Image Similarity Metrics Suitable for Infrared Video Stabilization during Active Wildfire Monitoring: A Comparative Analysis

Mario M Valero,Steven Verstockt,Eulàlia Planas,Lloyd Queen,Elsa Pastor,Dan Jimenez,Oriol Rios,Christian Mata

doi:10.3390/rs12030540

Mario M Valero, Steven Verstockt + Show 6 more

Open Access

https://doi.org/10.3390/rs12030540

Copy DOI

Abstract

Aerial Thermal Infrared (TIR) imagery has demonstrated tremendous potential to monitor active forest fires and acquire detailed information about fire behavior. However, aerial video is usually unstable and requires inter-frame registration before further processing. Measurement of image misalignment is an essential operation for video stabilization. Misalignment can usually be estimated through image similarity, although image similarity metrics are also sensitive to other factors such as changes in the scene and lighting conditions. Therefore, this article presents a thorough analysis of image similarity measurement techniques useful for inter-frame registration in wildfire thermal video. Image similarity metrics most commonly and successfully employed in other fields were surveyed, adapted, benchmarked and compared. We investigated their response to different camera movement components as well as recording frequency and natural variations in fire, background and ambient conditions. The study was conducted in real video from six fire experimental scenarios, ranging from laboratory tests to large-scale controlled burns. Both Global and Local Sensitivity Analyses (GSA and LSA, respectively) were performed using state-of-the-art techniques. Based on the obtained results, two different similarity metrics are proposed to satisfy two different needs. A normalized version of Mutual Information is recommended as cost function during registration, whereas 2D correlation performed the best as quality control metric after registration. These results provide a sound basis for image alignment measurement and open the door to further developments in image registration, motion estimation and video stabilization for aerial monitoring of active wildland fires.

Highlights

Forest fires have been studied through remote sensing techniques for decades
Variance-based global sensitivity analysis techniques described in Section 3.3 were used to assess the general response of various image similarity metrics to six variables of interest, namely: horizontal translation, vertical translation, rotation, scale, time and sampling frequency
An ideal image similarity measure should be highly sensitive to camera movement and robust in the presence of recording frequency variations and image content differences appearing over time

Summary

Introduction

Forest fires have been studied through remote sensing techniques for decades. A number of spaceborne sensors have successfully been used to analyze various fire aspects and post-fire effects [1]. A few successful experiences have been reported that use airborne monitoring systems in large-scale wildfires [17,26], the majority of developments in fire detection and monitoring occur via sensing prescribed fires, which are often restricted in areal extent as well as fire line radiative intensity [27]. In these cases, the remote sensor is usually placed in a fixed position or a hovering aircraft and it is deployed to collect high spatial resolution images with a moderate temporal resolution for the full duration of flaming combustion [28,29,30].

Objectives

Methods

Results

Discussion

Conclusion