Basic Lighting

Techniques for
Machine Vision
Daryl Martin
Technical Sales & Product Manager
Advanced illumination
 Class Learning Objectives
1) Understand that Dedicated, Object Appropriate
Lighting is critical for virtually all MV Applications.
2) Identify 9 “Guidelines” for applying MV Lighting.
3) Name 6 Lighting Techniques and describe when each
may be appropriate.
4) Identify the 6 steps in the “Vision Lighting Design
Method.”
5) Explain “Spectral Range” as applied to vision cameras.
6) List the 4 cornerstones of Vision Illumination.
7) Describe 4 ways light can interact with your object.
8) Identify the 3 primary inspection physical
environment constraints.
 Achieving Learning Objectives
1) Knowledge of:
- Lighting types and application advantages & disadvantages
- Vision camera sensor quantum efficiency & spectral range
- Illumination Techniques and their application fields relative to
surface flatness & surface reflectivity
2) Familiarity with the 4 (Contrast Enhancement)
Cornerstones of Vision Illumination:
- Geometry (Light Direction and Nature at the Source)
- Structure (pattern)
- Color (wavelength)
- Filtering (Light Characteristics)
3) Detailed Analysis of:
- Sample – Light Interactions with respect to your unique object
- Immediate Inspection Environment – Physical constraints and
requirements (critical for Robotics Apps)
 Primary Objective of Vision Lighting

What we really require is control of the lighting environment!

Why?
- Part (Object Feature) inspection & system appropriate lighting
- To the extent possible, standardization of components,
techniques, system implementation and operation
- Reproducibility of inspection results
- Robustness to handle sample variations of “all types”
 Topics
 Brief Review of Light as Applied to Machine Vision
 Compare / Contrast Lighting Sources
 Review Light / Sample and Light / Camera Interactions
 Review Basic Lighting Geometry Techniques - Examples
- Directional Bright Field vs. Dark Field
- Back Lighting
- Preview of Diffuse Lighting Techniques
 More Applications Examples
 Preview of Filtering: Pass and Polarization
 Preview of Color Lighting Analysis
 Preview of Near IR and UV Vision Light
 Lighting for Robotics (time and interest permitting)
 Machine Vision Definitions

 Machine Vision is the computer-based characterization

of a digital image from an electronic sensor.
 A digital image is a 1-D or 2-D array of picture elements
(pixels), each having an (X,Y) location and an intensity,
typically 0 – 255 gray scales, or 8-bit contrast.
 Contrast is the visible intensity difference between
dark (near 0) and light (near 255) pixels.
 In its most derivative form, then we are characterizing
light contrast patterns from an object.
 Vision Lighting Development

Art?

Science??

Or both?

Image Courtesy NASA - HST

Images Courtesy Wikimedia Commons Public Domain

 Vision Lighting Development

 Wave and Look (most common)

– Image the part while trying different sources at different
positions

 Scientific Analysis (most effective)

– Analyze the imaging environment and short-list the best
solution possibilities

 Test Lights! (saves time)

– Test on the bench then the floor to verify your analysis
 Object Feature-Appropriate Lighting
What is it?
The light type & technique, tailored for the specific application, that
allows the vision system do its job accurately, robustly AND reproducibly.

Why is it important to Machine Vision?

1) Provides for an accurate, consistent & robust lighting
environment.
2) Saves development time, effort & resources.

How do I get there?

1) Determine object features that uniquely ID parts as “bad” or
“good” (a.k.a. – “features of interest”)
2) Design lighting that creates consistent contrast between the
object’s features of interest and background.
3) Test the lighting method on many “bad” & “good” objects.
Review of Light for Vision
Illumination
 Characterizing Light for Vision
Light: Photons propagating as a transverse electromagnetic energy
wave - characterized by:
- Measured “Intensity”: Radiometric and Photometric
- Frequency: Varies inversely with wavelength (Hz – waves/sec)
- Wavelength: Expressed in nanometers (nm) or microns (um)

Photons:
Energy packets exhibiting properties of
waves and particles.

