When you are setting up a camera, the camera you choose depends on a number of factors like available space, lighting and temperature, the objects that you wish to inspect, the required speed and system costs.
Deciding on the appropriate lens for your machine vision camera requires you to do some math and consider important factors.
Below mentioned are the essentials that need to be considered at the time of applying machine vision technology:
Selecting the right machine vision camera
Let’s delve into the factors that make a difference while choosing the right machine vision camera:
Area Scan or Line Scan
If you require a camera to inspect a long object for which it is difficult to get the desired accuracy for the complete FOV, or a very high-resolution image which is not possible with the available areas scan cameras, you may need a line-scan camera.
These use a single line or sometimes two or three lines of pixels to capture the image.
These line cameras can quickly check the printing quality of the newspapers, sort letters and packages in the logistics department, and check food for damages. They also help in controlling the quality of plastic films, wafers, electronics, textiles and steel.
In case you want a detailed inspection, then the area scan cameras are an ideal option. The latest high-resolution high-speed cameras have eliminated the need for line scan cameras in most applications.
Area scan cameras have a square / rectangular sensor that includes multiple lines of pixels that capture the entire image at once.
This type of machine vision camera is used in quality assurance systems, pick and place process in robotics and code reading.
They are also equipped in dental scanners, microscopes and multiple other medical devices.
CMOS or CCD
In CMOS cameras, electronics are equipped into the surface of the sensor and these convert light into electronic signals.
This makes data transferring super simple. These sensors are cost-effective and have a higher dynamic range.
This allows them to capture high-lit license plates and shadowed persons in the car in a single image.
The CCD sensors can capture more light and have a low noise factor, high colour fidelity and high fill factor.
All this is possible because they do not have conversion electronics on the surface of the sensor.
These are the points that make CCD cameras an ideal option for low-light and low-speed applications.
Global or Rolling Shutter
Global shutter is an ideal choice for those who want to avoid the distortions at high speeds.
With this, the entire sensor surface is exposed to light all at once.
It is ideal for high-speed applications like pharma track and trace, packaging, logistics, transportation, and traffic.
The rolling shutter reads the image line by line. All the captured lines are restored in a single image.
In case the object is moving too fast or there are poor lighting conditions then the image gets distorted.
You can minimize the distortion by adjusting the exposure time and using flash. Typical applications are the inspection of automotive components, sub-assemblies, pharma bulk code reading etc.
This shutter type is less costly and only available for CMOS cameras.
Colour Camera or Monochrome Camera
Many vision systems use colour cameras as users are more comfortable with coloured images.
However, the main purpose of using colour cameras should be in the applications where colour is the main differentiation between OK and NG parts, or colour based identification of parts is necessary.
You may not always use colour cameras to recognize a few colours. You can easily do it using gray scale values of monochrome cameras.
Monochrome cameras provide better edge accuracies due to contrast and are good for applications such as dimensions measurement, surface inspections etc.
A few designers think that monochrome cameras are less expensive but that's not always the case.
There are other factors that need to be considered like the processing time, pass/fail accuracy, power consumption, resolution and cost.
The resolution of the machine vision camera is the number of pixels available on the sensor chip.
For example, if the sensor has 1000 pixels along the length and width equally, the total resolution will be = 1000 x 1000 = 1000,000 pixels = 1Mega Pixel.
Below calculations are useful to derive the required camera resolution:
The smallest feature should have 5px. (Text and code reading is different).
e.g. Object width = 50mm. Fixturing tolerance = 25mm.
Total = 75mm
Smallest feature (e.g. hole) = 1mm
5px = 1mm
1px = 0.2mm
375px = 75mm
Example: VGA would work.
Frame rate determines the number of images that a sensor can capture and transmit every second.
The human brain can detect 14 to 16 images every second and the frame rate of a movie is generally 24 rates per second.
For fast-moving applications, the machine vision camera is required to shoot in milliseconds.
On the contrary for microscopic applications, the camera requires low frame rates compared to the human eye.
The most compact size of the camera today is 29 mm on each side.
Some board level cameras offer compact form factors. These are ideal for size-restricted applications in the embedded field.
You need to choose the communication interface you will use.
This one has similar features to USB 2.0 with higher throughput and reliable transmission.
This one is an affordable communication interface having a throughput of up to 1 Gbit/s.
It is beneficial to have this one if you require long cable lengths.
It has a very high throughput of up to 6 Gbits/s and is ideal for high resolutions.
The ongoing trend toward higher resolution and speed due to ever-faster production cycles and new CMOS sensors call for cameras with increased transmission bandwidth.
10 Gigabit Ethernet allows for the transmission of 10 times more image data at 1.1 GB/s while at the same time providing all the advantages of Gigabit Ethernet interfacing without requiring a change in integration software.
Frequently Asked Questions
Are machine vision cameras and computer vision cameras the same?
No. The computer vision algorithms and then processes and interprets the image.
A machine vision system makes use of a camera to view any image, prior to instructing different parts in the system to act.
The computer vision does not need to be a part of a larger machine system and can be used alone.
What is a machine vision camera?
Machine vision is using single or multiple cameras to automatically inspect objects.
This system is generally used in the production or industrial sector.
It might be used in application on an assembly line.
After the operation is performed, the camera is then activated to capture and process the images.
What are the various kinds of machine vision cameras?
2D Area scan cameras
Line Scan cameras
We hope to have helped you select the right machine vision camera. Make sure to check out the above-mentioned factors that will help you make the right purchase.