The text on the video is not recognized in LabView on the MyRio module, what should I do?

Question

Machine learning

Good afternoon habrovtsy!

At the moment I am working with the NI MyRio module, and the LabView software platform. There was a problem while performing a task to recognize text in an image. In the block diagram in the Vision Assistant block, I converted the image to binary, removed defects, taught me to recognize text, everything works without problems:
I could not add code as a picture, LABVIEW is a graphic language!


But the video is no longer recognized:


The code in the flowchart:


Googled and tried a lot, nothing really helps.

Answer 1

Isn't it recognized on the video? Well, if you remove the while loop and run the tool once? If so, then perhaps Vision Acquisition continues to capture images even while the recognition is working (and it takes some time). Try to make a copy of the picture before sending it for recognition:

Also check the Error Out error output of the Vision Assistant (it is connected to the indicator in the screenshot above). If errors occur, you will see them.
Well, you must be absolutely sure that MachineEducation.abc is exactly the file that you trained. I still have doubts about the Cyrillic alphabet on the way. Try moving the project to a folder without Cyrillic characters, like С:\MyExperiment. Although if LabVIEW *.abc cannot load, then the error output will tell you about it.