Automatic Number Plate Recognition
Description
Automatic Number Plate Recognition (ANPR) systems have gained a lot attention due to its vast applications, including traffic management, law enforcement, toll collection and parking systems. ANPR can also be used to track stolen vehicles that have been used in criminal activities. ANPR systems face challenges in dusty environments, especially in developing countries where there are many untarred roads. ANPR systems also face tedious task due to the variety of plate formats (i.e., Plate size, background, character size, plate texture) especially where there is a busy environment and complex scenarios. To address some of these challenges we present license plate dataset captured at various locations in Ghana. The dataset is presented into one main folder which contains twelve (12) sub folders. The total number of images in the dataset is 11,326 images. It is grouped according to the regions in Ghana (6,202-Ashanti Region, 440-Brong Ahafo Region, 103 -Central Region, 665 - DV, 70 - Eastern Region, 3,084 -Greater Accra Region, 559 -Motorcycle, 24 - Northern Region, 10 -Upper East Region, 19 -Upper West Region, 07 - Volta Region, and 143 -Western Region). This dataset will contribute to the development of a more robust and efficient Number Plate Recognition system.
Files
Steps to reproduce
The dataset consists of images captured using three distinct mobile phones: iPhone XS Max, Techno Hot 9, and iPhone XR. These images are available in three formats: Portable Network Graphics (PNG), HEIC, and JPG, and have a resolutions of (720 x1280), (768 x 1024), (1024 x 768), and (3024 x 4032) pixels. there are twelve (12) classes specifically dedicated to license plates. The images were captured under diverse conditions and environments, including dusty environments and low-light conditions such as night time settings.