This repository contains a Flask API service designed to detect deviations in hallway videos using a YOLO model for object detection. Detected deviations are uploaded to Firebase storage and logged in a Firestore database.

• Object Detection: Utilizes YOLO for detecting predefined classes in video frames.
• Firebase Integration: Downloads videos from Firebase, processes them, and uploads detected frames back to Firebase storage.
• Deviations: Detected deviations is uploaded to firestore

• Python 3.x
• Firebase Admin SDK
• GCP account for firebase
• OpenCV
• YOLO model weights (yolov9c.pt)

1. Clone the repository: git clone Drone-deviation-service
2. Install flask and firebase sdk
3. Add the firebase credentials file in json format, and match the filename with the import in server.py. Get the file by going to service accounts -> generate new private key
4. Start the flask server: flask --app server.py run

• Ensure that your YOLO model weights (yolov9c.pt) are placed in the appropriate directory.
• Modify the class list in the predict_and_detect function as needed to suit your application.