All Issue

2022 Vol.11, Issue 3 Preview Page
Analysis of Street View Image Object Affecting Perceived Walkability Using Machine Learning

기계학습을 이용한 보행환경 정성적 평가에 영향을 미치는 거리영상 특성분석

이지윤, 강영옥, 김지연, 박지영

Jiyoon Lee, Youngok Kang, Jiyeon Kim, Jiyoung Park

2022. pp. 375~391
PDF Citation
Abstract
Walking, one of the green modes of transportation, is very important for the transition to a sustainable city. In addition, as it has been confirmed that a pleasant walking environment has a positive effect on health of local residents, many cities around the world are promoting the creation of an eco-friendly and people-centered walking environment as the top agenda in urban planning. As the awareness of the importance of the walking environment has increased, many studies have been conducted to identify the physical components that constitute the walking environment and to find out what kind of walking environment people consider good for walking. Existing methods for analyzing the qualitative walking environment based on surveys targeting residents or experts showed limitations in their representativeness. On the other hand, the recent development of high-resolution street view images and deep learning technology makes it possible to obtain detailed perceived walkability scores by using paired comparison data for street view images as a training set. However, perceived walkability score prediction based on deep learning technology has limitations in providing an answer to why such evaluation score was obtained. The purpose of this study is to analyze the characteristics of the urban landscape that affect the perceived walkability based on street view images. In this study, we tested various machine learning models with the perceived walkability score of the street image as the dependent variable and the semantic segmentation ratio value of the street view image as the independent variable. According to our study, the regression equation of the support vector machine was the most accurate. And in predicting the perceived walkability score, the importance of the object was in the order of roads, sidewalks, buildings, trees, and the sky. Our study showed that the higher the segmentation value, the higher the perceived walkability score for sidewalk, streetlight, road, grass, and tree. On the contrary, the higher the segmentation value, the lower the score for ashcans, mountains, trucks, and walls. This study is meaningful in that it was possible to identify important objects and their direction that affect the evaluation of the walking environment through the machine learning model, and to partially explain the perceived walkability score predicted by the deep learning model.
Keywords
Perceived walkability
Semantic segmentation
Machine learning
Explainable AI
보행은 녹색교통수단의 하나로 지속 가능한 도시로의 전환에 매우 중요하다. 또한 걷기 좋은 보행환경은 지역민의 건강증진에도 긍정적 효과를 나타냄이 확인되면서 세계 많은 도시들이 친환경적이고 사람중심의 보행환경 조성을 도시 계획에 있어 최우선 아젠더로 추진하고 있다. 보행환경의 중요성이 확대되면서, 보행환경을 구성하는 물리적 구성요소를 파악하기 위한 연구와 함께, 사람들이 걷기 좋다고 판단하는 보행환경이 어떤 것인지를 파악하고자 하는 연구도 이루어졌다. 정성적 보행환경을 분석하기 위한 방법으로 기존에는 주민이나 전문가를 통한 설문조사가 주를 이루면서 대표성의 한계를 보이고 있었지만, 최근 고해상도의 거리영상과 딥러닝 기술의 발전은 거리영상에 대한 쌍체비교 데이터를 훈련셋으로 하여 가로단위의 상세한 정성적 보행환경 평가점수 획득을 가능하게 하고 있다. 하지만 딥러닝 기술에 기반한 정성적 보행환경 평가 점수 예측은 왜 이러한 평가점수를 얻게 되었는지에 대한 해답을 제공하는데 한계가 있다. 본 연구의 목적은 거리영상을 기반으로 사람들이 걷기 좋다고 느끼는 보행환경에 영향을 미치는 도시경관의 특성을 분석하는 것이다. 이를 위해 거리영상의 정성적 보행환경 점수를 종속변수로, 거리영상의 시멘틱 세그먼테이션 비율 값을 독립변수로 하여 다양한 기계학습 모델을 실험하였다. 실험 결과 서포트 벡터 머신 회귀식이 가장 정확도가 높았고, 정성적 보행환경 점수를 예측함에 있어 도로와 보행로, 건물, 나무, 하늘 등의 순으로 객체 중요도가 도출되었다. 보행로(sidewalk), 가로등(streetlight), 길(road), 잔디(grass), 나무(tree) 등은 세그먼테이션 값이 높을수록 정성적 보행환경 평가를 높게 만들며, 쓰레기통(ashcan), 산(mountain), 트럭(truck), 담벼락(wall) 등은 세그먼테이션 값이 높을수록 평가점수를 낮게 만듦을 알 수 있었다. 본 연구는 기계학습모델을 통해 보행환경에 대한 정성적 평가에 영향을 주는 중요 객체와 객체의 방향성을 확인할 수 있었으며, 딥러닝 모델을 통해 예측한 보행환경 정성평가 점수를 일부 설명 가능하게 하였다는 점에 의의가 있다.
