|
Performance Evaluation of Thermal Infrared and Visible Image Fusion Methods using UAV Images based on VIFB framework
|
Ali Abdollahi   , Reza Shahhoseini * |
| University of Tehran |
|
|
Abstract: (23 Views) |
In the field of remote sensing, the fusion of visible and thermal images is a process in which spectral information from thermal imagery is combined with spatial detail from visible imagery to produce a thermally-informative image with enhanced spatial resolution. As the demand for higher spatial detail, shorter monitoring intervals, and real-time results increases, satellite-based remote sensing imagery becomes less practical. In such scenarios, UAV-based photogrammetric imagery presents a viable alternative. The aim of this study is to compare the results of several fusion methods, tailored to specific implementation platforms, based on established evaluation metrics, with a focus on UAV-acquired aerial data.
The dataset used in this study comprises 30 visible images captured using a Zenmuse X3 camera and 80 thermal images acquired using a Zenmuse XT camera mounted on a Matrice 100 UAV. The area of interest is a flat scene covering approximately 1.54 hectares, located within the Universidad del Valle, Colombia. The fusion process was conducted using 12 different algorithms within the VIFB framework. Each fused output was evaluated using 13 standard image fusion quality metrics. Subsequently, a relative standardization of the metrics was performed to rank the algorithms. Among the evaluated methods, the Hybrid-MSD, CBF, and HMSD-GF algorithms demonstrated superior performance, achieving relative scores of 3.82, 3.80, and 3.79, respectively, in the UAV-based thermal-visible image fusion task. Furthermore, following the fusion process, the grayscale output images were converted into surface luminance maps using a linear regression model. |
|
| Keywords: fusion of aerial images, fusion of images, visible image, thermal image, UAV-based photogrammetry, remote sensing |
|
|
|
Type of Study: Applicable |
Subject:
RS
Received: 2024/09/20 | Accepted: 2026/06/15 | ePublished ahead of print: 2026/06/17
|
|
|
|
|
|
|
| Send email to the article author |
|
|
