Measurement of visual parameters of landscape using projections of photographs in GIS

https://doi.org/10.1016/j.compenvurbsys.2016.09.005Get rights and content

Highlights

  • A quick and inexpensive way to measure visual parameters of landscape is proposed.

  • Apparent density is presented as a new visual parameter of landscape.

  • Visual angles/magnitudes or densities are used to assess landscape changes.

  • Panoramas can be unambiguously defined using just the coordinates of the viewpoint.

  • The method can be used for calibrating viewshed and visual magnitude calculations.

Abstract

Objective evaluations of the scenic or visual attributes of a landscape contribute to their thorough characterization and assessment. This means that the techniques used for calculating the visibility of the landscape and for studying spatial patterns (e.g. landscape metrics) must be complemented by considering the real view perceived by the observer. This could be done by measuring the visual parameters of the landscape using photographs, but there are currently no standard procedures as to how to do this. We therefore propose a method that establishes a common system of reference for the ground plan and panoramic photographs of a particular area, and for taking measurements on these images relating to visual angles, visual magnitudes and apparent densities. This process is quick and inexpensive and involves obtaining calibrated equirectangular projections of photographs (individual or combined in the form of panoramas), to which we assign a spherical coordinate system in a GIS. This allows us to define a panorama accurately and unambiguously with a sole reference namely the coordinates of the viewpoint from which the panorama was generated. This method also ensures that the same results are obtained even when different camera equipment is used to take the photograph(s). This method can be applied in different ways to characterize landscape aesthetics (impact assessments, evaluation of landscape preferences, estimation of the sky view factor, evolution of the landscape over time, management of urban panoramic views etc.) and is useful for the calibration of automated systems for the calculation of viewsheds and visual magnitudes.

Introduction

It is often necessary to measure the visual perception of existing elements and structures in the landscape, or of possible transformations of the landscape whose effects must be assessed. Ever since the concept of “valuable landscape” was enshrined in international agreements such as the European Landscape Convention (Council of Europe, 2000), procedures are required for the “objective” assessment of a landscape's aesthetic attributes (Uuemaa, Mander & Marja, 2013), associated with its “information functions” according to the framework proposed by De Groot (2006). However, the available “landscape metrics” for studying spatial patterns (for example using FRAGSTATS type processes) have proved insufficient when making assessments oriented towards the visual aspects of landscape (Uuemaa et al., 2013). These difficulties have led some authors to state that “the quantitative evaluation of the quality of a landscape remains a challenge” (Li & Mander, 2009, p. 40).

Various recent researchers have been working with some of these perception-related parameters, such as field of vision (Dupont et al., 2015, Palmer, 2015), visual angle (Foley and Matlin, 2009, Greater London Authority, 2012, Mairie de Paris, 2013) or the visual surface area occupied by an element of the landscape in the image. This last parameter is also known as visual magnitude and is defined as “the number of square minutes that a unit of landscape or a structure occupies in the field of vision” (Iverson, 1985, p. 16). For a particular viewpoint, these kinds of parameters can be estimated using a Geographic Information System (GIS). Significant progress has been made in the calculation of visual magnitudes using procedures related with the calculation of viewsheds (Grêt-Regamey et al., 2007, Chamberlain and Meitner, 2013), but at times visibility calculations are not totally accurate due to the effect of perspective (Sang et al., 2008, Nutsford et al., 2015) and may require adjustments to fit better with the observations in the field (Bishop & Hulse, 1994). These problems are frequently related to the difficulties inherent in correctly modelling the landscape effect of vegetation (Tomko, Trautwein & Purves, 2009).

One way to address these inaccuracies is to use photographs for measuring visual parameters. This additional input is of particular interest when dealing with sloping areas, where the effect of the slope makes the visual perception of the landscape very different from the topology that appears on the maps (although some GIS calculations, such as visual magnitude, take this into account). Photographs have been used for decades in the analysis, assessment, planning and design of the landscape (Dunn, 1976, Shuttleworth, 1980, Countryside Commission, 1988, Li and Mander, 2009, Bishop and Miller, 2007) and in a wide range of applications such as for example the assessment of the visual impact of forestry operations (Marc, 2008) or new buildings in rural areas (Hernández, García & Ayuga, 2004), the analysis of urban views (Greater London Authority, 2012), the study of visual quality (Arriaza, Canas-Ortega, Canas-Madueno & Ruiz-Aviles, 2004) or the assessment of the evolution of landscape over time (Puschmann & Dramstad, 2003). Many researchers have used simulations which display or evaluate the effect of specific changes (Sheppard, 1989, Lange and Bishop, 2005, Grêt-Regamey et al., 2007). Photographs can provide the GIS with additional territorial information that does not appear or is not updated in the land use categories (Ode, Tveit & Fry, 2010), and for evaluating the visual effect of certain specific elements that are not included in Digital Terrain Models.

