ENHANCING RAILWAY SECURITY WITH DEEP LEARNING-BASED PERSON RE-IDENTIFICATION AND TRACKING SYSTEMS

Abstract

The difficulties in recognizing and monitoring suspicious persons via several surveillance cameras provide a considerable concern for public safety in densely populated transportation systems, such as the Indian Railways. This work introduces a deep learning-based Person Re-Identification (Re-ID) system designed to enhance railway security via the automated identification and recognition of suspects in surveillance video. The suggested system utilizes Convolutional Neural Networks (CNNs) to extract strong characteristics related to body posture and face recognition, hence enhancing identification effectiveness, in contrast to standard methods that depend on manually produced features, which may be susceptible to errors.The system process entails creating a database of questionable individuals, employing convolutional neural network (CNN) models for feature extraction, trainin machine learning classifiers like Support Vector Machine (SVM) and Random Forest, and ultimately implementing a real-time video surveillance system for railway security personnel. Administrators may promptly react after the assessment of alerts derived from CCTV video. The system provides monitoring modules for both administrative and staff levels, developed using Python and MySQL. The suggested method significantly improves public safety in Indian Railways via automated and scalable intelligent monitoring, as seen by the Random Forest classifier's better accuracy and less false positives.