Cookies and biscuits are popular products that require quality control in various aspects before they are packaged and reach our plates. Consumers expect them to be baked to the right degree, have the correct color, consistent taste, be free from cracks or breaks, and have the correct quantity in the package. All these characteristics are examined using vision systems, ranging from simple to highly complex, depending on the specific inspection requirements.
Automated Vision Solutions and 2D/3D Imaging
Consumers continue to expect high quality. Those purchasing a package of biscuits desire consistent size, weight, and quality for each individual biscuit. The primary goal of the cookie manufacturer is to produce economically, utilizing resources and packaging wisely.
Today, advanced imaging software and 2D/3D imaging tools enable vision and measurement techniques that were previously unattainable.
In this article, we present several imaging and measurement options for cookies, which are also applicable to various other industries such as agrotech, pharmaceuticals, industry, defense, and more.
Inspecting 3,600 Cookies per Minute Using Vision Tools
One of the leading biscuit manufacturers benefits from an automated vision solution we implemented on their biscuit and cookie production line. The system is made of stainless steel, adhering to food industry standards.
A critical aspect that emerged is the need for a very fast and reliable response time that alerts to production process disruptions using data and learning from past experiences.
Additionally, cameras are used to measure the size and monitor the weight of the biscuits during the production process.
On the production line, 30 rows of baked biscuits pass under a moving conveyor, undergoing various inspections during this time—totaling 3,600 cookie inspections per minute.
The objective is to ensure all biscuits have the same height and weight.
The process becomes more complex when, later in the production line, two biscuits are combined into a sandwich with filling in between.
The goal of the monitoring and measurement process is to ensure each “cookie sandwich” has equal height and weight.
Each biscuit must be measured to calculate the required amount of filling to achieve the correct sandwich height. The filling is more expensive compared to the biscuit costs, so the idea is to minimize resource waste in creating the “cookie sandwich” and plan for the desired height and weight.
Biscuit size is easily controlled by the weight of raw materials and the pressures applied to the original dough sheet.
How Are Cookies Monitored Using Vision Systems?
The solution comprises three cameras that capture images of the biscuits from above and HALCON imaging software that analyzes the images to measure length and width, ensuring the biscuit is of the correct size.
A fourth camera operates in partial scan mode for higher speeds, capturing a laser line profile for the imaging software.
The software constructs a 3D image and obtains precise height measurements.
Using this process, an accuracy of ±0.17 mm is achieved for length, width, and thickness measurements.
The software’s customized interface includes a display of the average size measurements for the last 100 rows. These statistical data are stored by time and date for four years to meet customer traceability requirements. The operator can easily select from a dropdown menu to switch products. All this information is available on a dual-screen display located 100 meters away for monitoring by an engineer.
So far, we’ve discussed physical size inspections of cookies, but there are more complex parameters that can and should be examined.
How to Measure and Monitor Food Composition Using Vision Systems?
In many cookie products, the composition is also important—the relative proportion of various ingredients in the cookie or snack. This is crucial not only for the customer but also for the manufacturer to accurately manage the costs of snacks that depend on their relative composition. For example, in snacks like these:
Visual Imaging vs. Laboratory Testing
Several technological methods exist for monitoring the quality of food production processes. Unfortunately, many require laboratory analysis, which is labor-intensive and can be very slow, taking hours, days, or weeks from sampling to obtaining results. Additionally, the sample is often destroyed during testing.
Advantages of Hyperspectral Imaging (HSI) Combining Spectroscopy with Imaging Capability
Hyperspectral imaging (HSI) combines spectroscopy with imaging capabilities. This technique allows pixel-level analysis of complex or heterogeneous food.
As an imaging technique, spectral imaging does not require contact with the food and does not destroy or contaminate the sample.
With a good calibration model, analysis results can be provided in real-time.
For instance, by examining a cookie, spectral imaging can be applied to detect moisture levels in composite products, which are traditionally more challenging than single-component products. As shown in the following image:
Not only moisture levels in cookies and snacks can be observed but also fat content, sugar levels, various ingredients, etc., as seen in the following image:
Using Spectral Imaging for Color Measurement and Analysis
In other cases, we were asked to detect the baking level of cookies or buns. In this case, there is a direct correlation between the product’s color level and its baking level. Color reading is not trivial unless one is aware of the correct lighting and proper color analysis.
At AsioVision, we have accumulated extensive experience in this field. Sometimes, this can be done using standard color cameras, but in many cases, that is insufficient.
To demonstrate how spectral imaging can be used for color measurement and analysis, we measured the color of buns, each baked for a different duration. The baking time in the recipe was 5–6 minutes. The lightest-colored bun was baked for 3 minutes, with each subsequent bun baked one minute longer than the previous. Thus, the darkest and clearly burnt bun was baked for 8 minutes.
After analyzing the results, we obtained precise color parameter analyses that allow for both closed-loop feedback to the production line and quality control of already baked products. These results are obtained in an online setting.
