製作身份證數據集

1.代碼 

from PIL import Image, ImageDraw, ImageFont
import numpy as np
import cv2

# 定義寫字函數
def add_txt(image, size, draw_x, draw_y, txt):
    # 字體字號
    setFont = ImageFont.truetype('simhei.ttf', size)
    # 定義畫板
    draw = ImageDraw.Draw(image)
    # 繪製
    draw.text((draw_x, draw_y), txt, font=setFont, fill=(0, 0, 0))
    return image


def make_fake_id_card():
    ori_image = cv2.imread('./IDCard.png')

    print('==ori_image.shape:', ori_image.shape)
    ori_image = cv2.resize(ori_image, (0, 0), fx=0.4, fy=0.4)
    print('==resize ori_image.shape:', ori_image.shape)

    # 向圖片上寫字
    img = Image.fromarray(cv2.cvtColor(ori_image, cv2.COLOR_BGR2RGB))
    img = add_txt(img, 18, 90, 55, '某某姓名')
    img = add_txt(img, 16, 90, 87, '男')
    img = add_txt(img, 16, 190, 87, '漢')
    img = add_txt(img, 16, 90, 116, '2020')
    img = add_txt(img, 16, 160, 116, '2')
    img = add_txt(img, 16, 200, 116, '26')
    img = add_txt(img, 16, 90, 151, '北京市哈哈哈哈')
    img = add_txt(img, 16, 90, 172, '某某　606')
    img = add_txt(img, 18, 145, 234, '987654321')

    cv2.imwrite('./word.jpg', np.array(img)[..., ::-1])

def make_white_mask():
    # 生成一個空白的模板mask
    ori_image = cv2.imread('./IDCard.png')
    ori_image = cv2.resize(ori_image, (0, 0), fx=0.4, fy=0.4)
    mask_image = np.ones_like(ori_image)
    mask_image *= 255
    print(mask_image.shape)
    cv2.imwrite('./mask.jpg', mask_image)

    # 往空白模板上寫字(這裏只能用PIL寫，因爲OpenCV寫中文會亂碼)
    img = Image.fromarray(cv2.cvtColor(mask_image, cv2.COLOR_BGR2RGB))
    img = add_txt(img, 18, 90, 55, '某某姓名')
    img = add_txt(img, 16, 90, 87, '男')
    img = add_txt(img, 16, 190, 87, '漢')
    img = add_txt(img, 16, 90, 116, '2020')
    img = add_txt(img, 16, 160, 116, '2')
    img = add_txt(img, 16, 200, 116, '26')
    img = add_txt(img, 16, 90, 151, '北京市哈哈哈哈')
    img = add_txt(img, 16, 90, 172, '某某　606')
    img = add_txt(img, 18, 145, 234, '987654321')

    mask_image_txt = cv2.cvtColor(np.array(img), cv2.COLOR_RGB2BGR)
    cv2.imwrite('./mask_image_txt.jpg', mask_image_txt)
    gray = cv2.cvtColor(mask_image_txt, cv2.COLOR_BGR2GRAY)
    # 高斯模糊，製造邊緣模糊效果哦
    gray_Gaussianblur = cv2.GaussianBlur(gray, (3, 3), 0)

    # 使用閾值對圖片進行二值化
    th, res = cv2.threshold(gray_Gaussianblur, 200, 255, cv2.THRESH_BINARY)
    res_inv = cv2.bitwise_not(res)
    cv2.imwrite('./res_inv.jpg', res_inv)

    # 寫字的模板保留文字部分
    img_bg = cv2.bitwise_and(mask_image_txt, mask_image_txt, mask=res_inv)
    cv2.imwrite('./img_bg.jpg', img_bg)
    # 原圖保留除文字的其他部分
    img_fg = cv2.bitwise_and(ori_image, ori_image, mask=res)
    cv2.imwrite('./img_fg.jpg', img_fg)
    # 將兩張圖直接進行相加，即可
    final = cv2.add(img_bg, img_fg)
    cv2.imwrite('./final.jpg', final)

if __name__ == '__main__':
    # make_fake_id_card()
    make_white_mask()

1.背景圖 　　　　　　　　　　　　　　　　　　　　　　　　2.生成的