Insights from industry

Vision Inspection for Quality Control in Food Manufacturing

insights from industryJohn LawrenceFounderSightline Process Control

In this interview, AZoM talks to John Lawrence, founder of Sightline Process Control, about how vision inspection plays a huge role in quality control within food manufacturing.

Could you start by giving our readers an introduction to Sightline and its parent company KPM?

KPM is our parent company, headquartered in Milford, Boston. KPM’s goal was to put together a team of companies to provide comprehensive measurement systems and analytical capabilities for food manufacturers. There are six brands in total and each company has a different specialty. 

Sightline focuses on vision inspection, while other KPM companies focus on infrared inspection of dough, moisture measurement, and more. The theme is the same, however, and this whole family of companies provides measurement solutions to food manufacturers. 

The footprint is very large, and this group of companies has over 200 distributors, dozens of salespeople, and installations in over a hundred companies across 18 countries.

When we founded Sightline in 2009, a lot of vision systems were essentially proprietary black boxes. No one really understood how they worked and there was a lot of proprietary hardware, with most systems designed solely for one specific application. 

Our goal at Sightline was to create a new type of business system that was flexible, generic, and that you could use with any kind of product to measure anything using the same software and hardware tools. Our goal is to be open, non-proprietary, and highly flexible. 

We shipped our first system in 2009 and that customer is actually still our biggest client today. We joined KPM in 2019 to expand our footprint. Initially, we were very focused on the U.S. and Canada and KPM gave us the chance to become part of a bigger family and develop a more global reach. 

Since then, our target markets have essentially remained the same; we specialize in food products, bakery, poultry, meat, and more recently packaged foods. We can now inspect food inside packages or inspect the packages themselves, all with the same software.

What are some of the most common issues faced by Sightline’s customers?

Labor is one of the main issues faced by every company. Finding good people, training, and retaining those people is a universal challenge. 

After working with lots of Fortune 500 companies, I have also realized that for these companies and indeed every company, the brand is everything. A brand takes years to develop and it is important to make sure that that brand is well received in the market. 

It only takes seconds for that brand to dissolve when faced with some sort of catastrophic quality problem. Whether it is a catastrophic issue or erosion over time, companies are looking to make sure their brand is protected. 

Another key issue is compliance regulation. There are still a lot of things written down on paper and moving from that practice to digitization is a goal for a lot of companies. 

Consumers seem to be more fickle now and they are always looking for new and improved products. Changing and launching new products is a particular challenge for both R&D and manufacturing. 

Sightline Vision Inspection Systems

The ability to quantify and work with rigid numeric specifications for a new product and then port these to a manufacturing line and immediately start making these products to specification is very important for any company.

Finally, there is a big push towards Industry 4.0 and systems with a range of connected devices and sensors designed to measure digital values in real-time. 

Devices take these inputs and automatically make changes based on those measurements. Industry 4.0 is focused on connecting digital sensors with digital devices and using these tools to make manufacturing better and more automated.

How important is quality control in managing customer expectations and minimizing customer complaints?

In food manufacturing, quality control issues are most often manifested in customer complaints. For every individual customer complaint, there are probably 20 or 30 customers who experienced the same defect but did not complain. 

This is a problem because that defect has already been allowed to get out into the field, making it more difficult to measure effectively. 

Product waste is another important measurement. Product waste is often driven by quality parameters that have gone out of control, requiring products to be disposed of or put into poorly fitting packaging, creating more waste. 

These issues will lead to reduced sales over time because if consumers associate a brand with a lower quality product, they will start switching to the competitors. 

Companies often try to address these issues by increasing staffing levels, hiring, or redeploying staff to manually inspect their products. This is expensive, slow, and inaccurate. 

More so, it does not produce useful data because any reporting or documentation is essentially nonexistent; people are just doing their best to manually remove faulty products. 

Manual measurements are often meaningless because there is more variation between operators than the specification limit range and there is not enough data to make informed decisions around process quality. 

