uit-21521549 / detected-motorcyclist-not-wearing-a-helmet

Detected-motorcyclist-not-wearing-a-helmet.

In this article I want to detect motorcyclists who do not wear helmets with real time. Build a real-time system for traffic. I use Yolov8 for detect and Spark combine Kafka to take a data real time from camera trafic.

Step 1: I trained yolov8 with my dataset in google colab, then i take a model and use that for detection helmet and no-helmet.

I use ultralytics for collect and detect box in web, then i train with model yolov8. In folder train have result and evaluate.

Step 2: Transmit data using kafka.

First: start your sever kafka in your local.

In my file "producer.py", i have provided a data transfer mechanism using kafka.

Second: i get data from kafka with func readstream in spark.

By getting data from Spark, you can process data distributedly.

Third: i saved data in Postgresql.

I took pictures of people not wearing helmets when riding motorbikes and saved them in postgresql.

Step 3: get data from Postgresql.

You can get pictures from the pro and also the time of their violation.

Note: I think you should run in linux, u can esay control data.

DATASET AND MODEL CNN: https://drive.google.com/drive/folders/1JHSqZ1b2yVraRpq9oWihGYT0Gnrr0eVD