Machine vision appliances
Friday, 14 April, 2006
A new breed of vision system has arrived - Vision Appliances. Powerful and compact, yet friendly, cost effective and easy-to-use vision systems that have been designed for production personnel who are not vision experts and are busy looking after their day to day production processes. Vision Appliances are dispelling the once commonly held view that machine vision systems are complicated and require vision experts to deploy.
The top priorities listed by process engineers when evaluating a machine vision system are performance and ease-of-use. Performance is listed as the number one priority but ease-of-use is critical when operators have multiple duties and little time to support complex to use vision products on the factory floor. A simple interface that can be adjusted by the operator is the second most valued feature of a vision system and one that has proven to increase the acceptance of vision on the factory floor. Too often in the past vision systems have been purchased, installed and then turned off because they are too complicated, too difficult to setup and require an expert to maintain.
Why use Vision? Automated inspection vision systems allow manufacturers the confidence to know that all products coming off the production line are within tolerance, completely assembled, ready for shipment and meet the quality standards expected of today's manufacturers? Even if the production equipment is built for remarkable consistency there is still often a need to compile statistics for trends such as tool wear or as a tool for maintenance and servicing prediction.
[image] High speed bottle cap inspection
Machine Vision systems for automated inspection have been in use since the early 1980's as a way of offering precision, accuracy and repeatability in manufacturing environments. Many engineers in the past have been left with the impression that machine vision is too complex, expensive and difficult to implement and maintain. Vision Appliances are changing that impression markedly. As computer technology has developed, costs have decreased while performance has increased; manufacturers have seen the clear advantages of Vision Appliances in reducing costs, increasing accuracy and allowing their inspection and control processes to be automated. The final barrier has been ease-of-use and the new Vision Appliances are going a long way to remove this barrier.
Enter Vision Appliances
So what is a Vision Appliance [image]? Simply, it is a prepackaged, powerful machine vision solution with a simple and practical user interface.
Vision appliances are housed in a compact ruggedised enclosure that can be fitted on the factory floor and connected to one or more cameras mounted over the inspection point. Vision appliances allow a selection of cameras to be used so low resolution analog through to high resolution digital cameras can be used depending on the vision task. Vision appliances come preinstalled with intuitive software that is easily configured by process engineers or factory maintenance staff. Accessing your information can be done directly via a PC compliant Vision Appliance or via the internet as each vision appliance uses a web interface. This allows for real-time information to be accessed from a computer next to the manufacturing line or remotely.
The graphical user interface [image] found on Vision Appliances consists of a number of buttons and screens that force a user to step through a process of easy setup tasks. These setup tasks are broken down into a sequence where the operator firstly sets up the hardware, then sets up the inspection tasks and finally sets up rules for the result.
Setup hardware
The first step in setting up a vision appliance is to setup the hardware. When setting up hardware through the user interface Vision Appliances allow an operator to easily choose the camera operating parameters such as brightness, exposure, contrast, frame rate and others. These values are selected with a slide button so that users do not need to remember valid values to use.
The user interface [image] allows a user to easily connect to a sensor that triggers an inspection when the products are in place and to control a delay if required. The user interface will also allow the operator to setup a queue of the inspection results so that the Vision Appliance operates a reject mechanism to remove defective parts automatically and instantaneously when the defective part is in the reject position. These controls are all easily selected and manipulated by even an inexperienced operator.
Setup inspection
The next series of steps for setup lead the operator to locate inspection tools on the image to perform specific inspection tasks. This usually starts with the operator dropping a Landmark tool on the image to identify the exact position of the part to be inspected. Every inspection tool will position itself relative to this landmark automatically in case the part is not in the exact position or rotation every time.
Inspection tools [image] consist of tools to measure tolerances, presence and absence of features such as holes, to read and/or verify text or barcodes, to measure angles, to count features such as teeth of a gear, check for surface blemishes, to check label type codes and many more. The key to it all is that all of these inspection tools are easy to use, to locate and to set pass or fail tolerances for.
Setup decisions based on results
The final step in the setup procedure is to set rules for the results of the inspection task. For example a product may be inspected for three different features. The operator may choose to fail the part if two of the inspections fail and one passes. He may also choose to recycle the part if two inspections pass and only one fails, and so on. There are a myriad of combinations that can be easily setup.
All setup tasks in a Vision Appliance are described in common terms and do not require an engineer to understand them. They are broken down in to easy to understand and easy to control steps. The secret to the success of Vision Appliances is their ability to provide this ease of use without curtailing their performance and flexibility.