How can vision inspection help address these issues?

As a methodology, vision inspection involves taking images of every single product, analyzing those images, and extracting as many measurements as are mathematically possible. 

These measurements are not in pixels or RGB; they are in engineering units, calibrated units, inches, and millimeters, etc. This process is fairly straightforward in a lab, but on a production line where there are hundreds of rapidly moving objects per second, having to deal with different sizes, shapes or products can be challenging. 

The payback is great, however, because data is available for every single object manufactured. The data is stored, and it is repeatable, reliable, and accurate. 

Once that data is in place, it is possible to measure things that are virtually impossible to measure by hand, for example, the number of sesame seeds on a hamburger bun or the fat percentage on a piece of meat. 

Image Credit: Sightline Process Control

This removes the need for the additional labor I mentioned earlier. This approach also helps reduce waste in two ways: if you can provide operators with real-time data where they can make decisions before issues arise, it is possible to avoid product waste. 

Secondly, if you automatically remove products that you know are going to cause issues with packaging based on their size or shape, you can eliminate the source of waste downstream. Proper process control and quality control can stop defective products from ever getting out into the field, therefore helping to maintain brand equity.

What role does real-time and historical data play in quality and process control?

Access to real-time and historical data goes a long way to improving product quality and process efficiency. The goal is to achieve 100% inspection - measuring every single product and analyzing that data is a big part of understanding processes; for example, understanding why the first shift is always better than the third shift by looking at historical data, drilling down and reaching the right conclusions. 

How important is it to ensure that vision inspection systems use the right kind of cameras, lighting, and computing hardware?

It is important to note that there are many different types of cameras, color cameras, 3D cameras, thermal cameras, etc. Each camera is a source of images and those images need to be acquired in real-time, processed and turned into measurements. 

Once you have the right cameras, these have to be deployed into a factory, potentially moving 30 objects a second on a meter-wide conveyor.

There is a lot of infrastructures required to make sure any vision inspection system is turn-key for the customer and fits into the customer’s existing production line. The system – and its component cameras – must be able to function in the wet, dry, dusty, oily, and other challenging environments found in a lot of food plants. 

Cameras must also be able to survive any sanitation processes as well, with many of these using harsh chemicals to clean and sanitize equipment. 

Cameras have changed a lot over the past ten years. They have become smaller, faster, less expensive, and now offer higher resolutions than ever before. We are currently working with cameras that can measure kernels of wheat, down to two or three millimeters in size, and working with cameras able to resolve images as low as 40 microns.

Most importantly, they come in a variety of types now, for example, a color camera, a 3D camera, a thermal camera, an infrared camera, etc. 

Benchtop Inspection System. Image Credit: Sightline Process Control

Essentially, each of these types of cameras has the same footprint and the same commands, meaning that it is much easier today to mix and match different types of cameras and create a system that's collecting data from multiple sources. 

LED lighting has provided a number of key benefits for vision inspection, using much less power, offering increased brightness across a wide temperature range with no need to use glass. LED lights use 25% of the power of older lights, meaning it is much easier to manage the heat generated.

From a computing perspective, a typical i7 has more than enough processing power to accommodate all the mathematics involved – there is no need to buy proprietary or specialized hardware to do the heavy lifting. 

The image processing algorithms we use have been developed and optimized over many years and these are focused primarily on new measurements and new techniques for food products that are not currently being done by generic software packages. 

Overall, component costs have really come down in price and systems are generally much less expensive than they were 10 or 20 years ago.

What sort of things can be measured using vision inspection?

We take images and extract every mathematical measurement we can from them, meaning that regardless of the specific measurement a customer requires, we can usually integrate this into our system. 

These measurements typically come from the products themselves, but they can also come from the production line; for example, measurements of throughput, downtime, uptime, or changes over time – the methodology remains the same.

