På senare år har etableringen av 3D-applikationer på internet, för att visualisera geografiska områden och platser blivit allt vanligare. Exempel på sådana applikationer är tjänsterna Google Earth och Street View, som levereras av företaget Google. Den snabba utvecklingen av 3D-tjänsterna har medfört bättre bildupplösning och mer detaljerade satellitbilder. Detta har i sin tur fört med sig att integritets- och säkerhetsrelaterade frågeställningar relaterade till 3D-applikationerna har aktualiserats. Från flera håll i världen har Google stött på ifrågasättanden och protester. Skäl som påtalats i dessa sammanhang är bland annat integritetskränkande aspekter samt försvars- och säkerhetsrisker.Syftet med detta arbete är att undersöka de säkerhetsrisker som kan komma att uppstå i takt med att Google kontinuerligt etablerar sina 3D-tjänster i Sverige. Undersökningen görs för att få en helhetsbild över hur svenska företag och myndigheter upplever Googles tjänster samt den lagstiftning som är förknippad med dessa. Undersökningen, som utförts i enkätform, har riktats in på svenska företag och myndigheter samt ambassader belägna i Sverige, som bedöms vara en potentiell måltavla för terroristbrott alternativt grova rån.Undersökningen visar att svenska företag och myndigheter har en god kännedom om Google och dess 3D-tjänster. Det framgår även att en stor majoritet av de tillfrågade företagen och myndigheterna identifierat säkerhetsrisker förknippade med webbtjänsterna. Uppfattningen är att säkerhetsriskerna kommer att aktualiseras allt mer i framtiden. Detta eftersom teknikutvecklingen leder till allt mer detaljerade bilder och eftersom de nationella lagstiftningarna får allt svårare att hänga med när det gäller att förebygga säkerhetsrisker. / In recent years, the establishment of 3D-applications on the internet, in order to show geographic areas and places, has become more common. Examples of such applications are the web services Google Earth and Street View, supplied by the company Google. The rapid development of the 3D-sevices has resulted in better image resolution and more detailed satellite photos. This in turn has led to that privacy and security related to the 3D-services has been actualized. From several parts of the world, Google has faced contestations and protests. Reasons alleged in these cases are among others integrity aspects as well as defense and security reasons.The aim of this work is to examine the security related matters that can occur during Google´s continuous establishment of its 3D-sevices in Sweden. The examine is done to get an overall picture over how Swedish companies and government agencies consider Google´s services and the law associated with these. The examine has been done through a written questionaries’ that has been sent to Swedish companies and government agencies and embassies located in Sweden, that have been considered to be a potential target of terrorist crime or severe robberies.The examine show that Swedish companies and government agencies have good knowledge of Google and their 3D-services. A great majority of the respondents have identified security risks associated with the Google 3D-applications. The general view is that the technology development that lead to more and more detailed photo technology will lead to a greater actualization of the security related matters in the future. Also the fact that the difficulties to adjust the national laws in order to prevent the problems considers being a reason that will be actualized more in the future.

