Search

We study and analyse performance of a system for direct reflectance measurements from a drone. Key instruments of the system are upwards looking irradiance sensor and downwards looking imaging spectrometer. Requirement for both instruments is that they are radiometrically calibrated, the irradiance sensor has to be horizontally stabilized, and the sensors needs to be accurately synchronized. In our system, irradiance measurements are done with FGI Aerial Image Reference System (FGI AIRS), which uses novel optical levelling methodology and can compensate sensor tilting up to 15°. We performed SI-traceable spectral and radiance calibration of FPI hyperspectral camera at the National Physical Laboratory NPL (Teddington, UK). After the calibration, the radiance accuracy of different channels was between ±4% when evaluated with independent test data. Sensors response to radiance proved to be highly linear and was on average 0.9994 for all channels. The spectral response calibration showed side peaks on several channels that were due to the multiple orders of interference of the FPI and highlighted the importance of accurate calibration. The drone-based direct reflectance measurement system showed promising results with imagery collected over Jokioinen agricultural grass test site, Finland. AIRS-based image- and band wise image adjustment provided homogenous and seamless image mosaics even under varying illumination conditions and under clouds. [ABSTRACT FROM AUTHOR]

Light-weight 2D format hyperspectral imagers operable from unmanned aerial vehicles (UAV) have become common in various remote sensing tasks in recent years. Using these technologies, the area of interest is covered by multiple overlapping hypercubes, in other words multiview hyperspectral photogrammetric imagery, and each object point appears in many, even tens of individual hypercubes. The common practice is to calculate hyperspectral orthomosaics utilizing only the most nadir areas of the images. However, the redundancy of the data gives potential for much more versatile and thorough feature extraction. We investigated various options of extracting spectral features in the grass sward quantity evaluation task. In addition to the various sets of spectral features, we used photogrammetry-based ultra-high density point clouds to extract features describing the canopy 3D structure. Machine learning technique based on the Random Forest algorithm was used to estimate the fresh biomass. Results showed high accuracies for all investigated features sets. The estimation results using multiview data provided approximately 10% better results than the most nadir orthophotos. The utilization of the photogrammetric 3D features improved estimation accuracy by approximately 40% compared to approaches where only spectral features were applied. The best estimation RMSE of 239 kg/ha (6.0%) was obtained with multiview anisotropy corrected data set and the 3D features. [ABSTRACT FROM AUTHOR]