VERITAS: An Integrated Framework for Verification of Spatial Accuracy and Spectral Characteristics of Airborne Remote Sensing Sensors

The increasing use of airborne LiDAR, multispectral and hyperspectral sensors has created a growing demand for reliable methods to verify both spatial accuracy and radiometric quality of remotely sensed data. While individual calibration procedures are well established for specific sensor types, comprehensive approaches integrating geometric verification, radiometric calibration and artificial intelligence remain limited. This paper presents the conceptual framework of the project Development of AI-based tools for verifying the positional accuracy and spectral characteristics of remote sensing sensors (VERITAS), which aims to develop an integrated methodology for the verification of spatial accuracy and spectral characteristics of airborne remote sensing sensors. The proposed framework combines reference calibration targets, precise geospatial measurements, radiometric calibration protocols, hyperspectral reference databases and artificial intelligence techniques into a unified workflow. Particular emphasis is placed on the automatic identification of calibration targets in hyperspectral imagery using deep learning models trained on synthetic and real datasets, enabling accurate determination of their spatial characteristics and supporting automated quality assessment of remotely sensed data. The framework further incorporates methodologies for evaluating the positional accuracy of LiDAR point clouds and for assessing the spectral consistency of multispectral and hyperspectral observations using standardized quantitative metrics. The expected outputs include methodologies for spatial and radiometric calibration of airborne remote sensing sensors, AI-based tools for automatic identification and spatial delineation of calibration targets, a database of calibration protocols and reference spectral signatures, and procedures for quantitative assessment of sensor performance. The proposed framework establishes a unified methodological basis for the integrated spatial and radiometric verification of airborne LiDAR, multispectral and hyperspectral sensors, while enabling the development and validation of AI-based methods for automated interpretation of remote sensing data.