License Plate Detection Tutorial

Download notebook

Install Application Dependecies

!wget -nc https://raw.githubusercontent.com/georgia-tech-db/license-plate-recognition/main/requirements.txt
!pip -q --no-color install -r requirements.txt

File ‘requirements.txt’ already there; not retrieving.


�[1m[�[0mnotice�[1m]�[0m A new release of pip available: 22.2.2 -> 22.3.1
�[1m[�[0mnotice�[1m]�[0m To update, run: pip install --upgrade pip

Start EVA server

We are reusing the start server notebook for launching the EVA server.

!wget -nc "https://raw.githubusercontent.com/georgia-tech-db/eva/master/tutorials/00-start-eva-server.ipynb"
%run 00-start-eva-server.ipynb
cursor = connect_to_server()

File ‘00-start-eva-server.ipynb’ already there; not retrieving.

nohup eva_server > eva.log 2>&1 &

�[1m[�[0m�[34;49mnotice�[0m�[1;39;49m]�[0m�[39;49m A new release of pip available: �[0m�[31;49m22.2.2�[0m�[39;49m -> �[0m�[32;49m22.3.1�[0m
�[1m[�[0m�[34;49mnotice�[0m�[1;39;49m]�[0m�[39;49m To update, run: �[0m�[32;49mpython -m pip install --upgrade pip�[0m
Note: you may need to restart the kernel to use updated packages.

Register License Plate Extraction UDF

!wget -nc "https://raw.githubusercontent.com/georgia-tech-db/license-plate-recognition/main/ocr_extractor.py"
cursor.execute("DROP UDF LicensePlateExtractor;")
response = cursor.fetch_all()
print(response)
cursor.execute("""CREATE UDF IF NOT EXISTS LicensePlateExtractor
      INPUT  (frame NDARRAY UINT8(3, ANYDIM, ANYDIM))
      OUTPUT (labels NDARRAY STR(ANYDIM), bboxes NDARRAY FLOAT32(ANYDIM, 4),
              scores NDARRAY FLOAT32(ANYDIM))
      TYPE  Classification
      IMPL  'license_plate_extractor.py';
      """)
response = cursor.fetch_all()
print(response)

File ‘ocr_extractor.py’ already there; not retrieving.

@status: ResponseStatus.SUCCESS
@batch: 
                                                 0
0  UDF LicensePlateExtractor successfully dropped
@query_time: 0.021678909892216325
@status: ResponseStatus.SUCCESS
@batch: 
                                                                0
0  UDF LicensePlateExtractor successfully added to the database.
@query_time: 5.78774525411427

Download Images or Videos for License Plate Recognition

!wget -nc "https://raw.githubusercontent.com/georgia-tech-db/license-plate-recognition/main/test_image_1.png"
!wget -nc "https://raw.githubusercontent.com/georgia-tech-db/license-plate-recognition/main/test_image_2.png"
!wget -nc "https://raw.githubusercontent.com/femioladeji/License-Plate-Recognition-Nigerian-vehicles/master/test_images/car10.jpg"
!wget -nc "https://raw.githubusercontent.com/femioladeji/License-Plate-Recognition-Nigerian-vehicles/master/test_images/car6.jpg"
!wget -nc "https://im.ezgif.com/tmp/ezgif-1-c32008dd2a-jpg/ezgif-frame-001.jpg"
!wget -nc "https://i.ytimg.com/vi/p91epBEfISk/maxresdefault.jpg"