100,000 nm
 Visible Light Spectrum

UV Near IR

400 nm 500 nm 600 nm 700 nm

Human Visible Range

Decreasing Frequency Increasing Wavelength

Decreasing Photon Energy Increasing Photometric Output
Increasing Penetration Depth*

*For some sample materials, absorption bands may block longer wavelength penetration.
 Characterizing Light for Vision
Properties when interacting with media (objects):
- Diffusion (“dispersal”) through media.
- Reflection – When not viewing a light source directly, light must interact with
objects for us to see it! (Fundamental)
- Refraction (apparent bending) through media – longer wavelengths refract
less (i.e. - red light refracts < violet light).
- Diffraction (“bending”) around object edges (not that important).

Angle of Dispersion

White Light

n = 1.5 (glass)

n = 1.0 (air)
Courtesy Wikimedia Commons
Vision Lighting Sources
 Vision Lighting Sources

LED - Light Emitting Diode

Quartz Halogen – W/ Fiber Optics
Fluorescent
Xenon (Strobing)
Primary Vision Light Sources
 Intensity vs. Wavelength
Quartz Halogen / Tungsten
Mercury (Purple)
100
Daytime Sunlight

80 Fluorescent
Relative Intensity (%)

60
Xenon White
LED
40

20 Red
LED

0
300 400 500 600 700

Wavelength (nm)
 LED Types

T1 ¾, The Standard
Courtesy Sun LED

Surface Mount LEDs

Courtesy Sun LED

High Current LEDs

Courtesy Cree and Philips
Light - Sample Interaction
 Light - Sample Interaction
 Total Light In =
Reflected + Absorbed + Transmitted + Emitted (fluorescence) Light

• Measured Irradiance/Illuminance
Incident Illumination falls off as the inv. sq. of the
distance (I = 1 / r2)
Absorb
• 2X (WD) = ¼ the “intensity”

Reflect
WD
Emit (fluorescence) Transmit
 Light Interaction
Convergence of Concepts
(Object – Light – Lens**)
Contrast
Resolution
Spatial, Spectral
Focal Length / Field of View*
Focus
Working Distance / Stand-off*
Sensitivity

*Critical Parameters for Robotics Apps

**3-D Working Volume: Strong inter-relationship
You cannot solve vision problems working in a vacuum!
 Ambient Light
Any light other than the vision-specific lighting that
the camera collects.
Controlling and Negating Ambient Light
Turn off the ambient contribution
Most effective . . . Least Likely!
Build a shroud
Very effective, but time-consuming, bulky and expensive

Overwhelm the ambient contribution w/ high-power

lighting (Continuous-on or Strobe over-drive)
Effective, but requires more cost and complexity

Control it with pass filters

Very effective, but requires a narrow-band source light
 Pass Filters in Machine Vision
• Pass filters exclude or prefer
light based on wavelength. 715 nm Long
Pass
• Reduce sunlight and mercury
vapor light 4X
• Reduce fluorescent light 35X
Graphics courtesy of Midwest Optical, Palatine, IL

510 nm Short 660 nm Band

Pass Pass
 Pass Filter Example

Top Image: UV light

w/ strong Red 660 nm
“ambient” light.

Bottom Image: Same

UV and Red 660 nm
“ambient” light, with
510 nm Short Pass filter
applied.
 Lighting Contrast

It’s All About (creating) Contrast

Contrast!!

1) Maximum contrast
• features of interest (Signal)

2) Minimum contrast
• features of no interest (Noise)
3) Minimum sensitivity to normal variations (ROBUST)
• minor part differences
• presence of, or change in ambient lighting
• sample handling / presentation differences

Points 1 & 2 might solve some apps; # 3 can be critical!

 Lighting Cornerstones
How to change contrast?

Change Light Direction w/ Respect to Sample and

Camera (Geometry)
- 3-D spatial relationship - object, light & camera

Change Light Pattern (Structure)

- Light Head Type: Spot, Line, Dome, Sheet
- Illumination Type: B.F. - D.F. - Diffuse - B.L.

