TY - JOUR
T1 - Applications of hemispherical photographs in studies of forest ecology
AU - Lin, Teng Chiu
AU - Chiang, Jyh Min
PY - 2002/9
Y1 - 2002/9
N2 - Undercanopy hemispherical photographs provide permanent records of canopy structure, and as such have great potential for studies of forest canopy dynamics. Rapid computerized image analysis has facilitated studies of undercanopy light environments and the canopy leaf area index (LAI). Temporal and spatial variations in the undercanopy light environment and canopy LAI derived from hemispherical photographs have been used to study recovery rates of forest canopies after disturbance, seedling growth, habitat quality, forest productivity, and plant ecophysiology. The orientation and range of hemispherical photographs greatly affect estimates of light indices. A self-leveling mount equipped with an LED (light-emitting diode) should be used to level the hemispherical lens and to identify the range of the image. To obtain the best contrast possible, which is critical for image classification and the most important process in image analysis, photographs should be acquired either early in the morning just before sunrise, just after sunset but prior to dusk, or on overcast days. Obtaining images with high contrast can save time and increase the efficiency and accuracy of the image analysis. Consistency of individual analyzers is critical especially when photographs from different sites are to be compared; it could be improved by re-analyzing small numbers of photographs. Undercanopy light levels often show spatial autocorrelation and lack of independence; and therefore violate assumptions required for many statistical procedures; they must be treated explicitly. In addition, the light environment estimated from hemispherical photographs does not take into account sky conditions, so such data from different sites must be compared with care.
AB - Undercanopy hemispherical photographs provide permanent records of canopy structure, and as such have great potential for studies of forest canopy dynamics. Rapid computerized image analysis has facilitated studies of undercanopy light environments and the canopy leaf area index (LAI). Temporal and spatial variations in the undercanopy light environment and canopy LAI derived from hemispherical photographs have been used to study recovery rates of forest canopies after disturbance, seedling growth, habitat quality, forest productivity, and plant ecophysiology. The orientation and range of hemispherical photographs greatly affect estimates of light indices. A self-leveling mount equipped with an LED (light-emitting diode) should be used to level the hemispherical lens and to identify the range of the image. To obtain the best contrast possible, which is critical for image classification and the most important process in image analysis, photographs should be acquired either early in the morning just before sunrise, just after sunset but prior to dusk, or on overcast days. Obtaining images with high contrast can save time and increase the efficiency and accuracy of the image analysis. Consistency of individual analyzers is critical especially when photographs from different sites are to be compared; it could be improved by re-analyzing small numbers of photographs. Undercanopy light levels often show spatial autocorrelation and lack of independence; and therefore violate assumptions required for many statistical procedures; they must be treated explicitly. In addition, the light environment estimated from hemispherical photographs does not take into account sky conditions, so such data from different sites must be compared with care.
KW - Canopy structure
KW - Hemispherical photographs
KW - Leaf area index
KW - Undercanopy light environment
UR - http://www.scopus.com/inward/record.url?scp=0036761373&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=0036761373&partnerID=8YFLogxK
M3 - Article
AN - SCOPUS:0036761373
SN - 1026-4469
VL - 17
SP - 387
EP - 400
JO - Taiwan Journal of Forest Science
JF - Taiwan Journal of Forest Science
IS - 3
ER -