Dans un travail réalisé au Centre de Robotique et à l'Institut VEDECOM, nous nous sommes intéressés aux systèmes robustes de localisation visuelle en milieu urbain pour la voiture autonome. Obtenir une pose exacte à partir d'une mono-caméra est difficile et insuffisant en terme de précision pour la voiture autonome actuelle. Nous nous sommes concentrés sur l'utilisation de Systèmes d'Information Géographiques (SIG) pour concevoir une approche fiable, précise et absolue de localisation en milieu urbain.Le développement de SIG publics nous a apporté un nouvel horizon pour résoudre le problème de la localisation, mais ses informations, telles que les cartes topologiques, sémantiques, métriques, les Street Views, les cartes de profondeur, les cartes cadastrales 3D et les cartes en haute définition, doivent être bien analysées et organisées pour extraire les informations pertinentes pour une voiture autonome. Notre première tâche consistait à concevoir une base de données hors ligne accessible par un robot à partir d'un SIG public dense, à savoir Google Maps, qui a l'avantage d'avoir une couverture mondiale. Nous générons une représentation topométrique compacte de l'environnement urbain en extrayant quatre données utiles du SIG, y compris : les topologies, les géo-coordonnées, les Street Views panoramiques et les cartes de profondeur associées. Dans le même temps, un ensemble de données en ligne a été acquis par une mono-caméra équipée sur les véhicules de VEDECOM. Afin de rendre les Street View sphériques compatibles avec l'imagerie en ligne, une transformation basée sur l'interpolation d'image est introduite pour obtenir des images rectilignes à partir de Street Views.Nous proposons deux méthodes de localisation : l'une est une approche de vision par ordinateur basée sur l'extraction de caractéristiques, l'autre est une méthode d'apprentissage basée sur les réseaux de neurones convolutionnels (convnet). En vision par ordinateur, l'extraction de caractéristiques est un moyen populaire de résoudre le positionnement à partir d'images. Nous tirons parti de Google Maps et utilisons ses données topo-métriques hors ligne pour construire un positionnement grossier à fin, à savoir un processus de reconnaissance de lieu topologique puis une estimation métrique de pose par optimisation de graphe. La méthode a été testée en environnement urbain et démontre à la fois une précision sous-métrique et une robustesse aux changements de point de vue, à l'illumination et à l'occlusion. Aussi, les résultats montrent que les emplacements éloignés de Street Views produisent une erreur significative dans la phase d'estimation métrique. Ainsi, nous proposons de synthétiser des Street Views artificielles pour compenser la densité des Street View originales et améliorer la précision.Cette méthode souffre malheureusement d'un temps de calcul important. Étant donné que le SIG nous offre une base de données géolocalisée à l'échelle mondiale, cela nous motive à régresser des localisations globales directement à partir d'un convnet de bout en bout. La base de données hors ligne précédemment construite est encore insuffisante pour l'apprentissage d'un convnet. Pour compenser cela nous densifions la base d'origine d'un facteur mille et utilisons la méthode d'apprentissage par transfert pour faire converger notre régresseur convnet et avoir une bonne performance. Le régresseur permet également d'obtenir une localisation globale à partir d'une seule image et en temps réel.Les résultats obtenus par ces deux approches nous fournissent des informations sur la comparaison et la relation entre les méthodes basées sur des caractéristiques et celles basées sur le convnet. Après avoir analysé et comparé les performances de localisation des deux méthodes, nous avons également abordé des perspectives pour améliorer la robustesse et la précision de la localisation face au problème de localisation urbaine assistée par SIG. / In a work made at Centre de Robotique and Institut VEDECOM, we studied robust visual urban localization systems for self-driving cars. Obtaining an exact pose from a monocular camera is difficult and cannot be applied to the current autonomous cars. We mainly focused on fully leveraging Geographical Information Systems (GIS) to achieve a low-cost, robust, accurate and global urban localization.The development of public GIS's has brought us a new horizon to address the localization problem but their tremendous amount of information, such as topological, semantic, metric maps, Street Views, depth maps, 3D cadastral maps and High Definition maps, has to be well analyzed and organized to extract relevant information for self-driving cars. Our first task was to design a robotic accessible offline database from a dense public GIS, namely Google Maps, which has the advantage to propose a worldwide coverage. We make a compact topometric representation for the dynamic urban environment by extracting four useful data from the GIS, including topologies, geo-coordinates, panoramic Street Views, and associated depth maps. At the same time, an online dataset was acquired with a low-cost camera equipped on VEDECOM vehicles. In order to make spheric Street Views compatible with the online imagery, an image warping and interpolation based transformation is introduced to render rectilinear images from Street Views.We proposed two localization methods: one is a handcrafted-features-based computer vision approach, the other is a convolutional neural network (convnet) based learning technique. In computer vision, extracting handcrafted features is a popular way to solve the image based positioning. We take advantages of the abundant sources from Google Maps and benefit from the topometric offline data structure to build a coarse-to-fine positioning, namely a topological place recognition process and then a metric pose estimation by a graph optimization. The method is tested on an urban environment and demonstrates both sub-meter accuracy and robustness to viewpoint changes, illumination and occlusion. Moreover, we demonstrate that sparse Street View locations produce a significant error in the metric pose estimation phase. Thus our former framework is refined by synthesizing more artificial Street Views to compensate the sparsity of original Street Views and improve the precision.The handcrafted feature based framework requires the image retrieval and graph optimization. It is hard to achieve in a real-time application. Since the GIS offers us a global scale geotagged database, it motivates us to regress global localizations from convnet features in an end-to-end manner. The previously constructed offline database is still insufficient for a convnet training. We hereby augment the originally constructed database by a thousand factor and take advantage of the transfer learning method to make our convnet regressor converge and have a good performance. In our test, the regressor can also give a global localization of an input camera image in real time.The results obtained by the two approaches provide us insights on the comparison and connection between handcrafted feature-based and convnet based methods. After analyzing and comparing the localization performances of both methods, we also talked about some perspectives to improve the localization robustness and precision towards the GIS-aided urban localization problem.

