Scribd
Upload
20 views19 pages

ETM4272 Lecture01 Introduction

Original lecture presentation created by Ashan Weeraratne for B4-Mechatronics, FOT, USJ.

Uploaded by

Ashan Weeraratne
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
20 views19 pages

ETM4272 Lecture01 Introduction

Original lecture presentation created by Ashan Weeraratne for B4-Mechatronics, FOT, USJ.

Uploaded by

Ashan Weeraratne
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 19

ETM 4272: Machine Vision

LECTURER : ASHAN WE ERARATN E

ASHAN WEERARATNE FOT:USJ 1


Module Content
I. Camera & Imaging
II. Color Spaces & Contrast
III. Image Processing Operations
IV. Image Segmentation
V. Object Detection
VI. Optical Character Recognition
VII. Camera Calibration and Depth Estimation
VIII. Pose Estimation
IX. Deep Learning in Computer Vision

ASHAN WEERARATNE FOT:USJ 2


Tools & References
Tools Online Resources:
First Principles of Computer Vision - Colombia
OpenCV Computational Photography - Udacity
YTD: Nicolai Nielsen
Github Copilot

Books:
Handbook of Machine Vision – Alexander Hornberg

ASHAN WEERARATNE FOT:USJ 3


Evaluation
Continuous Assessment (OpenCV + Python)
o Lab Activities (2) (20 marks)
o Take Home Assignments (5) (20 marks)

End Examination (Double Sized A4 Page Allowed)


o MCQ (20 marks)
o Structured Essay (40 marks)

ASHAN WEERARATNE FOT:USJ 4


Lecture 01: Camera & Imaging
Lecture outcomes:

Be able to
o List and describe key factors and specifications which need to be considered when selecting
a camera and lens(es) for a machine vision application
o Describe how a digital image is formed, and represented

ASHAN WEERARATNE FOT:USJ 5


Sensors in Digital Imaging
-Charge Coupled Devices: CCD
o An array of capacitors
o Converts photons to free electrons
o The presence of free electrons is read-out
o Found in Medical Imaging, Astronomical Imaging,
High End Videography

-CMOS Sensors
o An array of photodiodes, one per pixel
o Considered to be of lesser quality than CCD
o Cheaper to produce in scale than CCD
o Faster read-out compared to CCD
o Found in Smartphones, Webcams, Consumer
Cameras

ASHAN WEERARATNE FOT:USJ 6


Digital Imaging & Color Sensing
Color sensing: The Bayer Filter: Demosaicing:
Photo sensors are sensitive to only A color filter which allows waves of Creating the total spread spectrum
to the intensity, not the color. only a certain color to pass through intensity from the RGGB readings
Four sensors with a preceding
Bayer filter is used to sense color.
Digital Images: Grayscale Representation
Grayscale Images:
- Records spread spectrum
illumination intensity at each
pixel
Digital Color Images: Representation
Video: In a nutshell
Image Sensor Specifications
-Sensor Type: CCD/CMOS
-Resolution: Number of pixels the sensor can capture
-Sensor Size: Physical dimensions of the sensor (CCD/CMOS array)
-Signal To Noise Ratio: Indicates how strong the signal is compared to the noise
-Exposure time: Minimum time required for single image to be captured
-Frame rate: The maximum rate at which images can be captured
-Color Scheme: RGB , Monochrome , NIR
-Data Interface: CSI, SPI, USB, Gigabit Ethernet, Camera Link
-Lens Mount: Type of mechanical interface for attaching lens(es)
-Spectral Response: Sensitivity to light at different wavelengths (different colors)

ASHAN WEERARATNE FOT:USJ 11


1, 2

Pinhole Camera Model

Why is a pinhole required to obtain an image of The effect of pinhole size on image sharpness
an object?

Pinhole filters out all unwanted reflections of the


object
Pinhole filters out light from other sources

ASHAN WEERARATNE FOT:USJ 12


Loss of Information During Capture
Loss: Depth Information

Cause: Image capture maps a 3D


space into a 2D space.

Due to loss of depth information


• Lines which are parallel in the real world, appear to
converge in the image, to the vanishing point.
• Angles will be distorted.

ASHAN WEERARATNE FOT:USJ 13


Lenses in Digital Imaging

Why Lenses:
- To gather more light
- To obtain magnification
What happens during zooming in?

What happens in a microscopic lens?

ASHAN WEERARATNE FOT:USJ 14


Lens Defocus & Focusing
What happens during defocus?
- A single point in the scene is spread out on the sensor.

What is the focus plane?


- The plane in which the image of a point is also a point, and not
a circle

What is meant by focusing on a particular area in a scene?


- The scene is brought into the focus plane.
- By moving the lenses(s) or moving the sensor
- Moving the ___ & ___?

What happens if blur circle of a point is smaller than a pixel ?


- There would not be any blur in the digital image
- Image would be considered to be in focus

ASHAN WEERARATNE FOT:USJ 15


Field of View
Note: A property dependent on the focal length of the lens. Expressed in degrees.

Types of lenses based of FOV

Standard: 50-70
Wide Angle: 80-120
Telephoto: 20-40
Super-Telephoto: 10-20
Fisheye: 180

ASHAN WEERARATNE FOT:USJ 16


Additional Videos: 1, 2, 3, 4, 5

Lens Specifications
Lens Types
Focal Length (Fixed / Range)
o Higher the focal length, field of view is narrower and objects are more magnified
o Lower the focal length, field of view is wider and objects are smaller
o Also specified as: Optical Power: 1/F (Measured in Diopters)
Sensor Size
Lens Aperture
Minimum Object Distance
o Closest distance at which lens can accurately focus on a subject
Field Of View
Interface: Determines how lens is mounted to camera

ASHAN WEERARATNE FOT:USJ 17


In Class Activity: Review Common Camera Specs
OV 2640 PI CAM HQ MARS 800S-46UM
Pixel Size: 2.2 x 2.2 um 1.5 x 1.5 um 2.74 x 2.74 um
Resolution: 1600 x 1200 (UXGA) 4056 x 3040 2840 x 2840
Stock Lens: ¼", 2.8mm 1/2.3" , No 2/3" , No
FOV: 650 Depends on Lens Depends on Lens
Frame Rate: 15 fps @ UXGA 60 fps @ 12 MP 47 fps @ 8MP
SNR: 15 dB 11.5 dB 38 dB
Data Interface: SPI CSI-2 USB 3
Price: ~$ 25 ~ $50 > $ 100

ASHAN WEERARATNE FOT:USJ 18


Take Home Activities: Lecture 01

- Create Github Account


- Sign-up for Github Student Developer Pack

End of Lecture 01 Optional Activities:

- OpenCV Bootcamp (Try to complete within 2 weeks)

Before lecture 02:

- Complete given activities

Lecture 02:
- Color Spaces

- Image Filters

- Image Segmentation

ASHAN WEERARATNE FOT:USJ 19

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy