Year: 2025
Type: article
Abstract: : Autonomous vehicles (AVs) rely heavily on computer vision and artificial intelligence (AI) for real-time decision-making. However, adverse conditions such as rain, fog, and low light present significant challenges to accurate object detection. This study investigates AI-based real-time object detection frameworks that enhance performance under such conditions. Using a convolutional neural network (CNN) architecture integrated with thermal and radar imaging, we conducted experiments on a synthesized dataset combining real-world and adverse weather imagery. Statistical analysis, including regression and predictive accuracy measures, shows a 19% improvement in mean average precision (mAP) and a 22% reduction in false negatives compared to baseline models. These findings indicate that multi-sensor fusion and advanced AI techniques significantly improve detection capabilities, supporting safer deployment of AVs. Keywords: AI, object detection, autonomous vehicles, real-time processing, adverse weather, CNN, radar imaging, thermal vision, predictive analysis, computer vision
Author: Murali Krishna Pasupuleti
Language: English
Subfield: Automotive Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: bronze
OpenAlex ID: W4410814893