Este trabalho apresenta uma plataforma para coleta e análise de imagens urbanas, que integra Interfaces de Programação de Aplicativos \"Application Programming Interfaces\" (APIs) de sistemas de busca de imagens, Sistemas de Informações Geográficas (SIGs), mapas digitais e técnicas de visão computacional. Esta plataforma, INACITY, permite que usuários selecionem regiões de interesse e capturem elementos de relevância para a arquitetura urbana, como, por exemplo árvores e buracos em ruas. A implementação da plataforma foi feita de maneira a permitir que novos módulos possam ser facilmente incluídos ou substituídos possibilitando a introdução de outras APIs de mapas, SIGs e filtros de Visão Computacional. Foram realizados experimentos com as imagens obtidas através do \"Google Street View\" onde árvores são capturadas em áreas de bairros inteiros em questão de minutos, um ganho significativo quando comparado com o procedimento manual para levantamento deste tipo de dado. Além disso, também são apresentados resultados comparativos entre os métodos de visão computacional propostos para a detecção de árvores em imagens com outros métodos heurísticos, em um conjunto onde as árvores estão marcadas manualmente e assim as taxas de precisão e de redescoberta de cada algoritmo podem ser avaliadas e comparadas. / This project presents a platform that integrates Application Programming Interfaces (APIs), image retrieval systems, Geographical Information Systems (GISes), digital maps and Computer Vision techniques to collect and analyse urban images. The platform, INACITY (an acronym for INvestigate and Analyse a City), empowers users allowing them to select a region over a map and see urban features inside that region that have relevance to the urban architecture context, for instance trees. The implementation is extensible and it is designed to make it easy to add or replace new modules, for instance, to add a new API to present a map, different GISes and other Computer Vision filters.

Este trabalho apresenta uma plataforma para coleta e análise de imagens urbanas, que integra Interfaces de Programação de Aplicativos \"Application Programming Interfaces\" (APIs) de sistemas de busca de imagens, Sistemas de Informações Geográficas (SIGs), mapas digitais e técnicas de visão computacional. Esta plataforma, INACITY, permite que usuários selecionem regiões de interesse e capturem elementos de relevância para a arquitetura urbana, como, por exemplo árvores e buracos em ruas. A implementação da plataforma foi feita de maneira a permitir que novos módulos possam ser facilmente incluídos ou substituídos possibilitando a introdução de outras APIs de mapas, SIGs e filtros de Visão Computacional. Foram realizados experimentos com as imagens obtidas através do \"Google Street View\" onde árvores são capturadas em áreas de bairros inteiros em questão de minutos, um ganho significativo quando comparado com o procedimento manual para levantamento deste tipo de dado. Além disso, também são apresentados resultados comparativos entre os métodos de visão computacional propostos para a detecção de árvores em imagens com outros métodos heurísticos, em um conjunto onde as árvores estão marcadas manualmente e assim as taxas de precisão e de redescoberta de cada algoritmo podem ser avaliadas e comparadas. / This project presents a platform that integrates Application Programming Interfaces (APIs), image retrieval systems, Geographical Information Systems (GISes), digital maps and Computer Vision techniques to collect and analyse urban images. The platform, INACITY (an acronym for INvestigate and Analyse a City), empowers users allowing them to select a region over a map and see urban features inside that region that have relevance to the urban architecture context, for instance trees. The implementation is extensible and it is designed to make it easy to add or replace new modules, for instance, to add a new API to present a map, different GISes and other Computer Vision filters.

