SimonsParkingAI (in process)

Parking AI system for Eafit univercity

Run

 pip install -r requieriments.txt 

 python main.py

add go to

http://localhost:8000/

linear pipeline for training and deploying a YOLO model for parking slot detection:

---

1️ Collect & Prepare Data

• Download a dataset or create it, (We can mix both thecniques for learning prouses)
• PKLot
• CNRPark-EXT
• Kaggle Parking Dataset
• Kaggle Parking Dataset2
• If needed, manually annotate images using LabelImg or Roboflow.
• Ensure data is in YOLO format (bounding box labels in .txt files).

---

2️ Organize Dataset

• Structure dataset:

  /dataset
  ├── images
  │   ├── train  (Training images)
  │   ├── val    (Validation images)
  │   ├── test   (Test images)
  ├── labels
  │   ├── train  (Bounding boxes for training)
  │   ├── val    (Bounding boxes for validation)
  │   ├── test   (Bounding boxes for test)
  ├── data.yaml
  

• Create data.yaml:

  train: dataset/images/train
  val: dataset/images/val
  test: dataset/images/test
  nc: 1  # Number of classes
  names: ['parking_slot']

---

3️ Train YOLO Model

Our yolo model https://huggingface.co/smartinezg/Yolo_parking/tree/main

• Install dependencies:

  pip install ultralytics

• Train the model:

  from ultralytics import YOLO
  model = YOLO("yolov8n.pt")  # Load pre-trained YOLOv8 model
  model.train(data="data.yaml", epochs=50, imgsz=640, batch=8)

• Save the trained model:

  model.export(format="onnx")  # Convert for deployment

---

4️ Validate & Test

• Run inference on a test image:

  results = model("test.jpg", save=True, conf=0.5)

• Visualize predictions:

  model.predict(source="dataset/images/test", show=True, conf=0.5)

---

5️ Deploy for Real-Time Detection

• Run real-time detection using a webcam:

  model.predict(source=0, show=True, conf=0.5)

• Integrate with FastAPI for a web service.

---

6 Final Deployment & Monitoring

• Deploy to Raspberry Pi in a web interface

References

https://www.youtube.com/watch?v=VZXdkOo3xNI

https://www.youtube.com/watch?v=F-884J2mnOY

PKLot-detection

Profesional proyect

no so imortant, but elegant

https://www.youtube.com/watch?v=MeSeuzBhq2E

https://www.youtube.com/watch?v=VZXdkOo3xNI

Our Chatgpt chat used for this