Change Spectrum (Color / Wavelength)

- Projecting Monochromatic, white vs. object / camera response
- Projecting Warm vs. cool color family light – object vs. background

Change Light Character (Filtering)

- Affecting the wavelength or character of light collected by the camera
Light - Pickup Device Interaction
 Sensors and Wavelength

IR Enhanced Analog
80 Digital Interline Transfer
Standard Analog
CMOS
60 UV Enhanced Analog
Human Scotopic (Night)
Absolute QE (%)

Human Photopic
IR Block (Short Pass)
40

20

0
300 400 500 600 700 800 900 1000

Wavelength (nm)
Lighting Geometry Techniques
 Basic Lighting Techniques

1 - Partial Bright Field

2 - Dark Field

3 - Back Lighting
 Advanced Lighting Techniques

4- Co-Axial Diffuse* 5 - Diffuse Dome 6 - Flat Diffuse

Full Bright Field

Collimated
Co-axial Multi-Axis / Combo
Back Lighting Dome + Dark Field Structured
Bright and Dark Field Laser/LED grids, lines
Addressable Rows Focused Linears

* It should be noted that strictly defined, Co-axial Diffuse lighting is a partial bright field Technique.
 Bright Field vs. Dark Field
Typical On-Axis Ring Light – Sample Geometry

Bright Field Dark Field

 Bright Field vs. Dark Field

Partial Bright Field Lights Dark Field Lights in Grey

in White Area Areas

45

Mirrored Surface

Scratch
 Bright Field vs. Dark Field Light
Bright Field Dark Field
- Specular surfaces reflect glare if light is - Emphasize Height, Edges
high-angle
- Diffuse Surfaces Bright
- Diffuse, flat and smooth surfaces reflect
evenly
- Flat Polished Surfaces Dark
- Shape and Contour Enhanced

Bright Field

Dark Field
Dark Field Example

- Angled light – 45 degrees or

less
- Used on highly reflective
surfaces
- OCR or surface defect
applications
 Dark Field Light

- Emphasize Height Changes

- Diffuse Surfaces are Bright
- Flat Polished Surfaces are Dark
- Shape and Contour are Enhanced
Back Lighting
- Edge or hole detection

- Useful on translucent
materials
Liquid fill levels
Glass/plastic cracks

- Part P/A

- Vision-Guided robotics – Pick

and Place

- Gauging – Including high-

accuracy measurements
 Back Lighting

Light Diffusion & Diffraction:

- Multiple angle light from back light
diffuser
- Bending around obstacles

Θ = λ / D, where Q is the diffraction

angle and D is opening width (Θ1> Θ2)

High-accuracy gauging:
- Use monochromatic light
- Shorter wavelengths best

Θ1 Θ2 Use collimation – parallel rays

D Longer λ light may penetrate

some objects better
Red Blue
 Back Lighting Example
Small Bottle – Determine Fill Level 660 nm Red Backlight
Consider colors and materials
properties also.
Longer wavelength isn’t always
best for penetration!

880 nm IR Backlight 470 nm Blue Backlight

 Collimated Backlight Illumination

No Collimation Collimation Film

End of Part 1 = Break Time!

(Please be back in 10 mins . . .)

Diffuse Dome
- Similar to the light on an
overcast day.
- Creates minimal glare.
 Diffuse Dome
Full Bright Field
- Surface texture and detail are de-emphasized
- Contrast is de-emphasized
- Useful for curved shiny parts
- Opposite effect of Dark Field
Co-Axial Diffuse Illumination
- Light directed at beam splitter
- Used on non-curved, reflective objects
 Co-Axial Diffuse Illumination

• Surface Texture Is Emphasized

• Angled Elevation Changes Are Darkened
Flat Diffuse
- Diffuse sheet directed downward
- Long WD and larger FOV
- Hybrid diffuse (dome and Co-Axial)
Diffuse
On-Axis BFDome
Ring
Flat Diffuse
Diffuse
On-Axis Co-Axial
DF Ring
Advantages - Disadvantages
 Technique vs. Object Surface
Surface Reflectivity
Matte Mixed Mirror Specular
Flat Axial Diffuse

Geometry Bright Field

Independent
Area
Flat Diffuse
Surface
Texture / Uneven Topography

Shape
Dark Field

Diffuse Dome / Cylinder

Curved
 Inspection Environment

Physical Constraints
- Access for camera, lens & lighting in 3-D (working volume)
- The size and shape of the working volume
- Min and max camera, lighting working distance and FOV

Part Characteristics
- Object stationary, moving, or indexed?
- If moving or indexed, speeds, feeds & expected cycle time?
- Strobing? Expected pulse rate, on-time & duty cycle?
- Is the part presented consistently in orientation & position?
- Any potential for ambient light contamination?

