Welcome to the repository for iOS app designed to recognize Turkish license plates using Machine Learning models and Optical Character Recognition (OCR). This app combines the power of Swift, Core ML, and OCR techniques to accurately detect and convert license plates into text format.
License Plate Detection: App utilizes Machine Learning models to detect and localize license plates within images. Model trained by 400 license plates.
OCR Conversion: Once a license plate is detected, app employs OCR to extract the alphanumeric text from the license plate image.
Turkey License Plates: It specialized in recognizing and processing license plates conforming to the Turkish license plate format, ensuring accurate and relevant results.
Use your device's camera to capture a clear image containing a Turkish license plate.
Recognition Process: The app will automatically process the image, detecting the license plate and converting it into text.
View Results: Instantly view the recognized license plate number on the screen.
There are 2 models in the project. Both of are generated by Create ML. CP_1.mlmodel is basic one with 200 lp. ADTLicensePlateDetector is complex one with 600 lp. Both model training data included both front and back of the cars in different variety.
For now it works only on Portrait mode.
ML Confidence level is + .8
Uses static MLModel.
We welcome contributions from the open-source community. Feel free to fork the repository, make improvements, and submit pull requests. Whether it's enhancing the recognition accuracy, adding support for additional license plate formats, or improving the user interface – your contributions are valuable!
This project is licensed under the MIT License.
