A Calibration Model for Field-of-View Effect in Thermal Infrared Remote Sensing Images From Uncrewed Aerial Vehicles

Quantitative thermal infrared (TIR) remote sensing with uncrewed aerial vehicles (UAVs) enables accurate measurement and analysis of ground temperature, temperature gradients, and related parameters. This technology is widely applied in areas such as agricultural monitoring, environmental assessment, and urban heat island studies. However, the field-of-view (FOV) effect arises from the use of central projection rather than orthographic projection in TIR imaging. This causes the edge pixels to represent larger ground areas than the center pixels under the same FOV, leading to temperature discrepancies. Such inconsistency can inevitably cause serious interference in subsequent quantitative analysis. To address this, this study proposes a calibration model by analyzing the 3-D issue in 2-D planes. Using the tangent theorem, a model is constructed to calculate the ground area ratios between nonorthorectified and center-orthorectified pixels, and calibrations are applied accordingly. The calibrated data agree well with ground measurements, achieving a coefficient of determination ( $R^{2}$ ) of 0.99, a root mean square error (RMSE) of 0.52 K, and a mean bias error (MBE) of −0.05 K. In addition, the calibrated images exhibit improved brightness uniformity at the edges, aligning more closely with orthophoto characteristics. The proposed calibration model significantly enhances the temperature consistency of TIR images, facilitating the generation of angle-independent temperature maps to support informed decision-making in various applications.