# DOWNLOAD ADDITIONAL IMAGES IF NEEDED AND LOAD THEM HERE

########################################################
###
### IMAGES
###
########################################################

cursor.execute('DROP TABLE IF EXISTS MyImages')
response = cursor.fetch_all()
print(response)

cursor.execute('LOAD IMAGE "test_image_1.png" INTO MyImages;')
response = cursor.fetch_all()
print(response)
cursor.execute('LOAD IMAGE "test_image_2.png" INTO MyImages;')
response = cursor.fetch_all()
print(response)
cursor.execute('LOAD IMAGE "car10.jpg" INTO MyImages;')
response = cursor.fetch_all()
print(response)
cursor.execute('LOAD IMAGE "car6.jpg" INTO MyImages;')
response = cursor.fetch_all()
print(response)
cursor.execute('LOAD IMAGE "ezgif-frame-001.jpg" INTO MyImages;')
response = cursor.fetch_all()
print(response)
cursor.execute('LOAD IMAGE "maxresdefault.jpg" INTO MyImages;')
response = cursor.fetch_all()
print(response)

cursor.execute("""SELECT LicensePlateExtractor(data)
                FROM MyImages""")
response = cursor.fetch_all()
print(response)

########################################################
###
### VIDEOS
###
########################################################

!wget -nc "https://raw.githubusercontent.com/apoorva-dave/LicensePlateDetector/master/video12.mp4"

# DOWNLOAD ADDITIONAL VIDEOS IF NEEDED AND LOAD THEM HERE

cursor.execute('DROP TABLE IF EXISTS MyVideos')
response2 = cursor.fetch_all()
print(response2)
cursor.execute('LOAD VIDEO "video12.mp4" INTO MyVideos;')
response2 = cursor.fetch_all()
print(response2)
cursor.execute("""SELECT LicensePlateExtractor(data)
                FROM MyVideos
                WHERE id = 10""")
response2 = cursor.fetch_all()
print(response2)

File ‘test_image_1.png’ already there; not retrieving.

File ‘test_image_2.png’ already there; not retrieving.

File ‘car10.jpg’ already there; not retrieving.

File ‘car6.jpg’ already there; not retrieving.

File ‘ezgif-frame-001.jpg’ already there; not retrieving.

File ‘maxresdefault.jpg’ already there; not retrieving.

@status: ResponseStatus.SUCCESS
@batch: 
                                       0
0  Table Successfully dropped: MyImages
@query_time: 0.03773099696263671
@status: ResponseStatus.SUCCESS
@batch: 
                            0
0  Number of loaded IMAGE: 1
@query_time: 0.05547582497820258
@status: ResponseStatus.SUCCESS
@batch: 
                            0
0  Number of loaded IMAGE: 1
@query_time: 0.014967684866860509
@status: ResponseStatus.SUCCESS
@batch: 
                            0
0  Number of loaded IMAGE: 1
@query_time: 0.014217793010175228
@status: ResponseStatus.SUCCESS
@batch: 
                            0
0  Number of loaded IMAGE: 1
@query_time: 0.01373684499412775
@status: ResponseStatus.SUCCESS
@batch: 
                            0
0  Number of loaded IMAGE: 1
@query_time: 0.09569284389726818
@status: ResponseStatus.SUCCESS
@batch: 
                            0
0  Number of loaded IMAGE: 1
@query_time: 0.022398902103304863
@status: ResponseStatus.SUCCESS
@batch: 
   licenseplateextractor.labels  \
0                  [TN4805566]   
1                   [INAQ3044]   
2                   [EGB62644]   
3                   [LEM446A4]   
4        [4II982A7A, 770L7726]   
5     [LPD7SRV616, IPD8NBZ548]   

                                                                          licenseplateextractor.bboxes  \
0                                                   [[[432, 648], [723, 648], [723, 698], [432, 698]]]   
1                                                   [[[298, 238], [398, 238], [398, 264], [298, 264]]]   
2                                                   [[[178, 430], [300, 430], [300, 462], [178, 462]]]   
3                                                   [[[263, 465], [383, 465], [383, 501], [263, 501]]]   
4  [[[2263, 398], [2347, 398], [2347, 436], [2263, 436]], [[216.02985749985467, 829.1194299994187],...   
5  [[[236, 372], [344, 372], [344, 396], [236, 396]], [[986, 536], [1100, 536], [1100, 560], [986, ...   

                   licenseplateextractor.scores  
0                           [0.765264012834225]  
1                          [0.5979598335637649]  
2                          [0.8205181373232121]  
3                         [0.37683450003837526]  
4  [0.005999226930693703, 0.041628318390644674]  
5      [0.3380678860228084, 0.8426790601376755]  
@query_time: 6.209830478997901
File ‘video12.mp4’ already there; not retrieving.

