Spatially Controlled Scene Illumination for Machine Vision

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional lighting schemes for machine vision applications often produce shadows and glints, leading to inadequate illumination, which affects the accuracy of edge detection and object recognition, especially when illumination sources are confined and close to the objects being illuminated.

Innovation Solution

The system employs multiple electromagnetic radiation sources positioned at distinct angles relative to the camera's viewing angle to provide spatially controlled brightness, using feedback signals from the camera to adjust illumination levels and reduce shadows and glints, ensuring uniform or varying brightness across the scene.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Illumination intensity

If conventional lighting schemes are used with limited illumination sources positioned close to the scene, then device complexity is reduced, but illumination quality deteriorates producing shadows and glints

Engineering Contradiction:
Improveillumination qualityVSAvoidnumber of illumination sources
Core Design Contradiction:
Illumination intensityVSDevice complexity

Solution Approach 1:

The illumination system is divided into multiple independent light sources (first illumination source, second illumination source, third illumination source) positioned at different locations. Each source illuminates different regions of the scene, allowing the system to eliminate shadows and glints by combining multiple illumination perspectives rather than relying on a single light source

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Different regions of the scene are illuminated with different characteristics by different light sources. The system selectively activates specific illumination sources based on which regions need illumination, creating locally optimized illumination quality across the entire scene while managing overall system complexity

Inventive Principle:
Principle #3Local quality

2Use of energy by stationary object

If illumination sources are positioned close to the scene to reduce power consumption, then energy efficiency improves, but illumination uniformity deteriorates due to shadows and glints

Engineering Contradiction:
Improvepower consumptionVSAvoidillumination uniformity
Core Design Contradiction:
Use of energy by stationary objectVSManufacturing precision

Solution Approach 1:

The illumination sources are positioned asymmetrically at different locations relative to the scene rather than symmetrically or in a single position. This asymmetric arrangement allows each light source to illuminate different aspects of the scene, reducing the formation of shadows and glints while maintaining proximity to the scene for energy efficiency

Inventive Principle:
Principle #4Asymmetry

Solution Approach 2:

The system transitions from considering only the vertical dimension (distance from scene) to incorporating horizontal spatial distribution of multiple light sources. By arranging sources at different lateral positions in addition to vertical positioning, the system achieves better illumination uniformity without increasing power consumption

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

3Reliability

If multiple illumination sources are used to eliminate shadows and glints, then illumination quality improves, but device complexity increases

Engineering Contradiction:
Improveedge detection accuracyVSAvoidsystem configuration
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The system captures an initial image to identify regions with shadows or glints, then uses this feedback information to selectively activate specific illumination sources that will eliminate the problematic regions. This feedback loop allows the system to achieve high illumination quality and edge detection accuracy while activating only the necessary subset of light sources, thereby managing system complexity

Inventive Principle:
Principle #23Feedback

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 reliability and accuracy of machine vision applications by minimizing shadows and glints, allowing for precise edge detection and object recognition, even in complex scenes with curved surfaces, and reduces the impact of shadows from three-dimensional structures.

Implementation Method 1

A first region of the scene is illuminated by a first electromagnetic radiation source at a first angle distinct from the viewing angle of the camera. A second region of the scene is illuminated by a second electromagnetic radiation source at a second angle distinct from the viewing angle of the camera.

Methodology Applied
Scientific EffectLight: Light

Data Source

PatentEP2543187B1Systems and methods for spatially controlled scene illumination
Publication Date: 2022.11.16 GOOGLE LLC
  • EP2543187B1 patent drawingFigure 1
  • EP2543187B1 patent drawingFigure 2
  • EP2543187B1 patent drawingFigure 3

AI summary

A scene illumination system is provided that produces spatially uniform or controlled brightness levels for machine vision applications. The system includes a camera, multiple light sources that preferentially illuminate different regions within the camera's field-of-view, and a processing unit coupled to the camera and light sources. Focal regions of the light sources within the camera's field-of-view are sampled to determine average regional brightness and compared to target brightness levels. The processing unit controls the light sources to increase or decrease illumination levels to converge toward the target brightness levels within the field-of-view. This modulation of the light sources may be repeated with successive video images until target brightness levels arc achieved. Once achieved, the iterative feedback control may be locked-in for some applications, while for others, the iterative process may continue periodically or continuously to account for different scenes or changes in lighting conditions.