Approche systémique et participative du diagnostic urbain

Social, economic and environmental phenomena as well as the presence of groups of stakeholders with diverging and contradictory interests highlight the complexity of urban projects. Traditional instruments of urban management are seldom adapted to these challenges, as they are often founded on sector- and expert-based approaches. Furthermore geographical information is paradoxically often available in a very large quantity, but not sufficiently relevant and synthetic for an understanding of urban complexity. In this context the goal of this thesis is to develop a method for the elaboration of relevant indicators systems from different stakeholders' cognitive representations of urban complexity; and thus to reinforce territorial diagnosis in the context of participative decision-making processes. However these differences of representations are often sources for conflicts. Each stakeholder uses his own representation in order to defend his point of view in a debate. The notice of these differences, from the "problem setting" phase of the decision-making process should contribute to a better structured debate and to a better management of the conflicts around a territorial project. This goal can be reached through consideration of all opinions and through an establishment of a discussion-space between different groups of representations. This allows the development of a coherent diagnosis framework which includes different stakeholders' cognition and perception of the territorial context. Our reflection is concretized by the developing of a prototype of interface that help a stakeholder to construct his own representation or urban phenomena, by organizing them in a dialectic (discussions) and didactic (complexity-training) manner. The phenomena-systems created are in the form of a causal model. The interpretation of this model permits to highlight critical phenomena through considering the preferred causal axes in order to output the stakeholder's goals. On this basis the stakeholder can debate and discuss with other stakeholders who also have constructed their own representations. From these individual representations, the question is how these representations can be integrated to form a collective representation of the territory. This process can be made in a deliberative manner. Before that, it is proposed to structure the debate by regrouping individuals models using analytical tools (classification, aggregation). Converging (collective reasons) and diverging elements of the urban complexity understanding are emphasized in the groups of models. The idea is thus to create a discussion-space for debates in order to find solutions. On the basis of different representation-groups, indicators-systems can be proposed. Indicators are selected in order to translate on one hand common representation of the main phenomena identified, on the other hand of causal relationships that result in strong differences. These indicators can contribute to a partial solution of conflicts by adding another, more objective point of view in the territorial context. Indicators are developed using spatio-thematical operators of geographical information systems. A concept of interface is proposed so as to facilitate the communication of systems of geographical indicators. In order to emphasis our approach, we have illustrated and applied the methodology of systemic urban-phenomena-modeling. We have conducted individual interviews with stakeholders who have participated in a diagnosis process in a neighborhood of Geneva and with stakeholders who are implied in suburban problematic in Québec. These discussions have shown the interest in our approach to help these stakeholders to formalize their representation of urban dynamics. In the same way, regrouping methods have proved their capability to reveal differentiated models of representation between groups of stakeholders. The method described presents an interest in participative decision-making and creates an advantageous framework for the interaction between these stakeholders around the representation of urban reality. The indicators-system contributes to this process and connects the representation of the territory and different stakeholders' cognitive representations according to a dynamic learning process. From this more solid and objectified basis the development of a diagnosis can succeed and the decision-making process can continue towards more operational phases.

User interaction with online geospatial systems

Jens Ingensand

User interaction with online geospatial systems is an inter-disciplinary research field that combines the two disciplines Human Computer Interaction (HCI) and geographic information sciences. In recent years online geospatial systems have been created for different purposes such as way-finding systems, online atlases and collaborative systems. In order to accommodate the fact that an increasing number of people have access to geospatial data and technology, research has begun to focus on the impact and the utilization of such systems. However few studies regarding the usability of online geospatial systems have been conducted and most of these studies were set up in artificial lab-environments with few mostly poorly described test-users. Furthermore to the best of our knowledge no formal research framework currently exists that allows for analyzing the interaction of real-world users with such systems and that takes into account the spatial nature of these systems. The goal of this thesis is to show that users, depending on their demographics, experience, skills and context show significant differences in their interaction strategies and performance. Another important factor that has been neglected in many recent laboratory-based usability studies is the user's computer. In this work, we want to show that users, depending on their computer's properties show significant differences regarding their interaction strategies, their performance and satisfaction. The work in this thesis consists of three parts: 1. the establishment of an evaluation framework that categorizes and defines the interaction of real-world users with geospatial systems. The four main entities of this framework are the user (in terms of demographics, skills, knowledge and context), the system (defined by the interface design, spatial data, spatial functions, technologies and architectures), the user interaction (interaction strategies, performance and perception of spatial data), and the satisfaction of using the system. 2. the development of methods and software tools for gathering experimental data based on the four entities of the evaluation framework 3. the validation of the evaluation framework with case studies involving systems that have been developed for specific groups of end-users. We have conducted three case studies where real-world users interact with three online geospatial systems. The important findings of those studies were that there are noticeable gender differences regarding spatial navigation, that user satisfaction appears to be biased by the user's previous experience with geographic information technologies and that the type of the input device used (e.g. mouse or touchpad) has a considerable influence on task-completion time and user satisfaction. The above results indicate that the user's demographic parameters and background have a significant effect on the user's spatial interaction strategies, performance and satisfaction. Furthermore we have found evidence that suggests that the user's perception of spatial data strongly depends on the scale. We conclude that the design and development of spatial interaction functions (spatial navigation, digitization, selection of objects and locations) in online geospatial systems should highly depend on the type of users that the systems are developed for. To further fuel research on usability aspects of online geospatial systems, we provide the necessary tools and methods for conducting remote- and laboratory based testing of such systems.

