wheresmycar

First attempt at utilizing YOLOv8 model for vehicle number plate detection.

[!NOTE] This project is for education and skills presentation purposes.

Motivation

Gain practical knowledge with machine learning technologies in real-world example.

The main goal was to gain hands-on experience in machine learning project utilizing PyTorch library and several technologies to improve software engineering skills.

About the project

Small Python package providing a class for object detection, utilizing YOLOv8 model which was trained to detect number plates on vehicles.

Documentation

plate_detector

plate_detector module

PlateDetector Objects

class PlateDetector()

Class for Vehicle Number Plate Detection based on pretrained YOLOv8 model.

load_model

def load_model(device: str) -> YOLO

Function to load pretrained model. params:

• device : device on which the model should run returns:
• <ultralytics.YOLO>: pretrained YOLOv8 model

model

@property
def model() -> YOLO

Access pretrained model returns:

• <ultralytics.YOLO>: pretrained YOLOv8 model

get_device

def get_device(enable_cuda: bool) -> str

Gets target device for inference. params:

• enable_cude : if True, will return cuda if available returns:
• either cuda or cpu

detect

def detect(target_path: str, conf: float = 0.5, **kwargs) -> Results

Get predictions on given input. params:

• target_path : path to directory with images or image's file path
• conf : minimum confidence threshold for detection returns:
• <ultralytics.engine.results.Results>: inference results