Vehicle Image Processing Reflection Detection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing object recognition devices face challenges in accurately detecting reflections in images captured by cameras mounted inside vehicles, leading to decreased recognition accuracy and inappropriate removal of features like lane lines, which affects warning systems.

Innovation Solution

An image processing device that distinguishes reflections from other features by analyzing the positional stability of features across images, using an obtaining unit to gather data and a detection unit to identify reflections based on movement direction, thereby accurately removing reflections from images.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If the object recognition device removes fixed feature points whose coordinates remain the same in sequentially captured images, then the device can eliminate some reflection artifacts, but the device cannot accurately distinguish reflections from stationary objects like lane lines, leading to false removal of essential features

Engineering Contradiction:
Improvereflection detection accuracyVSAvoidobject recognition reliability
Core Design Contradiction:
Measurement precisionVSReliability

Solution Approach 1:

The patent applies the dynamics principle by analyzing the movement characteristics of features across multiple sequentially captured images. The detection unit identifies that reflection artifacts maintain fixed coordinates while the vehicle moves forward, whereas stationary objects like lane lines exhibit predictable movement patterns consistent with vehicle motion. By dynamically evaluating coordinate changes relative to vehicle movement, the system distinguishes reflections from legitimate stationary features, resolving the contradiction between removing reflections and preserving essential features.

Inventive Principle:
Principle #15Dynamics

2Ease of operation

If the camera is installed inside the vehicle to capture images through the windshield, then the camera can provide a driver's perspective view, but reflections from interior objects degrade the quality of captured images

Engineering Contradiction:
Improvecamera installation convenienceVSAvoidimage quality
Core Design Contradiction:
Ease of operationVSMeasurement precision

Solution Approach 1:

The patent applies the extraction principle by isolating and removing reflection artifacts from the captured images while preserving the underlying scene information. The detection unit identifies reflection regions based on fixed coordinate characteristics, and the image processing unit extracts and removes these reflections, leaving the legitimate scene content intact. This allows the camera to remain installed inside the vehicle for convenient driver's perspective capture while eliminating the degrading reflection effects.

Inventive Principle:
Principle #2Taking out (Extraction)

3Measurement precision

If the object recognition device processes all feature points in sequentially captured images, then the device can maintain comprehensive object detection, but the processing time and computational load increase

Engineering Contradiction:
Improveobject detection accuracyVSAvoidimage processing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent applies the extraction principle by identifying and extracting only the fixed coordinate features that correspond to reflections, rather than processing all feature points. The detection unit selectively identifies reflection artifacts based on their invariant coordinate characteristics across sequential images, and the image processing unit removes only these specific features. This selective approach maintains comprehensive object detection accuracy while significantly reducing computational load and processing time compared to processing all feature points.

Inventive Principle:
Principle #2Taking out (Extraction)

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 enables precise detection and removal of reflections, enhancing the accuracy of object recognition and preventing false removal of essential features like lane lines, thus improving the reliability of warning systems.

Implementation Method 1

Reflections are phenomena in which light from objects inside the vehicle is reflected by a surface of the sheet of glass or the like and detected by a sensor of the camera installed inside the vehicle

Methodology Applied
Scientific EffectReflection: Reflection

Data Source

PatentEP2735996B1Image processing device and method for processing image
Publication Date: 2018.02.21 FUJITSU LTD
  • EP2735996B1 patent drawingFigure 1A~1B
  • EP2735996B1 patent drawingFigure 2
  • EP2735996B1 patent drawingFigure 3

AI summary

An image processing device includes: an obtaining unit configured to obtain image data regarding an image captured by an image capture device that moves along with a moving body and of which image capture direction is a certain traveling direction of the moving body; and a detection unit configured to detect a feature caused by a reflection from features based on information regarding positions of the features in the image data at a time when the moving body has moved in a direction different from the image capture direction.