Vehicle Detection using Artificial Intelligence

This project aims at developing a model that can detect vehicles. The project uses Deep learning approch to solve the problem as the objective is to perform real time vehicle detection. For this purpose YOLO v3 detector was used as it provides good accuracy with available resources. The project can

Project Title

Vehicle Detection using Artificial Intelligence

Project Area of Specialization

Artificial Intelligence

Project Summary

This project aims at developing a model that can detect vehicles. The project uses Deep learning approch to solve the problem as the objective is to perform real time vehicle detection. For this purpose YOLO v3 detector was used as it provides good accuracy with available resources. The project can be used for traffic surveillance, at toll plazas, parking lots etc.

Project Objectives

• Project that can reduce human interference while detecting vehicles.
• Providing easy and efficient solution for the problem.
• Using Artificial Intelligence instead of sensors for object detection.
• creating our own dataset instead of using the one available on internet.

Project Implementation Method

• Firstly, dataset needs to be created and for that Roboflow website is used for labeeling and performing augmentation on the images. 
• The images are further cloned using Darknet that generates text files for every image.
• Model is selected from doing literature review, which in this case is YOLO v3 model. 
• Dataset is uploaded on Google Drive
• Google Colab is used for training as it provides a GPU.
• Dataset is loaded and Yolo v3 model is cloned with the help of Darknet.
• After training is complete a video/image is given as testing data and it performs vehicle detection on it.
• Reslts such as detection time, accuracy, threshols values can be checked and compared.
• If desirable results are achieved model is loaded on Jetson nano kit otherwise, training is done again by increasing dataset. 
• After loading the model real time detection can be performed by activating the camera.

Benefits of the Project

• Can be used for traffic surveillance and montoring
• Can be used for detecting any car accident or theft
• Can count number of vehicles passing a specific point

Technical Details of Final Deliverable

The final project that will be delivered will be a trained model that can be loaded onto the Jetson nano kit along with the camera and LED that will start detecting vehicles once the camera is activated. Along with that the dataset and Colab notebook will be provided that will have weight file and config file along needed for training the model.

Final Deliverable of the Project

HW/SW integrated system

Core Industry

Transportation

Other Industries

Core Technology

Artificial Intelligence(AI)

Other Technologies

Sustainable Development Goals

Industry, Innovation and Infrastructure

Required Resources

If you need this project, please contact me on contact@adikhanofficial.com