Not able to recognize word using pytesseract.image_to_string

Question

I am having an image with rotated text. I am getting the position for this text and cropping it and then I rotated this text. Till this I can see my text displayed perfectly. But now if I send this rotated text to 'pytesseract.image_to_string', it is not recognizing the text. Here is the code.

croped = orig[y:y+h, x:x+w].copy()
cv2.imshow('cropped ',croped)
text_img = cv2.cvtColor(croped, cv2.COLOR_BGR2GRAY)
text_img_nt = cv2.bitwise_not(text_img)

cv2.imshow('text img not ',text_img_nt)
thresh = cv2.threshold(text_img_nt, 0, 255,cv2.THRESH_BINARY | cv2.THRESH_OTSU)[1]
coords = np.column_stack(np.where(thresh > 0))
angle = cv2.minAreaRect(coords)[-1]

if angle < -45:
     angle = -(90 + angle)
else:
     angle = -angle
            
(h, w) = text_img_nt.shape[:2]
center = (w // 2, h // 2)
M = cv2.getRotationMatrix2D(center, angle, 1.0)

rotated = cv2.warpAffine(text_img_nt, M, (w, h),flags=cv2.INTER_CUBIC, borderMode=cv2.BORDER_REPLICATE)
cv2.imshow('rotated',rotated)  
cv2.waitKey(100000)
config='-l eng --oem 1 --psm 3'
textRecognized = pytesseract.image_to_string(rotated, config = config, lang ='eng')
print(textRecognized)

Cropped Image Rotated Image

text recognized : "Ol Ey"

Is there any issue in my code? Any help will be much appreciated.

Answer 1

i did tried on the image you rotated, the only thing i do here is to invert the photo and i manage to get consistent result.

rotated = cv2.bitwise_not(rotated)