Beer Bottle Case Study
*Note: This case study is not an actual evaluation performed for a customer. Instead, we drew on various food and beverage applications CCS has previously solved.
About the Application
Beer bottle inspections using machine vision are especially demanding: they need to be adaptable to handle differences in packaging, whether that is bottle shape or label designs, and the bottle is naturally reflective, frequently leading to hot spots.
When customers want to determine the most reliable solution for the vision system, they can request a free lighting evaluation. Every evaluation begins with a customer filling out our application evaluation form to help our engineers recreate the inspection environment to the best of their ability.
Here are the conditions for this particular application:
| Please describe your part |
Variety pack of Samuel Adams beer bottles.
Glass is dark in color, labels will change in color (light blue, dark blue, red, brown)
|
| Does the part appearance vary? | Glass will stay the same, labels will change (light blue, dark blue, red, brown) and notch placement will change |
| What features need to be detected? |
Notch placement
Label color
Read barcode
|
| Camera Model | 5 MP |
| Camera Type | Color |
| Speed of the part? |
50,000pph |
| Desired exposure time or shutter speed | NA |
| How is the part presented to the camera? |
The labels will all be presented at the same location/orientation.
The inspection is taking place just after label placement.
They will be held upright in a fixture from the top of the bottle.
|
| Field of View (FOV) | 110cm X 7.5cm |
| Min. AND Max. Working Distance |
None |
| Can a back light be used? | No |
| How large can the light be without causing interference? | Light must be less than 200mm |
| Other Useful Information | NA |
The evaluation form listed the speed of the part as 50,000 parts per hour but not an exposure time. We will need to calculate the ideal exposure time to choose a light that can flash at the needed intensity. The following formula is used:
Based on our calculations, the exposure time needs to be about 30 μs.
The Challenge
High-speed inspections are challenging because the system needs to work quickly while maintaining high accuracy. It can require more technically advanced lighting solutions. Otherwise, the image will be blurred or too dark, obscuring details and making the application unreliable. This application will need an exposure time of 30 μs to image the beer bottle reliably.
Key Challenges
High-speed inspection
Space restriction
When solving applications with space and size restrictions, the primary challenge is to select and configure lighting and vision components that fit within the limited available space while still achieving optimal performance. Compact environments may restrict the placement of lights, limiting the angles and distances needed for ideal illumination and imaging.
Additionally, smaller form factors often demand more precise alignment and integration, increasing the complexity of setup and calibration. Overcoming these constraints can require specialized lighting to ensure the vision system operates effectively despite spatial limitations.
Designing vision systems without enough space is one of the most common pitfalls in machine vision lighting. To learn more about the mistakes engineers often make when choosing lights, we recommend watching our webinar Common Pitfalls in Machine Vision Lighting.
The Solution
At a high level, each evaluation has five phases: review the workpiece samples provided, recreate the application setup, test whether to use bright field or dark field illumination, test for the exact form factor to use, and then finally the wavelength to use.
Here’s how the Application Engineer went through this process for this particular application:
Step 2: Recreate the application setup
The inspection takes place just after label placement, and they will be held upright in a fixture from the top of the bottle. This makes testing easier because the Engineer can set the bottle in place without accounting for any obstructions like a conveyor belt. If there had been a fixture that could potentially disrupt the field of view, we might have requested the customer to send us one to verify the application’s reliability.
When doing high-speed applications, we have to ensure we can illuminate the area within the ideal exposure time. However, we recommend making it easier at the start by treating it like it is not a high-speed application. This is because you have more lighting options if the application can be solved with standard lights on overdrive compared to strobe-only lights. It is also easier to start with standard lights because you can test all lights in continuous mode. Otherwise, it will take extra time to set the current rating and strobe time for each light you test. Once you’ve determined the best lighting solution (including the light, light working distance, angle, etc.), you can test what version of intensity control is required to properly illuminate the high-speed image, whether it's overdriving the light or using a dedicated strobe overdrive light. So when you are setting up your testing station, start with a slower shutter speed.
“To make it easy, start by treating it like it’s not a high-speed application, using a slower shutter speed.”
