将此车牌图像翻译成文本

Question

I have a picture:

I want to turn the car plate into text.

After I do crop of the plates with python-opencv :

pytesseract.image_to_string(bbox, lang='eng', config='--psm 6') returns me: "TH

My code for cropping:

coordinates = [[0.6703869047619048, 0.39855072463768115],
               [0.7672123015873016, 0.39855072463768115],
               [0.7672123015873016, 0.47681159420289854],
               [0.6703869047619048, 0.47681159420289854],
               [0.6703869047619048, 0.39855072463768115]]
img = cv.imread(pic)
height, width, channels = img.shape
bbox_coordinates = [[int(coordinate[0] * width), int(coordinate[1] * height)] for coordinate in coordinates]

Then I turn it for better quality into greyscale:
bbox = cv.cvtColor(bbox, cv.COLOR_BGR2GRAY)

pytesseract.image_to_string(bbox, lang='eng', config='--psm 6') returns: To

Is there a way to get WA 6728 M ? (with or without usage of pytesseract 4 )

Answer 1

I try this picture: https://i.stack.imgur.com/HLgiF.jpg

def process_image(image):
    pytesseract.pytesseract.tesseract_cmd = 'C:\\Users\\abenhari\\AppData\\Local\\Tesseract-OCR\\tesseract.exe'
    img = Image.open(image)
    img = img.convert('L')
    pix = img.load()
    plt.imshow(img)
    plt.show()
    return pytesseract.image_to_string(img)

print(process_image("plat1.jpg"))