Automatic land cover classification of geo-tagged field photos by deep learning
Affiliation:
1. Ministry of Education Key Laboratory of Biodiversity Science and Ecological Engineering, Institute of Biodiversity Science, Fudan University, Shanghai 200433, China;2. Department of Microbiology and Plant Biology, Center for Spatial Analysis, University of Oklahoma, Norman, OK 73019, USA;3. Division of Climate Change, Institute for Agro-Environmental Sciences, NARO, Tsukuba, 305-8604, Japan;4. Department of Environmental Science, Policy and management, University of California, Berkeley, CA 94720, USA;5. National Center for Agro Meteorology, Seoul 08826, South Korea;6. National Institute of Agricultural Sciences, Rural Development Administration, Wanju, 55365, South Korea;7. Rubber Research Institute (RRI), Chinese Academy of Tropical Agricultural Sciences(CATAS), Hainan Province, 571737, P. R. China;8. Institute of Geographical Science and Natural Resource Research, Chinese Academy of Sciences, Beijing 100101, China;9. International Center for Agricultural Research in Dry Areas, Amman, 11195, Jordan;1. Department of Microbiology and Plant Biology, Center for Spatial Analysis, University of Oklahoma, Norman, OK 73019, USA;2. Key Laboratory of Land Surface Pattern and Simulation, Institute of Geographic Sciences and Natural Resources Research, CAS, Beijing 100101, China;3. LinkedIn, LLC, San Francisco, CA 94105, USA;4. Ministry of Education, Key Laboratory of Biodiversity Science and Ecological Engineering, Institute of Biodiversity Science, Fudan University, Shanghai 200433, China;5. Plankton Ecology and Limnology Laboratory, Department of Biology, University of Oklahoma, Norman, OK 73019, USA;1. Physical Geography, Trier University, Trier, Germany;2. Department of Civil Engineering, University of Coimbra, Coimbra, Portugal;1. International Institute for Applied Systems Analysis, Laxenburg, Austria;2. University of Freiburg, Freiburg, Germany
Abstract:
With more and more crowdsourcing geo-tagged field photos available online, they are becoming a potentially valuable source of information for environmental studies. However, the labelling and recognition of these photos are time-consuming. To utilise such information, a land cover type recognition model for field photos was proposed based on the deep learning technique. This model combines a pre-trained convolutional neural network (CNN) as the image feature extractor and the multinomial logistic regression model as the feature classifier. The pre-trained CNN model Inception-v3 was used in this study. The labelled field photos from the Global Geo-Referenced Field Photo Library (http://eomf.ou.edu/photos) were chosen for model training and validation. The results indicated that our recognition model achieved an acceptable accuracy (48.40% for top-1 prediction and 76.24% for top-3 prediction) of land cover classification. With accurate self-assessment of confidence, the model can be applied to classify numerous online geo-tagged field photos for environmental information extraction.