Ergonomics and Safety

- Man-in-the-loop for operator interaction?
- Safety related to strobing or intense lighting applications?
Applications Examples
 UPC Bar Code

Printing beneath cellophane

wrapped package

Off-Axis
On-Axis Broad
Dark
Bright
Co-Axial Area
Field
Field
Diffuse Linear
Ring
RingLight
Light
Illuminator
 Stamped Date Code

Recessed metal part

Reflective, textured, flat or
curved surface

Bright
Bright field
Dark Field spot
fieldLine light
ringlight
ring light
light
 Avoiding Surface Glare
Change Geometry – 3D spatial arrangement of Light,
Object, and Camera (preferred)
Strobe to overwhelm glare from ambient sources
Use polarization filters (least preferred)

Courtesy Wikimedia Commons

 Polarizing Filters in Vision

w/o Polarizers
w/o Polarizers

w/ Polarizers

Up to 2 ½
f/stops more
w/o Polarizers open! w/ Polarizers
 Avoiding Surface Glare
3-D Reflection Geometry: Light - Sample - Camera
Using Color and Wavelength
 Create Contrast with Color
Use Monocolor Light to Create Contrast

Use Like Colors or Families to Lighten Warm Cool

(red light makes red features brighter)
R V

Use Opposite Colors or Families to O B

Darken
(red light makes green features darker)
Y G
 Create Contrast with Color

Warm Cool

R V
Red Green

O B

Y G

Blue White

Consider how color affects both your object and its background!
White light will contrast all colors, but may be a compromise.
Wavelength vs. Composition
Imaging with Near IR and UV Light
Vision Lighting Spectrum

Typical CCD
Sensor

Human Visual
System

Graphics courtesy of Midwest Optical, Palatine, IL

 Imaging with Near IR (NIR)
• Infra-red (IR) light interacts with sample material
properties, often negating color differences.

White light – B&W Camera IR light – B&W Camera

 Imaging with Near IR (NIR)

• Near IR light can penetrate materials more easily because of

the longer wavelength.

Red 660 nm Back Light IR 880 nm Back Light

Imaging with UV Light

Fluorescing Printing

Fluorescing Polymers
(nylon)

Under 355 nm UV and

Strong Ambient

Application of a Short Pass

Filter
Review and Summary
 Vision Lighting Design Method
1) Determine the Exact Features of Interest
2) Analyze Part Access / Presentation
 Clear or obstructed, Moving / Stationary
 Min / Max WD range, Sweet Spot FOV, etc.
3) Analyze Surface Characteristics
 Texture
 Reflectivity / Specularity
 Effective Contrast – Object vs. background
 Surface flat, curved, combination
4) Understand Light Types and Applications Techniques
 Rings, Domes, Bars, Spots, Controllers, etc
 Bright Field, Diffuse, Dark Field, Back Lighting
5) Determine Cornerstone Issues
 3-D Geometry, Structure, Color & Filters
6) Eliminate Ambient Light Effects / Environmental Issues
9 Guidelines for Applying MV Lighting
1) Coordinated Lighting & Optics are crucial – when properly
selected, they provide the foundation for the MV system.
2) Develop the lighting solution early in the vision system design
process – on the bench first, if necessary.
3) Dedicated Lighting = Control of the Lighting Environment.
4) Develop Object Feature-Appropriate Lighting.
5) Apply the 4 Cornerstones for enhancing contrast of features.
6) Be aware of and block ambient light.
7) Consider that light MAY interact differently w/ respect to
surface texture, color, composition and incident wavelength.
8) Make the lighting solution robust AND reproducible.
9) Understand the Inspection Environment w/ respect to
Physical Constraints, Object Characteristics and Ergonomic /
Safety apects.
Contact Information

Daryl Martin
Technical Sales & Product Manager
Advanced illumination, Inc.
440 State Garage Road
Rochester, Vermont 05767
USA
Phone: +1 734-213-1312
Email: [email protected]
www.advancedillumination.com