Capstone_Jep_Final.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Capstone_Jep_Final.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/Applied-Machine-Learning-2022/final-project-jep-ua/blob/P-will/Capstone_Jep_Final.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install opencv-python\n",
        "!sudo apt install tesseract-ocr\n",
        "!pip install pytesseract==0.3.9 \n",
        "!pip install imutils\n",
        "!pip install scikit-image"
      ],
      "metadata": {
        "id": "xNqENVoR8ZpA"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "%cd ..\n",
        "!pip show torch"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "hKFySOfnKmnw",
        "outputId": "00cf13ff-2dc2-4801-8e33-14a1eef7c5ac"
      },
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Name: torch\n",
            "Version: 1.12.0+cu113\n",
            "Summary: Tensors and Dynamic neural networks in Python with strong GPU acceleration\n",
            "Home-page: https://pytorch.org/\n",
            "Author: PyTorch Team\n",
            "Author-email: packages@pytorch.org\n",
            "License: BSD-3\n",
            "Location: /usr/local/lib/python3.7/dist-packages\n",
            "Requires: typing-extensions\n",
            "Required-by: torchvision, torchtext, torchaudio, fastai\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#Installing libraries so we can use an MMOCR\n",
        "# Install mmcv-full thus we could use CUDA operators\n",
        "# !pip install mmcv-full -f https://download.openmmlab.com/mmcv/dist/cu111/torch1.10.0/index.html\n",
        "!pip install --upgrade mmcv-full==1.6.0 -f https://download.openmmlab.com/mmcv/dist/cu113/torch1.12.0/index.html\n",
        "\n",
        "# Install mmdetection\n",
        "!pip install mmdet\n",
        "\n",
        "# Install mmocr\n",
        "!git clone https://github.com/open-mmlab/mmocr.git\n",
        "%cd mmocr\n",
        "!pip install -r requirements.txt\n",
        "!pip install -v -e ."
      ],
      "metadata": {
        "id": "WIzxGAwULGUl"
      },
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "!git clone https://github.com/wzh191920/License-Plate-Recognition License-Plate-Recognition-master"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "0Fa9MLXL9jxP",
        "outputId": "11149e85-2c1b-4a12-d159-0b63d19d7f97"
      },
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "fatal: destination path 'License-Plate-Recognition-master' already exists and is not an empty directory.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import zipfile\n",
        "import os\n",
        "\n",
        "with zipfile.ZipFile('License-Plate-Recognition-master','r') as z:\n",
        "  z.extractall('./')\n",
        "\n",
        "os.listdir()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 356
        },
        "id": "HiEJP3Ef7oJo",
        "outputId": "c1d18f6f-5e39-41ef-f517-11e1db549a9b"
      },
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "error",
          "ename": "IsADirectoryError",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mIsADirectoryError\u001b[0m                         Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-8-ef36558141af>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      2\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mos\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      3\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 4\u001b[0;31m \u001b[0;32mwith\u001b[0m \u001b[0mzipfile\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mZipFile\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'License-Plate-Recognition-master'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m'r'\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0mz\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      5\u001b[0m   \u001b[0mz\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mextractall\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'./'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      6\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/lib/python3.7/zipfile.py\u001b[0m in \u001b[0;36m__init__\u001b[0;34m(self, file, mode, compression, allowZip64, compresslevel)\u001b[0m\n\u001b[1;32m   1238\u001b[0m             \u001b[0;32mwhile\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1239\u001b[0m                 \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1240\u001b[0;31m                     \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfp\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mio\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mopen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfile\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfilemode\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   1241\u001b[0m                 \u001b[0;32mexcept\u001b[0m \u001b[0mOSError\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1242\u001b[0m                     \u001b[0;32mif\u001b[0m \u001b[0mfilemode\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mmodeDict\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mIsADirectoryError\u001b[0m: [Errno 21] Is a directory: 'License-Plate-Recognition-master'"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from mmocr.utils.ocr import MMOCR"
      ],
      "metadata": {
        "id": "dH-3feKmXAO_"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#Final Code\n",
        "from PIL import Image\n",
        "import PIL\n",
        "import numpy as np\n",
        "import cv2\n",
        "import pytesseract\n",
        "import matplotlib.pyplot as plt\n",
        "import urllib.request\n",
        "import os\n",
        "import tarfile\n",
        "import shutil\n",
        "import tensorflow as tf\n",
        "from numpy.linalg import norm\n",
        "import sys\n",
        "import  os\n",
        "import json\n",
        "from google.colab.patches import cv2_imshow\n",
        "# from mmocr.utils.ocr import MMOCR\n",
        "\n",
        "\n",
        "SZ  =  20           #training picture length and width\n",
        "MAX_WIDTH  =  1000 #Original  image maximum width\n",
        "Min_Area  =  2000 #The   maximum allowable area of ​​the license plate area\n",
        "PROVINCE_START = 1000\n",
        "\n",
        "\n",
        "base_url = 'http://download.tensorflow.org/models/object_detection/'\n",
        "file_name = 'ssd_mobilenet_v1_coco_2018_01_28.tar.gz'\n",
        "\n",
        "url = base_url + file_name\n",
        "\n",
        "urllib.request.urlretrieve(url, file_name)\n",
        "\n",
        "os.listdir()\n",
        "\n",
        "dir_name = file_name[0:-len('.tar.gz')]\n",
        "\n",
        "if os.path.exists(dir_name):\n",
        "  shutil.rmtree(dir_name) \n",
        "\n",
        "tarfile.open(file_name, 'r:gz').extractall('./')\n",
        "\n",
        "os.listdir(dir_name)\n",
        "\n",
        "frozen_graph = os.path.join(dir_name, 'frozen_inference_graph.pb')\n",
        "\n",
        "with tf.io.gfile.GFile(frozen_graph, \"rb\") as f:\n",
        "  graph_def = tf.compat.v1.GraphDef()\n",
        "  loaded = graph_def.ParseFromString(f.read())\n",
        "\n",
        "cars_video = cv2.VideoCapture('/content/Mercedes Glk - 1406.mp4')\n",
        "height = int(cars_video.get(cv2.CAP_PROP_FRAME_HEIGHT))\n",
        "width = int(cars_video.get(cv2.CAP_PROP_FRAME_WIDTH))\n",
        "fps = cars_video.get(cv2.CAP_PROP_FPS)\n",
        "total_frames = int(cars_video.get(cv2.CAP_PROP_FRAME_COUNT))\n",
        "\n",
        "frames_read = 0\n",
        "\n",
        "r = 255\n",
        "g = 0\n",
        "b = 0\n",
        "scale = .5\n",
        "thickness = 1\n",
        "\n",
        "outputs = (\n",
