometry, the system extracts spatial information directly from the optical flow patterns generated by object motion. The image data serves its own purpose of revealing three-dimensional structure through temporal analysis.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The patent changes the analysis parameters from static image features to dynamic optical flow parameters. By examining how pixel positions and intensities change over time, the system derives three-dimensional information without requiring complex geometric models. This parameter transformation from spatial to spatio-temporal domain simplifies the computational approach.

Inventive Principle:
Principle #35Parameter changes

3Measurement precision

If optical flow analysis is performed on image sequences to extract spatial information, then measurement precision improves, but real-time processing capability deteriorates

Engineering Contradiction:
Improvespatial information accuracyVSAvoidreal-time processing speed
Core Design Contradiction:
Measurement precisionVSSpeed

Solution Approach 1:

The patent extracts only the essential optical flow information needed for 3D detection rather than performing complete image sequence analysis. By focusing specifically on detecting motion boundaries and optical flow discontinuities, the system obtains sufficient spatial information with reduced computational overhead, enabling real-time processing while maintaining accuracy.

Inventive Principle:
Principle #2Taking out (Extraction)

4Speed

If simplifying assumptions are made in optical flow analysis to improve real-time capability, then processing speed improves, but robustness in complex vehicle environments deteriorates

Engineering Contradiction:
Improvereal-time processing speedVSAvoidrobustness in vehicle environment
Core Design Contradiction:
SpeedVSReliability

Solution Approach 1:

The patent uses dynamic analysis of optical flow patterns to adapt to different vehicle environment conditions. Instead of relying on static simplifying assumptions, the system continuously analyzes motion patterns and adjusts its detection criteria based on the observed dynamics of the scene, maintaining robustness while achieving real-time performance.

Inventive Principle:
Principle #15Dynamics

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 enhances the accuracy and robustness of driver assistance systems, providing real-time information for safe navigation and hazard detection, improving lane recognition and terrain awareness, and supporting other sensors like yaw rate sensors, thus enhancing safety features such as Lane Departure Warning and pre-crash functions.

Implementation Method 1

at least one image sensor (12) is provided in order to record the vehicle surroundings

Methodology Applied
Scientific EffectLight reflection: Reflection

Implementation Method 2

the identifiable shifts in consecutive images provide information about the three-dimensional arrangement of the objects in the vehicle environment

Methodology Applied
Scientific EffectOptical flow:

Data Source

PatentEP2033165B1Method for picking up a traffic space
Publication Date: 2018.07.11 ROBERT BOSCH GMBH
  • EP2033165B1 patent drawingFigure 1~2
  • EP2033165B1 patent drawingFigure 3
  • EP2033165B1 patent drawingFigure 4~5

AI summary

The invention relates to a method for picking up a traffic space using a driver assistance system (1) comprising a monocular image sensor (12). The image sensor (12) produces chronologically successive images of the traffic space. The images in the image sequence are used to ascertain the visual flow and to examine it for discontinuities. Discontinuities found in the visual flow are attributed to objects in the traffic space.