AIRBORNE DETECTION OF SOUTHERN PINE-BEETLE DAMAGE USING KEY SPECTRAL BANDS

Citation
Ga. Carter et al., AIRBORNE DETECTION OF SOUTHERN PINE-BEETLE DAMAGE USING KEY SPECTRAL BANDS, Canadian journal of forest research (Print), 28(7), 1998, pp. 1040-1045
Citations number
22
Categorie Soggetti
Forestry
ISSN journal
00455067
Volume
28
Issue
7
Year of publication
1998
Pages
1040 - 1045
Database
ISI
SICI code
0045-5067(1998)28:7<1040:ADOSPD>2.0.ZU;2-H
Abstract
Damage by the southern pine beetle (SPB) (Dendroctonus frontalis Zimm. ) occurs frequently in the southeastern United States and can result i n tree death over large areas. A new technique for detection of SPB ac tivity was tested for shoaleaf pine (Pinus echinata Mill.) in the Cane y Creek Wilderness, Ouachita National Forest, Arkansas. Digital images with l-m pixel resolution were acquired from a light aircraft in 6- t o 10-nm bandwidths centered at wavelengths of 675, 698, and 840 nm. Th e 675-nm band was selected to yield a maximum contrast between yellow or brown versus green foliage. The 698-nm band was selected based on i ts high sensitivity to leaf chlorophyll content to enable detection of less severe chlorosis in more recently damaged trees. The 840-nm band was used as a reference band that is not sensitive to chlorophyll. Im ages acquired within each band were calibrated to percent reflectance based on the known reflectances of a gray scale placard-located on the ground. Individual trees with yellow to brown foliage were easily loc ated in the 675- and 698-nm images. Milder chlorosis in more recently damaged pines was detected by a normalized difference vegetation index (NDVI) that was derived from 698- and 840-nm reflectances. Although s tatistically significant, the contrast of recently infested trees vers us undamaged trees was generally visually poor in NDVI or color compos ite images. This was apparently a result of the inherent variability i n leaf chlorophyll content throughout the forest. The increased reflec tance near 700 nm characteristic of recent damage likely would be reso lved more easily in pine plantations of low species diversity. Images of a NDVI that was based on 675- and 840-nm reflectances produced the strongest contrast between heavily damaged and undamaged trees.