키워드
인지된 보행환경
시멘틱 세그먼테이션
기계학습
설명가능한 AI
References
  1. 김규리･이제선, 2016, “보행공간 요소에 대한 보행자의 인지 및 보행만족도에 관한 연구,” 한국도시설계학회, 17(3), 89-103.
  2. 김지연･강영옥, 2022, “거리영상 기반 보행환경의 정성적 평가 예측을 위한 딥러닝 모델 개발,” 대한공간정보학회지, 30(2), 45-56.
  3. 김준봉･오승철･서기성, 2016, “MLR 및 SVR 기반 선형과 비선형회귀분석의 비교: 풍속 예측 보정,” 전기학회논문지, 65(5), 851-856.
  4. 김희철･안건혁･권영상, 2014, “개인의 보행 확률에 영향을 미치는 거주지 환경요인,” 한국도시설계학회지, 15(3), 5-18.
  5. 박근덕･이수기, 2018, “근린환경특성과 일상보행활동 그리고 주관적 건강수준의 구조적 관계 분석: 경로모형의 적용,” 국토계획, 53(1), 255-272.
  6. 박소현･최이명･서한림･김준형, 2009, “주거지 보행환경 인지가 생활권 보행만족도에 미치는 영향에 관한 연구,” 대한건축학회 논문집-계획계, 25(8), 253-261.
  7. 박영은･이우성, 2022, “보행환경 평가지표를 활용한 통한 노인의 보행환경 인식 분석,” 한국공간디자인학회 논문집, 17(3), 200-212.
  8. 박지영･강영옥･김지연, 2022, “거리 영상과 시멘틱 세그먼테이션을 활용한 보행환경 평가 지표 개발,” 한국지도학회지, 22(1), 53-68.
  9. 박철영･이수기, 2016, “가로환경 특성이 보행자 교통사고에 미치는 영향 분석,” 한국도시설계학회지 도시설계, 17(3), 105-121.
  10. 성현곤･김태호･강지원, 2011, “구조방정식을 활용한 보행환경 계획요소의 이용만족도 평가에 관한 연구: 종로 및 강남일대를 대상으로,” 국토계획, 46(5), 275-288.
  11. 유승재･하정원･김혜준･기동환･이수기, 2021, “서울시 가로경관 이미지에 대한 주관적 인지에 영향을 미치는 가로환경 요인분석,” 국토계획, 56(2), 79-93.
  12. 이경환･안건혁, 2008, “지역 주민의 보행 활동에 영향을 미치는 근린 환경 특성에 관한 실증 분석: 서울시 12개 행정동을 대상으로,” 대한건축학회 논문집 - 계획계, 24(6), 293-302.
  13. 이수기･고준호･이기훈, 2016, “근린환경특성이 보행만족도에 미치는 영향 분석: 서울서베이 2013 년 자료를 중심으로,” 국토계획, 51(1), 169-187.
  14. 이수민･황기연, 2009, “보행환경요인이 보행안전에 미치는 영향분석,” 대한교통학회지, 27(1), 107-114.
  15. 이지원･유다은, 2021, “보행자의 주관적 인식과 성별이 보행환경 만족도에 미치는 영향분석,” 대한건축학회논문집, 37(10), 51-62.
  16. 조혜민･이수기, 2016, “보행목적별 보행활동시간에 영향을 미치는 근린환경 특성분석-주관적 인지환경과 객관적 측정환경의 차이를 중심으로,” 대한국토･도시계획학회, 51(4), 105-122.
  17. 행정안전부, 2012, ｢보행안전 및 편의증진에 관한 법률｣.
