- electric vehicle license plate
- 8 digit license plate
Generate Synthetic Korea License Plate.
-
You can use this generator when there is insufficient data on the license plate.
-
This image generator may be used in a character recognition model like 'CRNN'
-
I recommend pre-training with synthetic images and fine-tune with real data.
- You can create synthetic license plate pictures by selecting the plate of the desired type.
-
The name of the photo shows the letters and numbers on the license plate.
-
Hangul in the plate was translated into English with the following rules.
-
Region : 서울 -> A, 인천 -> B ...
-
Hangul : 나 -> sk, 너 -> sj, 다 -> ek, 도 -> eh ...
def random_bright(img):
img = cv2.cvtColor(img, cv2.COLOR_RGB2HSV)
img = np.array(img, dtype=np.float64)
random_bright = .5 + np.random.uniform()
img[:, :, 2] = img[:, :, 2] * random_bright
img[:, :, 2][img[:, :, 2] > 255] = 255
img = np.array(img, dtype=np.uint8)
img = cv2.cvtColor(img, cv2.COLOR_HSV2RGB)
return img
Ex)
Type 1 : Z13wn0965
Type 2 : Z22aj0246
Type 3 : B16an7086
Type 4 : A48sk2287
Type 5 : Z19tn7921
Type 6 : 112ah0833
Type 7 : X50fk9747
- os : Ubuntu 16.04.4 LTS or Windows 10
- Python : 3.5.2
File | Description |
---|---|
Generator_original.py | generate images without any image distortion/augmentation. |
Generator_augmentation.py | generate images with image augmentations such as random brightness. |
Generator_perspective.py | generate images with perspective transform. |