* This dataset supports online use.Click here to jump.

The CCPD dataset is called Chinese City Parking Dataset. It is a large, diverse, and carefully annotated license plate recognition dataset. It was established by the USTC team in 2018.Towards End-to-End License Plate Detection and Recognition: A Large Dataset and Baseline" has been accepted by ECCV 2018. This dataset is mainly collected in parking lots in Hefei, China, and contains license plate photos in a variety of complex environments, such as blur, tilt, rainy and snowy days, which makes the dataset more challenging in license plate recognition tasks. It is divided into two parts, CCPD 2019 and CCPD 2020 (also known as CCPD-Green), where CCPD 2019 contains ordinary license plates (blue license plates), while CCPD 2020 contains new energy license plates (green license plates).

CCPD 2019 has over 300k images with fine annotations. Segmentation files are in 'split/' folder. Images in CCPD-Base are split into training/validation sets. Subdatasets in CCPD (CCPD-DB, CCPD-Blur, CCPD-FN, CCPD-Rotate, CCPD-Tilt, CCPD-Challenge) are used for testing.

CCPD 2020 (also known as CCPD-Green) has eight-digit license plate numbers. Compared with CCPD 2019, CCPD 2020 only contains new energy license plate images, reflecting the growing trend of new energy vehicles.