Join or Sign in

Register for your free membership or if you are already a member,
sign in using your preferred method below.

To check your latest product inquiries, manage newsletter preference, update personal / company profile, or download member-exclusive reports, log in to your account now!
Login asmag.comMember Registration

How to select the right machine vision system

How to select the right machine vision system
Selecting the right vision system requires a basic understanding of the inner workings of the system and ultimately depends on the user’s own needs and objectives as well as environmental conditions inside the factory.
Machin vision has seen more and more deployments in factories around the world, given their ability to “see” stains, color nuances or other defects that are otherwise too small to be detected by the human eye. Selecting the right vision system requires a basic understanding of the inner workings of the system and ultimately depends on the user’s own needs and objectives as well as environmental conditions inside the factory.
Typically, components in a machine system include: cameras (along with lenses and image sensors), lighting, frame grabbers, and software. “There are two types of industrial camera: line-scan cameras and area-scan cameras. More line-scan cameras are used in machine vision systems, since they can generate higher-resolution images compared to area-scan cameras,” said Hongsuk Lee, Business Development Manager at SuaLab. “Frame grabbers capture individual, digital still frames from an analog or digital video stream. Video frames are usually captured in digital form as it needs to be transmitted.”
Lee added that a good machine vision system entails the combination of high quality hardware and software. “A good machine vision system is a system with high defect detection rate. To do this, good hardware must be installed to obtain good images, which should be analyzed with high-performance machine vision software,” he said. “No matter how good the hardware configuration is, if the image analysis capability is low, it cannot be a good machine vision system.”
Increasingly, machine vision systems leverage the power of deep learning whereby a system is fed with images of good samples and bad samples, and eventually it will make its own inferences, recognizing defects or bad spots on its own.
Lee cited his company’s technology applied to the textile industry as an example. “We are the first company in the world to commercialize a textile unmanned inspection system using deep learning technology. It has been difficult to apply the machine vision system to the textile industry for two reasons. First, the background of the textile product is complicated so it is difficult to detect the defect with a simple rule-based algorithm. In addition, because the product cycle is short, there is not enough time to optimize the algorithm. To solve this problem, we applied a self-customizing engine to machine vision system so that it can detect undetected textile defects and quickly respond to fast product switching,” he ssaid.

How to choose the right system

Since end user entities operate in different industries, they are bound by specific manufacturing rules or requirements. The machine vision system they choose therefore should be able to suit those needs and requirements. Bruno Menard, Program Manager for Embedded Vision at Teledyne DALSA, summarized the points to be looked at as follows:
  • Determine the task that needs to be performed by the vision system since different tasks may require different vision attributes. A machine vision system designed for one task may not be well suited for the other.
  • Define the key visual performance criteria to ensure the camera and lens are performing at the right levels. Factors such as the smallest object or defect to detect, measurement accuracy needed, the image size, speed of image capture and processing, and the need for color all affect camera and lens choices.
  • Environmental factors need to be considered since some cameras suit stationary views while others are more suitable for handling linear object motion. Temperature, humidity and vibration can impose a need for specific system fabrication and assembly practices. The physical space for installing the system can restrict camera and lens choices.
The end user should also consider from an economic perspective: If buying a machine vision system does not warrant a return on investment, then relying on traditional labor may be the better alternative.

Product Adopted:
Subscribe to Newsletter
Stay updated with the latest trends and technologies in physical security

Share to: