Machine Vision Basics 06 - Camera Selection
Summary
TLDRThis video script delves into the intricacies of machine vision for quality control, emphasizing the importance of camera and lighting selection. It covers the factors influencing camera choice, such as product size, defect detection, accuracy thresholds, and process speed. The script introduces two types of camera systems: smart cameras with integrated tools and algorithms, and vision system processors that handle image processing separately. It also highlights the capabilities of these systems in performing OCR and barcode reading, showcasing their versatility in industrial applications.
Takeaways
- đ Camera Selection: The process of selecting a camera for machine vision applications involves considering product size, defect size, accuracy requirements, and process speed.
- đ Accuracy Threshold: It's important to determine whether the inspection process needs to be 100% accurate or if a 90% accuracy is sufficient for the application.
- đ Process Speed: The number of objects inspected per minute is a critical factor in camera selection, which can range from 30 to 300 or more.
- đ Color vs Grayscale: The need for a color camera versus a grayscale camera is determined by the type of inspection and application requirements.
- đ ïž Smart Camera vs Vision System Processor: There are two types of camera systems - smart cameras with built-in tools and algorithms, and vision system processors that handle the processing while cameras relay images.
- đ Smart Camera Setup: Smart cameras have power and Ethernet cords and come with different lens sizes, filters, and built-in lighting.
- đŒïž Vision System Processor Layout: Vision system processors manage multiple cameras, with each camera connected via Ethernet and powered over Ethernet, acting as 'dummy' cameras.
- đ Real-time Monitoring: Data Logic's machine vision software offers real-time monitoring and statistics, including pass and fail rates, through a display module or software interface.
- đ€ OCR Capabilities: The machine vision cameras are capable of performing optical character recognition (OCR), useful for reading barcodes and characters on objects.
- đą Barcode Reading: While barcode readers are preferred for barcode-only tasks, machine vision cameras are useful for simultaneous object inspection and barcode reading.
- đ Data Presentation: The script mentions the use of pie charts and other statistics to represent the total counts, passes, and fails in the inspection process.
Q & A
What are the common factors to consider when selecting a machine vision camera?
-When selecting a machine vision camera, one should consider the product size, minimum defect size, required accuracy, threshold for passing, process speed, and whether a color or grayscale camera is needed.
What is the importance of accuracy in machine vision applications?
-Accuracy in machine vision is crucial as it determines the reliability of the inspection process. It can vary from needing to be 100% correct to being acceptable at 90% accuracy, depending on the application requirements.
How does the inspection speed affect the choice of machine vision system?
-The inspection speed, measured in objects inspected per minute, influences the choice of system. Higher speeds may require more advanced or multiple cameras to maintain accuracy and efficiency.
What is the difference between a smart camera and a vision system processor?
-A smart camera has all the tools and algorithms built-in, with power and Ethernet coming directly to the camera. A vision system processor, on the other hand, handles the processing while the cameras act as 'dumb' relays of images back to the processor.
Why would one choose a smart camera over a vision system processor?
-A smart camera might be chosen for its simplicity and self-contained operation, which can be beneficial for applications where a compact and integrated solution is preferred.
What are the capabilities of the new vision processor mentioned in the script?
-The new vision processor can support up to eight different cameras, allowing for more complex and multifaceted machine vision applications to be managed from a single processing unit.
How does a machine vision system with a processor display and monitor the inspection process?
-The system can display the inspection process on a floor module or through machine vision software, showing real-time monitoring, statistics, and historical image data for analysis.
What is OCR and how are machine vision cameras capable of performing it?
-OCR stands for Optical Character Recognition. Machine vision cameras capable of OCR can read and interpret text within images, which is useful for applications that require reading barcodes or other characters on objects.
Why would a barcode reader be preferred over a machine vision camera for reading barcodes?
-A barcode reader is typically preferred for just reading barcodes due to its simplicity and efficiency. Machine vision cameras are more versatile for applications where both object inspection and barcode reading are required simultaneously.
What additional insights can the provided script offer regarding machine vision applications?
-The script provides insights into the decision-making process for selecting camera types, the importance of accuracy and speed in inspections, the capabilities of smart cameras and vision processors, and the role of OCR in machine vision systems.
Outlines
đ· Camera Selection Criteria for Machine Vision Applications
This paragraph discusses the common approach to selecting cameras and lighting for machine vision applications. It emphasizes the importance of considering product size, defect detection, accuracy thresholds, process speed, and whether color or grayscale cameras are needed. It also touches on the type of inspection and the desired outcomes from the vision system. The paragraph introduces two types of camera systems: smart cameras with built-in tools and algorithms, and vision system processors that handle the processing while cameras act as simple image relays. The smart camera setup includes various lens sizes and filters, while the vision system processor can support multiple cameras and is responsible for all motor and drive control.
