🔥 Drowning Detection - Enhanced Version

This project is based on Drowning-Detection and has been improved for better compatibility, performance, and deployment on Kubernetes.

🚀 Key Modifications

• Updated Dependencies: Adjusted requirements.txt for compatibility with Python 3.12.
• Microservices Enhancement: Optimized for deployment in a Kubernetes environment.
• Improved Logging & Debugging: Enhanced logging for better observability.
• Performance Tweaks: Refactored some parts to improve efficiency.

📌 Original Repository

The original project was created by @randhana(https://github.com/randhana), and you can find it here: 🔗 Drowning-Detection on GitHub(https://github.com/randhana/Drowning-Detection-)

Overview

Introducing our drowning detection system, designed to detect and alert when someone is in danger of drowning. Our system uses computer vision algorithms to analyze video feeds from multiple cameras around a pool or other bodies of water.

The system works by using a deep learning-based object detection model to identify a person in the water. The model is trained on thousands of images and videos of people in water environments, allowing it to accurately recognize a human body in a variety of positions and lighting conditions. Once a person is detected, the system uses motion analysis algorithms to track their movements and monitor their safety.

Overall, the drowning detection system provides an innovative and effective solution for ensuring the safety of people in and around swimming pools. By leveraging advanced computer vision and machine learning technologies, the system can quickly and accurately identify potential drowning incidents and alert caregivers, ultimately helping to save lives.

Key Features:

• Computer Vision Algorithms: Analyze video feeds from multiple cameras.
• Deep Learning Model: Trained on thousands of images to detect human bodies in various positions and lighting conditions.
• Motion Analysis: Tracks movements to assess potential danger.

By providing rapid and accurate detection, this system helps to save lives and improve pool safety.

Installation

Follow the steps below to set up the Drowning Detection system:

Clone the Project

Clone with Test Sample Videos

git clone https://github.com/randhana/Drowning-Detection-.git

Clone Without Test Sample Videos (Lightweight Version)

git clone --branch lightweight https://github.com/randhana/Drowning-Detection-.git

Clone the Microservices Version of My Drowning Detection Project

git clone https://github.com/ngophong010/Microservices-DrowningDetection.git

Create a Virtual Environment

python -m venv env

Install Dependencies

pip install -r requirements.txt

Usage

To run the Drowning Detection program, follow these steps:

1. Create a "videos" Folder:

   • Inside the project directory, create a folder named videos.

2. Add Sample Videos:

   • Place video files into the videos folder.

3. Run the Program:

python DrownDetect.py --source video_file_name.mp4

4. Quit the Program:
   • Press the q key on your keyboard.

Watch our demonstration video showcasing the drowning detection system in action.

Additional Resources

For more details on YOLO object detection, visit their official website.

Microservices Architecture

This project uses Python, Kafka, Kubernetes, and a microservices-based architecture. Below are the instructions for setting up and deploying related components.

Install Kafka

Install Kafka with Helm Charts

A Helm chart to deploy Apache Kafka on Kubernetes.

Prerequisites

• Kubernetes 1.19+
• Helm 3.0+

Installation

helm install my-kafka ./my-kafka-chart

Install Kafka manually

1. Add the Bitnami Helm repository:

helm repo add bitnami https://charts.bitnami.com/bitnami

2. Install Kafka without persistence:

helm install kafka-local bitnami/kafka --set persistence.enabled=false,zookeeper.persistence.enabled=false

3. Deploy a Kafka client pod:

kubectl run kafka-local-client --restart='Never' --image docker.io/bitnami/kafka:3.3.1-debian-11-r19 \
--namespace drown-detector-microservice --command -- sleep infinity

4. Retrieve Kafka Client Password:

   • Run the following command to retrieve the password for the Kafka client:

     kubectl get secret kafka-local-user-passwords --namespace drown-detector-microservice -o jsonpath='{.data.client-passwords}' | base64 -d

   • Note the password as you will need it for authentication.

5. Create a client.properties File:

   • Inside your local environment, create a file named client.properties with the following content:

     security.protocol=SASL_PLAINTEXT
     sasl.mechanism=SCRAM-SHA-256
     sasl.jaas.config=org.apache.kafka.common.security.scram.ScramLoginModule required \
         username="user1" \
         password="<retrieved_password>";
     

   • Replace <retrieved_password> with the actual password you retrieved in the previous step.

6. Copy the client.properties File to the Kafka Client Pod:

kubectl cp --namespace drown-detector-microservice ./client.properties kafka-local-client:/tmp/client.properties

7. If you want to access the Kafka client:

kubectl exec --tty -i kafka-local-client --namespace drown-detector-microservice -- bash

If You want to use Helm charts: cd Helm_charts/Apache-kafka

Deploy Services

Drowning Detect Service

kubectl run drown-detect-svc --rm --tty -i --image phongngolam/drown-detect-service --restart Never --namespace drown-detector-microservice

Drowning Alert Service

kubectl run drown-alert-svc --rm --tty -i --image phongngolam/drown-alert-service --restart Never --namespace drown-detector-microservice

Kafka UI

1. Add the Kafka UI Helm repository:

helm repo add kafka-ui https://provectus.github.io/kafka-ui

2. Install Kafka UI:

helm install kafka-ui kafka-ui/kafka-ui \
--set envs.config.KAFKA_CLUSTERS_0_NAME=kafka-local \
--set envs.config.KAFKA_CLUSTERS_0_BOOTSTRAPSERVERS=kafka-local.drown-detector-microservice.svc.cluster.local:9092

3. Access the Kafka UI via port forwarding:

kubectl port-forward svc/kafka-ui 8080:8080