Discrimination of Spectrally-Close Crops Using Ground-Based Hyperspectral Data |
| |
| Authors: | K R Manjunath Shibendu Shankar Ray Sushma Panigrahy |
| |
| Institution: | (1) Agriculture, Terrestrial Biosphere & Hydrology Group, EPSA, Space Applications Centre, Ahmedabad, 380 015, India; |
| |
| Abstract: | Hyperspectral remote sensing, because of its large number of narrow bands, has shown possibility of discriminating the crops.
Current study was carried out to select the optimum bands for discrimination among pulses, cole crops and ornamental plants
using the ground-based Hyperspectral data in Patha village, Lalitpur district, Uttar Pradesh state and Kolkata, West Bengal
state. The field observations of reflectance were taken using a 512-channel spectroradiometer with a range of 325–1075 nm.
The stepwise discriminant analysis was carried out and separability measures, such as Wilks’ lambda and F-Value were used as criteria for identifying the narrow bands. The analysis showed that, the best four bands for pulse crop
discrimination lie mostly in NIR and early MIR regions i.e. 750, 800, 940 and 960 nm. Within cole crops discrimination is
primarily determined by the green, red and NIR bands of 550, 690, 740, 770 and 980 nm. The separability study showed the bands
420,470,480,570,730,740, 940, 950, 970, 1030 nm are useful for discriminating flowers. |
| |
| Keywords: | |
| 本文献已被 SpringerLink 等数据库收录！ |
|