        "  'num_detections:0',\n",
        "  'detection_classes:0',\n",
        "  'detection_scores:0',\n",
        "  'detection_boxes:0',\n",
        ")\n",
        "\n",
        "def wrap_graph(graph_def, inputs, outputs, print_graph=False):\n",
        "  wrapped = tf.compat.v1.wrap_function(\n",
        "    lambda: tf.compat.v1.import_graph_def(graph_def, name=\"\"), [])\n",
        "\n",
        "  return wrapped.prune(\n",
        "    tf.nest.map_structure(wrapped.graph.as_graph_element, inputs),\n",
        "    tf.nest.map_structure(wrapped.graph.as_graph_element, outputs))\n",
        "\n",
        "\n",
        "model = wrap_graph(graph_def=graph_def,\n",
        "                   inputs=[\"image_tensor:0\"],\n",
        "                   outputs=outputs)\n",
        "model.outputs\n",
        "\n",
        "\n",
        "def point_limit(point):\n",
        "    if point[0] < 0:\n",
        "        point[0] = 0\n",
        "    if point[1] < 0:\n",
        "        point[1] = 0\n",
        "\n",
        "\n",
        "def accurate_place(card_img_hsv, limit1, limit2, color):\n",
        "    row_num, col_num = card_img_hsv.shape[:2]\n",
        "    xl = col_num\n",
        "    xr = 0\n",
        "    yh = 0\n",
        "    yl = row_num\n",
        "    #col_num_limit = self.cfg[\"col_num_limit\"]\n",
        "    row_num_limit = 18\n",
        "    col_num_limit = col_num * 0.8 if color != \"green\" else col_num * 0.5#绿色有渐变\n",
        "    for i in range(row_num):\n",
        "        count = 0\n",
        "        for j in range(col_num):\n",
        "            H = card_img_hsv.item(i, j, 0)\n",
        "            S = card_img_hsv.item(i, j, 1)\n",
        "            V = card_img_hsv.item(i, j, 2)\n",
        "            if limit1 < H <= limit2 and 34 < S and 46 < V:\n",
        "                count += 1\n",
        "        if count > col_num_limit:\n",
        "            if yl > i:\n",
        "                yl = i\n",
        "            if yh < i:\n",
        "                yh = i\n",
        "    for j in range(col_num):\n",
        "        count = 0\n",
        "        for i in range(row_num):\n",
        "            H = card_img_hsv.item(i, j, 0)\n",
        "            S = card_img_hsv.item(i, j, 1)\n",
        "            V = card_img_hsv.item(i, j, 2)\n",
        "            if limit1 < H <= limit2 and 34 < S and 46 < V:\n",
        "                count += 1\n",
        "        if count > row_num - row_num_limit:\n",
        "            if xl > j:\n",
        "                xl = j\n",
        "            if xr < j:\n",
        "                xr = j\n",
        "    return xl, xr, yh, yl\n",
        "\n",
        "\n",
        "fourcc = cv2.VideoWriter_fourcc(*'mp4v')\n",
        "\n",
        "for i in range(0, int(cars_video.get(cv2.CAP_PROP_FRAME_COUNT))):\n",
        "  cars_video.set(cv2.CAP_PROP_POS_FRAMES, i)\n",
        "  ret, img = cars_video.read()\n",
        "  # img = cv2.imread(\"/content/License-Plate-Recognition-master/test/car5.jpg\")\n",
        "  if (not ret):\n",
        "    break \n",
        "  # img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)\n",
        "  # cv2_imshow(img)\n",
        "  h, w,_ = img.shape\n",
        "  input_images = [img]\n",
        "  tensor = tf.convert_to_tensor(input_images, dtype=tf.uint8)\n",
        "\n",
        "  detections = model(tensor)\n",
        "  # print(detections)\n",
        "  # for i, box in enumerate(detections[3][0]):\n",
        "  # print(detections[1][0])\n",
        "  if len([i for i in detections[1][0] if i == 3]):\n",
        "    # img = cv2.imread(\"/content/Corvette.jpeg\")\n",
        "    pic_hight, pic_width = img.shape[:2]\n",
        "    if pic_width > MAX_WIDTH:\n",
        "        pic_rate = MAX_WIDTH / pic_width\n",
