Information extraction from canopy spectral feature of Comnyza Canadensis (L.) Cronq

In this paper, we choosed De-xing copper mine as study area, collected canopy spectral data and determined the heavy metal concentration in Comnyza Canadensis (L.) Cronq leaves. These plants in background field have significant spectral feature, such as the low reflectance, the biggest absorption valley depth and the largest red valley area. Compared with spectral feature of Comnyza Canadensis (L.) Cronq in background field, the spectral reflectance of these plants in acid mine drainage contaminated area and in mine waste reservoir increase obviously. Some indexes, including the depth of red valley, the area of red valley and the water absorption valleys, decrease. While the red edge peak and its area increase and there are blue-shift obviously at red edge. The single factor analysis showed that most spectral parameters have a good correlation to the Pb and Cd concentration in leaves. The stepwise regression analysis results showed that spectral parameters are more closely related to the Pb concentration (r²=0.9338), and high estimation accuracy can be gained via stepwise regression. However, the accuracy is low for Cd concentration estimation by spectral parameters.