Extracted images and measurements are entered into our library. Our system, in essence, is a library of over 200 measurements developed over the past ten years. This library is continually expanding – almost every new customer requires a new measurement, so we develop it and put it in our standard package, so every customer has access to it.

Measurements must be rapidly performed in real-time, at production line speeds. One of the interesting things that we do is use multiple images to create a single measurement. 

All our cameras and the images that come from them are co-registered and overlaid onto one another. For example, if you had a thermal camera, you could find the hottest spot on an object and because that image is overlayed with a color image, you could find the height or the color of that hotspot. 

While every camera is independent, we fuse the images together, so if you find something of interest in one image, you can extract an appropriate measurement from another camera - we call this ‘image fusion.’

Our key goal is flexibility, you can have as many cameras as you require, and this can be expanded as necessary. If you start with a system that only has one camera, you can easily add additional cameras in the future. 

The core of our system is the measurement software, but we recognized early on that it was important to provide customers with a way of looking at the mountains of data that we generate. 

We have created tools and software to make sure that customers can easily access and analyze that data and every time we add a new feature or a new measurement, this is rolled out to every customer. 

How can these measurements and technologies be applied to specific customer applications in the field?

To date, we have focused our efforts on four primary application groups: food inspection, heat seal inspection, and package inspection. 

Within these groups, we have experience working with a wide range of products ranging from common food items to some other more unique products like pet products, packaged goods, and even letter mail. In 2019 over seven billion individual objects were measured using Sightline equipment.

By working with so many different products, we have built an extensive measurement library that can be applied to any product. Some of our most popular measurements include geometric measurements, which include everything needed to quantify the size and shape of products. 

This includes 3D measurements like the height, volume, and slopes of the surface and 2D measurements like length, width, and diameter. 

Many of these measurements can be presented in different ways; for example, height data can be presented as a minimum, a maximum, peak height, an average height, or even measurements at specific geometric locations, such as the edge, or the center. 

This flexibility ensures that our measurements always align with how the customer is currently measuring products, aligning with existing practices. 

I mentioned image fusion earlier and our equipment’s ability to use information from any one of our images to define a specific region of interest. We can analyze the same region of interest in another associated image. 

One of the best examples of where this has been applied comes from a customer who used our equipment to analyze steaks.

The customer used color data to separate an image into fat and lean meat regions, independently measuring each of these regions within the height data and providing thickness measurements for each so that the customer could make those comparisons within that segmented data. 

Color is another of our most popular measurements and there are a variety of color measurements available, including the color of the whole object, the color of specific regions, even the color of things like toppings or visual defects such as blisters and blemishes. 

Our color measurements are calibrated to industry-standard measurement units. They are reliable and repeatable, and this can provide significant improvements in consistent product quality when compared with traditional methods like color comparison charts, which are often subjective and dependent on operator interpretations.

Over-Line Inspection System.

Over-Line Inspection System. Image Credit: Sightline Process Control

Another key measurement category is something we call ‘blobs.’ Blobs allow users to identify specific features on their products, such as toppings, black spots, fat versus meat, etc., and obtain measurements of these specific features. 

Essentially, blobs are objects within an object that we can separately measure. Any feature that can be identified on the product, whether by color or specific height attributes, can be measured and quantified using our blob tools. 

Measurements available to any identified blobs include geometric and color measurements, as well as some specific measurements related to the count, total area, and distribution of that feature.

To provide some specific examples of our system in operation, we can look at baked goods. The most important measurements in these applications are geometric measurements like length, width, and diameter. 

When products have interior holes, such as bagels, donuts, or pretzels, we are also able to provide measurements of the size and shape of these holes. 

Height and volume are also very important measurements in this type of application and color readings can be absolutely critical for these types of products, as these will be directly linked to the process and to the overall quality of the product. 

Availability of color data to operators can improve the response time to any related issues and that in turn reduces product waste. 