        "        img = cv2.resize(img, (MAX_WIDTH, int(pic_hight*pic_rate)), interpolation=cv2.INTER_LANCZOS4)\n",
        "\n",
        "    # print(\"h,w:\", pic_hight, pic_width)\n",
        "    blur = 3\n",
        "    #高斯去噪\n",
        "    if blur > 0:\n",
        "        img = cv2.GaussianBlur(img, (blur, blur), 0)#图片分辨率调整\n",
        "    oldimg = img\n",
        "    img = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)\n",
        "    kernel = np.ones((20, 20), np.uint8)\n",
        "    img_opening = cv2.morphologyEx(img, cv2.MORPH_OPEN, kernel)\n",
        "    img_opening = cv2.addWeighted(img, 1, img_opening, -1, 0);\n",
        "\n",
        "    #找到图像边缘\n",
        "    ret, img_thresh = cv2.threshold(img_opening, 0, 255, cv2.THRESH_BINARY + cv2.THRESH_OTSU)\n",
        "    img_edge = cv2.Canny(img_thresh, 100, 200)\n",
        "    #使用开运算和闭运算让图像边缘成为一个整体\n",
        "    kernel = np.ones((5, 12), np.uint8)\n",
        "    img_edge1 = cv2.morphologyEx(img_edge, cv2.MORPH_CLOSE, kernel)\n",
        "    img_edge2 = cv2.morphologyEx(img_edge1, cv2.MORPH_OPEN, kernel)\n",
        "\n",
        "    #查找图像边缘整体形成的矩形区域，可能有很多，车牌就在其中一个矩形区域中\n",
        "    try:\n",
        "        contours, hierarchy = cv2.findContours(img_edge2, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE)\n",
        "    except ValueError:\n",
        "        image, contours, hierarchy = cv2.findContours(img_edge2, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE)\n",
        "    contours = [cnt for cnt in contours if cv2.contourArea(cnt) > Min_Area]\n",
        "    # print('len(contours)', len(contours))\n",
        "    #一一排除不是车牌的矩形区域\n",
        "    car_contours = []\n",
        "    for cnt in contours:\n",
        "        rect = cv2.minAreaRect(cnt)\n",
        "        area_width, area_height = rect[1]\n",
        "        if area_width < area_height:\n",
        "            area_width, area_height = area_height, area_width\n",
        "        wh_ratio = area_width / area_height\n",
        "        #print(wh_ratio)\n",
        "        #要求矩形区域长宽比在2到5.5之间，2到5.5是车牌的长宽比，其余的矩形排除\n",
        "        if wh_ratio > 0 and wh_ratio < 6: #2 and 5.5\n",
        "            car_contours.append(rect)\n",
        "            box = cv2.boxPoints(rect)\n",
        "            box = np.int0(box)\n",
        "            #oldimg = cv2.drawContours(oldimg, [box], 0, (0, 0, 255), 2)\n",
        "            #cv2.imshow(\"edge4\", oldimg)\n",
        "            #cv2.waitKey(0)\n",
        "\n",
        "    print('len(contours)', len(car_contours))\n",
        "\n",
        "    # print(\"精确定位\")\n",
        "    card_imgs = []\n",
        "    #矩形区域可能是倾斜的矩形，需要矫正，以便使用颜色定位\n",
        "    for rect in car_contours:\n",
        "        if rect[2] > -1 and rect[2] < 1:#创造角度，使得左、高、右、低拿到正确的值\n",
        "            angle = 1\n",
        "        else:\n",
        "            angle = rect[2]\n",
        "        rect = (rect[0], (rect[1][0]+5, rect[1][1]+5), angle)#扩大范围，避免车牌边缘被排除\n",
        "\n",
        "        box = cv2.boxPoints(rect)\n",
        "        heigth_point = right_point = [0, 0]\n",
        "        left_point = low_point = [pic_width, pic_hight]\n",
        "        for point in box:\n",
        "            if left_point[0] > point[0]:\n",
        "                left_point = point\n",
        "            if low_point[1] > point[1]:\n",
        "                low_point = point\n",
        "            if heigth_point[1] < point[1]:\n",
        "                heigth_point = point\n",
        "            if right_point[0] < point[0]:\n",
        "                right_point = point\n",
        "\n",
        "        if left_point[1] <= right_point[1]:#正角度\n",
        "            new_right_point = [right_point[0], heigth_point[1]]\n",
        "            pts2 = np.float32([left_point, heigth_point, new_right_point])#字符只是高度需要改变\n",
