ANOMALY DETECTION IN RAILWAY IMAGES USING UNSUPERVISED CLUSTERING OF INFRARED THERMOGRAPHY

Abstract

This paper presents a novel system for the unsupervised detection of anomalies in railway images using infrared thermography. The system comprises three main stages: image preprocessing, feature extraction, and clustering. In the image preprocessing stage, the infrared images are enhanced and normalized to improve the subsequent feature extraction. In the feature extraction stage, a set of relevant features is extracted from the preprocessed images. Finally, in the clustering stage, unsupervised clustering algorithms are employed to group the extracted features into different clusters, each representing a specific anomaly type. The proposed system is evaluated on a real-world dataset of railway images captured by a thermal camera. The experimental results demonstrate the effectiveness of the proposed system in detecting various types of anomalies with high accuracy.