Where our system really shines is when we start introducing blob measurements. For products like baguettes, we can identify splits as blobs, allowing us to individually analyze the color of the splits versus the rest of the product. 

We can count those splits and independently measure their size and shape in order to make sure that the baguette is meeting product quality specifications. 

For products with toppings such as seeds, we can apply our regular count and total coverage measurements. We can also go one step further and quantify the distribution of the seeds, ensuring that not only is the right number of sesame seeds added but that these are spread evenly around the product. 

Many of our customers also use blobs to identify defects in the products, for example, black spots, burns, or even holes right through the products.

In-Line Inspection System. Image Credit: Sightline Process Control

Poultry has been a great opportunity for us to showcase our products and our ability to adapt the equipment based on customer needs. When we entered the poultry market, we realized there would be several specific measurements that would be of interest to our customers. 

These tools include some special geometric measurements allowing our measurements to align with existing measurement practices in the facilities and a new tool to find specific key features. 

In one example application, our customers were very concerned with locating the keel of the chicken breast, which is the joint between the two chicken breasts. Using our new tools, we were able to share the location and orientation of that heel with operators and downstream equipment. 

We also developed a special coverage tool designed to show customers exactly how their products would fit onto a bun or any other template shape. All of these tools are being used to optimize product quality while providing customers with the best possible experience.

One final example involved packaged products. Our system features a wide range of available package inspection tools, including pattern matching, which can be used to ensure the correct labels are being applied onto the package, and text recognition, which can be used to ensure that things like date codes, lock codes, and any other texts are being printed correctly. 

We also have barcode verification tools to ensure that barcodes are readable and seal inspection tools. Seal inspection, in particular, has been a key area of interest with many of our customers. We currently offer two different techniques that can be used to analyze the seal of any packaging.

The first technique is a visual analysis of the seal, whereby we take a color image and we look at the seal to see if there is any foreign material or debris trapped inside that may impact its integrity. 

The second method introduces a thermal camera to our system shortly after the sealing equipment. This allows us to verify the seal integrity by confirming that the seal has reached the correct temperature and by visually checking that there are no breaks in the seal.

Can you give our readers some examples of the kinds of equipment provided by Sightline?

Sightline offers a range of products to suit any customer’s needs. All our systems come with the same software, so no matter which level of system you are interested in, you will still have full access to all our available measurement tools. 

Our first level offering is our benchtop systems. These systems are designed to be located near the production line or in a lab environment and they are ideal for improving upon existing QA processes. 

They still rely on operators to manually collect and run samples through the equipment, but our benchtop units provide faster and more repeatable measurements than traditional methods. 

Our next offering is our over-line units. Over-line units are an ideal way to introduce automated inspection to a facility and production line. These units can be installed over existing conveyors and come with everything required to start analyzing 100% of your production. 

Because this equipment can be installed over existing conveyors, these are generally very simple and easy to integrate into existing processes. 

Our flagship offering is our turn-key inline systems. Like the over-line units, these inline units come with a full vision solution, but they can also be configured to include conveyors and rejection or sorting mechanisms. 

The special design of our conveyors ensures excellent imaging performance, even allowing the installation of additional bottom and side cameras. This opens new options for even greater analysis of your products. 

Both our inline and over-line systems are available in a wide range of conveyor widths, ranging from 300 millimeters up to two meters wide. 

Variations on our designs already exist for a range of challenging site conditions, for example, working in high and low temperatures, washed down environments, or industries with specific sanitation requirements. Our equipment can be adapted to meet a wide range of needs.

Can you tell our readers more about the software package provided with Sightline’s solutions?

Our software options can be broken down into three main categories: set up tools, inspection tools, and reporting tools. Set up tools include the Measure Toolbox program – a toolbox containing everything a user might need to configure their system, packaged inside an easy-to-use interface. 

This toolbox allows users to modify product specifications, add new products, configure different measurements, and much more. 

In terms of inspection tools, we have our core software, Measura Inspection. Inspection is the primary user interface for our equipment and is responsible for translating camera images into usable data. 

