About the --dpi option on the tesseract command line

[xl c]

Hello , I am a foreign tesseract novice. I need your help, the following is google translate .

In the tesseract command line, there is an option called "--dpi", I don't understand what this option does. 

I found that modifying the "--dpi" option will make the ocr result of the same image different. 

I understand the meaning of the dpi option (one inch The number of inner pixels). 

So, I have the following questions: 

1.  I pass in the same picture, the number of pixels is the same, why does changing the --dpi affect the ocr result? ( Is this dpi The option is to control the zoom of the input image? Will the same image have different ocr results under different zoom ratios? )

2 . How should I fill in the appropriate value for this option "--dpi" ?

3 . The height of the text in my image is only about 14px , Will this text size be too small, resulting in low ocr accuracy? I can try to use resize in opencv to enlarge the picture, but the picture clarity has not changed, will this improve the quality of ocr?