Anomaly Detection for Hotspot Identification in Landsat Images

This paper presents a methodology for hotspot detection in multispectral images, utilizing the Reed-Xiaoli anomaly detection algorithm. By leveraging short-wave infrared data from Landsat, the Reed-Xiaoli algorithm identifies hotspots with adaptability to sensors similar to OLI in spectral coverage. The proposed approach is formulated as an a-contrario method, eliminating the need for manually set thresholds for hotspot detection and allows for the control of false detections. The application of this method extends to monitoring the activity status of cement plants, demonstrating robust performance across both daytime and nighttime images. The results show the efficacy of the proposed methodology in hotspot detection for monitoring the activity of industrial facilities such as cement plants.