Street-level Photographs Research Articles

Superimposing height-controllable and animated flood surfaces into street-level photographs for risk communication

Extreme flood events frequently threaten coastal and river communities, and communicating the potential impacts of such forecasts to their populations is crucial to protect property and human life. However, traditional methods to warn residents of forecasted flood events are often ignored or not fully understood. Recent works have produced 3D visualizations of flooding to better capture viewers’ attentions but tend to be expensive, visually unrealistic, or incapable of parameterizing water height. Here we propose an efficient and scientifically-grounded approach to generate realistic images and animations of a flood at any height composited with a photograph taken at street level. Using vehicular LIDAR point cloud and color photo data, we employ a convolutional neural network to generate a dense depth map across an image. We use 3D modeling software to automatically generate and render a water surface, along with its own depth map, at the appropriate height and orientation. The depth maps are used to composite the photo with the rendered water surface to generate the final images, and this process can be repeated to generate videos of rising coastal floodwaters with animated waves within minutes. These visualizations are striking, and the overall framework can be supported by any particular image collection or depth map construction methodology, making this an affordable and achievable approach to flood risk communication.

Automated survey of selected common plant species in Thai homegardens using Google Street View imagery and a deep neural network

Most previous studies of homegardens have used labor-intensive boots-on-the-ground plant surveys, owner questionnaires, and interviews, limiting them to at most a few hundred homegardens. We show that automated analysis of publicly available imagery can enable surveys of much greater scale that can augment these traditional data sources. Specifically, we demonstrate the feasibility of using the high-resolution street-level photographs in Google Street View and an object-detection network (RetinaNet) to create a large-scale high-resolution survey of the prevalence of at least six plant species widely grown in road-facing homegardens in Thailand. Our research team examined 4000 images facing perpendicular to the street and located within 10 m of a homestead, and manually outlined all perceived instances of eleven common plant species. A neural network trained on these tagged images was used to detect instances of these species in approximately 150,000 images constituting views of roughly one in every ten homesteads in five provinces of northern Thailand. The results for six of the plant species were visualized as heatmaps of both the average number of target species detected in each image and individual species prevalence, with spatial averaging performed at scales of 500 m and 2.5 km. Urban-rural contrasts in the average number of target species in each image are quantified, and large variations are observed even among neighboring villages. Spatial heterogeneity is seen to be more pronounced for banana and coconut than for other species. Star gooseberry and papaya are more frequently present immediately outside of towns while dracaena and mango persist into the cores of towns.

Built environment attributes and crime: an automated machine learning approach

This paper presents the development of an automated machine learning approach to gain an understanding of the built environment and its relationship to crime. This involves the automatic capture of street-level photographs using Google Street View (GSV), followed by the use of supervised machine learning techniques (specifically image feature recognition) to recognise features of the built environment. In this exploratory proof-of-concept work, 8 key features (building, door, fence, streetlight, tree, window, hedge, and garage) are considered and a worked case-study is demonstrated for a small geographical area (8300 square kilometres) in Northern England. A total of 60,100 images were automatically collected and analysed across the area where 5288 crime incidents were reported over a twelve-month period. Dependency between features and crime incidents are measured; however, no strong correlation has been identified. This is unsurprisingly considering the high number of crime incidents in a small geographic region (8300 square kilometres), resulting in an overlap between specific features and multiple crime incidents. Furthermore, due to the unknown precise location of crime instances, an approximation technique is developed to survey a crime’s local proximity. Despite the absence of a strong correlation, this paper presents a first-of-a-kind cross-discipline approach to attempt and use computation techniques to produce new empirical knowledge. There are many avenues of future research in this fertile and important area.

Fusing street level photographs and satellite remote sensing to map leaf area index

• Street-level photographs were fused with satellite data to map LAI. • The new method for LAI mapping was rapid and validated using field data. • The approach will allow high-resolution maps of LAI to be generated rapidly. Leaf area index (LAI) is an important structural parameter of vegetation, and is used in many models of climate and ecosystem services. Maps of LAI are typically produced by relating satellite remote sensing with field-based measurements of LAI, but such field measurements are time consuming to collect over large areas. In this study we develop a rapid and scalable method for mapping LAI by fusing high-resolution freely-available Sentinel 2 satellite imagery with ground measurements of LAI extracted from a large publicly available database of street-level panoramic photographs. The use of existing street-level photographs allowed large numbers of training data to be automatically obtained. The method developed here was validated against a field dataset collected using established techniques. The use of existing online databases of street-level photographs will allow rapid mapping of LAI in many situations worldwide. The technique developed here may be particularly useful in cities, which have high heterogeneity in vegetation, and high densities of street level photographs collected along road networks. The approach could be applied to map LAI at high resolution across very large areas, for national- or continental-scale comparison.