All of our systems feature touch screen interfaces through which users can select which products to run while being given valuable feedback in the form of measurement data, trend graphs, and Pareto charts. 

Sightline measura® Software

Our reporting tools are comprised of two offerings. The Measura Dashboard software is usually displayed on large screens within the production facility to provide real-time feedback performance to users on site. 

Meanwhile, the Measura Analytics software is a fully featured reporting tool, which can be used to access the database, generate formatted reports, or export raw data. All reports are highly configurable and can be set up to run automatically and emailed to your team. 

We also have a variety of methods for linking our data to existing data collection systems and our SQL database via SPC or other technologies. 

In what ways do Sightline’s product offerings offer a return on investment to users? Can you give us some examples of how this can be measured?

Return on investment, or payback, is an important consideration that we all have to deal with whenever we're buying capital equipment. With the vision inspection system, because you can replace manual inspection, there is a huge payback from a labor perspective. 

Instead of paying people for three shifts a day, seven days a week, for example, a vision inspection system can replace that with one capital expenditure. 

Waste is another form of payback. You can eliminate the waste generated by processing faults by staying on track, meeting specifications, monitoring trend graphs and real-time data and adjusting as required. 

This will reduce process waste and packaging waste. Being able to automatically remove objects which could potentially jam up packaging is an ideal means of reducing waste because they are never going to get to the packaging processes.

Vision systems are the future. We can measure every single product's top, bottom, color size, shape, and temperature at up to a hundred objects per second. We can quantify products, track products, measure products and reject products in real-time as they are being manufactured. 

These things can all be done simultaneously with measurements that are not possible without a vision inspection system. This goes a long way to reducing costs, as well as protecting your brand as defective products will not reach customers.

We can add virtually any new measurement to our systems, as long as it is mathematically and physically possible. We have done these hundreds of times around the world and we have a long history of success with any kind of food product that you can imagine. 

About John Lawrence

John LawrenceJohn Lawrence, the founder of Sightline Process Control, has been involved in machine vision and automated testing for over 25 years. 
John began his career as a Production Engineer, then worked as an Automation Engineer, an Applications Engineer, Operations Manager, and then as a Vice President of Sales and Marketing for a machine vision company. 
In 2009, John founded Sightline with a mission to change the way people think about vision systems and automated inspection. His vision was to create a better, more affordable inspection system that was not another proprietary ‘black box’. These systems would be designed using non-proprietary components, open-source software, and flexible analysis tools to ensure longevity and avoid obsolescence. 
This approach has proven successful: Sightline now has systems installed in 20+ countries globally, including systems in some of the largest manufacturing facilities in the world, serving some of the world’s top foodservice brands. 
Today, John continues to be actively involved with product development, application testing, and even programming from time-to-time. He is still excited by the challenge of new applications for vision inspection and automation. 

Disclaimer: The views expressed here are those of the interviewee and do not necessarily represent the views of Limited (T/A) AZoNetwork, the owner and operator of this website. This disclaimer forms part of the Terms and Conditions of use of this website.


Please use one of the following formats to cite this article in your essay, paper or report:

  • APA

    Sightline Process Control Inc.. (2021, March 29). Vision Inspection for Quality Control in Food Manufacturing. AZoM. Retrieved on April 22, 2021 from

  • MLA

    Sightline Process Control Inc.. "Vision Inspection for Quality Control in Food Manufacturing". AZoM. 22 April 2021. <>.

  • Chicago

    Sightline Process Control Inc.. "Vision Inspection for Quality Control in Food Manufacturing". AZoM. (accessed April 22, 2021).

  • Harvard

    Sightline Process Control Inc.. 2021. Vision Inspection for Quality Control in Food Manufacturing. AZoM, viewed 22 April 2021,

Tell Us What You Think

Do you have a review, update or anything you would like to add to this article?

Leave your feedback