Submitted by Cássia Santos (cassia.bcufg@gmail.com) on 2016-12-16T11:16:34Z No. of bitstreams: 2 Dissertação - Allan Kardec Lopes - 2016.pdf: 3329186 bytes, checksum: 643e80a250306bf9b1899cc724206ff5 (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Approved for entry into archive by Jaqueline Silva (jtas29@gmail.com) on 2016-12-16T16:49:55Z (GMT) No. of bitstreams: 2 Dissertação - Allan Kardec Lopes - 2016.pdf: 3329186 bytes, checksum: 643e80a250306bf9b1899cc724206ff5 (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Made available in DSpace on 2016-12-16T16:49:55Z (GMT). No. of bitstreams: 2 Dissertação - Allan Kardec Lopes - 2016.pdf: 3329186 bytes, checksum: 643e80a250306bf9b1899cc724206ff5 (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) Previous issue date: 2016-11-23 / Fundação de Amparo à Pesquisa do Estado de Goiás - FAPEG / Urban environments, such as streets, roads and buildings, always require management and maintenance to better use. In this sense, computational tools to assist their managers are always desirable. Furthermore, these tools generally decrease spending in order to automate several tasks. This research presents an approach to recognition of pole utility in streets mapped by images from Google Street View. Features such as color, texture and shape were examined in order to find the best set of information that represents the objects of interest. The recognition was performed by a neural network type Multilayer Perceptron trained with the Levenberg-Marquardt algorithm. The results show a higher accuracy in recognition when used in combination, mode RGB and texture properties as features to represent the structures present in the images. / Ambientes urbanos, tais como ruas, estradas e construções, sempre demandam gerenciamento e manutenção para que sejam melhor utilizados. Nesse sentido, ferramentas computacionais que auxiliem seus gestores são sempre desejáveis. Por outro lado, tais ferramentas geralmente diminuem os gastos tendo em vista que automatizam várias tarefas. Esta pesquisa apresenta uma abordagem para o reconhecimento de postes da rede elétrica em imagens de ruas mapeadas pelo Google Street View. Características como cor, textura e forma foram pesquisadas com o objetivo de se encontrar o melhor conjunto de informações que represente os objetos de interesse. O reconhecimento foi realizado por uma rede neural do tipo Multilayer Perceptron treinada com o algoritmo Levenberg-Marquardt. Os resultados obtidos demonstram uma acurácia superior no reconhecimento quando se utiliza, de forma combinada, a moda RGB e propriedades de textura como características para representar as estruturas presentes nas imagens

Many municipalities around Sweden use public access spatial data from Artportalen to track and fight invasive species, but anyone can upload data to Artportalen and not all of it becomes verified. They could potentially enhance their efforts by incorporating street view technology for more reliable data. One municipality that uses Artportalen data is Jönköping, which suffers from Reynoutria japonica. I randomly selected 50 areas around Jönköping with reports of R. japonica and went through them using Street View. The purpose of this study is to test if Street View can be used to improve the reliability of Artportalen data and if it can fully replace other methods to do so. I counted each individual of R. japonica I could identify and compared it to the number of reports in each area. Among the 50 areas only 26 were reachable in Google Street View and some plants were impossible to identify if they were R. japonica or not. When looking at the areas that could be reached it showed no difference between Reports in Artportalen and what was found in Street View if you included the unidentifiable plants. When excluding unidentified plants or when including areas that could not be reached it showed a significant difference. Using Google Street View to complement Artportalen data for R. japonica comes with benefits and limitations. Many areas were impossible to reach, but where accessible, it proved effective in identifying misreports and finding unreported plants without the need to visit the location. Street View data can enhance the reliability of distribution data used for monitoring invasive species without ever needing to travel, however it cannot fully replace other methods of enhancement and for the best reliability, a combination of Street View and on-site visits is necessary. / Många kommuner runt om i Sverige använder offentligt tillgängliga geodata från Artportalen för att spåra och bekämpa invasiva arter, men vem som helst kan ladda upp data till Artportalen och inte all data blir verifierad. Kommunerna skulle potentiellt kunna förbättra sina insatser genom att