Case Report

Split Viewer

Econ. Environ. Geol. 2022; 55(5): 551-561

Published online October 31, 2022

https://doi.org/10.9719/EEG.2022.55.5.551

© THE KOREAN SOCIETY OF ECONOMIC AND ENVIRONMENTAL GEOLOGY

Automated Analyses of Ground-Penetrating Radar Images to Determine Spatial Distribution of Buried Cultural Heritage

Moonhee Kwon1,2, Seung-Sep Kim1,3,*

1Department of Astronomy, Space Science and Geology, Chungnam National University, Daejeon 34134, Korea
2Archaeology Division, National Research Institute of Cultural Heritage, Daejeon 34122, Korea
3Department of Geological Sciences, Chungnam National University, Daejeon 34134, Korea

Correspondence to : *seungsep@cnu.ac.kr

Received: September 21, 2022; Revised: October 28, 2022; Accepted: October 28, 2022

This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/3.0) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided original work is properly cited.

Abstract

Geophysical exploration methods are very useful for generating high-resolution images of underground structures, and such methods can be applied to investigation of buried cultural properties and for determining their exact locations. In this study, image feature extraction and image segmentation methods were applied to automatically distinguish the structures of buried relics from the high-resolution ground-penetrating radar (GPR) images obtained at the center of Silla Kingdom, Gyeongju, South Korea. The major purpose for image feature extraction analyses is identifying the circular features from building remains and the linear features from ancient roads and fences. Feature extraction is implemented by applying the Canny edge detection and Hough transform algorithms. We applied the Hough transforms to the edge image resulted from the Canny algorithm in order to determine the locations the target features. However, the Hough transform requires different parameter settings for each survey sector. As for image segmentation, we applied the connected element labeling algorithm and object-based image analysis using Orfeo Toolbox (OTB) in QGIS. The connected components labeled image shows the signals associated with the target buried relics are effectively connected and labeled. However, we often find multiple labels are assigned to a single structure on the given GPR data. Object-based image analysis was conducted by using a Large-Scale Mean-Shift (LSMS) image segmentation. In this analysis, a vector layer containing pixel values for each segmented polygon was estimated first and then used to build a train-validation dataset by assigning the polygons to one class associated with the buried relics and another class for the background field. With the Random Forest Classifier, we find that the polygons on the LSMS image segmentation layer can be successfully classified into the polygons of the buried relics and those of the background. Thus, we propose that these automatic classification methods applied to the GPR images of buried cultural heritage in this study can be useful to obtain consistent analyses results for planning excavation processes.

Keywords buried cultural heritage exploration, ground-penetrating radar, image feature extraction, image segmentation, machine learning

매장 문화재 공간 분포 결정을 위한 지하투과레이더 영상 분석 자동화 기법 탐색

권문희1,2 · 김승섭1,3*

1충남대학교 우주지질학과
2국립문화재연구소 고고연구실
3충남대학교 지질환경과학과

요 약

지구물리탐사기법은 매장 문화재 조사에 필요한 높은 해상도의 지하 구조 영상 생성과 매장 유구의 정확한 위치 결정하는 데 매우 유용하다. 이 연구에서는 경주 신라왕경 중심방의 고해상도 지하투과레이더 영상에서 유구의 규칙적인 배열이나 선형 구조를 자동적으로 구분하기 위하여 영상처리 기법인 영상 특징 추출과 영상분할 기법을 적용하였다. 영상 특징 추출의 대상은 유구의 원형 적심과 선형의 도로 및 담장으로 캐니 윤곽선 검출(Canny edge detection)과 허프 변환(Hough Transform) 알고리듬을 적용하였다. 캐니 윤곽선 검출 알고리듬으로 검출된 윤곽선 이미지에 허프 변환을 적용하여 유구의 위치를 탐사 영상에서 자동 결정하고자 하였으나, 탐사 지역별로 매개변수를 달리해서 적용해야 한다는 제약이 있었다. 영상 분할 기법의 경우 연결 요소 분석 알고리듬과 QGIS에서 제공하는 Orfeo Toolbox (OTB)를 이용한 객체기반 영상분석을 적용하였다. 연결 요소 분석 결과에서, 유구에 의한 신호들이 연결된 요소들로 효과적으로 인식되었지만 하나의 유구가 여러 요소로 분할되어 인식되는 경우도 발생함을 확인하였다. 객체기반 영상분석에서는 평균이동(Large-Scale Mean-Shift, LSMS) 영상 분할을 적용하여 각 분할 영역에 대한 화소 정보가 포함된 벡터 레이어를 우선 생성하였고, 유구를 포함하는 영역과 포함하지 않는 영역을 선별하여 훈련 모델을 생성하였다. 이 훈련모델에 기반한 랜덤포레스트 분류기를 이용해 LSMS 영상분할 벡터 레이어에서 유구를 포함하는 영역과 그렇지 않은 영역이 자동 분류 될 수 있음을 확인하였다. 이러한 자동 분류방법을 매장 문화재 지하투과레이더 영상에 적용한다면 유구 발굴 계획에 활용가능한 일관성 있는 결과를 얻을 것으로 기대한다.

