Extracting data from a table with known labels with tesseract

Question

I am trying to use Tesseract to create a small Windows application that allows the user to:

• Take a screenshot of the monitor and cut a smaller portion containing a table (the table always has the same format, and the labels are consistent. The numerical data are different each time).
• Provide the screenshot to Tesseract to extract the data. My strategy is to remove the vertical and horizontal lines in the table, extract the entire text, and collect the numerical values corresponding to the labels I want to capture.
• Finally, generate a text output based on the extracted data.

The app works fine, but there are still many errors in data extraction. Sometimes, some values are not extracted at all because the label is not correctly recognized. Other times, even if the labels are recognized correctly and the data are extracted, the numbers are incorrect. Also I noticed that the error quote is higher on my work PC, probably because the screen resolution (and so the screenshot) is lower than my home PC.

I am wondering if there is a more reliable way to accomplish my goal.

Below I attached some images of the App to give you an idea, an example of the table and the python script I am using for OCR.

Thank you very much for your help!!!

tesseract v5.4.0.20240606

Python 3.13.1

def preprocess_image(image_path):
    image = cv2.imread(image_path)

    gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)

    thresh = cv2.threshold(gray, 0, 255, cv2.THRESH_BINARY_INV + cv2.THRESH_OTSU)[1]

    horizontal_kernel = cv2.getStructuringElement(cv2.MORPH_RECT, (50, 1))
    detect_horizontal = cv2.morphologyEx(thresh, cv2.MORPH_OPEN, horizontal_kernel, iterations=2)
    cnts = cv2.findContours(detect_horizontal, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)[-2]
    for c in cnts:
        cv2.drawContours(thresh, [c], -1, (0, 0, 0), 2)

    vertical_kernel = cv2.getStructuringElement(cv2.MORPH_RECT, (1, 15))
    detect_vertical = cv2.morphologyEx(thresh, cv2.MORPH_OPEN, vertical_kernel, iterations=3)
    cnts = cv2.findContours(detect_vertical, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)[-2]
    for c in cnts:
        cv2.drawContours(thresh, [c], -1, (0, 0, 0), 5)

    result = cv2.bitwise_and(image, image, mask=thresh)
    result[thresh == 0] = (255, 255, 255)  # Set background to white

    return result

def extract_text(image):
    """Extract text from the processed image using Tesseract."""
    return pytesseract.image_to_string(image, lang="eng", config="--psm 6")