  18. Awad, M. and Khanna, R., 2015, Support vector regression, in Awad, M. and Khanna, R., eds., Efficient Learning Machines, Berkeley, CA: Apress, 67-80.
  19. Biau, G. and Scornet, E., 2016, A random forest guided tour, Test, 25(2), 197-227.
  20. Biljecki, F. and Ito, K., 2021, Street view imagery in urban analytics and GIS: A review, Landscape and Urban Planning, 215, 104217.
  21. Bijmolt, T.H. and Wedel, M., 1995, The effects of alternative methods of collecting similarity data for multidimensional scaling, International Journal of Research in Marketing, 12(4), 363-371.
  22. Blečić, I., Cecchini, A., and Trunfio, G.A., 2018, Towards automatic assessment of perceived walkability, International Conference on Computational Science and Its Applications, Cham: Springer, 351-365.
  23. Chen, T. and Guestrin, C., 2016, Xgboost: A scalable tree boosting system, Proceedings of the 22nd acm sigkdd international conference on knowledge discovery and data mining, San Francisco, CA, USA, 785-794.
  24. Dubey, A., Naik, N., Parikh, D., Raskar, R., and Hidalgo, C.A., 2016, Deep learning the city: Quantifying urban perception at a global scale, European conference on computer vision, Cham: Springer, 196-212.
  25. Frank, L.D., Schmid, T.L., Salis J.F., Chapman, J., and Saelems, B.E., 2005, Linking Objectively Measured Physical Activity with Objectively Measured Urban Form: Findings from SMARTRAQ, American Journal of Preventive Medicine, 28(2), 117-125.
  26. Guan, W., Chen, Z., Feng, F., Liu, W., and Nie, L., 2021, Urban perception: sensing cities via a deep interactive multi-task learning framework, ACM Transactions on Multimedia Computing, Communications, and Applications (TOMM), 17(1s), 1-20.
  27. Joglekar, S., Quercia, D., Redi, M., Aiello, L.M., Kauer, T., and Sastry, N., 2020, FaceLift: a transparent deep learning framework to beautify urban scenes, Royal Society Open Science, 7(1), 190987.
  28. Kim, J.H., Lee, S., Hipp, J.R., and Ki, D., 2021, Decoding urbal landscapes: Google street view and measurement sensitivity, Computer, Environment and Urban Systems, 88, 101626.
  29. Kim, S., Park, S., and Lee, J.S., 2014, Meso-or micro-scale? Environmental factors influencing pedestrian satisfaction, Transportation Research Part D: Transport and Environment, 30, 10-20.
  30. Li, X. and Ratti, C., 2018, Mapping the spatial distribution of shade provision of street trees in Boston using Google Street View panoramas, Urban Forestry & Urban Greening, 31, 109-119.
  31. Li, Y., Yabuki, N., Fukuda, T., and Zhang, J., 2020, A big data evaluation of urban street walkability using deep learning and environmental sensors-a case study around Osaka University Suita campus, The Cognitive City(AI), 2, 319-328.
  32. Lu, X., Lin, Z., Jin, H., Yang, J., and Wang, J.Z., 2014, Rapid: Rating pictorial aesthetics using deep learning, Proceedings of the 22nd ACM International Conference on Multimedia, Orlando, Florida, USA, 457-466.
  33. Lu, X., Lin, Z., Shen, X., Mech, R., and Wang, J.Z., 2015, Deep multi-patch aggregation network for image style, aesthetics, and quality estimation, Proceedings of the IEEE International Conference on Computer Vision, Santiago, Chile, 990-998.
  34. Mason, C.H. and Perreault Jr, W.D., 1991, Collinearity, power, and interpretation of multiple regression analysis, Journal of Marketing Research, 28(3), 268-280.
  35. Mateo-Babiano, I., 2016, Pedestrian’s needs matter: Examining Manila’s walking environment, Transport Policy, 45, 107-115.
  36. Melkumova, L.E. and Shatskikh, S.Y., 2017, Comparing Ridge and LASSO estimators for data analysis, Procedia Engineering, 201, 746-755.
  37. Min, W., Mei, S., Liu, L., Wang, Y., and Jiang, S., 2019, Multi-task deep relative attribute learning for visual urban perception, IEEE Transactions on Image Processing, 29, 657-669.
