With the development of web 2.0 technology (a term popularized by Tim O'Reilly and Dale Dougherty at the O'Reilly Media Web 2.0 Conference in late 2004), that promotes user-generated content and interoperability for end user, variety of applications have been developed for sharing data by crowdsourcing. These applications come under the category of Volunteered Geographic Information (VGI) a term first used by (Goodchild, 2007). Among these applications of VGI, the most outstanding application is the OpenStreetMap (OSM) project for sharing geospatial information. The objective of OpenStreetMap is, creating a free digital map of the world through the involvement of volunteers in a manner similar to software development in Open Source projects. The OpenStreetMap project is recently being employed for mapping the spatial details in disaster-stricken areas by Humanitarian OpenStreetMap Team (HOT). HOT opens a mapping task of the areas hit by disasters for volunteer mappers. The purpose is to facilitate the access of relief organizations and expedite the relief activities in those areas. Regarding the crowd-sourced data, there is a growing concern about their quality because the volunteers that contribute the data lack the sufficient cartographic training and the quality cannot be guaranteed. In this dissertation, a comparative approach was employed for evaluating the positional accuracy of OSM dataset of Lecco and Abbadia Lariana. The analysis compared the position of point features on the OSM dataset of Lecco and Abbadia Lariana with the position of reference points recorded in a local survey using a GPS device with a centimetric level of precision. This study goes through the conventional procedure of positional accuracy of OSM dataset by finding the average Euclidean distances between point features for OSM datasets, inspired by (Girres and Touya, 2010; Stark, 2011; Jackson et al., 2013; Mashhadi et al., 2014, the average Euclidean distance for linear features: Girres and Touya, 2010; Fan et al., 2014), and extends this for the positional accuracy of Bing Aerial imagery and Mapbox satellite imagery of Lecco and Abbadia Lariana. The study consists of a local and remote review of the OSM data quality, Bing Aerial imagery and Mapbox satellite imagery. The current positional accuracy of the OSM dataset of Lecco and Abbadia was analyzed and the results presented in statistical terms. The issues with the imageries were statistically observed. It was noticed that among the three layers, OSM, Bing aerial imagery and Mapbox satellite imagery, the Mapbox satellite imagery layer is more close to the actual position than the other two. Moreover, it was observed that the sample of the Euclidean distances of Bing imagery data layer points have an average and a standard deviation of almost double that of the OSM dataset and Mapbox imagery under consideration. Which means that the Euclidean distances of Bing imagery have large variations in the high and low positional accuracy of the data sample points. In the end some recommendations are proposed for the future work regarding the data quality assessment of the area under consideration.
Con lo sviluppo della tecnologia web 2.0 (un termine introdotto da Tim O'Reilly e Dale Dougherty alla conferenza O'Reilly Media Web 2.0 alla fine del 2004), che promuove contenuti e interoperabilità generati dagli utenti per l'utente finale, sono state sviluppate varie applicazioni per condividere i dati da crowdsourcing. Queste applicazioni rientrano nella categoria dell'Informazione Geografica Volontaria (VGI), un termine usato per la prima volta da Goodchild (2007). Tra queste applicazioni di VGI, l'applicazione più importante è il progetto OpenStreetMap (OSM) per la condivisione delle informazioni geospaziali. L'obiettivo di OpenStreetMap è quello di creare una mappa digitale gratuita del mondo attraverso il coinvolgimento di volontari in un modo simile allo sviluppo del software nei progetti Open Source. Il progetto OpenStreetMap è recentemente impiegato per mappare i dati spaziali nelle aree colpite da disastri dallo Humanitarian OpenStreetMap Team (HOT). HOT richiede a mappatori volontari di cartografare le aree colpite da disastri. Lo scopo è facilitare l'accesso delle organizzazioni occupate nei soccorsi e accelerarne le attività in tali aree. Per quanto riguarda i dati provenienti dal crowdsourcing, la loro qualità può non non può essere garantita perché i volontari che contribuiscono alla mappatura non dispongono della formazione cartografica sufficiente. In questa tesi è stato utilizzato un approccio comparativo per valutare l'accuratezza del posizionamento dei dati di OSM di Lecco e Abbadia Lariana. L'analisi si è basat sul confronto fra la posizione delle caratteristiche del