Connectivity
Vision Appliances communicate with factory SCADA systems, PLC's, management networks, robots and SPC equipment through standard industrial interfaces and protocols making them easy to deploy and connect. These include a web server for access via an Internet browser, several Ethernet and Serial RS232 protocols including Ethernet/IP and Modbus, direct support for several PLC's and other standards.
Other connectivity features of Vision Appliances include direct and opto-isolated digital inputs and outputs; 10/100/1000 Base T Ethernet IEEE 802.3 compliant with support for DHCP and static IP addressing; USB compliant ports; IEEE 1394 Firewire (400Mbps); RS-232 serial port; parallel port; Keyboard, mouse and VGA display; status LEDs.
Common problem - specific solutions
Vision Appliances are suitable for a great majority of inspections required by manufactures. Some common applications include:
Gauging
There are three key factors required to ensure your gauging vision appliance is a success: they are repeatability, accuracy and precision. Repeatability defines how close together repeated measurements are or refers to the variance of the measurements. Accuracy defines how close the measurements are to the true value. You can think of accuracy as the mean difference between the repeated measurements and the true value. Precision is the number of digits to which the measurement can be read.
Vision Appliances were designed to precisely and accurately measure parts. Examples include measuring critical dimensions on a part, verifying the presence and size of holes in a machined part, measuring diameter and circularity of holes, verifying the correct number and size of pills in a blister pack. They can take thousands of measurements per second, measure attributes such as circularity or surface markings that are hard or impossible for contact gauging or optical systems to handle. Vision Appliances use in-built techniques such as sub-pixel interpolation to increase the native resolution of the cameras and so increase measurement performance.
Label verification
Label verification vision appliances are specifically built to determine acceptable variations in the manufacturing process or environments which are easily captured using a "show and learn" technique. Once trained using an easy to setup training template the resulting model is used during inspection to determine if a label is within the learned specification.
Vision appliances have proven to be very popular in food and beverage industries where product presentation is essential. Vision appliances are used for rapidly checking the placement and quality of labels on bottles, boxes, cans and other ridged packaged goods. They have improved the efficiency of manufacturing lines and guaranteed a new level of quality and consistency in label verification.
With the ability to fully configure inspection (size and shape), "don't care" regions and selectable reference areas for label alignment, it is possible for vision appliances to be configured and deployed with the minimum of delay and at a low or no integration cost. Label verification vision appliances check labels for flags, tears and missing portions of labels. They can also detect a variety of defects including smears, spatter and contamination. Label verification vision appliances have built in compensation for lighting variations. This feature can prove useful in changing environments, or for factories that operate 24 hours a day, 7 days a week.
Multiple camera inspection appliances
Vision appliances have also been built to service manufacturers who need high performance vision systems. Designed for high speed applications with up to 60 parts per second these vision appliances have the capacity to process up to 3 independent image views simultaneously. High performance vision appliances often have a local display option for viewing runtime images and results while also offering remote access for product monitoring. By utilising the multiple camera inputs it is possible to inspect the front, back and top surfaces simultaneously. The VA50 [image] and VA51 fro iPD (www.goipd.com) are examples of vision appliances that have been developed specifically for multiple camera inspection.
What to look for when selecting a Vision Appliance -
- Ease of use to ensure correct application. Make sure that all personnel who will be using the machine vision product have been quickly trained in use of the appliance interface.
- Connectivity to other plant floor automation, and to information systems, business programs for ease of integration.
- Automatic compensation for lighting and control variations, and for lens and perspective distortion, to provide greater inspection reliability.
- Industrially hardened housing to withstand demanding plant floor environments.
- Input/Output capabilities so that the vision system can make decisions and control accordingly what happens when parts or products are accepted or rejected.
- Historical/archiving capabilities that enable users to review data accumulated from the vision system, identify the causes of any ongoing manufacturing problems and adjust them accordingly to increase quality in the most effective manner possible.
Where to from here
The Machine Vision industry is only just developing to a point where you don't have to spend your whole budget and shut down your line for days in order to get a vision system installed. With the expected decrease in costs, and rapid development of technology, vision appliances have found a way to handle many manufacturer's inspection problems without the risk that has been traditionally associated with machine vision projects.
Advanced robotics in tomorrow's factory
Addressing the production challenges of complexity, customisation and openness.
Cracking the nut: robotic automation at Freedom Fresh
SCARA robots from Shibaura Machine have found a place in helping to package macadamia nuts.
Food plant expansion sustained by central robotic palletising system
A palletising system with eight robotic cells has been installed at Unilever's food factory...