Optical character recognition (OCR) is used in conjunction with deep learning AI to verify correct text, wire labels and components in a manufacturing assembly. On the setup tab the user can adjust the following parameters:
- Enable Character Filter: This filter removes all non alpha numeric characters (characters that are not letters and numbers) from OCR validation.
- Enable UPPERCASE Filter: This filter enables UPPERCASE text comparison during OCR validation. Sometimes industrial devices are marked with different font size upper case letters than can confuse OCR.
- Text Location Accuracy: This parameter sets text placement location accuracy by using the size of the OCR bounding box. The bounding box size is multiplied by the number selected. 1x means the OCR bounding box of the assembly must be within the initial trained position. 10x means the OCR bounding box must be within ten times the size of the bounding box.
- Minimum Character Size: This parameter sets the minimum number of width and height pixels for a valid OCR string.