Optimised U-Net for Land Use–Land Cover Classification Using Aerial Photography

Abstract

<p>Convolutional Neural Networks (CNN) consist of various hyper-parameters which need to be specifed or can be altered when defning a deep learning architecture. There are numerous studies which have tested diferent types of networks (e.g. U-Net, DeepLabv3+) or created new architectures, benchmarked against well-known test datasets. However, there is a lack of real-world mapping applications demonstrating the efects of changing network hyper-parameters on model performance for land use and land cover (LULC) semantic segmentation. In this paper, we analysed the efects on training time and classifcation accuracy by altering parameters such as the number of initial convolutional flters, kernel size, network depth, kernel initialiser and activation functions, loss and loss optimiser functions, and learning rate. We achieved this using a well-known top performing architecture, the U-Net, in conjunction with LULC training data and two multispectral aerial images from North Queensland, Australia. A 2018 image was used to train and test CNN models with diferent parameters and a 2015 image was used for assessing the optimised parameters. We found more complex models with a larger number of flters and larger kernel size produce classifcations of higher accuracy but take longer to train. Using an accuracy-time ranking formula, we found using 56 initial flters with kernel size of 5×5 provide the best compromise between training time and accuracy. When fully training a model using these parameters and testing on the 2015 image, we achieved a kappa score of 0.84. This compares to the original U-Net parameters which achieved a kappa score of 0.73.</p>