Underground Parking Lot License Plate Recognition and Vehicle Attributes Image Dataset

The current transportation sector faces challenges such as parking difficulties and low management efficiency, especially in urban underground parking lots where vehicles frequently enter and exit, and there is an urgent need for license plate recognition. Existing license plate recognition systems perform poorly under complex environments and lighting changes, failing to meet real-time monitoring needs. Therefore, this dataset is constructed to provide high-quality vehicle image data to help improve the accuracy and efficiency of license plate recognition. The dataset includes license plate images from different lighting and angles, collected using high-definition cameras, covering various car models and environments. To ensure data quality, we adopted multiple rounds of annotation and expert review to ensure consistency and accuracy of annotations. The data is stored in JPG format, organized by time and location for easy retrieval and analysis. The core advantage of this dataset lies in its high annotation accuracy, with annotation consistency exceeding 95%, addressing the limitations of traditional datasets in vehicle attribute recognition. By introducing new data augmentation techniques, the robustness of models is enhanced, and experiments show that models trained with this dataset have a 15% improvement in recognition rate.