@status: ResponseStatus.SUCCESS
@batch: 
                                       0
0  Table Successfully dropped: MyVideos
@query_time: 0.035812282003462315
@status: ResponseStatus.SUCCESS
@batch: 
                            0
0  Number of loaded VIDEO: 1
@query_time: 0.09962842799723148
@status: ResponseStatus.SUCCESS
@batch: 
   licenseplateextractor.labels  \
0                 [TSO7FX353L]   

                         licenseplateextractor.bboxes  \
0  [[[361, 765], [685, 765], [685, 841], [361, 841]]]   

  licenseplateextractor.scores  
0         [0.5929218894477851]  
@query_time: 5.568858061917126

Annotate Model Output on Image

import cv2
from pprint import pprint
from matplotlib import pyplot as plt
from pathlib import Path

def annotate_image_ocr(detections, input_image_path, frame_id):
    color1=(0, 255, 150)
    color2=(255, 49, 49)
    white=(255, 255, 255)
    thickness=4

    frame = cv2.imread(input_image_path)
    height, width = frame.shape[:2]

    if frame_id == 0:
        frame= cv2.copyMakeBorder(frame, 0, 100, 0, 100, cv2.BORDER_CONSTANT,value=white)

    print(detections)
    plate_id = 0

    df = detections
    df = df[['licenseplateextractor.bboxes', 'licenseplateextractor.labels']][df.index == frame_id]

    x_offset = width * 0.1
    y_offset = height * 0.4

    if df.size:
        dfLst = df.values.tolist()
        for bbox, label in zip(dfLst[plate_id][0], dfLst[plate_id][1]):
            x1, y1, x2, y2 = bbox
            x1, y1, x2, y2 = int(x1[0]), int(x1[1]), int(x2[0]), int(x2[1])
            # object bbox
            cv2.rectangle(frame, (x1, y1), (x2, y2), color1, thickness) 

            # object label
            cv2.putText(frame, label, (int(x1), int(y1 - height * 0.1)), cv2.FONT_HERSHEY_SIMPLEX, (6 * width)/2400, color2, 
                                  int((12 * width)/2400), cv2.LINE_AA) 

            # Show every  frame
            plt.imshow(frame)
            plt.show()

            p = Path(input_image_path)
            output_path = "{0}_{2}{1}".format(p.stem, p.suffix, "output")

            cv2.imwrite(output_path, frame)

dataframe = response.batch.frames
annotate_image_ocr(dataframe, 'test_image_1.png', frame_id = 0)
annotate_image_ocr(dataframe, 'test_image_2.png', frame_id = 1)
annotate_image_ocr(dataframe, 'car10.jpg', frame_id = 2)
annotate_image_ocr(dataframe, 'car6.jpg', frame_id = 3)
annotate_image_ocr(dataframe, 'ezgif-frame-001.jpg', frame_id = 4)
annotate_image_ocr(dataframe, 'maxresdefault.jpg', frame_id = 5)

  licenseplateextractor.labels  \
0                  [TN4805566]   
1                   [INAQ3044]   
2                   [EGB62644]   
3                   [LEM446A4]   
4        [4II982A7A, 770L7726]   
5     [LPD7SRV616, IPD8NBZ548]   

                        licenseplateextractor.bboxes  \
0  [[[432, 648], [723, 648], [723, 698], [432, 69...   
1  [[[298, 238], [398, 238], [398, 264], [298, 26...   
2  [[[178, 430], [300, 430], [300, 462], [178, 46...   
3  [[[263, 465], [383, 465], [383, 501], [263, 50...   
4  [[[2263, 398], [2347, 398], [2347, 436], [2263...   
5  [[[236, 372], [344, 372], [344, 396], [236, 39...   

