Stereo Vision for Low-Texture Objects Using Color Texture

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for three-dimensional detection of objects with low texture or color are inefficient, particularly for moving objects, as they require complex constructions to maintain structured illumination, which complicates the process and limits throughput.

Innovation Solution

The use of at least two lighting devices emitting different colors, aligned from various directions to create a strong color texture on the object's surface, allowing for clear assignment of correspondence points and three-dimensional reconstruction without the need for structured illumination, using line scan cameras aligned transversely to the scanning direction and arranged in epipolar geometry.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If structured illumination is used to create color texture on the object surface, then correspondence point assignment is improved, but device complexity increases due to the need for moving structured illumination systems

Engineering Contradiction:
Improvecorrespondence point assignmentVSAvoidstructured illumination system
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent extracts the structured illumination requirement by using the object's own surface properties (natural texture, color variations, geometric features) instead of projecting external structured patterns. This eliminates the need for complex moving projection systems while maintaining the ability to identify correspondence points through feature matching algorithms.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The object serves itself by providing natural surface features that can be used for stereo matching. The surface texture, color variations, and geometric characteristics of the object itself become the reference patterns for correspondence point assignment, eliminating the need for external structured illumination.

Inventive Principle:
Principle #25Self-service

2Measurement precision

If structured illumination is used for three-dimensional detection, then measurement precision is improved, but productivity decreases due to complex device construction and movement requirements

Engineering Contradiction:
Improvethree-dimensional detection accuracyVSAvoiddetection throughput
Core Design Contradiction:
Measurement precisionVSProductivity

Solution Approach 1:

The patent removes the structured illumination component entirely, using only the object's natural surface features for three-dimensional detection. This simplification enables higher detection throughput by eliminating the mechanical movement and calibration requirements of structured projection systems.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent replaces the mechanical structured illumination system with a static multi-camera setup that captures images from different viewpoints. The three-dimensional information is extracted through computational stereo vision algorithms rather than mechanical projection and scanning, significantly improving productivity.

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

3Measurement precision

If multiple cameras are used for stereo imaging, then three-dimensional reconstruction quality is improved, but device complexity increases

Engineering Contradiction:
Improvethree-dimensional reconstruction qualityVSAvoidcamera system configuration
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent combines multiple cameras into a unified stereo imaging system with fixed geometric relationships. The cameras are positioned and calibrated together as an integrated unit, simplifying the overall device structure while maintaining high three-dimensional reconstruction quality through coordinated multi-view capture.

Inventive Principle:
Principle #5Merging (Combining)

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

Enables real-time, high-throughput three-dimensional detection of objects with low texture or color, particularly suitable for metallic surfaces and objects with locally strong curvature, by generating a robust color texture that aids in correspondence point assignment and stereo reconstruction, simplifying the device construction by eliminating the need for moving structured illumination.

Implementation Method 1

at least two images of the object are created by means of at least one color camera from different recording directions

Methodology Applied
Scientific EffectLight reflection: Reflection

Implementation Method 2

at least two lighting devices, which are directed towards a recording area from different directions and emit light in different colors

Methodology Applied
Scientific EffectLight emission: Light

Data Source

PatentEP3309506B1Device and method for the spatial acquisition of the surface of an object
Publication Date: 2021.12.29 ROBERT BOSCH GMBH
  • EP3309506B1 patent drawingFigure 1~2
  • EP3309506B1 patent drawingFigure 3

AI summary

The invention relates to a device for spatially capturing the surface of an object using stereo images, comprising at least one color camera aligned with a recording area for positioning the object, wherein the at least one color camera enables the capture of at least two images of the object at a triangulation angle. The invention further relates to a method for spatially capturing the surface of an object using stereo images. The device and the method enable the assignment of correspondence points even for objects with low inherent texture.