PyOCR cannot detect clearly?

[wood_6636]

As you can see in the down below captures, the image cannot be nicely detected to the final result, could anyone tell me how to have a better recognition by using PyOCR ?

Image Source:


Final Detected image:

---

def codeocr():
    #My Code
    img=cv2.imread('/Users/Woodylin/Desktop/Python Learnings/Bank_Fubon_Mort_Scrapping/img_source.png')
    dst=cv2.fastNlMeansDenoisingColored(img,None,10,10,7,21)
    ret,thresh=cv2.threshold(dst,127,255,cv2.THRESH_BINARY_INV) 
  
 

   
    cv2.imwrite("/Users/Woodylin/Desktop/Python Learnings/Bank_Fubon_Mort_Scrapping/final.png",thresh)

    from PIL import Image
    import sys
    import pyocr
    import pyocr.builders

    tools = pyocr.get_available_tools()
    if len(tools) == 0:
        print("No OCR tool found")
        sys.exit(1)
    tool = tools[0]

    result = tool.image_to_string(
        Image.open('/Users/Woodylin/Desktop/Python Learnings/Bank_Fubon_Mort_Scrapping/final.png'),
        builder=pyocr.builders.TextBuilder()
    )

    return result