  38. Nagata, S., Nakaya, T., Hanibuchi, T., Amagasa, S., Kikuchi, H., and Inoue, S., 2020, Objective scoring of streetscape walkability related to leisure walking: Statistical modeling approach with semantic segmentation of Google Street View images, Health & Place, 66, 102428.
  39. Quercia, D., Aiello, L.M., Schifanella, R., and Davies, A., 2015, The digital life of walkable streets, Proceedings of the 24th international conference on World Wide Web, Florence, Italy, 875-884.
  40. Salesses, P., Schechtner, K., and Hidalgo, C.A., 2013, The collaborative image of the city: mapping the inequality of urban perception, PLOS ONE, 8(7), e68400.
  41. Santani, D., Ruiz-Correa, S., and Gatica-Perez, D., 2018, Looking south: Learning urban perception in developing cities, ACM Transactions on Social Computing, 1(3), 1-23.
  42. Saritas, M.M. and Yasar, A., 2019, Performance analysis of ANN and Naive Bayes classification algorithm for data classification, International Journal of Intelligent Systems and Applications in Engineering, 7(2), 88-91.
  43. Stewart, N., Brown, G.D., and Chater, N., 2005, Absolute identification by relative judgment, Psychological Review, 112(4), 881-911.
  44. Xu, M., Watanachaturaporn, P., Varshney, P.K., and Arora, M.K., 2005, Decision tree regression for soft classification of remote sensing data, Remote Sensing of Environment, 97(3), 322-336.
  45. Xu, Y., Yang, Q., Cui, C., Shi, C., Song, G., Han, X., and Yin, Y., 2019, Visual Urban Perception with Deep Semantic-Aware Network, International Conference on Multimedia Modeling, Cham: Springer, 28-40.
  46. Yoo, K., Ryu, D., Lee, C., Nam, K., and Kang, Y., 2021, Creation of a paired comparison set for preference evaluation based on crowdsourcing, Proceedings of Korea Industrial Information Systems Research, Incheon, Korea, 2-6.
  47. Wang, P., Chen, P., Yuan, Y., Liu, D., Huang, Z., Hou, X., and Cottrell, G., 2018, Understanding convolution for semantic segmentation, 2018 IEEE Winter Conference on Applications of Computer Vision (WACV), Piscataway, NJ: IEEE, 1451-1460.
  48. Wang, R., Liu, Y., Lu, Y., Zhang, J., Liu, P., Yao, Y., and Grekousis, G., 2019, Perceptions of built environment and health outcomes for older Chinese in Beijing: A big data approach with street view images and deep learning technique, Computers, Environment and Urban Systems, 78, 101386.
  49. Zhang, F., Zhou, B., Liu, L., Liu, Y., Fung, H.H., Lin, H., and Ratti, C., 2018, Measuring human perceptions of a large-scale urban region using machine learning, Landscape and Urban Planning, 180, 148-160.
  50. Zhou, H., He, S., Cai, Y., Wang, M., and Su, S., 2019, Social inequalities in neighborhood visual walkability: Using street view imagery and deep learning technologies to facilitate healthy city planning, Sustainable Cities and Society, 50, 101605.
  51. Zou, H. and Hastie, T., 2005, Regularization and variable selection via the elastic net, Journal of the Royal Statistical Society: Series B (statistical methodology), 67(2), 301-320.
  52. Zou, W., Zhang, D., and Lee, D.J., 2021, A new multi-feature fusion based convolutional neural network for facial expression recognition, Applied Intelligence, 52(3), 2918-2929.
  53. Cityscapes Dataset, https://www.cityscapes-dataset.com/
  54. ADE20K, https://groups.csail.mit.edu/vision/datasets/ADE20K/
Information
  • Publisher :The Association of Korean Geographers
  • Publisher(Ko) :한국지리학회
  • Journal Title :Journal of the Association of Korean Geographers
  • Journal Title(Ko) :한국지리학회지
  • Volume : 11
  • No :3
  • Pages :375~391
  • DOI :https://doi.org/10.25202/JAKG.11.3.6
Journal Informaiton Journal of the Association of Korean Geographers Journal of the Association of Korean Geographers
Online Submission
  • NRF
  • KISTI Current Status
  • KISTI Cited-by
  • orcid
  • open access
  • ccl
Journal Informaiton Journal Informaiton - close