Leveraging Deep Convolutional Neural Network for Point Symbol Recognition in Scanned Topographic Maps
Point symbols on a scanned topographic map (STM) provide crucial geographic information. However, point symbol recognition entails high complexity and uncertainty owing to the stickiness of map elements and singularity of symbol structures. Therefore, extracting point symbols from STMs is challengin...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-03-01
|
| Series: | ISPRS International Journal of Geo-Information |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2220-9964/12/3/128 |
| _version_ | 1797611353914998784 |
|---|---|
| author | Wenjun Huang Qun Sun Anzhu Yu Wenyue Guo Qing Xu Bowei Wen Li Xu |
| author_facet | Wenjun Huang Qun Sun Anzhu Yu Wenyue Guo Qing Xu Bowei Wen Li Xu |
| author_sort | Wenjun Huang |
| collection | DOAJ |
| description | Point symbols on a scanned topographic map (STM) provide crucial geographic information. However, point symbol recognition entails high complexity and uncertainty owing to the stickiness of map elements and singularity of symbol structures. Therefore, extracting point symbols from STMs is challenging. Currently, point symbol recognition is performed primarily through pattern recognition methods that have low accuracy and efficiency. To address this problem, we investigated the potential of a deep learning-based method for point symbol recognition and proposed a deep convolutional neural network (DCNN)-based model for this task. We created point symbol datasets from different sources for training and prediction models. Within this framework, atrous spatial pyramid pooling (ASPP) was adopted to handle the recognition difficulty owing to the differences between point symbols and natural objects. To increase the positioning accuracy, the k-means++ clustering method was used to generate anchor boxes that were more suitable for our point symbol datasets. Additionally, to improve the generalization ability of the model, we designed two data augmentation methods to adapt to symbol recognition. Experiments demonstrated that the deep learning method considerably improved the recognition accuracy and efficiency compared with classical algorithms. The introduction of ASPP in the object detection algorithm resulted in higher mean average precision and intersection over union values, indicating a higher recognition accuracy. It is also demonstrated that data augmentation methods can alleviate the cross-domain problem and improve the rotation robustness. This study contributes to the development of algorithms and the evaluation of geographic elements extracted from STMs. |
| first_indexed | 2024-03-11T06:26:43Z |
| format | Article |
| id | doaj.art-1032e439613747c385c518470b44f803 |
| institution | Directory Open Access Journal |
| issn | 2220-9964 |
| language | English |
| last_indexed | 2024-03-11T06:26:43Z |
| publishDate | 2023-03-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | ISPRS International Journal of Geo-Information |
| spelling | doaj.art-1032e439613747c385c518470b44f8032023-11-17T11:28:27ZengMDPI AGISPRS International Journal of Geo-Information2220-99642023-03-0112312810.3390/ijgi12030128Leveraging Deep Convolutional Neural Network for Point Symbol Recognition in Scanned Topographic MapsWenjun Huang0Qun Sun1Anzhu Yu2Wenyue Guo3Qing Xu4Bowei Wen5Li Xu6Institute of Geospatial Information, Information Engineering University, Zhengzhou 450001, ChinaInstitute of Geospatial Information, Information Engineering University, Zhengzhou 450001, ChinaInstitute of Geospatial Information, Information Engineering University, Zhengzhou 450001, ChinaInstitute of Geospatial Information, Information Engineering University, Zhengzhou 450001, ChinaInstitute of Geospatial Information, Information Engineering University, Zhengzhou 450001, ChinaInstitute of Geospatial Information, Information Engineering University, Zhengzhou 450001, ChinaInstitute of Geospatial Information, Information Engineering University, Zhengzhou 450001, ChinaPoint symbols on a scanned topographic map (STM) provide crucial geographic information. However, point symbol recognition entails high complexity and uncertainty owing to the stickiness of map elements and singularity of symbol structures. Therefore, extracting point symbols from STMs is challenging. Currently, point symbol recognition is performed primarily through pattern recognition methods that have low accuracy and efficiency. To address this problem, we investigated the potential of a deep learning-based method for point symbol recognition and proposed a deep convolutional neural network (DCNN)-based model for this task. We created point symbol datasets from different sources for training and prediction models. Within this framework, atrous spatial pyramid pooling (ASPP) was adopted to handle the recognition difficulty owing to the differences between point symbols and natural objects. To increase the positioning accuracy, the k-means++ clustering method was used to generate anchor boxes that were more suitable for our point symbol datasets. Additionally, to improve the generalization ability of the model, we designed two data augmentation methods to adapt to symbol recognition. Experiments demonstrated that the deep learning method considerably improved the recognition accuracy and efficiency compared with classical algorithms. The introduction of ASPP in the object detection algorithm resulted in higher mean average precision and intersection over union values, indicating a higher recognition accuracy. It is also demonstrated that data augmentation methods can alleviate the cross-domain problem and improve the rotation robustness. This study contributes to the development of algorithms and the evaluation of geographic elements extracted from STMs.https://www.mdpi.com/2220-9964/12/3/128point symbol recognitionscanned topographic mapdeep learninggeneralization abilitydata augmentation |
| spellingShingle | Wenjun Huang Qun Sun Anzhu Yu Wenyue Guo Qing Xu Bowei Wen Li Xu Leveraging Deep Convolutional Neural Network for Point Symbol Recognition in Scanned Topographic Maps ISPRS International Journal of Geo-Information point symbol recognition scanned topographic map deep learning generalization ability data augmentation |
| title | Leveraging Deep Convolutional Neural Network for Point Symbol Recognition in Scanned Topographic Maps |
| title_full | Leveraging Deep Convolutional Neural Network for Point Symbol Recognition in Scanned Topographic Maps |
| title_fullStr | Leveraging Deep Convolutional Neural Network for Point Symbol Recognition in Scanned Topographic Maps |
| title_full_unstemmed | Leveraging Deep Convolutional Neural Network for Point Symbol Recognition in Scanned Topographic Maps |
| title_short | Leveraging Deep Convolutional Neural Network for Point Symbol Recognition in Scanned Topographic Maps |
| title_sort | leveraging deep convolutional neural network for point symbol recognition in scanned topographic maps |
| topic | point symbol recognition scanned topographic map deep learning generalization ability data augmentation |
| url | https://www.mdpi.com/2220-9964/12/3/128 |
| work_keys_str_mv | AT wenjunhuang leveragingdeepconvolutionalneuralnetworkforpointsymbolrecognitioninscannedtopographicmaps AT qunsun leveragingdeepconvolutionalneuralnetworkforpointsymbolrecognitioninscannedtopographicmaps AT anzhuyu leveragingdeepconvolutionalneuralnetworkforpointsymbolrecognitioninscannedtopographicmaps AT wenyueguo leveragingdeepconvolutionalneuralnetworkforpointsymbolrecognitioninscannedtopographicmaps AT qingxu leveragingdeepconvolutionalneuralnetworkforpointsymbolrecognitioninscannedtopographicmaps AT boweiwen leveragingdeepconvolutionalneuralnetworkforpointsymbolrecognitioninscannedtopographicmaps AT lixu leveragingdeepconvolutionalneuralnetworkforpointsymbolrecognitioninscannedtopographicmaps |