The identified string is repeated
Case 1: Detection error, the same string is detected twice.
For example, "12 plan"
is recognized as: "12" "2 plan". This can be crossed by the x coordinate and the last character of the first result is the same as the second one. Judgment if the first characters of the results are the same
Case 2: Recognition error. 11 is recognized as 111.
There may be too much blank area around the text. The text area is cropped through the outline. Setting det to False can improve the recognition rate.
The string in this line is too long and the recognition is incomplete.
The recognition result is empty
The aspect ratio of the resized image is 1:10, which can improve the detection and recognition rate
. Appropriate padding can increase the pixels on the top and bottom of the image.
Only part of the string detected
Same as above, but some text may still be missed and not recognized.
Poor single digit recognition rate
Crop the digital area through the outline. Setting det to False can improve the recognition rate.