주요어 매장문화재탐사, 지하투과레이더, 영상특징추출, 영상분할, 기계학습

Article

Case Report

Econ. Environ. Geol. 2022; 55(5): 551-561

Published online October 31, 2022 https://doi.org/10.9719/EEG.2022.55.5.551

Copyright © THE KOREAN SOCIETY OF ECONOMIC AND ENVIRONMENTAL GEOLOGY.

Automated Analyses of Ground-Penetrating Radar Images to Determine Spatial Distribution of Buried Cultural Heritage

Moonhee Kwon1,2, Seung-Sep Kim1,3,*

1Department of Astronomy, Space Science and Geology, Chungnam National University, Daejeon 34134, Korea
2Archaeology Division, National Research Institute of Cultural Heritage, Daejeon 34122, Korea
3Department of Geological Sciences, Chungnam National University, Daejeon 34134, Korea

Correspondence to:*seungsep@cnu.ac.kr

Received: September 21, 2022; Revised: October 28, 2022; Accepted: October 28, 2022

This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/3.0) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided original work is properly cited.

Abstract

Geophysical exploration methods are very useful for generating high-resolution images of underground structures, and such methods can be applied to investigation of buried cultural properties and for determining their exact locations. In this study, image feature extraction and image segmentation methods were applied to automatically distinguish the structures of buried relics from the high-resolution ground-penetrating radar (GPR) images obtained at the center of Silla Kingdom, Gyeongju, South Korea. The major purpose for image feature extraction analyses is identifying the circular features from building remains and the linear features from ancient roads and fences. Feature extraction is implemented by applying the Canny edge detection and Hough transform algorithms. We applied the Hough transforms to the edge image resulted from the Canny algorithm in order to determine the locations the target features. However, the Hough transform requires different parameter settings for each survey sector. As for image segmentation, we applied the connected element labeling algorithm and object-based image analysis using Orfeo Toolbox (OTB) in QGIS. The connected components labeled image shows the signals associated with the target buried relics are effectively connected and labeled. However, we often find multiple labels are assigned to a single structure on the given GPR data. Object-based image analysis was conducted by using a Large-Scale Mean-Shift (LSMS) image segmentation. In this analysis, a vector layer containing pixel values for each segmented polygon was estimated first and then used to build a train-validation dataset by assigning the polygons to one class associated with the buried relics and another class for the background field. With the Random Forest Classifier, we find that the polygons on the LSMS image segmentation layer can be successfully classified into the polygons of the buried relics and those of the background. Thus, we propose that these automatic classification methods applied to the GPR images of buried cultural heritage in this study can be useful to obtain consistent analyses results for planning excavation processes.

Keywords buried cultural heritage exploration, ground-penetrating radar, image feature extraction, image segmentation, machine learning

매장 문화재 공간 분포 결정을 위한 지하투과레이더 영상 분석 자동화 기법 탐색

권문희1,2 · 김승섭1,3*

1충남대학교 우주지질학과
2국립문화재연구소 고고연구실
3충남대학교 지질환경과학과

Received: September 21, 2022; Revised: October 28, 2022; Accepted: October 28, 2022

