Multi-models Based on Yolov8 for Identification of Vehicle Type and License Plate Recognition

Embedded systems with cameras and deep learning techniques have been shown to be flexible and good at finding different targets in the areas of intelligent monitoring and urban mobility. These use cases are present in diverse situations and regions. The collection of pertinent data from the deployment site is of utmost importance. This study introduces an innovative methodology for a comprehensive system that integrates vehicle category identification with license plate recognition using the YOLOv8 algorithm. The system comprises three main components: vehicle type detection and recognition, detection of the license plate, and detection of the license plate characters and numbers. The suggested approach intends to enhance the identification system’s applicability in the unique context of Iraqi vehicles, particularly on roadways and in cities and their environments. The dataset used in this study was obtained from various areas inside Iraq. The detection system employed in our research successfully identified three distinct vehicle classes as well as detected and recognized license plates in both Arabic and English. The mean average precision achieved for the aforementioned tasks was 97.5%, 98.94%, 98.6%, and 98.4%, respectively. Through the use of visual data, such as images and videos, our system successfully identified license plates with reduced dimensions. It is posited that our technology has the potential to be used in densely populated areas in order to cater to the substantial requirements for improved visual acuity in smart urban environments.