- 2D vision sensors
- Applications
2D Vision sensor applications
Pattern recognition | Object position | ||
Shape recognition | Number of objects | ||
Rotational position | Sorting tasks |
Object area | Object contrast | ||
Object width / height | Rotational position | ||
Inside / outside radius | Object position | ||
Roundness / rectangularity | Number of objects | ||
Number of holes | Sorting tasks |
Application examples
Metal injection molding quality
Objective: Customer wants to check the presence of 18 flats located on the outside of a gearbox. If a flat is missing then the injection molding process was not completed correctly. The customer also wants to know the location of a potentially missing flat.
Application details:
- 1500 mm requested mounting distance
- Part is 800 mm wide
- Static application - robot unloads gearbox from injection molding press, presents to the camera for inspection
- Trigger - the customer's PLC triggers the camera when ready for an inspection
- Integration - camera needs to provide a digital output once a good part is analyzed
Solution:
- IR camera with external IR ring light due to part size, distance to part and nearby window with sunlight
- Anchor contour on gearbox housing to allow for flight variations in the positioning when shown to the camera
- BLOB analysis with 18 ROIs
- Use Logic Layer to get digital output if all 18 ROIs are found
- Hardware used: O2D522 (wide angle) + O2D917 (external IR ring light) + E2D506 (ring light mount)
Bottle filling operation quality
Objective: Check liquid level in a glass bottle
Application details:
- 400 mm mounting distance
- Glass bottle height is 100 mm
- Static application
Solution:
- Anchor tracking contour on the bottle label show in image 1 to the right. The image for anchor tracking used an erosion filter to make black lettering appear dark versus the clear background
- The second image was taken using a dilation filter to make the liquid level line appear whiter versus reflections on the bottle and multiple ROIs are stacking on top of each other at a resolution acceptable for the customer
- The width of the liquid level line in each ROI is measured. The ROI with the widest object is where the liquid level is located as seen in the image below.
Packaging box quality
Objective: Find edge of box and find first fold. Measure the distance between the two. Be able to detect if this distance is off by 1 mm.
Application details:
- 450 mm camera mounting distance
- Box width is about 600 mm
- Dynamic application, 3 m/s line speed
- Triggered via the customer's PLC
Solution:
- Detected edge and fold on the box with the contour tool
- Used distance between two points block in the Logic Layer to get the distance
- Distance is compared to known good distance for final evaluation of pass / fail
Below is an example of calculating the distance between two points using our Logic Layer. This feature acts like a mini PLC where processing data can be done on the camera, not at the PLC.
Filter quality
Objective: Check the filter for open channels. Detect how many of them are filled with glue.
Application details:
- 750 mm mounting height
- Static application - filters are presented to the camera one at a time
- Camera is triggered by an external part presence sensors
- Required information from the camera:
- Number of channels found
- Number open
- Number closed
Solution:
- External red bar light is used to get best contrast on the channels
- Use contour model and draw ROI's to match the shape on the filter
- Number of expected channels is known per ROI
- Count number of channels found
- Compare to known amount on a good filter
Tapped hole detection
Objective: Check that the correct holes are tapped in a metal block.
Application details:
- Part is a 50x50x25 machined aluminum block
- Ability to mount camera at any angle
- Detect presence of a tap plus confirm the correct location was tapped
- Trigger via PLC interface
Solution:
- Mount infrared camera on an angle to catch direct reflections of the tapped aluminum inside the hole
- Pass/fail the application based on an area measurement using the BLOB analysis
- Object number tells you the location of the hole that passes the criteria