Multihyperspectral Microscopic Imaging for the Precise Identification of Pollen |
| |
Authors: | Xiaolin Bi Jiaqi Xian Lijun Wei Hao Yin Migao Li |
| |
Affiliation: | 1. Guangdong Provincial Key Laboratory of Optical Fiber Sensing and Communications, Guangzhou, Guangdong, China;2. Department of Optoelectronic Engineering, Jinan University, Guangzhou, Guangdong, China;3. Guangzhou Liss Optical Instrument Co., Guangzhou, Guangdong, China |
| |
Abstract: | ![]() We report a method for precisely identifying and locating the pollen of different species in a microscope based on multihyperspectral imaging. Since pollen species are complicated and difficult to identify by one feature, both fluorescence and transmission spectral images of the samples were obtained and used for the identification together with their morphologic features. Three different species of pollens were used as test samples for investigation. The results showed that the accuracy reached 99.5%, which was a significant improvement comparing with the single spectral detection. Because more information was taken into consideration, this method has great potential for the precise identification of biological specimens. |
| |
Keywords: | Fisher algorithm multihyperspectral imaging pollen identification |
|
|