                   licenseplateextractor.scores  
0                           [0.765264012834225]  
1                          [0.5979598335637649]  
2                          [0.8205181373232121]  
3                         [0.37683450003837526]  
4  [0.005999226930693703, 0.041628318390644674]  
5      [0.3380678860228084, 0.8426790601376755]

  licenseplateextractor.labels  \
0                  [TN4805566]   
1                   [INAQ3044]   
2                   [EGB62644]   
3                   [LEM446A4]   
4        [4II982A7A, 770L7726]   
5     [LPD7SRV616, IPD8NBZ548]   

                        licenseplateextractor.bboxes  \
0  [[[432, 648], [723, 648], [723, 698], [432, 69...   
1  [[[298, 238], [398, 238], [398, 264], [298, 26...   
2  [[[178, 430], [300, 430], [300, 462], [178, 46...   
3  [[[263, 465], [383, 465], [383, 501], [263, 50...   
4  [[[2263, 398], [2347, 398], [2347, 436], [2263...   
5  [[[236, 372], [344, 372], [344, 396], [236, 39...   

                   licenseplateextractor.scores  
0                           [0.765264012834225]  
1                          [0.5979598335637649]  
2                          [0.8205181373232121]  
3                         [0.37683450003837526]  
4  [0.005999226930693703, 0.041628318390644674]  
5      [0.3380678860228084, 0.8426790601376755]

  licenseplateextractor.labels  \
0                  [TN4805566]   
1                   [INAQ3044]   
2                   [EGB62644]   
3                   [LEM446A4]   
4        [4II982A7A, 770L7726]   
5     [LPD7SRV616, IPD8NBZ548]   

                        licenseplateextractor.bboxes  \
0  [[[432, 648], [723, 648], [723, 698], [432, 69...   
1  [[[298, 238], [398, 238], [398, 264], [298, 26...   
2  [[[178, 430], [300, 430], [300, 462], [178, 46...   
3  [[[263, 465], [383, 465], [383, 501], [263, 50...   
4  [[[2263, 398], [2347, 398], [2347, 436], [2263...   
5  [[[236, 372], [344, 372], [344, 396], [236, 39...   

                   licenseplateextractor.scores  
0                           [0.765264012834225]  
1                          [0.5979598335637649]  
2                          [0.8205181373232121]  
3                         [0.37683450003837526]  
4  [0.005999226930693703, 0.041628318390644674]  
5      [0.3380678860228084, 0.8426790601376755]

  licenseplateextractor.labels  \
0                  [TN4805566]   
1                   [INAQ3044]   
2                   [EGB62644]   
3                   [LEM446A4]   
4        [4II982A7A, 770L7726]   
5     [LPD7SRV616, IPD8NBZ548]   

                        licenseplateextractor.bboxes  \
0  [[[432, 648], [723, 648], [723, 698], [432, 69...   
1  [[[298, 238], [398, 238], [398, 264], [298, 26...   
2  [[[178, 430], [300, 430], [300, 462], [178, 46...   
3  [[[263, 465], [383, 465], [383, 501], [263, 50...   
4  [[[2263, 398], [2347, 398], [2347, 436], [2263...   
5  [[[236, 372], [344, 372], [344, 396], [236, 39...   

                   licenseplateextractor.scores  
0                           [0.765264012834225]  
1                          [0.5979598335637649]  
2                          [0.8205181373232121]  
3                         [0.37683450003837526]  
4  [0.005999226930693703, 0.041628318390644674]  
5      [0.3380678860228084, 0.8426790601376755]

  licenseplateextractor.labels  \
0                  [TN4805566]   
1                   [INAQ3044]   
2                   [EGB62644]   
3                   [LEM446A4]   
4        [4II982A7A, 770L7726]   
5     [LPD7SRV616, IPD8NBZ548]   

                        licenseplateextractor.bboxes  \
0  [[[432, 648], [723, 648], [723, 698], [432, 69...   
1  [[[298, 238], [398, 238], [398, 264], [298, 26...   
2  [[[178, 430], [300, 430], [300, 462], [178, 46...   
3  [[[263, 465], [383, 465], [383, 501], [263, 50...   
4  [[[2263, 398], [2347, 398], [2347, 436], [2263...   
5  [[[236, 372], [344, 372], [344, 396], [236, 39...   

