Machine Vision Advances Boost Productivity from Agriculture to Paper Mills
| By: Winn Hardin, Contributing Editor
pro•duc•tiv•i•ty [proh-duhk-tiv-i-tee], noun: The quality, state, or fact of being able to generate, create, enhance, or bring forth goods and services.
By definition, machine vision is reshaping the global manufacturing industry by helping companies create and generate products that could not have been manufactured in previous eras. Think of semiconductors as one example of a product that could not be designed or manufactured in large quantities without the assistance of automated microscopy, defect detection, and many more machine vision-enabled production machines.
Machine vision also enhances existing production processes, allowing trees to be more efficiently made into lumber, saving the environment while providing necessary raw materials, as well as cutting waste, rework and associated energy consumption from manufactured products ranging from cars to camcorders by identifying defects before a “bad” product passes further down a manufacturing line. All of these accomplishments make companies more efficient and productive, which in turn help them to employ more people worldwide.
Today, using automation and machine vision, companies are not forced to move all manufacturing to a region of the world with the lowest labor cost. Around the globe, machine vision is helping companies keep production next to customer markets, allowing them to better respond to changing customer demands, spread the wealth of employment, raise the standard of living across all regions, and reduce the negative environmental impact of shipping and logistics.
Will these benefits taper off, or continue to increase? As recent developments reveal, advances in machine vision technology that make machine vision systems more efficient directly impact the productivity of the customer in a mutually beneficial relationship that isn’t likely to disappear any time soon.
Hardware Improvements for Hard Returns
Machine vision automation allows companies to produce more, faster, and at less cost than manual operations. With this in mind, technological advances that make machine vision systems faster and more powerful convey productivity improvements to its users.
For example, recent work at the machine vision’s North American trade association, AIA, helps machine vision suppliers’ guidelines for improving system performance while retaining the compatibility that is critical to a successful global automation industry. Examples include the new USB3 Vision standard, which followed close on the heels of the first USB 3.0 standard release from the electronics industry; new Camera Link HS IP core solutions that allow hardware designers to use common FPGA cores to increase the bandwidth of one industry’s highest speed network standards; and new pixel format naming conventions that further increase compatibility among different hardware manufacturers through the GigE Vision standard.
Machine vision suppliers are not afraid to leverage developments from outside their industry and supporting associations, too. For example, NorPix (Montreal, Quebec, Canada) a provider of high-speed digital video recording solutions, recently introduced a discreet hardware module for high-speed, high-quality JPEG compression.
“We incorporated NVIDIA graphics cards into the new module, and we’ve shown that now a single computer can record video from 10 cameras simultaneously in high definition at 30 frames per second,” explains NorPix President Luc Nocente. “This allows you to record more cameras at once, or fewer cameras for longer periods of time without buying a huge computer. Our customers from pharmaceuticals to paper mills use these solutions to video high-speed processes, either before fielding a machine vision system or after fielding the system. In both cases, the video helps them to understand where defects originate and to design a solution. The ability to put multiple cameras around a piece of production equipment and slow down the process in video is crucial to helping our customers understand how they can improve their process, or improve the performance of an existing machine vision quality-assurance system.”
Recently, the Imaging Solutions Group (Fairport, New York), a specialist in custom camera and imaging solutions, delivers on-cameras FPGAs to enhance automated fluorescent systems used in medical screening and drug discovery as well as multi-spectral imaging systems for checking agricultural crop health.
“By utilizing unique functions found in ISG cameras, companies that make automated fluorescent microscopes and screening machines can operate faster and complete an automated inspection in a shorter period of time,” explains Kerry Van Iseghem, co-founder of ISG. “ISG LightWise smart cameras can integrate imaging algorithms to detect and count fluorescent results within an image and keep track of results without the need for PC involvement. Multi-Spectral Imaging Systems can be optimized via ISG LightWise Smart cameras. By utilizing on-board CPU and FPGA’s along with large image buffers, unique algorithms can be used to optimize Multi-Spectral Imaging Systems. For instance, agricultural applications can use multi-spectral techniques to determine moisture and plant growth quickly with smart cameras.”
Everything Doesn’t Have to Be Hard
Running a production line faster is one easy way to measure productivity enhancements, but machine vision technologies are spreading beyond traditional manufacturing industries.
“Of course, the default definition is ROI [return on investment], or increasing throughput,” explains Michael Cyros, chief commercial officer at Allied Vision Technologies (Newburyport, Massachusetts). “But there are many other points about productivity that I like to consider. For example, providing enhanced-quality images to the doctor, aiding sportscasters and TV viewers with clearer images combined with scene tracking, guiding a surgeon more accurately, helping ensure that a customer is happy with the exact fit of their new glasses, and on and on. So, ‘increased productivity’ has a much wider meaning for me than the traditional definitions of the measurement of productivity.
“The main point I wanted to make is that ‘productivity’ improvements aren’t just related to a quantifiable or clearly objective measurement. There is also a very large area for subjective interpretation of ‘productivity,’ and this is where the non-industrial application markets really benefit from ‘productivity’ enhancements,” Cyros said.
How does using machine vision for golf swing analysis increase productivity? “Is it helping a player become more ‘productive’ in their game to reach the hole in fewer shots?” Cyros asks. “No. The real answer there is it helps the sales people in the golf pro supply shops make more sales with less investment because the customer will already know exactly what they’d like to purchase to improve their game and overall experience. So, productivity must be measured also in very subjective, as well as objective, terms.”