        "            pts1 = np.float32([left_point, heigth_point, right_point])\n",
        "            M = cv2.getAffineTransform(pts1, pts2)\n",
        "            dst = cv2.warpAffine(oldimg, M, (pic_width, pic_hight))\n",
        "            point_limit(new_right_point)\n",
        "            point_limit(heigth_point)\n",
        "            point_limit(left_point)\n",
        "            card_img = dst[int(left_point[1]):int(heigth_point[1]), int(left_point[0]):int(new_right_point[0])]\n",
        "            card_imgs.append(card_img)\n",
        "            #cv2.imshow(\"card\", card_img)\n",
        "            #cv2.waitKey(0)\n",
        "        elif left_point[1] > right_point[1]:#负角度\n",
        "            \n",
        "            new_left_point = [left_point[0], heigth_point[1]]\n",
        "            pts2 = np.float32([new_left_point, heigth_point, right_point])#字符只是高度需要改变\n",
        "            pts1 = np.float32([left_point, heigth_point, right_point])\n",
        "            M = cv2.getAffineTransform(pts1, pts2)\n",
        "            dst = cv2.warpAffine(oldimg, M, (pic_width, pic_hight))\n",
        "            point_limit(right_point)\n",
        "            point_limit(heigth_point)\n",
        "            point_limit(new_left_point)\n",
        "            card_img = dst[int(right_point[1]):int(heigth_point[1]), int(new_left_point[0]):int(right_point[0])]\n",
        "            card_imgs.append(card_img)\n",
        "            #cv2.imshow(\"card\", card_img)\n",
        "            #cv2.waitKey(0)\n",
        "    #开始使用颜色定位，排除不是车牌的矩形，目前只识别蓝、绿、黄车牌\n",
        "    colors = []\n",
        "    for card_index,card_img in enumerate(card_imgs):\n",
        "        green = yello = blue = black = white = 0\n",
        "        card_img_hsv = cv2.cvtColor(card_img, cv2.COLOR_BGR2HSV)\n",
        "        #有转换失败的可能，原因来自于上面矫正矩形出错\n",
        "        if card_img_hsv is None:\n",
        "            continue\n",
        "        row_num, col_num= card_img_hsv.shape[:2]\n",
        "        card_img_count = row_num * col_num\n",
        "\n",
        "        for i in range(row_num):\n",
        "            for j in range(col_num):\n",
        "                H = card_img_hsv.item(i, j, 0)\n",
        "                S = card_img_hsv.item(i, j, 1)\n",
        "                V = card_img_hsv.item(i, j, 2)\n",
        "                if 11 < H <= 34 and S > 34:#图片分辨率调整\n",
        "                    yello += 1\n",
        "                elif 35 < H <= 99 and S > 34:#图片分辨率调整\n",
        "                    green += 1\n",
        "                elif 99 < H <= 124 and S > 34:#图片分辨率调整\n",
        "                    blue += 1\n",
        "                \n",
        "                if 0 < H <180 and 0 < S < 255 and 0 < V < 46:\n",
        "                    black += 1\n",
        "                elif 0 < H <180 and 0 < S < 43 and 221 < V < 225:\n",
        "                    white += 1\n",
        "        color = \"no\"\n",
        "\n",
        "        limit1 = limit2 = 0\n",
        "        if yello*2 >= card_img_count:\n",
        "            color = \"yello\"\n",
        "            limit1 = 11\n",
        "            limit2 = 34#有的图片有色偏偏绿\n",
        "        elif green*2 >= card_img_count:\n",
        "            color = \"green\"\n",
        "            limit1 = 35\n",
        "            limit2 = 99\n",
        "        elif blue*2 >= card_img_count:\n",
        "            color = \"blue\"\n",
        "            limit1 = 100\n",
        "            limit2 = 124#有的图片有色偏偏紫\n",
        "        elif black + white >= card_img_count*0.7:#TODO\n",
        "            color = \"bw\"\n",
        "        # print(color)\n",
        "        colors.append(color)\n",
        "        # print(blue, green, yello, black, white, card_img_count)\n",
