基本代码
create_text_model_reader ('auto', 'Universal_0-9A-Z_Rej.occ', TextModel)
* Optionally specify text properties
set_text_model_param (TextModel, ['min_char_height','polarity'], [30, 'dark_on_light'])
find_text (ImageResult, TextModel, TextResultID)
* Return character regions and corresponding classification results
get_text_object (Characters, TextResultID, 'all_lines')
get_text_result (TextResultID, 'class', Class)
局限性
基于传统分类器的识别不稳定，对于对比度不高的容易误判
