Jump to Main Content
Automated identification of intensive animal production locations from aerial photography
- Sheffield, KJ, Hunnam, JC, Cuzner, TN, Morse‐McNabb, EM, Sloan, SM, Nunan, J, Smith, J, Harvey, W, Lewis, H
- Australian veterinary journal 2018 v.96 no.9 pp. 323-331
- aerial photography, algorithms, animal production, automation, biosecurity, cost effectiveness, data collection, databases, disease outbreaks, image analysis, industry, infrastructure, irrigation, land use change, landscapes, remote sensing, roadsides, sheep, surveys, Australia
- OBJECTIVE: Successful control of an emergency animal disease outbreak requires the timely and accurate identification of properties of interest. The identification of commercial piggeries within study areas in the Goulburn–Murray Irrigation District in Victoria, Australia, is used to demonstrate the innovative application of object‐based image analysis (OBIA) techniques for the identification of intensive animal production land uses, to improve the accuracy of existing datasets. METHODS: Characteristics of infrastructure and landscape features were combined to form a commercial piggery identification algorithm. These criteria were applied to recent aerial photography that had been classified using OBIA techniques. The results were then compared with three datasets containing known commercial piggery locations and visually checked by roadside surveys. RESULTS: The OBIA technique identified 21 potential piggery locations across three study areas, 14 of which were identified in existing databases. Of the 7 additional sites, 4 were dairy properties, 1 was a cropping and sheep property and 2 were previously undocumented piggery locations. CONCLUSIONS: The OBIA approach has potential of OBIA for identifying the locations of commercial piggeries. Further development and testing will determine how generic this approach is in terms of industry type and operation size. The method described is cost‐effective, automated and repeatable, and could be used to regularly update existing databases by analysing newly acquired aerial imagery to identify possible land use changes. This would improve the reliability of currently available data and increase the effectiveness of a biosecurity response during an emergency animal disease outbreak.