With the decrease in Arctic sea ice extent and the changes in ice characteristics, there is an immediate concern for the effects of these changes on the ecosystems of the Arctic. In the last three years, innovative approaches have been applied to address this concern. Unmanned aircraft, in addition to manned aircraft, are being flown in the Arctic to collect images of the ice. The unmanned flights have occurring from all eight of the Arctic countries. Analysis of the tens of thousands of images that are collected requires automation to allow for systematic and comprehensive analysis of the information gathered. We report here on the analysis of images and on improvements in the ability of software to assist in the identification of ice seals. Two significant benefits of this automation assistance are the reduction of researcher-to-researcher differences in the identification of seals, and the consistent analysis of seal attributes including length and size estimates. The software developments are a part of a larger, open source effort that is just starting to provide software that improve biologists’ ability to interpret the large amounts of image data becoming available for mammal surveys. In the case of seals, the simultaneous acquisition of sea-ice and population data may support new means of assessing habitat requirements and threats posed by loss of ice.
This publication was presented at the following: