3D Scanning Dr. Francesco Banterle, francesco.banterle@isti.cnr.it - - PowerPoint PPT Presentation

▶
3d scanning
SMART_READER_LITE
LIVE PREVIEW

3D Scanning Dr. Francesco Banterle, francesco.banterle@isti.cnr.it - - PowerPoint PPT Presentation

Aug 16, 2023 149 likes •837 views

3D Scanning Dr. Francesco Banterle, francesco.banterle@isti.cnr.it banterle.com/francesco What is 3D Scanning? 3D scanning is the process of measuring 3D information; and it is the very first step when creating a complete 3D model. 3D

slide-1
SLIDE 1

3D Scanning

  • Dr. Francesco Banterle,

francesco.banterle@isti.cnr.it banterle.com/francesco

slide-2
SLIDE 2

What is 3D Scanning?

  • 3D scanning is the process of measuring 3D

information; and it is the very first step when creating a complete 3D model.

slide-3
SLIDE 3

3D Scanning Outputs

  • Each device outputs measure 3D information
  • differently. The main outputs are:
  • 3D sparse points
  • Range maps
  • 3D volumes
slide-4
SLIDE 4

3D Scanning Outputs: Sparse Points

slide-5
SLIDE 5

3D Scanning Outputs: Sparse Points

  • Each point can have attributes:
  • An RGB color
  • …
  • Metadata: position and orientation of the origin, and

scale

slide-6
SLIDE 6

3D Scanning Outputs: Range Maps

Each pixel in the image encodes the distance between the surface and center of the camera

slide-7
SLIDE 7

3D Scanning Outputs: Range Maps

  • Metadata:
  • Camera extrinsics: position and rotation
  • Camera intrinsics: field of view, size of pixels in mm
  • Scale of distances
  • From Metadata:
  • we can obtain 3D points!
slide-8
SLIDE 8

3D Scanning Outputs: Range Maps

Image Plane Field of View Camera Center Surface 3D point d

slide-9
SLIDE 9

3D Scanning Outputs: Range Maps

slide-10
SLIDE 10

3D Scanning Outputs: Range Maps

  • A range map is already a 3D model… but it will be

surely incomplete

  • A single acquisition IS NOT enough to

reconstruct an entire object

  • Multiple shots are needed…
  • How many?
  • Which ones to choose?
slide-11
SLIDE 11

3D Scanning Outputs: Range Maps

slide-12
SLIDE 12

3D Scanning Outputs: 3D Volumes

  • 3D space is discretized into a regular grid or volume
  • Each cube in the grid is called voxel (volume pixel) or

a cube encodes a value in the range [0, 1].

Volume Voxel

slide-13
SLIDE 13

3D Scanning Outputs: 3D Volumes

  • Metadata:
  • size of the pixel in mm for each slice
  • distance in mm between a slice and another
  • scale of the normalized values (typically

encoded as 16-bit values)

slide-14
SLIDE 14
  • A sagittal plane is an anatomical plane that

divides the body into right and left parts

3D Scanning Outputs: 3D Volumes

slide-15
SLIDE 15

3D Scanning Outputs: 3D Volumes

  • A coronal plane is an anatomical plane that

divides the body into ventral and dorsal parts

slide-16
SLIDE 16

3D Scanning Outputs: 3D Volumes

  • An axial plane is an anatomical plane that divides

the body into superior and inferior parts

slide-17
SLIDE 17

3D Scanning Taxonomy

slide-18
SLIDE 18

3D Scanning Taxonomy

Non-Contact Contact

slide-19
SLIDE 19

3D Scanning Taxonomy

Contact Non-Contact

Robot Gantry

Slicing

slide-20
SLIDE 20

3D Scanning Taxonomy

Contact Non-Contact

Robot Gantry

Slicing

slide-21
SLIDE 21

3D Scanning Taxonomy:
 Robot Gantry

Object is “probed” at different locations

slide-22
SLIDE 22

3D Scanning Taxonomy:
 Robot Gantry

  • Highly accurate (micron)
  • Moderate-high costs: $2,000 - $15,000
  • Slow scanning; labor intensive!
  • Ideal for: rigid and non-fragile objects
  • Uses: manufacturing control, art/design, reverse

