earticle

영어

The vehicle-related challenges are escalating, such as vehicle thefts, traffic violations, and outdated license registrations. Computer visions using artificial intelligence can be utilized as a tool to find solutions. However inaccuracies in license plate recognition can occur due to environmental factors such as camera dust, camera location, and vibration causing blurry images and distracting. To solve these challenges, we propose an automatic license plate recognition system. Proposed system operates in two main stages: license plate detection and recognition. Initially, it detects and isolates the license plate from the vehicle image. Following this, the proposed system determines whether the plate is single or double-lined and proceeds to accurately classify and arrange the characters on the plate in their correct order. Utilizing the YOLOv7 object detection model and trained on the Karplate dataset, in addition the model is training on a data augmentation that represent the environmental factors, where we notes that the YOLOv7 _training is more accurate in recognizing the car plates and accurately recognizes various license plate layouts, presenting a significant results in vehicle identification technology.