Study on the Relationship between the Forest CanopyClosure and Hyperspectral Signatures

摘要

Forest canopy density is an ideal representative ofrnthe forest habitat situations. It can directly or indirectly depictrnthe canopy structure and gap size in the forestland, thus couldrnbe applied to assessment of wildlife's diversit y. Sincernpopulation survey of vegetation and wildlife diversities is arnkey issue for sustainable forest ecosystem management, manyrnresearch efforts have been focused on forest canopy densityrnusing multispectral data in the last two decades. Unfortunately,rnprediction of canopy density using large scaling remoternsensing data remains a challenging issue. Due to recentrnadvances in hyperspectral image sensors hyperspectralrnimagery is now available for environmental monitoring. In thisrnpaper, we conduct experiments to monitor complicatedrnenvironments of forestland that can be captured by usingrnhyperspectral imagery and further be analyzed to test arnprediction model of forest canopy density. The results showrnthat 95% of canopy density could be well described by using 2rndifference vegetation indices (DVIs), which are difference ofrnblue and green reflectances rband_100-rband_150 and difference ofrn2 short wave infrared reflectancse rband_406-rband_410 With thernwavelengths of band no. 100, 150, 406, and 410 specified byrn462.39 nm, 534.40 nm, 918.22 nm and 924.41 nm respectively.