punto si dati di OSM di Lecco e Abbadia Lariana con la posizione dei punti di riferimento registrati in un rilievo realizzato con un dispositivo GPS di precisione. Sono state valutate le distanze euclidee tra i punti individuati sui layer di OSM e sulle immagini satellitari di Bing Aerial, rispetto ai punti rilevati (Girres e Touya, 2010; Stark, 2011; Jackson et al., 2013; Mashhadi et al. , 2014, Girres e Touya, 2010; Fan et al., 2014). Lo studio consiste in una validazione locale (in situ) e remota (attraverso le immagini satellitari) della qualità dei dati OSM. L'attuale accuratezza posizionale dei dati di Lecco e Abbadia è stata analizzata in termini statistici. I problemi con le immagini sono stati osservati statisticamente. È stato notato che tra i tre strati informativi, OSM, immagini aeree di Bing e immagini satellitari di Mapbox, le immagini satellitari di Mapbox presentano una maggiore accuratezza posizionale. Inoltre, è stato osservato che il campione delle distanze euclidee dei punti di strato di dati Bing ha una media e una deviazione standard di quasi il doppio di quello del set di dati OSM e delle immagini Mapbox in esame. Alla fine sono proposte alcune raccomandazioni per il futuro lavoro in materia di valutazione della qualità dei dati della zona in esame.
Validation of OpenStreetMap data : a case study in Lecco and Abbadia Lariana
ULLAH, ZAKIR
2016/2017
Abstract
With the development of web 2.0 technology (a term popularized by Tim O'Reilly and Dale Dougherty at the O'Reilly Media Web 2.0 Conference in late 2004), that promotes user-generated content and interoperability for end user, variety of applications have been developed for sharing data by crowdsourcing. These applications come under the category of Volunteered Geographic Information (VGI) a term first used by (Goodchild, 2007). Among these applications of VGI, the most outstanding application is the OpenStreetMap (OSM) project for sharing geospatial information. The objective of OpenStreetMap is, creating a free digital map of the world through the involvement of volunteers in a manner similar to software development in Open Source projects. The OpenStreetMap project is recently being employed for mapping the spatial details in disaster-stricken areas by Humanitarian OpenStreetMap Team (HOT). HOT opens a mapping task of the areas hit by disasters for volunteer mappers. The purpose is to facilitate the access of relief organizations and expedite the relief activities in those areas. Regarding the crowd-sourced data, there is a growing concern about their quality because the volunteers that contribute the data lack the sufficient cartographic training and the quality cannot be guaranteed. In this dissertation, a comparative approach was employed for evaluating the positional accuracy of OSM dataset of Lecco and Abbadia Lariana. The analysis compared the position of point features on the OSM dataset of Lecco and Abbadia Lariana with the position of reference points recorded in a local survey using a GPS device with a centimetric level of precision. This study goes through the conventional procedure of positional accuracy of OSM dataset by finding the average Euclidean distances between point features for OSM datasets, inspired by (Girres and Touya, 2010; Stark, 2011; Jackson et al., 2013; Mashhadi et al., 2014, the average Euclidean distance for linear features: Girres and Touya, 2010; Fan et al., 2014), and extends this for the positional accuracy of Bing Aerial imagery and Mapbox satellite imagery of Lecco and Abbadia Lariana. The study consists of a local and remote review of the OSM data quality, Bing Aerial imagery and Mapbox satellite imagery. The current positional accuracy of the OSM dataset of Lecco and Abbadia was analyzed and the results presented in statistical terms. The issues with the imageries were statistically observed. It was noticed that among the three layers, OSM, Bing aerial imagery and Mapbox satellite imagery, the Mapbox satellite imagery layer is more close to the actual position than the other two. Moreover, it was observed that the sample of the Euclidean distances of Bing imagery data layer points have an average and a standard deviation of almost double that of the OSM dataset and Mapbox imagery under consideration. Which means that the Euclidean distances of Bing imagery have large variations in the high and low positional accuracy of the data sample points. In the end some recommendations are proposed for the future work regarding the data quality assessment of the area under consideration.| File | Dimensione | Formato | |
|---|---|---|---|
|
Thesis report.pdf
non accessibile
Descrizione: Thesis Report
Dimensione
9.77 MB
Formato
Adobe PDF
|
9.77 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/136485