요 약

지구물리탐사기법은 매장 문화재 조사에 필요한 높은 해상도의 지하 구조 영상 생성과 매장 유구의 정확한 위치 결정하는 데 매우 유용하다. 이 연구에서는 경주 신라왕경 중심방의 고해상도 지하투과레이더 영상에서 유구의 규칙적인 배열이나 선형 구조를 자동적으로 구분하기 위하여 영상처리 기법인 영상 특징 추출과 영상분할 기법을 적용하였다. 영상 특징 추출의 대상은 유구의 원형 적심과 선형의 도로 및 담장으로 캐니 윤곽선 검출(Canny edge detection)과 허프 변환(Hough Transform) 알고리듬을 적용하였다. 캐니 윤곽선 검출 알고리듬으로 검출된 윤곽선 이미지에 허프 변환을 적용하여 유구의 위치를 탐사 영상에서 자동 결정하고자 하였으나, 탐사 지역별로 매개변수를 달리해서 적용해야 한다는 제약이 있었다. 영상 분할 기법의 경우 연결 요소 분석 알고리듬과 QGIS에서 제공하는 Orfeo Toolbox (OTB)를 이용한 객체기반 영상분석을 적용하였다. 연결 요소 분석 결과에서, 유구에 의한 신호들이 연결된 요소들로 효과적으로 인식되었지만 하나의 유구가 여러 요소로 분할되어 인식되는 경우도 발생함을 확인하였다. 객체기반 영상분석에서는 평균이동(Large-Scale Mean-Shift, LSMS) 영상 분할을 적용하여 각 분할 영역에 대한 화소 정보가 포함된 벡터 레이어를 우선 생성하였고, 유구를 포함하는 영역과 포함하지 않는 영역을 선별하여 훈련 모델을 생성하였다. 이 훈련모델에 기반한 랜덤포레스트 분류기를 이용해 LSMS 영상분할 벡터 레이어에서 유구를 포함하는 영역과 그렇지 않은 영역이 자동 분류 될 수 있음을 확인하였다. 이러한 자동 분류방법을 매장 문화재 지하투과레이더 영상에 적용한다면 유구 발굴 계획에 활용가능한 일관성 있는 결과를 얻을 것으로 기대한다.

주요어 매장문화재탐사, 지하투과레이더, 영상특징추출, 영상분할, 기계학습

    Fig 1.

    Figure 1.Stacked depth slices of ground-penetrating radar (GPR) data collected in the center of the Silla Kingdom, Gyeongju, South Korea. The GPR image features associated the buried ancient roads, walls and building structures are indicated by red, blue, and green boxes, respectively. The inset satellite map of the Korean Peninsula shows the study location.
    Economic and Environmental Geology 2022; 55: 551-561https://doi.org/10.9719/EEG.2022.55.5.551

    Fig 2.

    Figure 2.Comparison of image features automatically extracted from the stacked GPR time slice. (a) Original GPR time slice image, (b) Result from Canny edge detection. (c) Result from Hough transform for line detection, and (d) Result from Hough transform for circle detection.
    Economic and Environmental Geology 2022; 55: 551-561https://doi.org/10.9719/EEG.2022.55.5.551

    Fig 3.

    Figure 3.Linear and circular features extracted from the GPR depth slices of the entire study area by using Hough transform (a) Linear features obtained from Hough transform for line detection, (b) Circular features obtained from Hough transform for circle detection, (c) Enlarged image of the linear features extracted from the GPR slices, and (d) Enlarged image of the circular features extracted from the GPR image shown in (a).
    Economic and Environmental Geology 2022; 55: 551-561https://doi.org/10.9719/EEG.2022.55.5.551

    Fig 4.

    Figure 4.Image segmentation result by analyzing the GPR depth slices shown in Figure 3a. (a) Color-coded image of connected components detected from all the survey area, (b) Enlarged image of the connected components displayed in (a), and (c) Enlarged image of the original GPR time slice for the same area with (b).
    Economic and Environmental Geology 2022; 55: 551-561https://doi.org/10.9719/EEG.2022.55.5.551

    Fig 5.

    Figure 5.Data preparation for image classification with random forest approach. (a) Example of selected layers for training. Class 1 with pink-filled polygons indicates the areas where the buried heritages are expected, whereas class 2 with green-filled polygons is associated with the areas where no buried targets are expected (e.g., soil layer). (b) Polygons selected for training (dark green) and validation (dark red) to build a classification model.
    Economic and Environmental Geology 2022; 55: 551-561https://doi.org/10.9719/EEG.2022.55.5.551

    Fig 6.

    Figure 6.Image classification result for the study area. The polygon layers of the 2017 GPR data are successfully classified into class 1 (i.e., buried heritage) and class 2 (i.e., background). Only class 1 polygons classified in the 2017 data are filled with green. The same classification model developed from the 2017 GPR data as shown in Figure 5b is applied to classify the 2016 data. Although the 2016 data are not incorporated into the classification model, the polygons associated with the buried heritages are consistently classified as class 1 (purple-filled polygons).
    Economic and Environmental Geology 2022; 55: 551-561https://doi.org/10.9719/EEG.2022.55.5.551
    KSEEG
    Feb 29, 2024 Vol.57 No.1, pp. 1~91

    Stats or Metrics

    Share this article on

    • kakao talk
    • line

    Related articles in KSEEG

    Economic and Environmental Geology

    pISSN 1225-7281
    eISSN 2288-7962
    qr-code Download