Detection of Cannabis Plants by Hyper-spectral Remote Sensing Means

The dramatic increase of drug use, mostly hashish and marijuana, reinforces the demand for drug prevention and the need for accurate and updated information on cannabis fields. This study sought to evaluate the ability of Hyperspectral spectroscopy to discriminate cannabis from different scales and land use. The study was conducted in three stages: 1) Examination of the cannabis spectrum under laboratory controlled conditions from a short distance with field spectrometer and Hyperspectral camera under artificial light; 2) Remote sensing of the cannabis from an oblique view using static imaging spectrometer from 25m and 80m; and 3) Airborne Hyperspectral pushbroom sensor (AISA Eagle 400-1000 nm). This method of down/up scaling was found to be useful in understanding the meaning of spectral discrimination. Results of Principal Component Analysis (PCA) show that the spectral signal of cannabis (leaf and canopy) varied with distance from the sensor, however spectral bands with the most influence are in the range of 530-550, 670-680 nm and 705-720nm.