använda Street View teknologi för att få mer tillförlitliga data. En kommun som använder sig av data från Artportalen är Jönköping, som lider av den invasiva växten Reynoutria japonica. Syftet med den här rapporten är att testa om Street View kan användas för att komplimentera data från Artportalen och om det kan helt byta ut andra metoder att göra samma sak. Jag valde slumpmässigt ut 50 områden runt Jönköping med rapporter om R. japonica och gick igenom dem med Street View. Jag räknade varje individ av R. japonica jag kunde identifiera och jämförde det med antalet rapporter i varje område. Bland de 50 områdena var det bara 26 som kunde nås med Google Street View och vissa växter var omöjliga att identifiera om de var R. japonica eller inte. När man tittade på de områden som kunde nås visade det ingen skillnad i antal växter mellan rapporterna i Artportalen och vad som hittades i Street View om man inkluderade de oidentifierbara växterna. När man exkluderade oidentifierade växter eller när man inkluderade områden som inte kunde nås visade det en signifikant skillnad. Att använda Google Street View för att komplettera Artportalen-data för R. japonica har både fördelar och begränsningar. Många områden var omöjliga att nå, men där de var 2 tillgängliga visade det sig effektivt för att identifiera felrapporter och hitta orapporterade växter utan att behöva besöka platsen. Street View-data kan förbättra tillförlitligheten av distributionsdata som används för att övervaka invasiva arter utan att man någonsin behöver resa, men det kan inte fullt byta ut andra metoder att öka tillförlighet och för bästa resultat är en kombination av Street View och platsbesök nödvändig.

The present work proposes methods to calculate street-level exposures to solar radiation. The methods comprise a combination of different software algorithms, online databases and real-time standard measurements of solar radiation. Firstly, the use of the free access image database “Google Street View” to reconstruct urban geometries is illustrated. Google Street View represents an enormous source of information readily available for its general use in the field of urban atmospheric studies. With the aid of existing software packages, it was possible to reconstruct urban geometries as projected fisheye images of the canyon upper-hemispheric view, and to model total-shortwave solar irradiance within an urban canyon. The models allowed the calculation of relative street-canyon irradiance as a fraction of that received under a full-sky view, depending on the visibility of the solar disc and the reduced sky view factor. The combined use of the ideal models with real-time data allows for the calculation of street-canyon irradiance under any cloud conditions. Validation of these techniques was obtained by comparing the calculations against in situ measurements of irradiance from a local street canyon. The existing software, however, does not allow the calculation of spectral irradiance, required for inferring, for example, the biological effects of solar radiation. The use of spectral radiative transfer software was explored to provide spectral irradiance, but commonly available models do not include the effects of horizon obstructions. The approach presented here followed the same general guidelines used to calculate total-shortwave irradiance. The spectral models required a spectral partitioning of global irradiance into direct and diffuse components, allowing the independent analysis of horizon obstruction effects on these components at each wavelength. To partition global irradiance, two equations were developed for the calculation of the diffuse-to-global irradiance ratio (DGR) under cloudless conditions: one based on simplified radiative transfer theory, and an empirical fit for local conditions. Afterwards, the effects of horizon obstructions were evaluated in combination with real-time measurements of unobstructed global spectral irradiance. A set of simulated obstructions were used to validate the models. Finally, it was observed that neglecting the anisotropic distribution of the diffuse component of solar radiation in these simple models could produce large uncertainties in some situations. A practical solution for including the anisotropy of diffuse radiation was proposed, requiring images from an unobstructed digital sky camera. The combination of tools described here will allow calculation of total and spectral global irradiance upon a flat horizontal surface whatever the local field of view. This is possible at any geographical location were the urban geometries can be described, either by manually obtaining digital photographs, or through the Google Street View database, and where there is a reasonably local standard measurement of radiation.