                   licenseplateextractor.scores  
0                           [0.765264012834225]  
1                          [0.5979598335637649]  
2                          [0.8205181373232121]  
3                         [0.37683450003837526]  
4  [0.005999226930693703, 0.041628318390644674]  
5      [0.3380678860228084, 0.8426790601376755]

  licenseplateextractor.labels  \
0                  [TN4805566]   
1                   [INAQ3044]   
2                   [EGB62644]   
3                   [LEM446A4]   
4        [4II982A7A, 770L7726]   
5     [LPD7SRV616, IPD8NBZ548]   

                        licenseplateextractor.bboxes  \
0  [[[432, 648], [723, 648], [723, 698], [432, 69...   
1  [[[298, 238], [398, 238], [398, 264], [298, 26...   
2  [[[178, 430], [300, 430], [300, 462], [178, 46...   
3  [[[263, 465], [383, 465], [383, 501], [263, 50...   
4  [[[2263, 398], [2347, 398], [2347, 436], [2263...   
5  [[[236, 372], [344, 372], [344, 396], [236, 39...   

                   licenseplateextractor.scores  
0                           [0.765264012834225]  
1                          [0.5979598335637649]  
2                          [0.8205181373232121]  
3                         [0.37683450003837526]  
4  [0.005999226930693703, 0.041628318390644674]  
5      [0.3380678860228084, 0.8426790601376755]

Annotate Model Output on Video

def annotate_video(detections, input_video_path):
    color1=(207, 248, 64)
    color2=(255, 49, 49)
    thickness=4

    vcap = cv2.VideoCapture(input_video_path)
    width = int(vcap.get(3))
    height = int(vcap.get(4))
    fps = vcap.get(5)
    fourcc = cv2.VideoWriter_fourcc('m', 'p', '4', 'v') #codec

    p = Path(input_video_path)
    output_video_path = "{0}_{2}{1}".format(
        p.stem,  
        p.suffix, 
        "output")
    video=cv2.VideoWriter(output_video_path, fourcc, fps, (width,height))

    frame_id = 0
    # Capture frame-by-frame
    # ret = 1 if the video is captured; frame is the image
    ret, frame = vcap.read() 

    x_offset = width * 0.2
    y_offset = height * 0.8

    while ret:

        df = detections
        df = df[['licenseplateextractor.bboxes', 'licenseplateextractor.labels']][df.index == frame_id]

        if df.size:

            dfLst = df.values.tolist()
            for bbox, label in zip(dfLst[frame_id][0], dfLst[frame_id][1]):
                x1, y1, x2, y2 = bbox
                x1, y1, x2, y2 = int(x1[0]), int(x1[1]), int(x2[0]), int(x2[1])

                # object bbox
                cv2.rectangle(frame, (x1, y1), (x2, y2), color1, thickness) 

                # object label
                cv2.putText(frame, label, (int(x_offset), int(y_offset)), cv2.FONT_HERSHEY_SIMPLEX, 3, color2, thickness * 4, cv2.LINE_AA) 
                y_offset = y_offset + height * 0.2
        
                video.write(frame)
                
                # Show every fifth frame
                if frame_id % 5 == 0:
                    plt.imshow(frame)
                    plt.show()
        
        frame_id+=1
        ret, frame = vcap.read()

    video.release()
    vcap.release()

dataframe2 = response2.batch.frames
annotate_video(dataframe2, "video12.mp4")

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

README.md

README.md

License Plate Detection Tutorial

Install Application Dependecies

Start EVA server

Register License Plate Extraction UDF

Download Images or Videos for License Plate Recognition

Annotate Model Output on Image

Annotate Model Output on Video

Files

README.md

Latest commit

History

README.md

File metadata and controls

License Plate Detection Tutorial

Install Application Dependecies

Start EVA server

Register License Plate Extraction UDF

Download Images or Videos for License Plate Recognition

Annotate Model Output on Image

Annotate Model Output on Video