Recognition and Semantic Information Extraction for Map Based on Deep Learning

EasyChair Preprint 10633

Abstract

Geospatial information contained in maps plays an important role in geographic information data acquisition, map understanding, intelligent mapping and other applications. In terms of map recognition and geospatial information extraction from maps, traditional methods that heavily rely on human or human-computer interaction for semantic recognition can no longer meet the real-time needs. In this paper, we first analysed the composition and characteristics of maps, and then systematically illustrated the semantic understanding methods of map image recognition, target detection of geographic features and semantic segmentation of geographic features based on deep learning architecture, which is crucial to intelligent map recognition and mapping.

Keyphrases: CNN, Geographic Feature Detection, Map Recognition, map, map reading, object detection, recognition, semantic segmentation

@booklet{EasyChair:10633,
  author    = {Yong Wang and Kaixuan Du and Xianghong Che and Ruiyuan Ma and Fu Ren},
  title     = {Recognition and Semantic Information Extraction for Map Based on Deep Learning},
  howpublished = {EasyChair Preprint 10633},
  year      = {EasyChair, 2023}}