Ziyun Yan , Lei Ma , Xuan Wang , Yongil Kim , Liqiang Zhang
{"title":"基于遥感大数据和先进变压器深度学习模型的高精度人口估计","authors":"Ziyun Yan , Lei Ma , Xuan Wang , Yongil Kim , Liqiang Zhang","doi":"10.1016/j.rsase.2025.101638","DOIUrl":null,"url":null,"abstract":"<div><div>High-precision population estimation is crucial for sensing where and how people live, which consequently supports sustainable development goals. Yet, there isn't a systematic theory that explains how geospatial big data works in population estimation studies, and deep learning models are eagerly applied in such fields as social sciences (e.g., population estimates) due to the recent prosperity of artificial intelligence (AI). The Shapley Additive Explanations (SHAP) tool was used in this study to check how well machine learning models and geospatial big data could be interpreted quantitatively for the population estimation process. The results show significant disparities among artificial intelligence models for population estimates, not only in estimate accuracy but also in dependencies on geospatial data. It was found that the classic Random Forest model relies too much on derived urban morphological features. The advanced transformer deep learning model, which can understand scenes, does much better and can directly get population-related semantics from satellite imagery. Subsequently, the high-precision population estimates were promised by integrating CNN's local and Transformer's global interpretation abilities. This study firstly implements the advanced Transformer model in population estimates and provides interpretability evidence within the deep learning framework. It was expected to become a typical application demonstration of AI in the social sciences.</div></div>","PeriodicalId":53227,"journal":{"name":"Remote Sensing Applications-Society and Environment","volume":"39 ","pages":"Article 101638"},"PeriodicalIF":4.5000,"publicationDate":"2025-06-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"High-precision population estimates by remote sensing big data and advanced transformer deep learning model\",\"authors\":\"Ziyun Yan , Lei Ma , Xuan Wang , Yongil Kim , Liqiang Zhang\",\"doi\":\"10.1016/j.rsase.2025.101638\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<div><div>High-precision population estimation is crucial for sensing where and how people live, which consequently supports sustainable development goals. Yet, there isn't a systematic theory that explains how geospatial big data works in population estimation studies, and deep learning models are eagerly applied in such fields as social sciences (e.g., population estimates) due to the recent prosperity of artificial intelligence (AI). The Shapley Additive Explanations (SHAP) tool was used in this study to check how well machine learning models and geospatial big data could be interpreted quantitatively for the population estimation process. The results show significant disparities among artificial intelligence models for population estimates, not only in estimate accuracy but also in dependencies on geospatial data. It was found that the classic Random Forest model relies too much on derived urban morphological features. The advanced transformer deep learning model, which can understand scenes, does much better and can directly get population-related semantics from satellite imagery. Subsequently, the high-precision population estimates were promised by integrating CNN's local and Transformer's global interpretation abilities. This study firstly implements the advanced Transformer model in population estimates and provides interpretability evidence within the deep learning framework. It was expected to become a typical application demonstration of AI in the social sciences.</div></div>\",\"PeriodicalId\":53227,\"journal\":{\"name\":\"Remote Sensing Applications-Society and Environment\",\"volume\":\"39 \",\"pages\":\"Article 101638\"},\"PeriodicalIF\":4.5000,\"publicationDate\":\"2025-06-23\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Remote Sensing Applications-Society and Environment\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://www.sciencedirect.com/science/article/pii/S2352938525001910\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q2\",\"JCRName\":\"ENVIRONMENTAL SCIENCES\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Remote Sensing Applications-Society and Environment","FirstCategoryId":"1085","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S2352938525001910","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"ENVIRONMENTAL SCIENCES","Score":null,"Total":0}
High-precision population estimates by remote sensing big data and advanced transformer deep learning model
High-precision population estimation is crucial for sensing where and how people live, which consequently supports sustainable development goals. Yet, there isn't a systematic theory that explains how geospatial big data works in population estimation studies, and deep learning models are eagerly applied in such fields as social sciences (e.g., population estimates) due to the recent prosperity of artificial intelligence (AI). The Shapley Additive Explanations (SHAP) tool was used in this study to check how well machine learning models and geospatial big data could be interpreted quantitatively for the population estimation process. The results show significant disparities among artificial intelligence models for population estimates, not only in estimate accuracy but also in dependencies on geospatial data. It was found that the classic Random Forest model relies too much on derived urban morphological features. The advanced transformer deep learning model, which can understand scenes, does much better and can directly get population-related semantics from satellite imagery. Subsequently, the high-precision population estimates were promised by integrating CNN's local and Transformer's global interpretation abilities. This study firstly implements the advanced Transformer model in population estimates and provides interpretability evidence within the deep learning framework. It was expected to become a typical application demonstration of AI in the social sciences.
期刊介绍:
The journal ''Remote Sensing Applications: Society and Environment'' (RSASE) focuses on remote sensing studies that address specific topics with an emphasis on environmental and societal issues - regional / local studies with global significance. Subjects are encouraged to have an interdisciplinary approach and include, but are not limited by: " -Global and climate change studies addressing the impact of increasing concentrations of greenhouse gases, CO2 emission, carbon balance and carbon mitigation, energy system on social and environmental systems -Ecological and environmental issues including biodiversity, ecosystem dynamics, land degradation, atmospheric and water pollution, urban footprint, ecosystem management and natural hazards (e.g. earthquakes, typhoons, floods, landslides) -Natural resource studies including land-use in general, biomass estimation, forests, agricultural land, plantation, soils, coral reefs, wetland and water resources -Agriculture, food production systems and food security outcomes -Socio-economic issues including urban systems, urban growth, public health, epidemics, land-use transition and land use conflicts -Oceanography and coastal zone studies, including sea level rise projections, coastlines changes and the ocean-land interface -Regional challenges for remote sensing application techniques, monitoring and analysis, such as cloud screening and atmospheric correction for tropical regions -Interdisciplinary studies combining remote sensing, household survey data, field measurements and models to address environmental, societal and sustainability issues -Quantitative and qualitative analysis that documents the impact of using remote sensing studies in social, political, environmental or economic systems