Characterizing community-wide housing attributes using georeferenced street-level photography.

New methods are needed to efficiently characterize built environment attributes and residential behaviors to improve exposure assessment in epidemiologic research, given limitations of available databases and approaches. Window opening and presence of air conditioning (AC) units predict indoor air quality and thermal comfort, but data are not widely available. In this study, we tested the utility of a GIS-based tool for rapidly assessing open windows and window/wall air conditioning units in the city of Chelsea, Massachusetts using georeferenced street-level photographs and crowdsourced online surveys. We characterized open windows and window/wall AC units for 969 parcels in the winter and 1 213 parcels in the summer, requiring approximately 40 person-hours per season. In the winter, 21% of parcels surveyed had a window or wall AC unit and 19% had an open window. In the summer, 69% had a window or wall AC unit and 53% had an open window. We demonstrated an efficient method for rapidly characterizing open windows and window/wall AC units across an entire city. This tool can help characterize exposures for epidemiological research, engage community members, and inform local land use planning and decision-making.

Can Field Crews Telecommute? Varied Data Quality from Citizen Science Tree Inventories Conducted Using Street-Level Imagery

Street tree inventories are a critical component of urban forest management. However, inventories conducted in the field by trained professionals are expensive and time-consuming. Inventories relying on citizen scientists or virtual surveys conducted remotely using street-level photographs may greatly reduce the costs of street tree inventories, but there are fundamental uncertainties regarding the level of data quality that can be expected from these emerging approaches to data collection. We asked 16 volunteers to inventory street trees in suburban Chicago using Google Street ViewTM imagery, and we assessed data quality by comparing their virtual survey data to field data from the same locations. We also compared virtual survey data quality according to self-rated expertise by measuring agreement within expert, intermediate, and novice analyst groups. Analyst agreement was very good for the number of trees on each street segment, and agreement was markedly lower for tree diameter class and tree identification at the genus and species levels, respectively. Interrater agreement varied by expertise, such that experts agreed with one another more often than novices for all four variables assessed. Compared to the field data, we observed substantial variability in analyst performance for diameter class estimation and tree identification, and some intermediate analysts performed as well as experts. Our findings suggest that virtual surveys may be useful for documenting the locations of street trees within a city more efficiently than field crews and with a high level of accuracy. However, tree diameter and species identification data were less reliable across all expertise groups, and especially novice analysts. Based on this analysis, virtual street tree inventories are best suited to collecting very basic information such as tree locations, or updating existing inventories to determine where trees have been planted or removed. We conclude with evidence-based recommendations for effective implementation of this type of approach.

Tropel: Crowdsourcing Detectors with Minimal Training

This paper introduces the Tropel system which enables non-technical users to create arbitrary visual detectors without first annotating a training set. Our primary contribution is a crowd active learning pipeline that is seeded with only a single positive example and an unlabeled set of training images. We examine the crowd's ability to train visual detectors given severely limited training themselves. This paper presents a series of experiments that reveal the relationship between worker training, worker consensus and the average precision of detectors trained by crowd-in-the-loop active learning. In order to verify the efficacy of our system, we train detectors for bird species that work nearly as well as those trained on the exhaustively labeled CUB 200 dataset at significantly lower cost and with little effort from the end user. To further illustrate the usefulness of our pipeline, we demonstrate qualitative results on unlabeled datasets containing fashion images and street-level photographs of Paris.

An Adapted Wayfinding System for Pedestrians with Cognitive Disabilities

This paper presents a novel wayfinding system adapted to people with cognitive disabilities. It adapts to the user in terms of route calculation, instructions delivery, and interface design. To do so, the system divides the calculated route into atomic instructions and uses street-level photographs at the decision points. To evaluate this approach, we compared it with a commercial navigation application on a field trial with a sample of users (N= 18). From the evaluation, we concluded that our system improves users’ performance in terms of the number who reached the destination and were able to identify it correctly.

