Journal cover Journal topic
Proceedings of the ICA
Journal topic
Articles | Volume 4
Proc. Int. Cartogr. Assoc., 4, 109, 2021
https://doi.org/10.5194/ica-proc-4-109-2021
Proc. Int. Cartogr. Assoc., 4, 109, 2021
https://doi.org/10.5194/ica-proc-4-109-2021

  03 Dec 2021

03 Dec 2021

Automatic vectorization of point symbols on archive maps using deep convolutional neural network

Gergely Vassányi and Mátyás Gede Gergely Vassányi and Mátyás Gede
  • Eötvös Loránd University Budapest, Department of Cartography and Geoinformatics, Hungary

Keywords: MRCNN, vectorization, map symbols, Python

Abstract. Archive topographical maps are a key source of geographical information from past ages, which can be valuable for several science fields. Since manual digitization is usually slow and takes much human resource, automatic methods are preferred, such as deep learning algorithms. Although automatic vectorization is a common problem, there have been few approaches regarding point symbols. In this paper, a point symbol vectorization method is proposed, which was tested on Third Military Survey map sheets using a Mask Regional Convolutional Neural Network (MRCNN). The MRCNN implementation uses the ResNet101 network improved with the Feature Pyramid Network architecture and is developed in a Google Colab environment. The pretrained network was trained on four point symbol categories simultaneously. Results show 90% accuracy, while 94% of symbols detected for some categories on the complete test sheet.

Publications Copernicus
Download
Citation