Spatial Calibration of the Hyperspectral Line Scanner by the Bundle Block Adjusting Method

Sa？etak Close-range hyperspectral imaging of objects encompasses the collection and processing of the spectral characteristics of the reflected electromagnetic (EM) radiation from object's surface at such distances from the sensor that they cannot be considered to be at optical infinity. Hyperspectral sensors used in close-range imaging applications are primarily calibrated in the spectral domain while their spatial calibration is questionable or not implemented at all. Data obtained by hyperspectral scanners are widely used, however, unlike most commonly used handheld spectrometers, the calibrated line scanner, besides the high-quality spectral data, poses the big potential in providing the accurate spatial information, too. To geolocate the spectral information gathered by the sensor, the mathematical model of the imaging the scene to image plane has to be derived and implemented, taking into consideration the expected systematic errors that influence the measuring results. That's why in this paper a new algorithm for calibrating the push-broom line scanner is derived and presented. The testing was done by hyperspectral line scanner ImSpector V9 + PixelFly