The Label Match Inspection Tool can quickly compare a live image from the camera to a saved master image. This is performed by detecting identical image features between the master image and the live image. The percentage of a match can be adjusted manually to fit any label matching needs.
What are features?
Features are uniquely identified characteristics within an image. The higher you set your Min. Feature Count, the closer your live image must match with your master image.
Simulated image of how Features are detected on a master image.
Simulated image of Label Match passing with >100 matching features.
Simulated image of Label Match failing with < 100 matching features.
Where do I set my feature matching threshold?
You can change the Min. Feature Count to define how closely you want a live image must match your master image. This will be unique for every use-case and require some adjustment to make sure your matches are as accurate as possible.
How do I optimize my Label Matching Inspection?
You can verify your feature matching sensitivity by performing a run on a set of saved images from a previous run.
- Have a target master image to check against.
- Have an album of previous runs saved. The larger the sample size of run examples will help identify the best settings to use.
Step 1: Set a Min. Feature Count and select Done. If this is the first time Label Match is being configured, the default value will be good baseline to test against your run.
Step 2: Navigate to Inspection Results and select Run.
Step 3: Compare expected results.
If too many images pass, increase the Min. Feature Count.
If not enough images pass, decrease the Min. Feature Count or try a different Master Image. Restart Step 1 if required for adjustment.
Tip: You can see a summary graph on the Results page which will help quickly identify trends in your inspection results. You can also click on the individual images to see exactly how many features were identified.