Unmanned aerial vehicles (UAV) are being used for low altitude remote sensing for\nthematic land classification using visible light and multi-spectral sensors. The objective of this\nwork was to investigate the use of UAV equipped with a compact spectrometer for land cover\nclassification. The UAV platform used was a DJI Flamewheel F550 hexacopter equipped with GPS\nand Inertial Measurement Unit (IMU) navigation sensors, and a Raspberry Pi processor and camera\nmodule. The spectrometer used was the FLAME-NIR, a near-infrared spectrometer for hyperspectral\nmeasurements. RGB images and spectrometer data were captured simultaneously. As spectrometer\ndata do not provide continuous terrain coverage, the locations of their ground elliptical footprints\nwere determined from the bundle adjustment solution of the captured images. For each of the\nspectrometer ground ellipses, the land cover signature at the footprint location was determined\nto enable the characterization, identification, and classification of land cover elements. To attain a\ncontinuous land cover classification map, spatial interpolation was carried out from the irregularly\ndistributed labeled spectrometer points. The accuracy of the classification was assessed using spatial\nintersection with the object-based image classification performed using the RGB images. Results\nshow that in homogeneous land cover, like water, the accuracy of classification is 78% and in mixed\nclasses, like grass, trees and manmade features, the average accuracy is 50%, thus, indicating the\ncontribution of hyperspectral measurements of low altitude UAV-borne spectrometers to improve\nland cover classification.
Loading....