engineering

  • Output data: sparse 3D points
slide-23
SLIDE 23

3D Scanning Taxonomy

Contact Non-Contact

Robot Gantry

Slicing

slide-24
SLIDE 24

3D Scanning Taxonomy:
 Slicing

slide-25
SLIDE 25

3D Scanning Taxonomy:
 Slicing

slide-26
SLIDE 26

3D Scanning Taxonomy:
 Slicing

slide-27
SLIDE 27

3D Scanning Taxonomy:
 Slicing

  • It can be accurate and precise; if slicing is automatic
  • Slow scanning
  • Ideal for:
  • rigid and non-deformable objects
  • breakable objects
  • Uses: biology, reverse engineering
  • Output data: a 3D volume (in this case we can have a per voxel

color)

slide-28
SLIDE 28

3D Scanning Taxonomy

Non-Contact Optical Active Passive

Magnetic Acoustic X-Ray

slide-29
SLIDE 29

3D Scanning Taxonomy

Non-Contact Optical Active Passive

Magnetic Acoustic X-Ray

slide-30
SLIDE 30

3D Scanning Taxonomy

Non-Contact Optical Active Passive

Magnetic Acoustic X-Ray

slide-31
SLIDE 31

3D Scanning Taxonomy: Optical - Active

  • Main blocks:
  • A calibrated camera
  • A light source —> that’s why it’s active!
slide-32
SLIDE 32

3D Scanning Taxonomy: Optical - Active: Structured Light

Cameras Projector

slide-33
SLIDE 33

3D Scanning Taxonomy: Optical - Active: Structured Light

slide-34
SLIDE 34

3D Scanning Taxonomy: Optical - Active: Structured Light

Breuckmann GmbH Cost: €70,000-80,000 Accuracy: 0.1 mm

slide-35
SLIDE 35

3D Scanning Taxonomy: Optical - Active: Structured Light

Microsoft Kinect v1 Cost: €100 Accuracy: 2-5 mm

slide-36
SLIDE 36

3D Scanning Taxonomy: Optical - Active: Laser-based

Laser Line Camera

slide-37
SLIDE 37

3D Scanning Taxonomy: Optical - Active: Laser-based

Laser Camera Surface Z β d

α

slide-38
SLIDE 38

3D Scanning Taxonomy: Optical - Active: Laser-based

Konica Minolta Range 7 Cost: $80,000 Accuracy: 40 micron

slide-39
SLIDE 39

3D Scanning Taxonomy: Optical - Active: Laser-based

Konica Minolta Vivid 910 Cost: $15,000 (second hand) Accuracy: 0.2-0.3mm

slide-40
SLIDE 40

3D Scanning Taxonomy: Optical - Active: Laser-based

NextEngine Cost: $2,000 Accuracy: 0.2-0.5mm

slide-41
SLIDE 41

3D Scanning Taxonomy: Optical - Active: Time-of-flight

Transmitter Detector Clock

slide-42
SLIDE 42

3D Scanning Taxonomy: Optical - Active: Time-of-flight

Microsoft Kinect v2 Cost: €200 Accuracy: 2-5 mm It is meant for small environments: 2-3m radius

slide-43
SLIDE 43

3D Scanning Taxonomy: Optical - Active: Time-of-flight

Cost: €50,000 - 100,000 Accuracy: 5-10 mm It is meant for large environments: 1-30m radius

slide-44
SLIDE 44

3D Scanning Taxonomy:
 Optical - Active

  • It can be accurate and precise
  • Ideal for: rigid object with diffuse optical properties;

i.e., it does not work well for specular surfaces and dark materials

  • Uses: reverse engineering, cultural heritage, metrology

(if calibrated), body scanning, etc.

  • Costs: from $200 to $100,000
  • Output data: a range map
slide-45
SLIDE 45

3D Scanning Taxonomy

Non-Contact Optical Active Passive

Magnetic Acoustic X-Ray

slide-46
SLIDE 46

3D Scanning Taxonomy: Optical - Passive

  • Main blocks:
  • One ore more calibrated camera(s)
  • No control on lighting —> that’s why it’s passive!
slide-47
SLIDE 47

3D Scanning Taxonomy: Optical - Passive: Stereo

  • It is based on the same principle of human stereo

vision:

  • two cameras that captures the real-world from

two slightly different positions

  • Our brains does it automatically though
slide-48
SLIDE 48

3D Scanning Taxonomy: Optical - Passive: Stereo

Surface d Left Camera Right Camera α β Z

slide-49
SLIDE 49

3D Scanning Taxonomy: Optical - Passive: Stereo

Left Camera Right Camera

slide-50
SLIDE 50

3D Scanning Taxonomy: Optical - Passive: Stereo

Left Camera Right Camera

slide-51
SLIDE 51

3D Scanning Taxonomy: Optical - Passive: Stereo

slide-52
SLIDE 52

3D Scanning Taxonomy: Optical - Passive: Stereo

slide-53
SLIDE 53

3D Scanning Taxonomy:
 Optical - Passive

  • It can be accurate and precise
  • Many images are required
  • Ideal for: objects with diffuse optical properties
  • Uses: reverse engineering, cultural heritage, body

capturing, metrology (if calibrated)

