A Fully Automated Computer Vision System for the Biometric Identification of African Penguins (Spheniscus demersus) on Robben Island

African penguins (Spheniscus demersus) carry a unique, unchanging configuration of spots on their chests. We have developed a biometric vision system that locates, extracts and employs this pattern information from photo and video to provide individual identification by remote, completely non-invasive observation. We will demonstrate the potential of our prototype system on Robben Island to detect the movements of all the penguins in a colony. In static mode the system identifies about 20% of the penguins passing the camera (the chests of most of the 80% that are not identified are occluded by other birds). The proportion of birds that are incorrectly identified is less than one in one thousand. As birds tend to use the same paths every day it is probable that all birds using that path are seen and identified within a month. In dynamic mode the camera actively seeks out penguins capturing images of their complete un-occluded chests; thus significantly increasing the proportion of passing birds identified. We will also present early results on the movements of penguins from the initial deployment of the system as a 24/7 monitoring device on Robben Island. Finally, we will discuss how the system could be used in other species.