        "        #cv2.imshow(\"color\", card_img)\n",
        "        #cv2.waitKey(0)\n",
        "        if limit1 == 0:\n",
        "            continue\n",
        "        #以上为确定车牌颜色\n",
        "        #以下为根据车牌颜色再定位，缩小边缘非车牌边界\n",
        "        xl, xr, yh, yl = accurate_place(card_img_hsv, limit1, limit2, color)\n",
        "        if yl == yh and xl == xr:\n",
        "            continue\n",
        "        need_accurate = False\n",
        "        if yl >= yh:\n",
        "            yl = 0\n",
        "            yh = row_num\n",
        "            need_accurate = True\n",
        "        if xl >= xr:\n",
        "            xl = 0\n",
        "            xr = col_num\n",
        "            need_accurate = True\n",
        "        card_imgs[card_index] = card_img[yl:yh, xl:xr] if color != \"green\" or yl < (yh-yl)//4 else card_img[yl-(yh-yl)//4:yh, xl:xr]\n",
        "        if need_accurate:#可能x或y方向未缩小，需要再试一次\n",
        "            card_img = card_imgs[card_index]\n",
        "            card_img_hsv = cv2.cvtColor(card_img, cv2.COLOR_BGR2HSV)\n",
        "            xl, xr, yh, yl = accurate_place(card_img_hsv, limit1, limit2, color)\n",
        "            if yl == yh and xl == xr:\n",
        "                continue\n",
        "            if yl >= yh:\n",
        "                yl = 0\n",
        "                yh = row_num\n",
        "            if xl >= xr:\n",
        "                xl = 0\n",
        "                xr = col_num\n",
        "        card_imgs[card_index] = card_img[yl:yh, xl:xr] if color != \"green\" or yl < (yh-yl)//4 else card_img[yl-(yh-yl)//4:yh, xl:xr]\n",
        "    #以上为车牌定位\n",
        "    #以下为识别车牌中的字符\n",
        "    predict_result = []\n",
        "    roi = None\n",
        "    card_color = None\n",
        "    for i, color in enumerate(colors):\n",
        "        if color in (\"blue\", \"yello\", \"green\", \"bw\"):\n",
        "            card_img = card_imgs[i]\n",
        "            gray_img = cv2.cvtColor(card_img, cv2.COLOR_BGR2GRAY)\n",
        "            #黄、绿车牌字符比背景暗、与蓝车牌刚好相反，所以黄、绿车牌需要反向\n",
        "            if color == \"green\" or color == \"yello\":\n",
        "                gray_img = cv2.bitwise_not(gray_img)\n",
        "            ret, gray_img = cv2.threshold(gray_img, 0, 255, cv2.THRESH_BINARY + cv2.THRESH_OTSU)\n",
        "                      \n",
        "            # gray = cv2.cvtColor(car_license_plate, cv2.COLOR_RGB2GRAY)\n",
        "            blur = cv2.GaussianBlur(gray_img, (3,3), 0)\n",
        "            thresh = cv2.threshold(blur, 0, 255, cv2.THRESH_BINARY_INV + cv2.THRESH_OTSU)[1]\n",
        "            \n",
        "            # Morph open to remove noise and invert image\n",
        "            kernel = cv2.getStructuringElement(cv2.MORPH_RECT, (3,3))\n",
        "            opening = cv2.morphologyEx(thresh, cv2.MORPH_OPEN, kernel, iterations=1)\n",
        "            invert = 255 - opening\n",
        "            # Perform text extraction utilizing MMOCR\n",
        "            # data = pytesseract.image_to_string(card_imgs[i], lang='eng', config='--psm 10')\n",
        "            # detection+recognition\n",
        "            from mmocr.utils.ocr import MMOCR\n",
        "            mmocr = MMOCR(det='TextSnake', recog='SAR')\n",
        "            data = mmocr.readtext(card_imgs[i], print_result=True,)\n",
        "            data\n",
        "            # if len(data) > 12 or len(data) < 4:\n",
        "            #     continue\n",
        "            # else:\n",
        "            #   cv2_imshow(card_imgs[i])\n",
        "            #   print(data)\n"
      ],
      "metadata": {
        "id": "uCSKP1DLb0k9"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "_6oUdHTTYyEx"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}