Hyperspectral cameras record a 2D projection of the measured object by scanning lines and measuring spectra of each scan-line. This delivers a data-cube consisting of a stack of images of the object’s reflectance throughout a range of wavelengths. Such data cubes are generally large and targeted data evaluation is necessary for phenotyping purposes .
Multispectral cameras record a selection of a series of wavelength bands, but do not record contineous spectra as hyperspectral cameras do. Thus, they collect less data, but they are faster and easier to operate than hyperspectral cameras.
Vegetation indices  are calculated from a number of wavelengths – typically two or three spectral bands – from the measured spectra. They characterise plant properties such as physiological or structural characteristics, pigment content or responses to biotic and abiotic stress. Furthermore, data mining and machine learning approaches can serve for extracting detailed data out of the data cubes.
 Govender, M.; Chetty, K.; Bulcock, H. (2007): A review of hyperspectral remote sensing and its application in vegetation and water resource studies. In: Water Sa
33 (2). http://www.ajol.info/index.php/wsa/article/view/49049
 Jackson, Ray D.; Huete, Alfredo R. (1991): Interpreting vegetation indices. In: Preventive Veterinary Medicine 11 (3), S. 185–200. http://www.sciencedirect.com/science/article/pii/S0167587705800042