Automatic track recognition of footprints for identifying cryptic species

摘要: The recognition of tracks plays an important role in ecological research and monitoring, tracking tunnels are a cost-effective method for indexing species over large areas. Traditionally, collected by system, analysis is carried out manual identification procedure experienced wildlife biologists. Unfortunately, human experts unable to reliably distinguish morphologically similar species. We propose new using image analysis, which allows automatic tracks, apply the identifying cryptic small-mammal demonstrate footprints three invasive rat with morphology that co-occur New Zealand, including detection recent invasion rat-free island. Automatic footprint successfully identified >70% footprints, >83% cards. With appropriate changes recognition, could be broadly applicable any taxa can tracked. Identification level gives better estimates presence composition communities.