Management of avian populations near anthropogenic infrastructures, specifically wind farms, has been hampered due to biased bird fatalities estimates. Currently, these estimations are based on field surveys performed by humans, which is a method with low efficiency and accuracy. Detection dogs have been used for decades to assist humans, and their use for wildlife surveys is of increasing interest to scientists and wildlife managers. We evaluate the accuracy rates of human and dog-handler teams in real field conditions to address if dogs could be used instead of humans for bird carcass searches. Furthermore, to verify the efficiency of detection dogs (determined by the time spent to detect each bird carcass) searching for bird carcasses, we investigate the influence of several factors that affect the performance of dogs (carcass decomposition condition, distance to the target and weather conditions). Results indicate that dogs are more accurate than humans, independently of vegetation density. Furthermore, carcass decomposition condition, distances to the carcass and weather conditions significantly affect the efficiency of working dogs. The influence of these factors on detection time was minor. Results demonstrate the usefulness of dogs in field surveys to improve bird-strike mortality estimates at wind farms and other anthropogenic structures that cause bird fatalities worldwide.