Exploratory Completeness Analysis of Mapillary for Selected Cities in Germany and Austria

Mapillary, founded in early 2014, is a Web 2.0 application that allows voluntary users to contribute street level photographs from all over the world. This paper analyses the growth in uploaded data and contributor numbers over time, and assesses data completeness by comparing Mapillary with Google Street View and OpenStreetMap data for selected cities in Germany and Austria. Results show that as of now Google Street View generally provides a better coverage in the cities in which it is offered, but that Mapillary has the potential to reach higher coverage along off-road segments, such as footpaths and bicycle trails, or even railroads.

Made for walking: density and neighborhood form

This book looks at urban neighborhoods and the characteristics that promote walking. Twelve neighborhoods, two from Canada, ten from the United States, are presented as case studies for walkability. The characteristics that make them walkable are detailed and street level photographs are provided.

From Spaceship Earth to Google Ocean: Planetary Icons, Indexes, and Infrastructures

What sort of image does the planet Earth possess at the opening of the 21st century? If in the 1960s, the Whole Earth, the planet as seen from space, became a cold war, proto-environmentalist icon for a fragile ocean planet, in the 2010s, Google Earth, the globe encountered as a manipulable virtual object on our computer screens, has become an index for multiple and socially various interpretations and interventions; its thicket of satellite images, text legends, and street level photographs can all be tagged, commented upon, modified. In this essay, I examine a kindred image-object, Google Ocean, asking what sort of representation of the planetary sea is in the making in our digital days. Stirring up the century-old classification of signs by semiotician Charles Sanders Peirce, I argue that Google Ocean is a mottled mash of icons, indexes, and symbols of the marine and maritime world as well as a simultaneously dystopian and utopian (that is to say, heterotopian) diagram of the sea - though one that floats in a media ecology that tends to occlude its infrastructural history and conditions of possibility.

Finding food: Issues and challenges in using Geographic Information Systems (GIS) to measure food access

A significant amount of travel is undertaken to find food. This paper examines challenges in measuring access to food using Geographic Information Systems (GIS), important in studies of both travel and eating behavior. It compares different sources of data available including fieldwork, land use and parcel data, licensing information, commercial listings, taxation data, and online street-level photographs. It proposes methods to classify different kinds of food sales places in a way that says something about their potential for delivering healthy food options. In assessing the relationship between food access and travel behavior, analysts must clearly conceptualize key variables, document measurement processes, and be clear about the strengths and weaknesses of data.

High resolution urban morphology data for urban wind flow modeling

The application of urban forestry methods and technologies to a number of practical problems can be further enhanced by the use and incorporation of localized, high resolution wind and temperature fields into their analysis methods. The numerical simulation of these micrometeorological fields will represent the interactions and influences of urban structures, vegetation elements, and variable terrain as an integral part of the dynamics of an urban domain. Detailed information of the natural and man-made components that make up the urban area is needed to more realistically model meteorological fields in urban domains. Simulating high resolution wind and temperatures over and through an urban domain utilizing detailed morphology data can also define and quantify local areas where urban forestry applications can contribute to better solutions. Applications such as the benefits of planting trees for shade purposes can be considered, planned, and evaluated for their impact on conserving energy and cooling costs as well as the possible reconfiguration or removal of trees and other barriers for improved airflow ventilation and similar processes. To generate these fields, a wind model must be provided, as a minimum, the location, type, height, structural silhouette, and surface roughness of these components, in order to account for the presence and effects of these land morphology features upon the ambient airflow. The morphology of Sacramento, CA has been characterized and quantified in considerable detail primarily for wind flow modeling, simulation, and analyses, but can also be used for improved meteorological analyses, urban forestry, urban planning, and other urban related activities. Morphology methods previously developed by Ellefsen are applied to the Sacramento scenario with a high resolution grid of 100 m × 100 m. The Urban Morphology Scheme defines Urban Terrain Zones (UTZ) according to how buildings and other urban elements are structured and placed with respect to each other. The urban elements within the 100 m × 100 m cells (one hectare) are further described and digitized as building height, building footprint (in percent), reflectivity of its roof, pitched roof or flat, building's long axis orientation, footprint of impervious surface and its reflectivity, footprint of canopy elements, footprint of woodlots, footprint of grass area, and footprint of water surface. A variety of maps, satellite images, low level aerial photographs, and street level photographs are the raw data used to quantify these urban properties. The final digitized morphology database resides in a spreadsheet ready for use on ordinary personal computers.