Mindmap
Keywords
đĄMachine Vision
đĄCamera Selection
đĄDefect Size
đĄAccuracy
đĄProcess Speed
đĄColor Camera
đĄGrayscale Camera
đĄSmart Camera
đĄVision System Processor
đĄOptical Character Recognition (OCR)
đĄBarcode Reader
Highlights
Camera selection involves considering product size, defect size, accuracy, and process speed.
Accuracy thresholds can vary, with some processes requiring 100% correctness, while others might accept 90% accuracy.
The choice between color and grayscale cameras depends on the inspection type and application requirements.
Smart cameras contain tools and algorithms internally and are powered via Ethernet.
Vision system processors handle the processing for 'dummy' cameras, which relay images back to the processor.
Data Logic's new processor supports up to eight cameras, showcasing technological advancement in vision systems.
Vision system processors can output to display modules for real-time monitoring on the factory floor.
Machine vision software allows for detailed inspection, including previous images and real-time statistics.
Pie charts and statistics provide insights into total counts, pass, and fail rates in the inspection process.
OCR capabilities in machine vision cameras enable optical character recognition for applications requiring text reading.
Barcode reading is a common application for OCR, with the choice between barcode readers and vision cameras depending on inspection needs.
Machine vision systems can be used for simultaneous object inspection and barcode reading, useful for complex applications.
The transcript discusses the importance of selecting the right camera and lighting for machine vision applications.
Different lens sizes and built-in filters are available for smart cameras to suit various inspection needs.
The layout of a smart camera system includes power and Ethernet cords, with all processing done within the camera.
Vision system processors are connected to cameras via Ethernet, with the processor performing all the image processing tasks.
The transcript provides a sample mesh from a program, indicating the practical application of machine vision in real-world scenarios.
The importance of gray level and pixel considerations in machine vision is briefly mentioned towards the end of the transcript.
Transcripts
each machine vision application is
different but the way we go about
picking cameras and lighting kind of all
are the same so what goes into a camera
selection will take the products and the
size of it your minimum defect size that
you want to have on the object you're
intersecting required accuracy
what's your threshold what do you want
to be able to pass the does it need to
be a hundred percent correct every time
can it something be 90 percent accurate
and still pass your process speed for
example how many of these objects are
you inspecting per minute is it 30 is it
a hundred is it three hundred are you
required to have a color camera we're
going to lead with a grayscale camera
what is a color camera required what
type of inspection are you doing what
type of application is is where you're
looking for what do you want to
accomplish what are you looking to get
out of this machine vision camera and
then just your rules of vision and then
we'll go through all of those and then
we'll discuss our available cameras and
then we'll get your object and do some
testing and then just go from there with
DES logic we have two different types of
cameras we have a smart camera and then
we have a vision system processor or our
smart camera this is where all your
tools and all your algorithms are inside
of the smart camera you're gonna have a
power in a ethernet cord coming from the
camera everything is running through the
actual camera and you can kind of see
what the typical system of a smart
camera kind of what the layout is all
these cameras come with different lens
sizes different filters built-in
lighting and the next type of system we
have is a vision system processor this
is actually where all the motor and all
the drive and all the tools and all of
that good stuff is actually going to be
run through a vision processor the
cameras will have ethernet cords running
to them they'll be powered over Ethernet
but these cameras are actually what you
would call dummy cameras
they have nothing no type of horsepower
running on them what they're doing
they're just relaying the image back to
the vision system processor and the
processor is doing all the work
data logic actually we have a brand new
processor that came out several months
ago but it can actually we can actually
have eight different cameras running
from one vision processor and this is
kind of the typical system layout when
it comes to a vision system processor
this is kind of what you'll see you can
get this output on a display module on
your floor or if you pull up the data
logic machine vision software you can
take a look at it like this but you can
you see you have the object you're
inspecting you have previous images down
below and you also have real-time
monitoring and statistics over here on
the bottom left with the pie chart you
have the total counts you have your
passes and your fails we also do OCR
these machine vision cameras are capable
of performing an OCR optical character
recognition we do do some OCR
applications as well
these cameras are able to read barcodes
but typically if we're reading a barcode
just a barcode we want to stick with a
barcode reader typically reading
barcodes comes in if you want to inspect
an object that has a barcode on it and
you have to inspect the object and read
the barcode at the same time comes in
very useful then for the types of
applications like that
and here's the sample mesh from the
program I think we might win over this a
few times and fuelie up slides that up
we went over earlier regarding gray
level in pixels
5.0 / 5 (0 votes)