Performance improvement of character recognition in industrial applications using prior knowledge for more reliable segmentation

Highlights：

• The algorithms are developed to be real-time capable and robust to noise or illumination changes.

• We show how to improve classification accuracy and speed using prior knowledge, e.g., number of lines or characters, in character segmentation.

• Prior knowledge in segmentation also improves the reliability of the segmentation algorithm.

• Comparison of different combinations of features and classifiers to show differences in accuracy and processing time.