Memoria para optar al título de Ingeniero Civil Eléctrico / El presente trabajo tiene como objetivo principal la predicción de robos de vehículos en la ciudad Santiago de Chile, mediante la confección de un modelo de Redes Neuronales que se alimente de información gráfica, socioeconómica y temporal de un sector geográfico de robo. Para ello se utiliza la base de edatos de robos de vehículos de la Asociación de aseguradoras de autos de Chile (AACH), y a cada muestra se le asignan cuatro imágenes de Google Street View, 15 características socioeconómicas extraídas del Instituto Nacional de Estadística (INE) y de la Infraestructura de Datos Geospaciales de Chile (IDE), en conjunto con la temperatura en la fecha del incidente y la cantidad de incidentes cercanos que ocurren en el mes anterior. Esta memoria está enmarcada en el proyecto Fondef ID16I10222, liderado por los profeso- res Richard Webe y Ángel Jiménez del departamento de Ingeniería Industrial de la Univer- sidad de Chile. Para alcanzar el objetivo señalado al comienzo ,se propone un modelo de red neuronal alimentado por joint features y deep features; esto es que la red neuronal aprende de dis- tintos formatos de información como lo son datos estadísticos e imágenes. Además, en el caso de las imágenes, no son utilizadas directamente sino que se extraen carecterísticas de ellas, mediante una segunda red convolucional denominada Alexnet. Para entrenar la red es necesario disponer de zonas seguras, para lo cual se organiza la base de datos de ro- bos por fecha y se agrupa por mes, períodos en los cuales se denotan como zonas seguras aquellas que en el presente mes no posean a la redonda de 500 metros una ubicación de robos. El principal resultado esperado de la aplicación de este modelo es la obtención de proba- bilidades por zonas en una grilla de 0.01 grados de latitud y longitud sobre Santiago de Chile que puede ser representado como un heatmap de riesgo en la ciudad. Del modelo presentado, utilizando el conjunto completo de características se obtiene un 97.7% de accuracy, un 95% de precisión , un 98.9% de recall y un 97% de F1. Además al hacer reducción de características mediante Recursive feature selection y seleccionar las principales 1000 características, se obtiene un 92.96 % de accuracy, un 93.46 % de precisión, un 92.09 % de recall y un 92.77 % de F1. Con esto se concluye que el desempeño del modelo propuesto efectivamente permite una alta tasa de clasificación y además permite la creación de una representación gráfica. / Fondef 16I10222

Invasiva främmande arter är ett allvarligt hot mot biologisk mångfald. Kartläggning och övervakning av invasiva arter är viktigt för att kunna arbeta förebyggande mot spridningen och reducera effekterna. Parkslide är en invasiv art som idag är spridd över södra Sverige och riskerar att utgöra ett hot mot inhemska arter samt orsaka skada på infrastruktur på grund av dess starka invasiva förmåga. Syftet med denna studie är att undersöka förekomst av parkslide i relation till tätorter samt att jämföra två olika metoder för kartläggning av parkslide i Blekinge län. I studien användes data från SLU Artdatabankens Artportal från åren 2020–2021 för att jämföra förekomsten av parkslide i närheten av (0–5 km) och längre bort (5–10 km) från tätort. Fortsatt undersöktes Google Street View (GSV) som en metod för kartläggning av parkslide genom att jämföra antalet observationer med antalet inrapporteringar i Artportalen. Slutligen gjordes en jämförelse mellan inrapporterad storlek (m2) på bestånd i Artportalen och från en fältundersökning. Resultatet av studien visade på en trend att antalet inrapporteringar ökar nära tätort, men trenden är inte statistiskt fastställd. Fortsatt visade resultatet att antalet observationer i GSV stämde överens med antalet inrapporteringar i Artportalen. Till sist visade resultatet att medelvärdena för storlek i Artportalen och från fältundersökningen inte skiljer sig signifikant från varandra. Dock var variationen i dessa data stor och visade ibland på stora skillnader i storlek, vilket sannolikt minskade den statistiska säkerheten. Sammantaget visar denna studie att inrapportering genom Artportalen kan ge en indikation på hur spridningen av parkslide ser ut men att data kan behöva verifieras genom fältstudier. Dessutom visar studien att GSV kan vara en användbar metod för att kontrollera inrapporteringar på ett resurseffektivt sätt. / Invasive alien species is a major threat to endemic biodiversity. Mapping and monitoring of invasive species is important in