cv-analysis-service/vidocp/table_parsig.py

from itertools import count

import cv2
import numpy as np
import pdf2image
from matplotlib import pyplot as plt


def parse(image: np.array):

    gray_scale = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
    th1, img_bin = cv2.threshold(gray_scale, 150, 225, cv2.THRESH_BINARY)
    img_bin = ~img_bin

    line_min_width = 4
    kernel_h = np.ones((1, line_min_width), np.uint8)
    kernel_v = np.ones((line_min_width, 1), np.uint8)

    img_bin_h = cv2.morphologyEx(img_bin, cv2.MORPH_OPEN, kernel_h)
    img_bin_v = cv2.morphologyEx(img_bin, cv2.MORPH_OPEN, kernel_v)

    img_bin_final = img_bin_h | img_bin_v

    _, labels, stats, _ = cv2.connectedComponentsWithStats(~img_bin_final, connectivity=8, ltype=cv2.CV_32S)

    return labels, stats


def parse_tables_in_pdf(pages):
    return zip(map(parse, pages), count())


def annotate_image(image, stats):
    for x, y, w, h, area in stats[2:]:
        if w > 10 and h > 10:
            cv2.rectangle(image, (x, y), (x + w, y + h), (255, 0, 255), 2)

            for i, (s, v) in enumerate(zip(["x", "y", "w", "h"], [x, y, w, h])):
                anno = f"{s} = {v}"
                xann = int(x + 5)
                yann = int(y + h - (20 * (i + 1)))
                cv2.putText(image, anno, (xann, yann), cv2.FONT_HERSHEY_SIMPLEX, 0.7, (255, 0, 255), 2)

    return image


def annotate_tables_in_pdf(pdf_path, page_index=1):

    page = pdf2image.convert_from_path(pdf_path, first_page=page_index + 1, last_page=page_index + 1)[0]
    page = np.array(page)

    _, stats = parse(page)
    page = annotate_image(page, stats)

    fig, ax = plt.subplots(1, 1)
    fig.set_size_inches(20, 20)
    ax.imshow(page)
    plt.show()