EPFL2010

Registration of single landscape photographs with 3D landscape models

Timothée Produit

Terrestrial imagery found in public databases is an alternative and complementary data source for environmental studies. Compared to usual data source, such as airborne images, terrestrial pictures may have higher temporal and spatial resolutions. They are especially valuable in studying the past, when airborne acquisitions were more rare or inexistent. Nevertheless, terrestrial image databases are typically imperfectly georeferenced, if at all. Hence, it is difficult to retrieve images of interest in a database. Moreover, georeferencing processes proposed for these images are either totally manual or rely on the existence of a collection of similar and overlapping images. The first is time-consuming and the second is appropriate only for particular sets of images. An alternative and more general georeferencing procedure is the registration of a picture with a landscape model. In this thesis, we propose to use as the reference a 3D virtual globe and our aim is to retrieve the orientation and the location of the camera. The location and orientation of a picture can be computed with correspondences that a user recognizes and clicks in a picture and in an orthoimage. However, these two distinct types of images are difficult to match because of the distortions resulting from the different viewpoints. Hence, we provide a navigation tool to scroll through a 3D model to ease this task. Indeed, to find the orientation of a picture, users typically exploit the skyline which appears only in a 3D view. In imitation, we develop a novel skyline-matching technique, based on Dynamic Time Warping, to compute the fine orientation of a picture. In a second phase, we assess whether the similarities between a 3D synthetic image and real images can be used to compute the orientation of a picture. The matching of real images with a model is challenging. We show how images shared on the web can be inserted in the landscape model to ease the georeferencing. Finally, all of the presented pose estimation schemes assume that every location of the map is as likely to be a shooting spot. However, we show that geographic indicators related to accessibility and morphology are correlated with a database of picture locations. Exploiting these relations, our last contribution is a map of the attractiveness. In this thesis, we observed that pose estimation from the automatic registration of a real image with a landscape model cannot compete with the accuracy reached by an operator or a standard photogrammetric workflow. However, the proposed methods can greatly reduce the operator workload for the manual georeferencing by finding automatically correspondences with a 3D landscape model. Second, they are a valid alternative if the required conditions for Structure-from-Motion are not respected. Finally, the accuracy obtained is sufficient to explore image databases with spatial queries and to add geographic information in the pictures. Registration of landscape images with landscape models has previously received limited attention. The tools developed in this thesis can ease the work of archivists (in the management and provision of efficient image databases), environmental scientists (for the extraction of georeferenced information from pictures) and the general public (to browse collections of augmented images inserted in a virtual globe).

EPFL2015

Biogeoinformatics for the management of Farm Animal Genetic Resources (FAnGR)

Solange Catherine Gaillard

In the context of both severe selection in farm animals and potential effects of climate change, it is crucial to implement a sustainable management of the breeding practice, supported by a judicious use of geographic information technologies. Based on this observation, this thesis advocates the use of biogeoinformatics (the combined use of biology, geographic information and informatics) to cope with the challenges encountered by the livestock sector. Indeed, although biogeoinformatics can provide key insights for FAnGR (Farm Animal Genetic Resources) management, the variety and complexity of tasks involved hinders a wider usage of this type of analyses. The thesis shows how novel dedicated tools are likely to facilitate the adoption of biogeoinformatics by animals scientists and by stakeholders in the livestock sector, while investigating three main challenges related to FAnGR management, namely i) erosion of genetic diversity, ii) effects of climate change on the breeding activity, and iii) pressure on typical cultural breeding practices such as high alpine grazing. On this basis, the thesis is organised around three axes:

Preserving locally adapted breeds In order to prevent the erosion of genetic diversity, locally adapted breeds should be monitored to prevent their extinction. To this end, we developed the open source GENMON WebGIS platform, able to monitor FAnGR and to evaluate the degree of endangerment of livestock breeds. The system integrates various sources of information that are linked with the help of geographic information: pedigree and introgression, geographical concentration of animals, cryo-conservation and sustainability of breeding activities. The score can be visualised on a map and allows a fast and regional identification of breeds in danger.
Preserving locally adapted genetic variations Considering the pace at which genetic diversity is being eroded, it has become urgent to identify and then preserve important genetic variations linked to locally adapted phenotypes. In this context, the recent SamÎ²ada software was designed to search for signatures of local adaptation through the study of genomeâenvironment association. However, preâ and postprocessing of data for this analysis can be labourâintensive, and, we therefore developed the R.SamBada R package providing a pipeline for landscape genomic analyses. Based on SamÎ²ada, it spans from the retrieval of environmental conditions at sampling locations to gene annotation using the Ensembl genome browser. As a result, it grants access to biogeoinformatic analyses to researchers with no skills in geography.
Preserving a traditional farming technique suited for local breeds The preservation of locally adapted breeds is also strongly linked with the conservation of traditional farming techniques, which in Switzerland, include the grazing of high alpine pastures during summer. One major effect of this transhumance on cows is that milk production declines due to food shortage and climatic stress. Here, we developed a new mathematical model to fit a lactation curve for mountain-pastured cows, and tested the influence of environmental, physiological, and morphological factors on the production using five million monthly milk records from Braunvieh cows. When compared to physiological factors, environmental variables show a limited impact on milk production at alpine pastures, precipitation in spring being the most important.

EPFL2020