order to prevent spreading and reduce impacts. The invasive plant Japanese knotweed is spread across the south of Sweden and is threatening to harm native species and cause damage to infrastructure, due to its aggressive invasive capacity. The aim of this study was to investigate the presence of Japanese knotweed in relation to urban areas and to compare two different methods of mapping Japanese knotweed in Blekinge County, Sweden. Data from SLU Swedish Species Information Centre (Artdatabankens Artportal) from the years 2020–2021 was used to compare the presence of Japanese knotweed near to (0–5 km) and farther away from (5–10 km) urban areas. In addition, Google Street View (GSV) was used as a method for mapping Japanese knotweed by comparing the number of observations inGSV to the number of reported findings in Artportalen. Finally, a comparison was made between the patch size (m2) reported in Artportalen versus from a field study.The result of the study shows a trend that the number of reported findings increase near urban areas, although the trend is not statistically significant. The result alsoshows that the number of observations in GSV match the number of reported findings in Artportalen. Finally, the result indicates that the mean patch size in Artportalen versus that from the field study does not significantly differ from each other. However, the variation in the data was substantial and sometimes showed large differences in patch size, which likely reduced statistical power. In conclusion, the result of this study shows that reporting through Artportalen can give an indication of the spread of Japanese knotweed, but that the findings may need to be verified by field studies. Furthermore, the study shows that GSV can be a useful method to verify reported findings in a resource-efficient way.

Sambandet mellan den fysiska miljön och brottslighet har sedan länge varit en relevant fråga inom den kriminologiska diskursen. Den föreliggande studien ämnar vidare undersöka huruvida fysisk oordning i urbana bostadsområden kan studeras genom Google Street View, ett webbaserat instrument för virtuella observationer. Syftet med studien är att undersöka om virtuellt observerad och självrapporterad uppfattad grad av fysisk oordning i bostadsområdet mäter samma fenomen, men även om virtuellt observerad fysisk oordning kan förklara skillnader i självrapporterad utsatthet för egendomsbrott. Genom att utföra virtuella observationer av fysisk oordning med hjälp av Google Street View i tjugo centralt belägna bostadsområden i Malmö visar resultaten att observerad och självrapporterad grad av fysisk oordning är starkt korrelerade och förefaller mäta samma fenomen. Resultaten visar även att observerad nivå av fysisk oordning genom Google Street View till viss del kan förklara variansen av utsatthet för egendomsbrott mellan bostadsområden. Avslutningsvis framhålls i studien att virtuella observationer genom Google Street View är ett lovande samt potentiellt kostnadseffektivt tillvägagångssätt för att undersöka graden av fysisk oordning i urbana bostadsområden. Användandet av Google Street View kantas dock av flera begränsningar som både framhålls och diskuteras grundligt i denna studie. / The correlation of physical environment and crime has been an ever relevant topic in the criminological discourse. This study attempts to unravel whether physical disorder in inner-city urban neighborhoods may be studied through Google Street View as a virtual observational tool. The aims of the study is to examine whether virtually observed and self-reported perceived level of neighborhood disorder measure the same phenomenon, and whether virtually observed physical disorder may explain variations of self-reported victimization of property crimes. By conducting virtual observations of physical disorder in twenty inner-city neighborhoods of Malmö through Google Street View, the results of the study propose that virtually observed and self-reported perceived level of disorder is strongly correlated and thus seems to measure the same phenomenon to a great extent. The results of the study also imply that observed physical disorder through Google Street View also accounts for neighborhood differences in victimization of property crimes. The study concludes that virtual observation through Google Street View is a promising and potentially cost-effective alternative approach when auditing neighborhood physical disorder. The methodology does however suffer by limitations which is highlighted and thoroughly discussed.