  • Output data: sparse 3D points or range maps
slide-54
SLIDE 54

3D Scanning Taxonomy

Non-Contact Optical Active Passive

Magnetic Acoustic X-Ray

slide-55
SLIDE 55

3D Scanning Taxonomy: Magnetic - Magnetic Resonance Imaging (MRI)

Hydrogen atoms in our body are made to emit a radio signal (using a magnetic field) that is detected by the scanner.

Jan Ainali 2008 from wikipedia

Philips MRI Scanner

slide-56
SLIDE 56

3D Scanning Taxonomy:
 Magnetic - MRI

  • T1 weighted images are generated by using short

(15ms and 500ms) time to echo (TE) and time of repetition (TR)

  • T2 weighted images are generated by using long

(>80ms and >2000ms) TE and TR (also less noise than T1)

  • TE is the time between the initial pulse and the echo
  • TR is the time between two excitation pulse
slide-57
SLIDE 57

3D Scanning Taxonomy:
 Magnetic - MRI

  • T1: tissues with high fat content (e.g., white matter)

appear bright and compartments filled with water appears dark:

  • ideal for showing anatomy features
  • T2: compartments filled with water (e.g. CSF

compartments) appear bright and tissues with high fat content (e.g. white matter) appear dark:

  • ideal for highlighting pathologies (more water!)
slide-58
SLIDE 58

3D Scanning Taxonomy:
 Magnetic - MRI

slide-59
SLIDE 59

3D Scanning Taxonomy:
 Magnetic - MRI

  • No hazard, but it requires no metal implant in the

patient’s body

  • It takes long time for a scan; e.g., 15-30 mins
  • Costs: they start at $1 million
  • Ideal for: soft tissues, ligaments, tendons, etc.
  • Uses: medical imaging, and cultural heritage
  • Output data: a 3D volume
slide-60
SLIDE 60

3D Scanning Taxonomy

Non-Contact Optical Active Passive

Magnetic Acoustic X-Ray

slide-61
SLIDE 61

3D Scanning Taxonomy: X-Ray - Computer Tomography (CT)

CT works by taking X-ray images from different angles to produce cross- sectional images

David P. Fulmer 2012 from wikipedia

GE LightSpeed CT scanner

slide-62
SLIDE 62

3D Scanning Taxonomy:
 X-Ray - CT

  • Hazard for the patient
  • It takes long time; e.g., 30 secs - 5 mins
  • Costs: they start at $85,000 - $500,000
  • Ideal for: bones (Ca absorbs X-rays), lungs (contain gas;

lower absorption than tissues), chest, and ER (for time)

  • Uses: medical imaging, and cultural heritage
  • Output data: a 3D volume
slide-63
SLIDE 63

3D Scanning Taxonomy:
 X-Ray - CT

slide-64
SLIDE 64

3D Scanning Taxonomy

Non-Contact Optical Active Passive

Magnetic Acoustic X-Ray

slide-65
SLIDE 65

3D Scanning Taxonomy: Acoustic: Medical Ultrasound

Daniel W. Rickey 2006 from wikipedia

A probe sends pulses of ultrasounds (>20,000Hz) The sound echoes off the tissue; with different tissues reflecting varying degrees of sound

slide-66
SLIDE 66

3D Scanning Taxonomy: Acoustic: Medical Ultrasound

Skin Ultrasound probe Ultrasound

slide-67
SLIDE 67

3D Scanning Taxonomy:
 Acoustic

  • It is real-time!
  • Quality-wise it can have speckle; i.e., granular

noise

  • Ideal for: medical imaging of soft tissues (tendons,

muscles, joints, internal organs, etc.); especially during pregnancy. It can not penetrate bones

  • Output data: a 3D volume (limited to a pyramid)
slide-68
SLIDE 68

Acknowledgements

  • Some images and text are based on work by:
  • Dr. Matteo Dellepiane:
  • http://vcg.isti.cnr.it/~dellepiane/
  • Dr. Marco Callieri:
  • vcg.isti.cnr.it/~callieri/
  • Wikipedia