You are on page 1of 18

GANDHINAGAR INSTITUTE OF TECHNOLGY

Information Technology Department


Academic Year: 2023 - 2024

AI-Car Detection

Presented By Guide Name:


Devani Prince(200120116118) Ms. Ankita Patel
Outline
• Introduction
• Project Summery
• Aim and Objectives
• Tools & Technologies
• Implementation
• Functional
• Non Functional
• Outcomes
• Conclusion
• Future Enhancement
• Progress report with result Pictures
Introduction
• AI is making big changes in jobs, car detection is the use for
the safety and security purpose.
• Our project, " AI-Car Detection " uses smart computers and
cameras for safety and security.
• Our aim is reducing risks and ensuring safety rules are
followed. We're creating a smart system to detect car use.
• Using advanced AI that learns from visuals, we prevent
accident and ensure rule compliance.
• Combining tech and security rules, our project shows how
technology enhances security. We truly care about car dirver
safety.
• Through AI-driven car monitoring, we're dedicated to
improving roads safety and security for all.
Project Summary
• This project focuses on creating an advanced car detection
system using YOLOv5, an cutting-edge object detection
algorithm.
• The primary aim is to train YOLOv5 with specific car-related
data. This enables the system to precisely recognize car usage
in different sides.
• With a custom-trained model, the system accurately detects
whether car, crucial for safety, in various scenarios.
• The YOLOv5 algorithm is trained using a varied dataset
containing images of car in different environments, teaching it
to recognize car shapes and positions.
• Once trained, the system efficiently analyzes images or video
streams in real-time. It promptly identifies instances where car
regulations are not followed.
• By utilizing YOLOv5 and custom data, the project creates a
strong car detection system. This solution automates car
checks.
Aim and Objectives
Aim:
The aim of the Car Detection System is to enhance safety and
mitigate risks in various contexts.

Objectives:
• Real-time Monitoring: Enable real-time analysis of images or
videos to swiftly identify instances of any other vehicle usage.

• Reducing A: Contribute to reduced injuries by ensuring usage


in various settings.

• Adaptability: Ensure the system functions effectively in


varying conditions and environments.
• Privacy and Ethics: Implement the system ethically,
respecting privacy rights and maintaining a balance between
safety, security and privacy.

• User-Friendly Interface: If human interaction is involved,


design an intuitive interface for clear communication.

• Data Collection and Analysis: Gather and analyze data for


usage patterns and insights to enhance system performance.

• Continuous Improvement: Build flexibility for regular


system updates and advancements, adapting to feedback and
evolving technology.
Tools & Technologies
• Programming Languages: Utilize languages like Python for
implementing algorithms and creating a functional system.

• YOLOv5: Employ the YOLO (You Only Look Once) object


detection framework to achieve accurate and real-time helmet
detection.

• Data Annotation Tools: Use annotation tools like LabelImg


for annotating images with car labels, crucial for training the
model.

• Hardware: Depending on the application, consider


appropriate hardware like GPUs for faster model training and
real-time inference.
• Computer Vision Libraries: Utilize libraries like OpenCV for
image processing, manipulation, and feature extraction.

• Deep Learning Frameworks: Implement deep learning


models using frameworks such as TensorFlow or PyTorch for
training and inference.
Implementation

1 Dataset Collection
Gather a diverse dataset of
Data Preprocessing 2 car-related images,
categorizing them into
Prepare and preprocess the wearing and non-wearing
dataset, ensuring consistency, instances.
annotation, and augmentation
to improve model
generalization.
3 Model Training
Train the YOLOv5 model
using the prepared dataset,
Real-time Integration 4
adjusting parameters for
Integrate the trained model optimal performance in car
into a real-time system that detection and optimizing for
analyzes incoming video accuracy and speed.
streams or images, promptly
flagging instances of non-
compliance and generating
timely alerts.
Functional Requirements
1. Car Detection: The system must accurately detect whether
individuals are car in real-time video streams or images.

2. Real-time Analysis: The system should perform analysis and


detection in real-time.

3. Bounding Box Visualization: The system should visually


mark the detected car with bounding boxes in images or video
frames.

4. Platform Independence: The system should be designed to


work on multiple platforms, such as desktop, web, or mobile.
Non-Functional Requirements
1. Accuracy: The system's detection accuracy should be high,
minimizing false positives and negatives to ensure reliable
results.

2. Cost effectiveness: The implementation and maintenance of


the system should be economically affordable, ensuring
efficient resource utilization.

3. User-Friendliness The system should be easy to use for car


drivers, require minimal training for effective operation.

4. Scalability: The solution should be scalable, capable of


handling increased data volume and multiple camera feeds
while maintaining performance.
Outcomes

• Improved Safety and security


Our system has helped reduce accidents
and create a safer and security working
environment for all drivers.

• Increased Productivity
Our system provides a non-intrusive way
to monitor drivers' safety,

• Cultural Shift:
The system will encourage people to always
follow safety rules, creating a culture where
safety is a priority
Conclusion
• Our car detection system is all about safety and security first.
It's like having a digital helper that uses smart eyes to spot car
in pictures and videos.

• Think about how much safer and security this makes our
Roade! It helps us make sure that everyone follows the rule to
prevent accidents and improve safety and security.

• Because it's quick and easy to use, it's a reliable safety partner.
We're dedicated to improving it even more, so it can keep safe
people on road.
Future Enhancement
• Expanded Detection: Improve the system to detect not only
car usage but also other vehicles, ensuring comprehensive
compliance with safety and security regulations.

• Intelligent Alerts: Enhance alerts by integrating AI to


provide context-aware notifications, helping drivers
understand specific safety violations and corrective actions.

• Weather Adaptability: Enhance the system's robustness by


training it to work effectively under different weather
conditions, accounting for rain, snow, and varying lighting.
• Remote Monitoring and Reporting: Develop a centralized
monitoring dashboard that allows supervisors to remotely
monitor car compliance across multiple sites and
generate reports.

• Localized Language Support: Offer support for multiple


languages and regional variations in notifications and alerts,
ensuring effective communication in diverse
driving environments.
Result Pictures

You might also like