In cooperation with the Agricultural institute of Slovenia we included in our offers hyperspectral imaging. The main parts of the hyperspectral system form two cameras: VNIR covers wavelengths from 400 to 980 nm, while the second, SWIR, records short-wave infrared wavelengths, from 950 to 2500 nm. Average spectral resolution is 5 nm, which allows for identification of individual observed elements and their states. Spatial resolution depends on flying speed and altitude. We can provide spatial resolutions below 0.5 m.
Hyperspectral imaging can be used in a wide variety of applications, for example in archeology, agronomy, biology and ecology, geology, and urbanism,...
Ortorectified and geographically less accurate images – accuracy up to 5 m, without image classification.
Ortorectified and georeferenced images with accuracies up to 0.5 m, with or without image classification.
Analysis of thematic images – raster or vector records of desired elements.
The shape and chemical composition of objects under study (e.g. plants, rock formations, materials) affects reflected light , which is recorded by the cameras. Because of each elements distinct and often unique chemical composition, the reflected spectrum is often element specific. The reflected spectra is often referred to as a spectral signature.
Figure 1: Characteristic spectral signatures, in the visible to near-infrared spectrum, of common elements.
Characteristic spectral signatures enable us to identify plant species, minerals, materials and vegetation states, e.g. nutritional status, water stress, diseases and pest infestations. If the differences in spectral signatures are large enough identification and quantification are quite straightforward. As is evident from the spectral signatures in Figure 1, we can easily distinguish water, traffic infrastructure and vegetation. Often even the differences between plant species are large enough for comparatively easy identification. Determining the physiological status of plants, water stress or amount of organic matter in the upper layer of soil is more complicated. In these cases ground truthing (measuring spectral signatures at short distances) protocols have to be prepared, in order to facilitate reliable identification and quantification in airborne images.
Hyperspectral imaging is suitable for all projects where a good spectral and spatial resolution is required. Both resolutions are superior to currently available satellite images, and therefore expedient for classifying and accurately defining individual spatial elements.
Figure 2: Section of an ortorectified VNIR image (each pixel carries data from 160 spectral bands)
Figure 3: Section of an ortorectified SWIR image (each pixel carries data from 288 spectral bands)