The document discusses computer vision technologies including stereo vision cameras, line scan cameras, area scan cameras, LIDAR sensors, and automotive radar. It notes their applications in areas like automated inspection, autonomous vehicles, and robotics. It then covers future development trends like higher resolutions, processing speeds, and sensor miniaturization. Challenges in computer vision are also discussed, such as not fully understanding biological vision and difficulties emulating it computationally.
[2024]Digital Global Overview Report 2024 Meltwater.pdf
三維機器視覺的最新應用與趨勢 part 2
1. Ta-Te Lin
Dept. of Bio-Industrial Mechatronics Engineering
National Taiwan University
AOI & ROBOTIC
APPLICATIONS
Source: https://www.google.com/atap/project-tango/
2. Ta-Te Lin
Dept. of Bio-Industrial Mechatronics Engineering
National Taiwan University
AOI & ROBOTIC
APPLICATIONS
Stereo Vision Camera
3. Ta-Te Lin
Dept. of Bio-Industrial Mechatronics Engineering
National Taiwan University
Vehicle Collision Avoidance
19.6〫
43
AOI & ROBOTIC
APPLICATIONS
4. Ta-Te Lin
Dept. of Bio-Industrial Mechatronics Engineering
National Taiwan University
Vehicle Collision Avoidance
AOI & ROBOTIC
APPLICATIONS
5. Ta-Te Lin
Dept. of Bio-Industrial Mechatronics Engineering
National Taiwan University
AOI & ROBOTIC
APPLICATIONS
Stereo Line Scan Camera
Source: http://www.chromasens.de/
Specifications Range
Optical Resolution 5~70 m
FOV 35~500 mm
Height Resolution 1~10 m
Height Range 0.7~52 mm
Working Distance 71.9~796.9 mm
Max. Transport Speed 0.1~1.45 mm
Line Rate Up to 60 kHz
Number of Pixels Max. 7300
6. Ta-Te Lin
Dept. of Bio-Industrial Mechatronics Engineering
National Taiwan University
AOI & ROBOTIC
APPLICATIONS
Stereo Line Scan Camera
Source: http://www.chromasens.de/
Metal Surface Wire Bonds
Ball Grid Array PCB
7. Ta-Te Lin
Dept. of Bio-Industrial Mechatronics Engineering
National Taiwan University
AOI & ROBOTIC
APPLICATIONS
Stereo Line Scan Camera
Source: http://www.chromasens.de/
8. Ta-Te Lin
Dept. of Bio-Industrial Mechatronics Engineering
National Taiwan University
AOI & ROBOTIC
APPLICATIONS
3D Area Scan Camera
Source: http://www.shape-drive.com/
Specifications Range
Optical
Resolution
10~150 m
Height
Resolution
1~50 m
Height Range 2~100 mm
Working
Distance
10~400 mm
Acquisition Time 0.8 or 1.4 s
Number of
Pixels
2.3 M pixels
9. Ta-Te Lin
Dept. of Bio-Industrial Mechatronics Engineering
National Taiwan University
AOI & ROBOTIC
APPLICATIONS
3D Area Scan Camera
Source: http://www.shape-drive.com/
10. Ta-Te Lin
Dept. of Bio-Industrial Mechatronics Engineering
National Taiwan University
AOI & ROBOTIC
APPLICATIONS
LIDAR Sensor
Source: http://www.chromasens.de/
64 lasers/detectors
360 degree FOV (azimuth)
0.08 degree angular resolution
26.8 degree vertical FOV
<2 cm distance accuracy
5~15 Hz FOV update
50 m range for pavement
120 m range for cars and forliage
> 1.3 M points per second
300~900 rpm spin rate
100 MBPS UDP Ethernet packets
905 nm wavelength
~10 ns pulse width
Class 1 – eye safe laser
11. Ta-Te Lin
Dept. of Bio-Industrial Mechatronics Engineering
National Taiwan University
AOI & ROBOTIC
APPLICATIONS
LIDAR for Autonomous Navigation
Source: http://google.com/+GoogleSelfDrivingCars
12. Ta-Te Lin
Dept. of Bio-Industrial Mechatronics Engineering
National Taiwan University
Source: www.universalrobotics.com
AOI & ROBOTIC
APPLICATIONS
13. Ta-Te Lin
Dept. of Bio-Industrial Mechatronics Engineering
National Taiwan University
Source: http://www.viscom.com/
AOI & ROBOTIC
APPLICATIONS
Solder joint inspection
Solder paste inspection
Wire bond inspection
SMD inspection
14. Ta-Te Lin
Dept. of Bio-Industrial Mechatronics Engineering
National Taiwan UniversitySource: http://bosch-automotive.com
AOI & ROBOTIC
APPLICATIONS
76–77 GHz
Range: 100 ~ 250 m
FOV: 45 ~ 150
15. Ta-Te Lin
Dept. of Bio-Industrial Mechatronics Engineering
National Taiwan University
16. Ta-Te Lin
Dept. of Bio-Industrial Mechatronics Engineering
National Taiwan University
FUTURE DEVELOPMENT
Regarding the Market
Regarding the Vision Industry
Regarding Applications
Regarding Academic Researches
Regarding the Practical Technology
17. Ta-Te Lin
Dept. of Bio-Industrial Mechatronics Engineering
National Taiwan University
Human Vision vs. Machine Vision
FUTURE DEVELOPMENT
18. Ta-Te Lin
Dept. of Bio-Industrial Mechatronics Engineering
National Taiwan University
FUTURE DEVELOPMENT
Human Vision vs. Machine Vision
Function Human Vision Machine Vision
Spatial resolution Excellent, Non-linear, Analog Poor, but improving,
Linear
Grey-scale resolution Limited Relatively large (4000
shade or more)
Dynamic range Excellent, Non-linear Limited, Typically linear
Color resolution Limited Discriminate millions of
colors
Complex color pattern
perception
Excellent Currently poor but
improving
Three Dimensional
Perception
Automatic, Powerful, but
poor in Measurement
Precision
Complex but can make
precise measurements
Learning Learns by example and
deduction
Generally by programming
Processing speed, simple
and defined tasks
Relatively slow Can be very fast
Processing speed,
complex but ambiguity
Relatively fast Very slow, or not possible
Sensitive to illumination Less Yes
Processing Massively parallel, Highly
non-linear, Intuitive
Serial, Usually linear and
fixed logic
19. Ta-Te Lin
Dept. of Bio-Industrial Mechatronics Engineering
National Taiwan University
FUTURE DEVELOPMENT
Promising Future Trends
– Smarter algorithms
– Higher resolution
– Faster speed
– Sensor miniaturization
– Proliferation to daily life
– Merging with computer graphics
– Virtual reality
– Data compression
20. Ta-Te Lin
Dept. of Bio-Industrial Mechatronics Engineering
National Taiwan University
FUTURE DEVELOPMENT
Challenges in Computer Vision
– The way biological vision works is still largely
unknown and therefore hard to emulate on
computers.
– Attempts to ignore biological vision and
reinvent a sort of silicon-based vision have
not been so successful as initially expected.
Olivier Faugeras
21. Ta-Te Lin
Dept. of Bio-Industrial Mechatronics Engineering
National Taiwan University
THANK YOU
… The study of vision must therefore include not
only the study of how to extract from images the
various aspects of the world that are useful to us,
but also an inquiry into the nature of the internal
representations by which we capture this
information and thus make it available as basis for
decisions about our thoughts and action.
David Marr, Vision