摘要：As thermal infrared imaging technology has improved, it has increasingly been used for estimating sizes of wildlife populations. The greatest bias of thermal infrared surveys is the lack of known detection rates to adjust for visibility bias. As with visual surveys, a measure of detection rate is needed to provide unbiased estimates. We assessed the detection rate of white-tailed deer (Odocoileus virginianus) using a thermal infrared sensor (1.2 - 5.9 /an) from an aerial platform. Similar characteristics between thermal signatures of people and deer allowed us to use people in a reclined or horizontal position as surrogates for deer. We conducted a census of 2.56 km2 within which 20 people were randomly placed. We detected 75.0% of the people (n = 20) across the area and 93.8% of the people (n= 16) when the effect of water was taken into consideration. Thermal signatures of people and deer occupying flooded areas were likely masked by the surrounding thermal signature of water. We found the method worked well in bottomland hardwood forests under dry conditions. As with visual aerial population counting methods, detection rates for an area should be developed to provide unbiased estimates.
