Opencv распознает числа с ошибками. Как поднять точность?

import numpy as np
import cv2
import time
import imutils
 
from imutils import contours
from imutils import grab_contours
from imutils.perspective import four_point_transform
from PIL import Image
import pytesseract
 
pytesseract.pytesseract.tesseract_cmd = r'C:\Program Files\Tesseract-OCR\tesseract.exe'
 
webcam = cv2.VideoCapture('video.mp4')
t = time.time()
 
i = 0
while (1):
    i += 1
    _, imageFrame = webcam.read()
    imageFrame = cv2.resize(imageFrame, (700, 500))
    imageFrame = cv2.rotate(imageFrame, cv2.ROTATE_90_CLOCKWISE)
    image = imageFrame[200:285, 70:440]
    cv2.imshow("frame", image)
 
    #img = cv2.GaussianBlur(image, (3, 3), 0)
    #edge = cv2.Canny(img, 255, 0)
    gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
    gray = cv2.bilateralFilter(gray, 11, 5, 5)
    thresh = cv2.threshold(gray, 0, 255, cv2.THRESH_BINARY_INV | cv2.THRESH_OTSU)[1]
    kernel = cv2.getStructuringElement(cv2.MORPH_ELLIPSE, (1, 5))
    thresh = cv2.morphologyEx(thresh, cv2.MORPH_OPEN, kernel)
    thresh2 = cv2.resize(thresh, (0, 0), fx=1, fy=1)
    thresh3 = cv2.threshold(thresh2, 0, 255, cv2.THRESH_BINARY_INV | cv2.THRESH_OTSU)[1]
 
    zeroscreen = np.zeros((85, 370, 1), np.uint8)
 
    cnts = cv2.findContours(thresh3.copy(), cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE)
    cnts = imutils.grab_contours(cnts)
    digitCnts = []
    for c in cnts:
        x, y, w, h = cv2.boundingRect(c)
        if (w >= 0 and w <= 100) and (h >= 0 and h <= 100):
            cv2.rectangle(image, (x, y), (x + w, y + h), (0, 0, 0), 1)
            digitCnts.append(c)
 
    digits = []
 
    for c in digitCnts:
        cv2.drawContours(zeroscreen, digitCnts, -1, (255, 255, 255), 2)
        zer = cv2.resize(zeroscreen, (0, 0), fx=0.6, fy=1)
            #cv2.imshow('7.png', zer)
        if time.time() - t >= 1.5:
            t = time.time()
            cv2.imwrite("frames\\frame " + str(i) + ".jpg", zer)
            cv2.imwrite("7.jpg", zer)
            # print(digits)
            zer = cv2.imread("7.jpg")
            cv2.imshow('zer', zer)
            text = pytesseract.image_to_string(zer, config='--psm 13 --oem 3 -c tessedit_char_whitelist=0123456789')
            print(text)
            custom_config = r'--oem 3 --psm 6 outputbase digits'
            # print(pytesseract.image_to_string(zer, config=custom_config))
 
    if cv2.waitKey(10) & 0xFF == ord('q'):
        webcam.release()
        cv2.destroyAllWindows()
        break

import sys
import numpy as np
import cv2
import uuid
import pytesseract
import re
 
 
pytesseract.pytesseract.tesseract_cmd = r"C:\Program Files\Tesseract-OCR\tesseract.exe"
 
img = cv2.imread("123.jpg", 1)
# img = cv2.resize(img, None, fx=1.2, fy=2)  # Увеличение изображения в 9 раз
data = pytesseract.image_to_string(img, config='outputbase digits')
 
summa = list(map(float, data.split()))
list_rec = sum((float(summa[i]) for i in range(0, int(len(summa)))))
 
print("Массив - ", summa)
print("Сумма  - ", list_rec)
 
cv2.imshow('contours', img)
cv2.waitKey()
cv2.destroyAllWindows()

data = pytesseract.image_to_string(img, config='outputbase digits', lang="chi_sim")
data = pytesseract.image_to_string(img, config='outputbase digits', lang="eng")
data = pytesseract.image_to_string(img, config='outputbase digits', lang="rus")

    gray=imageFrame[:,:,1]
    thresh = cv2.threshold(gray, 150, 255, cv2.THRESH_BINARY)[1]