Machine Vision

Machine vision is the automatic extraction of information from digital images. A typical environment would be a manufacturing production line where hundreds of products are flowing down the line in front of a smart camera. Manufacturers use vision systems instead of human inspectors because it’s faster, more consistent and doesn’t get tired.  The camera captures the digital image and analyzes it against a pre-defined set of criteria. If the criteria are met, the object can proceed. If not, the object will be re-routed off the production line for further inspection.

For example, a beverage manufacturer would typically have human inspectors watching thousands of bottles move down a production line. The workers would need to ensure every bottle cap was secured correctly, every label was on straight and contained the correct information and every bottle was filled to the appropriate level. With machine vision, this entire repetitive process can be automated.

Vision solutions are used heavily in conjunction with robots to increase their effectiveness and overall value for the business. These types of robots resemble a human arm with a camera mounted at the “hand” position. The camera acts as the robot’s “eyes,” guiding it to complete the assigned task.

A vision system has five key components that can be configured either as separate components or integrated into a single smart camera. The correct configuration depends on the application and its complexity. The five key components are:

• **Lighting:**This critical aspect of a machine vision system illuminates the part to be inspected, allowing its features to stand out so that the system can see them as clearly as possible.
• Lens: Captures the image and presents it to the sensor in the form of light.
• **Sensor:**Converts light into a digital image for the processor to analyze.
• **Vision Processing:**Consists of algorithms that review the image and extract the required information.
• Communication: The resulting data is communicated out to the world in a useful manner.

There are four main benefits:: reducing errors, increasing yield, track parts and products as well as comply with regulations.

Reduce Errors

Cameras for inspection ensure fewer bad parts enter the market which can cause costly recalls and tarnish a company’s reputation.

Prevent mislabeled products whose label doesn’t match the content. These defects create unhappy customers, have a negative impact on your brand reputation, and pose a serious safety risk – especially with pharmaceutical products and food items for customers with allergies.

Increase Yield

Turn additional available material into a saleable product.

Avoid scrapping expensive materials and rebuilding parts.

Reduce downtime by detecting product routing errors that can cause system disruptions.

Tracking Parts and Products

Uniquely identify products so they can be tracked and traced throughout the manufacturing process.

Identify all pieces in the process, reducing stock and ensuring the product will be more readily available for just-in-time (JIT) processes.

Avoid component shortages, reduce inventory, and shorten delivery time.

Comply with Regulations

To compete in some markets, manufacturers must comply with various regulations.

In pharmaceuticals, a highly regulated industry, machine vision is used to ensure product integrity and safety by complying with government regulations such as 21CFR Part 11 and GS1 data standards.

Machine vision is better-suited to repetitive inspection tasks in industrial processes than human inspectors. Machine vision systems are faster, more consistent, and work for a longer period of time than human inspectors, reducing defects, increasing yield, tracking parts and products, and facilitating compliance with government regulations to help companies save money and increase profitability.

Four common machine vision applications: Measurement, Counting, Location and Decoding

Measurement

A common machine vision application is automated measurement. After an image of a part is captured, the software compares the measurement to a required tolerance. For example, in factory automation, machine vision systems are used to measure the gap in a spark plug, ensuring it meets the tolerance required to function properly. In the packaging and labeling industry, machine vision is used to measure the fill level of a water bottle, ensuring it was filled to the correct height.

Counting

Another common machine vision application is counting – looking for a specific number of parts or features on a part to verify that it was manufactured correctly. In the electronics manufacturing industry, for example, machine vision is used to count various features of printed circuit boards (PCBs) to ensure that no component or step was missed in production.

Location

Machine vision can be used to locate the position and orientation of a part and to verify proper assembly within specific tolerances. Location can identify a part for inspection with other machine vision tools, and it can also be trained to search for a unique pattern to identify a specific part. In the life sciences and medical industries, machine vision can locate test tube caps for further evaluation, such as cap presence, cap color, and measurement to ensure correct cap position.

Decoding

Machine vision can be used to decode linear, stacked, and 2D symbologies. It can also be used for optical character recognition (OCR), which is simultaneously human- and machine-readable. In factory automation, machine vision is used to sort products on a production line by decoding the symbol on the product. The symbols themselves can also be verified by machine vision-based verification systems to ensure that they comply with the requirements of various symbology standards organizations.

The most common reasons to consider when selecting a smart camera include:

• Ready to deploy solution
• Options for built-in light, lens (including auto-focus) and accessories
• No need to build additional panel at the machine
• Normally, can be installed on existing machines without many modifications
• Single-camera application or independent multi-point inspections
• Cost-effective
• Scalable to Vision Systems
• Single and dual-core processors
• Camera resolutions from 0.3MP to 12MP
• Programming software comes with the camera

The most common reasons to consider when selecting a vision system include:

• High-speed
• Single-camera or multi-camera options, including correlated or independent image processing
• Robust and dedicated processing hardware for machine vision
• Video output directly from the controller
• Customizable and advanced programming
• Camera Link cameras for high-speed transmission, lighting power and control via a single camera cable
• Flexible and powerful system to address the majority of applications
• More fieldbus and I/O options
• Complete camera options (0.3MP to 20MP), including 3D
• Dual- and quad-core multi-thread processors
• Programming software comes with the system

The most common reasons to consider when selecting PC-based cameras include:

• Flexibility to choose your own processing hardware. Install Omron programming software on your PC or IPC
• GigE cameras
• Limited I/O and interfaces
• Speed-dependent on hosting hardware. Normally combining Vision and other tasks in the same hardware, where one can impact the other
• You purchase a camera and a software license

The most common reasons to consider when selecting PC-based vision cameras/industrial cameras include:

• Flexibility to choose camera interface (GigE Vision, Camera Link, USB3 Vision, Coax, etc.) and a multitude of sensors from Sony, CMOSIS, e2v, OnSemi, etc.)
• Desire to create proprietary application software or use third-party software.
• Limited I/O and interfaces
• Flexible processing hardware. Capable of running on high-powered PC systems, single-board computers, Linux OS, etc.
• Lower hardware cost, but increased engineering hours to develop (or integrate) software/application solution