Fully convolutional networks for building and road extraction: Preliminary results
Authors:
- 3 November 2016
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1591-1594
Abstract
Available big geoscientific data and modern powerful computation hardware have laid a solid foundation for the prevailing deep learning models in the field of image classification, detection and segmentation. In these models, fully convolutional networks achieve unprecedented success in image segmentation tasks [6]. In this paper, we apply the contemporary image segmentation models in the context of extracting buildings and roads from high spatial resolution imagery. We estimate the influence of filter stride, learning rate, input data size, training epoch and fine-tuning on model performance. Selected Massachusetts road and building datasets are used for training, validation, and testing the performance of the models with different parameters. As a result of combining shallow fine-grained pooling layer outputs with the deep final-score layer or abandoning coarse-grained pooling layers, the extraction precision rate of the best modified model improves significantly to over 78%.Keywords
This publication has 9 references indexed in Scilit:
- Semantic Image Segmentation via Deep Parsing Network2015
- Conditional Random Fields as Recurrent Neural Networks2015
- Effective semantic pixel labelling with convolutional networks and Conditional Random Fields2015
- Object-based land cover mapping and comprehensive feature calculation for an automated derivation of urban structure types at block levelRemote Sensing of Environment, 2014
- Mapping Urban Tree Species Using Very High Resolution Satellite Imagery: Comparing Pixel-Based and Object-Based ApproachesISPRS International Journal of Geo-Information, 2013
- Quantifying tree mortality in a mixed species woodland using multitemporal high spatial resolution satellite imageryRemote Sensing of Environment, 2013
- A comparison of pixel-based and object-based image analysis with selected machine learning algorithms for the classification of agricultural landscapes using SPOT-5 HRG imageryRemote Sensing of Environment, 2012
- Remote sensing of impervious surfaces in the urban areas: Requirements, methods, and trendsRemote Sensing of Environment, 2012
- A Stochastic Framework for the Identification of Building Rooftops Using a Single Remote Sensing ImageIEEE Transactions on Geoscience and Remote Sensing, 2007