[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-119511-en":3,"doc-seo-119511-105":30,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":27,"seo_description":14,"update_tm":28,"read_time":29},119511,13056703019662,"Evangeline","https://ap-avatar.wpscdn.com/avatar/be000253a8e92610077?_k=1778726343310543188",8,"Research & Report","A Machine Learning Algorithm to Detect Fog from Space","Fog detection is critical for traffic safety, and machine learning using satellite data has been studied, yet global-scale work with synthetic data remains limited. This thesis assesses machine learning feasibility for fog detection in satellite imagery using a synthetic MODIS dataset generated via RTTOV, driven by varied atmospheric conditions. The dataset is tabular under Independent Pixel Approximation, so each pixel is processed independently. Multiple XGBoost variants and feed-forward DNNs are trained and evaluated, and one synthetic-trained model is tested on a real MODIS image. Oversampling improves recall (up to 0.87) but introduces substantial false positives and low precision (0.27). No model replaces existing fog detection methods, though adding labeled real satellite data is proposed as future work for spatial analysis.","A Machine Learning Algorithm to Detect Fog from Space  \nMPALG, MPCAS  \nNILS JOHNSSON KEVIN SVENSSON  \nDEPARTMENT OF SPACE, EARTH AND ENVIRONMENT  \nCHALMERS UNIVERSITY OF TECHNOLOGY Gothenburg, Sweden 2024  \n[www.chalmers.se](www.chalmers.se)  \nDegree project report 2024  \nA Machine Learning Algorithm to Detect Fog from Space  \nNILS JOHNSSON  \nKEVIN SVENSSON  \nDepartment of Earth, Space and Environment Chalmers University of Technology Gothenburg, Sweden 2024  \nA Machine Learning Algorithm to Detect Fog from Space NILS JOHNSSON  \nKEVIN SVENSSON  \n© NILS JOHNSSON, KEVIN SVENSSON, 2024 .  \nSupervisor: Chiara Ceccobello, AI Sweden  \nSupervisor: Ronald Scheirer, SMHI  \nExaminer: Patrick Eriksson, Department of Space, Earth and Environment  \nDegree project report 2024  \nDepartment of Space, Earth and Environment Chalmers University of Technology SE-412 96 Gothenburg  \nSweden  \nTelephone +46 31 772 1000  \nCover: Car in traffic on a foggy road. Photo by chris-mueller / iStock.  \nTypeset in LATEX  \nGothenburg, Sweden 2024  \nA Machine Learning Algorithm to Detect Fog from Space NILS JOHNSSON, KEVIN SVENSSON  \nDepartment of Space, Earth and Environment Chalmers University of Technology  \nAbstract  \nFog detection is important for traffic safety. Detecting fog using machine learning on satellite data has been researched before, but not on a global scale using synthetic data. The aim of the thesis is to use a synthetic dataset of simulated MODIS satellite data to determine the viability of machine learning algorithms for detecting fog in satellite images. The synthetic dataset we use is simulated using a fast radiative transfer model called RTTOV by inputting various atmospheric information for different conditions. The dataset is tabular and no spatial or temporal relationship exists between the data points meaning each pixel is treated independently. We use the synthetic data to train and evaluate numerous machine learning models including various implementations of XGBoost and feed forward deep neural networks. We also apply a model trained on synthetic data to a real MODIS image. We demonstrate that classification models can achieve good recall values on synthetic data when oversampling fog in the training data, the best being 0.87 recall with a deep neural network. However, we find that this comes at the cost of a large amount of false positives evident by the low precision value of 0 .27. It is concluded that no model performed satisfactory results for replacing existing methods of fog detection. We identify the acquisition of supplemental labeled real satellite images as a possibility for future improvement, allowing for spatial analysis which is impossible with the independent pixels of the synthetic dataset alone. However, this is a non-trivial task due to the challenges in obtaining and labeling a sufficiently large and diverse dataset of real satellite images.  \nKeywords: MODIS, fog, machine learning, nowcasting.  \nAcknowledgements  \nWe would like to thank our supervisors, Chiara Ceccobello and Ronald (Rolle) Scheirer, for their guidance, support and encouragement throughout this project.  \nWe also thank our examiner, Patrick Eriksson, for his helpful feedback and assistance.  \nThank you all for your contributions.  \nNils Johnsson, Kevin Svensson, Gothenburg, May 2024  \nvi  \nList of Acronyms  \nBelow is the list of acronyms that are used throughout this thesis listed in alphabetical order:  \nAI  \nCOT  \nDNNGBDTs IPA MODISRTTOV SMHI  \nTOA XGBoost  \nArtificial Intelligence  \nCloud Optical Thickness  \nDeep Neural Network  \nGradient Boosting Decision Trees  \nIndependent Pixel Approximation  \nModerate Resolution Imaging Spectroradiometer Radiative Transfer for TOVS  \nSwedish Meteorological and Hydrological Institute Top-of-atmosphere  \nExtreme Gradient Boosting  \nContents  \nList of Acronyms vi  \n1 Introduction 1  \n1.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1  \n1.2 Dataset . . . . . . . . . . . . . . .","cbCaihuYWcwscIKV","https://ap.wps.com/l/cbCaihuYWcwscIKV","pdf",2556325,1,63,"English","en",105,"# Contents\n## List of Acronyms\n## 1 Introduction\n## 1.1 Background\n## 1.2 Dataset\n## 1.3 Purpose\n## 1.4 Goals\n## 1.5 Limitations\n## 2 Theory\n## 2.1 The MODIS instrument\n## 2.2 Synthetic data with RTTOV\n## 2.3 Machine learning models\n## 3 Methods\n## 3.1 Python testing environments in Jupyter Notebook\n## 3.2 Data loading and pre-processing\n## 3.3 Comparing and tuning different machine learning models\n## 3.4 Evaluation metrics\n## 3.5 Exploring regression models on synthetic data\n## 3.6 Exploring classification models on synthetic data\n## 3.7 Classifying pixels in real MODIS images\n## 4 Results\n## 4.1 Model architectures and parameters\n## 4.2 Regression models\n## 4.3 Classification models\n## 4.4 Influence of surface temperature and pressure on classifier performance","[{\"question\":\"What data source and simulation method does the thesis use for fog detection?\",\"answer\":\"The work uses a synthetic dataset simulating MODIS satellite data. The synthetic data are generated with the fast radiative transfer model RTTOV using different atmospheric inputs.\"},{\"question\":\"How are machine learning models trained and evaluated in this study?\",\"answer\":\"The thesis trains and evaluates multiple machine learning models on the synthetic tabular dataset, including different XGBoost implementations and feed-forward deep neural networks. A synthetic-trained model is also applied to a real MODIS image for testing.\"},{\"question\":\"What trade-off does the thesis find when using oversampling for fog in training?\",\"answer\":\"Oversampling fog improves recall substantially, reaching 0.87 with a deep neural network. However, this increases false positives, producing low precision of 0.27.\"}]","A Machine Learning Algorithm to Detect Fog from Space | PDF",1785724714,159,{"code":4,"msg":31,"data":32},"ok",{"site_id":24,"language":23,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"a-machine-learning-algorithm-to-detect-fog-from-space","",{"@graph":36,"@context":85},[37,54,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,51],{"item":41,"name":42,"@type":43,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":50},"https://docshare.wps.com/document/research-report/",3,{"item":52,"name":13,"@type":43,"position":53},"https://docshare.wps.com/document/a-machine-learning-algorithm-to-detect-fog-from-space/119511/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":23,"description":14,"dateModified":62,"datePublished":62,"encodingFormat":61,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-03",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What data source and simulation method does the thesis use for fog detection?","Question",{"text":75,"@type":76},"The work uses a synthetic dataset simulating MODIS satellite data. The synthetic data are generated with the fast radiative transfer model RTTOV using different atmospheric inputs.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How are machine learning models trained and evaluated in this study?",{"text":80,"@type":76},"The thesis trains and evaluates multiple machine learning models on the synthetic tabular dataset, including different XGBoost implementations and feed-forward deep neural networks. A synthetic-trained model is also applied to a real MODIS image for testing.",{"name":82,"@type":73,"acceptedAnswer":83},"What trade-off does the thesis find when using oversampling for fog in training?",{"text":84,"@type":76},"Oversampling fog improves recall substantially, reaching 0.87 with a deep neural network. However, this increases false positives, producing low precision of 0.27.","https://schema.org",{"og:url":52,"og:type":87,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":89,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,135],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":46,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":46,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":46,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":46,"category_name":137,"show_sort_weight":106,"slug":138},19,"General","general"]