What Is a Machine Vision Camera?
A machine vision camera, also known as an industrial camera, is a key part of a 2D image processing system. Its main function is to capture images, which are then processed by a combination of hardware and software. The information obtained is prepared for various applications.
A typical example of an image processing application in a manufacturing system is quality control, presence control and completeness control. This involves analyzing a specific feature of a part that is produced on an assembly line. In this way, it can be checked whether the part meets the quality criteria or, if necessary, must be sorted out.
Camera Variants of the 2D Image Processing Systems
Line Cameras
Line cameras are primarily used in continuous inspection processes. They are particularly suitable for applications where objects move on conveyors or endless material is processed. Unlike surface cameras, line cameras do not capture the image at once, but line by line. To create a complete image, the object must move through the camera’s field of view during the capture process. The software then compiles the individual lines into an overall picture. In applications like these, they are significantly faster than conventional 2D cameras and are therefore particularly suitable for high-speed applications.
Their performance often exceeds that of 2D cameras, especially in continuously running processes. The image quality depends heavily on factors such as smooth movement of the object, the time of capture, the line resolution and the exposure time. Typical application examples include quality control of textiles, paper, fabrics and other continuous materials where seamless and precise image capture is required.
Surface cameras
Unlike the line camera, which captures an image line by line, the surface camera uses an image sensor that captures the entire image at once. This means that it is able to create a complete two-dimensional image with just one image capturing.
Surface cameras are widely used wherever immediate, complete imaging is required – such as in industrial quality control, medical imaging, surveillance systems and many other areas where high precision and fast image processing are required. They are particularly suitable for stationary objects, as there is no relative movement between the camera and the object.
Their advantage lies in capturing detailed and high-resolution 2D images precisely. Surface cameras are used in particular for applications where very high image quality is required.
Differentiation for Surface Cameras: Machine Vision Cameras and Smart Cameras
Machine Vision Cameras
Image evaluation is performed either via the machine vision controller and the image processing software or as a stand-alone solution using third-party software
Multiple cameras can be connected to a single machine vision controller
Faster process times thanks to high computing power of the machine vision controller
Suitable for very high resolution inspection tasks
Compact camera design
Camera Selection Starts with the Image Chip
What’s an Image Chip?
Monochrome or Color Camera? Which Do I Use When?
Actual Image
Image Capture with a Monochrome Camera
A monochrome camera can distinguish between gray-scale values.
Image Capture with a Color Camera
A color camera is able to differentiate color values from objects.
Operating Orinciple of CMOS Sensors with Global or Rolling Shutter
CMOS image sensors have two exposure methods that control how an image is captured and read. These procedures determine the exposure time and thus the amount of light that is converted into electrons as a value in the camera sensor. A distinction is made between global shutter and rolling shutter:
Global Shutter
| Entire image area is exposed simultaneously |
| Suitable for static as well as dynamic applications |
| No image distortion on moving objects |
Rolling Shutter
| Lines are exposed with a time offset |
| For static applications |
| Image distortions due to fast object movements (rolling shutter effect) |
| Capturing still images |
The Rolling Shutter Effect
Left: Global shutter, Right: Rolling shutterIn industrial image processing, a distinction is made between monochrome and color cameras. Monochrome cameras capture grayscale and focus on the differences in brightness in the image. This makes them particularly suitable for applications that require fine contrasts and details, such as when inspecting surfaces or measuring objects.
A color camera is able to distinguish objects from each other and from the background. The red, green and blue filters on the pixels can capture a color spectrum of up to 16.7 million colors. This makes it possible to detect objects with varying colors that would not be distinguishable with monochrome cameras.
Other Sensor Properties
Image sensors or image chips differ in many features, including sensor size, resolution, pixel size, frame rate, light sensitivity and dynamic range. Depending on the resolution, sensors of different sizes are used in industrial image processing. Larger models typically offer higher performance, but are less suitable for compact camera systems with limited installation space.
The market is tending toward ever smaller sensor sizes due to increasingly better manufacturing processes that minimize the disadvantages of smaller image chips. However, as the sensor size decreases, the space for the individual pixels also decreases. The larger a pixel, the more light it can absorb – reducing the additional light requirement of the application accordingly.
Especially in industrial applications with short exposure times, such as in fast dynamic processes, a balanced ratio between the number of pixels and pixel size is therefore crucial for reliable image quality.
Because exposure times are often short in image processing, e.g. in fast dynamic applications, particular attention must be paid to the balance between the number and size of pixels.
Resolution
Frame Rate
Exposure Time
The Right Resolution for Every Application
| Resolution | Accuracy | Examples |
|---|---|---|
| 1.6 MP | Applications that do not require extremely high resolution | Optical character recognition, assembly control, presence check |
| 5 MP | Applications requiring medium level of detail | Inspection of packaging |
| 12 MP | Applications requiring high precision | Inspection of fine mechanical parts |
| 24 MP | Applications requiring very high resolution and attention to detail | Checking PCBs for faulty components |
Main Components of 2D Image Processing Systems
Lens
Together with the camera and illumination, lenses are crucial to the quality of the images. Lenses enable precise imaging, which not only facilitates the analysis of the finest structures. Choosing the right high-quality lens can optimize both detection and processing time.
The Machine Vision Camera Interface
Gigabit Ethernet (GigE)
Fast transfer of large amounts of image data
Easy integration thanks to protocol standard
- Multiple cameras can be operated in a network
It is also possible to connect the machine vision camera via a cable using PoE (Power over Ethernet), which means that both power supply and data transfer take place via a single connection.
Applications of Machine Vision Cameras
Position Check
Robot Positioning
Parts Measurement
Quality Control
Presence Check
Process Monitoring
Code Reading
Reliable Solution for Cross-Industry Applications
Automotive Industry
Quality inspection of car interior doors
Quality inspection of engine blocks
Position detection for automated tightening
Electronics Industry
Position check of PCBs
Checking the alignment of components
Inspection of plug connectors and cables
Packaging Industry
Check packages for damage, contamination or missing labels
Label inspection of packaging
Minimum shelf life test on PET bottles
Food Industry
Orientation of beverage cans
Label check on packaging
Tethered cap inspection