Step 1: Review all samples
We start by going through the samples we've received from the customer. In this application, all the samples provided include the application-cricitical target information: the notch, barcode, and label color. Moreover, there is one of each possible label color, and the notches are in different locations, so there is a good amount of sample variation to test how robust the system will be.
If there were defects to inspect for, we may have needed more samples. Our pharmaceutical case study goes into the importance of having enough sample variety for such applications.
Step 3: Test bright field vs. dark field
We typically use our HPR2 ring light to decide if bright field or dark field works best. Knowing that we’re reading barcodes and checking notches, even without testing with the HPR2, it is clear the application needs bright field instead of dark field.
If you’d like to better understand testing bright field vs. dark field and determining which is better for your application, we recommend watching our webinar Get Inside the Mind of a Lighting Expert.
Step 4: Determine form factor
The Engineer decided to start with an LFXV flat dome light because it typically works best on rounded items. But since it required the light to be extremely close to the sample and there was a hot spot, she went on to try several other form factors:
LDR2 ring light (without polarizer)
LDL2 bar light (with/out a polarizer)
The LDL2, with a polarizer, set vertically at a slight angle was the best solution, so then she moved on to determine how to reach the 30 μs exposure time. Here is the general process to follow when figuring out the necessary intensity control:
Set the exposure time of the camera to the ideal time calculated previously.
Overdrive light at its maximum intensity (with respect to the duty cycle and maximum intensity rating).
Set the strobe time of the light to 85-90% shorter than the exposure time. (This is doing our due diligence to make sure 100% of the flash is captured within the exposure time.)
If the image is not bright enough (like in this application), switch to a dedicated strobe light and set it up the same way you did for Step 3. (If you don’t have the right model in your lab, request a demo from a lighting manufacturer or send your sample to them for testing.)
In most cases, your intensity will be bright enough if you step up to a strobe light. If it’s still not enough light, you need to consider slowing your inspection speed or consulting a lighting manufacturer about a custom strobe light built to reach your specific intensity.
Confirm the initial solution works with the new intensity control method (overdrive or dedicated strobe light) and adjust as needed.
Regarding Step 6, the solution might need to be adjusted if you change to a dedicated strobe light. For example, a strobe-only version may not be available in the initial solution’s wavelength, size, or form factor. If this happens, adjustments such as moving the position or angle of the light are common, but sometimes you have to find a new solution with the products available. A custom high-strobe light can be considered as well.
For this application, there wasn’t an exact match for the standard light used in the initial solution, so the Engineer tested with the standard strobe lights available and switched to the size that achieved the best result. She also found that the polarizer blocked too much light, so the image was not bright enough for the darker sample. She removed the polarizers, but the light had to be placed horizontally and angled down to remove the glare that was now present. This created a uniform image with the right intensity.
Also in this application, our Engineer set the strobe time to 30 μs for the majority of the inspection but then determined the shortest exposure time possible before the image became too dark. The image begins to fail at 20 μs, so we know the solution she chose at 27μs will work (90% of 30μs).
Watch the engineer talk through the process of testing which intensity control works for the application:
Step 5: Determine wavelength
For wavelength, we know it will be white because we have a color camera, simplifying this part of the test. Lights dedicated for strobe overdrive tend to be available in red and white only—other wavelengths would be a custom. If you were to overdrive a general-purpose light, you’d have more wavelength options (blue, green, infrared, etc.).
The Result
The total time for the evaluation process depends on the application and the number of active evaluations at a given time. Our engineers can typically provide a solution within two weeks after receiving the sample.
This application evaluation, performed in about 1.5 hours, determined the following about the best lighting solution:
(1) LDL-PF-102X30SW at an exposure time of 27μs can perform all inspections
Our engineers summarize the test results in an evaluation report. They include details about the specific lights, components, and configuration they used to get the best image (see example here).
Sample image from the report comparing with polarizers (left) and without (right) at 30μs strobe time.These reports help prove out lighting needs and concepts. Customers can use the results to determine feasibility, whether to adjust the project scope, and ultimately build a more accurate vision system.
No application is the same. No challenge is the same.
So no solution is the same.
These light models won’t necessarily solve another blister pack application. We always recommend testing yourself or getting an evaluation done for your application’s unique needs.