Deep Learning for Building Density Estimation in Remotely Sensed Imagery Uzaktan Algilanan Görüntülerde Bina Yogunlugu Kestirimi Ǐin Derin Ögrenme


Suberk N. T., Ates H. F.

4th International Conference on Computer Science and Engineering, UBMK 2019, Samsun, Turkey, 11 - 15 September 2019, pp.423-428 identifier identifier

  • Publication Type: Conference Paper / Full Text
  • Doi Number: 10.1109/ubmk.2019.8907133
  • City: Samsun
  • Country: Turkey
  • Page Numbers: pp.423-428
  • Keywords: remote sensing, deep learning, building density estimation
  • Istanbul Medipol University Affiliated: Yes

Abstract

This paper is about point-wise estimation of building density from remote sensing optical imagery using deep learning methods. Convolutional neural network (CNN) based deep learning approaches are used for this work. Pre-trained VGG-16 and FCN-8s deep architectures are adapted to the problem and fine-tuned with additional training. Estimated values are used to generate building heat maps in urban areas. Comparative simulation results of the two architectures reveal that accurate density estimation is possible without the need for detailed maps of building locations during supervised training.