Abstract
Hyperspectral imaging system can be used to measure the object in a continuous waveband, which can capture the spatial information and spectral information simultaneously. So hyperspectral images can not only reflect the external characteristics of the object through the spatial information, but also reflect its internal qualities of the spectral information. Based on this advantage, hyperspectral imaging has been widely used in the quality detection of agricultural products. Firstly, this paper summarizes the application of different imaging modes under different conditions based on hyperspectral imaging. Then it sums up the methods of spectral preprocessing and their applications in hyperspectral systems, the multiplicative scatter correction, the standard normal variable, the savitzky-golay smoothing, median-filter and the spectral differential all can correct the spectrum effectively in diverse backgrounds. Again, in this paper some common methods of hyperspectral data reduction are summarized either, the methods of principal component analysis, partial least squares, optimum index factor, successive projection algorithm and load factor are all widely used in reduction of hyperspectral data in agricultural products, these methods mentioned above can decrease the data dimension by feature extraction or feature selection, not only to simplify the computational process but to optimize conclusions through reduce redundancy information.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call