However, there are no standardized procedures in the literature for measuring visual parameters in photographs, although some scholars have combined landscape metrics with visualizations in order to address perception issues (Ode, Fry, Tveit, Messager & Miller, 2009). This is firstly due to the wide variety of techniques used. Authors in this field normally detail the type of camera and the focal length they use (Downes & Lange, 2014), and the field of vision (Palmer, 2015); on some occasions they precisely define fixed positions, fields of vision and focal lengths as a reference so that images covering exactly the same field can be obtained in the future (Puschmann & Dramstad, 2003). There is also wide disparity in the way the photograph is processed. In some cases, the dimensions and the pixel size of the final image are considered and the apparent size of the photographed objects is calculated (Tomko et al., 2009). In other cases, the visual magnitude of certain elements is measured manually by superimposing a grid on the photographs (Schroeder, 1988, Grimmond et al., 2001), by counting the pixels (García-Moruno, Montero-Parejo, Hernández-Blanco & López-Casares, 2010) or using electronic planimeters (Marc, 2008). However, in all these cases the procedures are generally not comparable as they depend on the technical parameters of each photograph.

In this article we propose a method for the objective analysis of a panorama obtained from a certain viewpoint, using a quick and inexpensive procedure, which establishes a common system of references between the panoramic photographs and the cartographic projection in the same study area. For this purpose, we used various free software applications for creating panoramas and cartographic projections. These enabled us to identify particular structures and elements simultaneously in both representations of the view and to measure the different visual parameters (such as visual angles and visual magnitudes) with great accuracy and precision. We also applied these procedures in the measurement of a new visual parameter called apparent density.

In this way we provide a systematic method with objective results that are comparable for different viewpoints. This method can be applied to a wide variety of processes of visual analysis of the landscape and its scenic attributes.

Section snippets

Materials and methods

The measurement procedures can be applied both to single photographs and to compositions of various pictures in panoramic format. In this section we will be describing the general procedure to be applied in the latter case (Fig. 1), providing additional instructions for the single photograph case when necessary. For each stage, we will present the general procedure for application of the method with any combination of panorama-creation and GIS software, and then mention some specific details

Measurement of visual angles and magnitudes

Fig. 7 shows the results of the application of the methods to a panoramic view taken from a viewpoint near the village of Ardales. The panoramic image is made up of six photographs of 10 megapixels each, taken with a handheld Olympus E-510 camera and a Zuiko 14–42 mm lens. Within this panoramic view we have cropped an area of 140° × 60°. Various vertical control lines were defined in each photograph using the Hugin programme interface, so as to ensure that the yaw, pitch and roll parameters were

Discussion

The method we are presenting allows us to analyse the image obtained from a particular viewpoint in an objective way, regardless of the camera equipment used, by means of spherical projections and the establishment of a common system of reference between the ground plan and the panoramic photographs. The method enables the measurement of different visual parameters and introduces a new magnitude parameter: apparent density. The method can be applied systematically and inexpensively to various

Conclusions

Although the measurement of visual parameters of the landscape from photographs offers a wide array of possibilities for the objective evaluation of its scenic attributes, as of today there are no systematic and homogeneous procedures for performing this task. For this reason we propose a method for the objective analysis of the image and its components, by means of spherical projection of the photographs and the establishment of a common system of reference for the photographs and the

Acknowledgments

This research was possible thanks to funding provided by the Spanish University Teacher Training Programme (Programa de Formación de Profesorado Universitario – FPU, Ministry of Education) (FPU-AP-2009-4864), and by the University of Málaga (Programa de fortalecimiento de las capacidades de I + D + I en las Universidades 2014–2015). We would like to thank three anonymous reviewers for their inspiring comments on earlier versions of this paper, and Nigel Walkington for his assistance with the

References (56)

  • J.F. Palmer

    Effect size as a basis for evaluating the acceptability of scenic impacts: Ten wind energy projects from Maine, USA

    Landscape and Urban Planning

    (2015)
  • S. Pardo-García et al.

    A geospatial indicator for assessing urban panoramic views

    Computers, Environment and Urban Systems

    (2015)
  • H.W. Schroeder

    Visual impact of hillside development: Comparison of measurements derived from aerial and ground-level photographs

    Landscape and Urban Planning

    (1988)
  • H. Shang et al.

