International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 11 Issue: 10 | Oct 2024
p-ISSN: 2395-0072
www.irjet.net
SAFE ROAD AI: REAL-TIME ACCIDENT DETECTION FROM MULTI-ANGLE CRASH VIDEOS Piyush Kumar1, Dipti Ranjan Tiwari2 1Master of Technology, Computer Science and Engineering, Lucknow Institute of Technology, Lucknow, India 2Assistant Professor, Department of Computer Science and Engineering, Lucknow Institute of Technology,
Lucknow, India ---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - Road accidents pose a serious threat to public
safety, often leading to fatalities or severe injuries, and causing tremendous emotional and physical distress to individuals and their families. Many accidents, particularly those on highways, in remote areas, or during nighttime, go unreported for extended periods, which delays the arrival of necessary assistance. This lack of timely medical intervention can worsen the outcomes, underscoring the need for efficient coordination between emergency services and healthcare facilities. To tackle these challenges, the innovative SAFE ROAD AI system has been developed. This system harnesses cutting-edge deep learning algorithms to simultaneously analyze multiple crash videos, significantly improving the accuracy and speed of accident detection. By incorporating techniques like data preprocessing, feature extraction, and model training, SAFE ROAD AI employs a modified deep convolutional neural network (D-CNN) to effectively and reliably identify accidents. A mobile application built with React Native has been integrated into the system, allowing for automatic notifications to be sent to relevant authorities as soon as an accident is detected. This ensures swift responses and timely intervention, potentially saving lives by enabling faster medical and emergency aid. Advanced accident detection technologies such as SAFE ROAD AI are essential in reducing the severity of road accidents and mitigating their impact on society. Through the adoption of innovative solutions, we can move towards safer roads and more secure communities for everyone. Key Words: Road accidents, Public safety, Fatalities, Accident detection, Deep learning algorithms, Crash video analysis, Modified deep convolutional neural network (D-CNN)
1.INTRODUCTION Road accidents remain one of the leading causes of injury and death worldwide, posing a significant threat to public safety. The ability to detect accidents in real-time and respond swiftly is crucial in reducing fatalities and minimising the severity of injuries. However, traditional methods of accident reporting often rely on delayed responses, especially in remote areas or during nighttime incidents, leading to critical delays in providing life-saving assistance. In recent years, advancements in artificial intelligence (AI) and deep learning have opened new © 2024, IRJET
|
Impact Factor value: 8.315
|
possibilities for improving road safety. Leveraging these technologies, the development of automated systems capable of detecting accidents in real-time has gained momentum. The SAFE ROAD AI system is one such innovative solution designed to address this critical need. By analysing multi-angle crash videos through the use of deep learning algorithms, SAFE ROAD AI enhances the accuracy and speed of accident detection, offering the potential to revolutionise emergency response systems.
Figure-1: Crash of Vehicles This research focuses on the design and implementation of the SAFE ROAD AI system, which utilises a modified deep convolutional neural network (D-CNN) to process crash footage from multiple angles, improving detection precision. The system is integrated with a mobile application, which automatically alerts relevant authorities when an accident is identified, ensuring immediate response and reducing delays in emergency assistance. Through this research, we explore how AI-driven technologies can significantly contribute to enhancing road safety and minimising the impact of accidents.
1.1.Overview of the Project The project "Real-Time Accident Detection from Multi-Angle Crash Videos" is a groundbreaking initiative that focuses on the development of an intelligent system with the primary goal of detecting accidents in real-time. By employing advanced DL techniques such as Convolutional Neural Networks (CNNs) and Multi-layer Perceptron (MLPs), this system aims to revolutionize road safety and emergency response procedures. ISO 9001:2008 Certified Journal
|
Page 583