    Visual thresholds for detection, recognition and visual impact in landscape settings

    Journal of Environmental Psychology

    (2000)
  • E. Uuemaa et al.

    Trends in the use of landscape spatial metrics as landscape indicators: A review

    Ecological Indicators

    (2013)
  • S. Alpers

    The art of describing: Dutch art in the seventeenth century

    Simiolus Netherlands Quarterly for the History of Art

    (1984)
  • P. Arias Sierra

    Periferias y nueva ciudad: el problema del paisaje en los procesos de dispersión urbana

    (2003)
  • M. Brown et al.

    Recognising panoramas

  • Council of Europe

    European Landscape Convention

  • Countryside Commission

    A review of recent practices and research in landscape assessment

    (1988)
  • C. De Seta

    Ritratti di città: dal Rinascimento al secolo XVIII

    (2011)
  • M. Downes et al.

    What you see is not always what you get: A qualitative, comparative analysis of ex ante visualizations with ex post photography of landscape and architectural projects

    Landscape and Urban Planning

    (2014)
  • M.C. Dunn

    Landscape with photographs: Testing the preference approach to landscape evaluation

    Journal of Environmental Management

    (1976)
  • Environmental Resources Management Australia

    Ararat wind farm, Victoria: Landscape & visual assessment report

    (2009)
  • G.I. Evenden

    Cartographic projection procedures for the UNIX environment: A user's manual

    (1990)
  • T.G. Feeman

    Portraits of the earth: A mathematician looks at maps

    (2002)
  • H.J. Foley et al.

    Sensation and perception

    (2009)
  • L. García-Moruno et al.

    Analysis of lines and forms in buildings to rural landscape integration

    Spanish Journal of Agricultural Research

    (2010)
  • Cited by (14)

    • A geospatial model of nature-based recreation for urban planning: Case study of Paris, France

      2022, Land Use Policy
      Citation Excerpt :

      New data sources, such as street view images, unmanned aerial vehicle images, and Google Earth images, are making such assessments possible. These data can be applied to delineate and classify urban environments at high accuracy and large scale (Pardo-García and Mérida-Rodríguez, 2017). However, these approaches still constitute a research frontier, especially at larger scales.

    • Geospatial analysis of lake scenery as an indicator for the environment: The City of Van (Turkey) and its surroundings

      2021, Environmental and Sustainability Indicators
      Citation Excerpt :

      In this regard, Zhao et al. (2013) pointed out that most of the visual indicators for predicting landscape preferences are criticized due to lack of standardization in methodology, non-transparent application of values and lack of replicability. Enhanced data handling capacities of GIS and fine spatial resolution of contemporary digital surface models (i.e. DSMs) have enabled researchers to build virtual landscapes and analyse visibility of landscapes in a quantitative manner (Pardo-Garcia and Merida-Rodriguez 2017; Chmielewski et al., 2018). Digital analysis of visibility is particularly important for tourism planning and urban development since these are influenced by scenic attributes of landscapes.

    • Can computer game landscapes target new audiences for landscape quality assessment?

      2019, Applied Geography
      Citation Excerpt :

      Visual impressions of landscapes were originally recorded through narratives, sketches and paintings and this is still important in geographical research (Turkington, 2010). Site photographs represented a major step forward with respect to capturing the view and they remain important in visual landscape assessment (Arriaza, Canas-Ortega, Canas-Madueno, & Ruiz-Aviles, 2004; Pardo-Garcia & Merida-Rodriguez, 2017; (Swetnam & Tweed, 2018). However, the 3D nature of the landscape is still constrained by the 2D capacity of such recording devices.

    • Environmental heterogeneity as a bridge between ecosystem service and visual quality objectives in management, planning and design

      2017, Landscape and Urban Planning
      Citation Excerpt :

      Such efforts may be greatly assisted by cost-effective computer-based tools to test large numbers of scenarios, such as the HERCULES model for urban spatial heterogeneity (Cadenasso, Pickett, & Schwarz, 2007) or the RASCAL model simulating spatio-temporal complexityin shallow aquatic ecosystems (Larsen & Harvey, 2010). Such initiatives provide rich foundation not only for advancing the research on ecological importance of heterogeneity, but also for testing its compatibility with aesthetic criteria (Pardo-Garcia & Merida-Rodriguez, 2017). In summary, the diverse evidence from ecological and landscape aesthetic studies suggests that environmental heterogeneity provides a powerful basis for achieving visual quality together with other ES within both natural and human-dominated regions.

    View all citing articles on Scopus
    View full text