Dynamic and Mobile GIS: Investigating Changes in Space and Time - Chapter 8

Part III Display and Visualisation These three chapters forming ‘Part III – Display and Visualisation’ of ‘Dynamic and Mobile GIS: Investigating Changes in Space and Time’ examine the challenges of visualisation within the constraints of mobile GIS, particularly the small-screen of a typical PDA, but also the limitations of using only GPS as a positioning system. It is noted that in all three chapters it is the visualisation of transportation networks that is under consideration. In Chapter 8, Malisa Plesa and William Cartwright remind us that the 240 x 320 pixel resolution and 16-bit colour display of the typical PDA restrict the appearance of a displayed map. So, although inexperienced users express a preference for ‘realistic’ visualisation, photorealism, for example, is very difficult to achieve within the PDA environment. A series of tests, including a navigation task, were given to ten subjects. These tests required representing a downtown area realistically (using oblique photography) and non-realistically and determining their usability. The tests showed that the non-realistic map was more effective. For example, the slight tonal variations required to display photography could not be shown on the PDA, limiting its utility for the display of realistic imagery. The non-realistic map was able to utilise a series of colours that achieved the maximum contrast and legibility available from this form of map delivery. As well as the results obtained showing that non-photorealistic graphics are more effective than photorealistic graphics in the context of 3D mobile cartography, non-realistic representations are easier to develop than realistic scenes, so are more suited to delivery on mobile devices. Plesa and Cartwright’s test, although involving transport networks, was for pedestrians; the representation of buildings and other neighbourhood features was as important as that for the road network. The extra rapid decision making required for vehicle navigation, compared to pedestrians, will almost entirely limit the display to the road network. Suchith Anand, Mark Ware and George Taylor, in Chapter 9, consider the generalisation of transport networks specifically for Mobile GIS with PDA display. Transportation networks are considered ideal candidates for using ‘schematisation’ to ease the interpretation of information by generalisation. Schematic maps are diagrammatic representations based on linear abstractions of networks. The authors of Chapter 9 have looked at the generalisation techniques that can be applied to generate schematic maps from large-scale digital geographic data, and among their conclusions is that, to be effective, a route map must show all turning points on a road. Route finding using mobile GIS not only requires an adequate representation of the transport network, but also adequate positioning. Reliable and accurate © 2007 by Taylor & Francis Group, LLC 140 Dynamic and Mobile GIS: Investigating Changes in Space and Time determination of the current user position is considered a prerequisite not only for car navigation (automotive) applications but also for a wide variety of mobile GIS applications. The algorithm, described by Britta Hummel in Chapter 10, is not specifically for car navigation applications and so can be integrated into any mobile GIS requiring positioning. But the challenge, in navigation along a mapped network, is not just to say where you are on the Earth’s surface (which GPS can do adequately), but where you are in the map. Hummel presents a map matching method that exclusively relies upon information from a standard GPS receiver. But, the method exploits the vehicle position and orientation history, information about road network topology, driving restrictions and the assumed driving direction for each road element, to allow correct positioning on the network. This part of the book, although ostensibly addressing visualisation in the dynamic and mobile GIS environment in an unrestricted manner, has concentrated on the visualisation of the transportation network and positioning within it. The reader might feel that there are other visualisation challenges for mobile GIS, and indeed there are, but given the way mobile GIS will be applied (i.e. in dynamic circumstances), the visualisation of the transport network provides a real visualisation challenge, and is a necessary first step to whatever follows. © 2007 by Taylor & Francis Group, LLC Chapter 8 An Evaluation of the Effectiveness of Non-Realistic 3D Graphics for City Maps on Small-Screen Devices Malisa Ana Plesa and William Cartwright RMIT University, Australia 8.1 Introduction When we think of navigating through a city we need to determine where we are, where we wish to get to, and mentally calculate how to get there. We use landmarks, signs and other relevant built and natural environmental cues to assist our wayfinding. But we generally rely upon maps to provide an overview of the space we wish to move through and as a tool to assist navigation. Once, these were paper products, designed, delivered and consumed under the ‘umbrella’ of a print mindset. But, now, paper maps are just one of the choices in the plethora of locational and wayfinding tools available. We are offered maps on mobile telephones, Personal Digital Assistants (PDAs), on television monitors, in kiosks and even as illuminated billboards. However, in many instances the way in which the maps are designed for use with contemporary information delivery tools, and especially those that deliver geographical information via small-screens, is no different to the designs for their paper map cousins. ‘Just’ applying the ‘rules’ of paper maps, or maps on large-screen computer graphics, cannot ensure a usable mapping product for navigation and wayfinding. A different approach needs to be explored. The use of Non-Photorealistic Rendering (NPR) for 3D cartography displays on mobile devices was explored. A research ‘gap’ was identified, and the following questions were found to be unanswered: 1. Are non-photorealistic computer graphics more effective than photorealistic graphics for the delivery of three-dimensional, spatial information on mobile devices? 2. What is the potential of non-photorealistic computer graphics, combined with three-dimensional cartography, for the representation of spatial information on mobile devices? One way to address these questions is to create and test a prototype built on Döllner and Walther’s (2003) theory. This prototype consisted of two 3D maps: one non-photorealistic map that utilised similar design strategies as those outlined by Döllner and Walther (2003), and one aerial photograph that provided an example of a ____________________________________________________________________________________ Dynamic and Mobile GIS: Investigating Changes in Space and Time. Edited by Jane Drummond, Roland Billen, Elsa João and David Forrest. © 2006 Taylor & Francis © 2007 by Taylor & Francis Group, LLC 142 Dynamic and Mobile GIS: Investigating Changes in Space and Time photorealistic map. They were created as static views and the extent of each map focused on the same urban area, situated just south of the Melbourne CBD, Australia. The prototype was delivered on a Personal Digital Assistant (PDA) and tested in situ with a representative target population. This chapter provides a background to the application of non-realistic computer graphics to mobile devices and describes the results from a research project formulated to find an answer to the above questions. 8.2 Photorealism vs. non-photorealism Since the introduction of computer graphics, the ultimate goal has been to achieve photographic realism (Schumann et al., 1996; Durand, 2002; Gooch and Gooch, 2001). Computer graphics has subsequently realised a need for the generation of abstract imagery, which has led to a rapidly growing interest in NPR (Markosian et al., 1997; Goldstein, 1999; Gooch and Gooch, 2001). Non-Photorealistic Rendering is focused on emphasising the most relevant components of an image, while suppressing unimportant details. This helps to overcome the expense related to the creation of realistic imagery, while still providing users with the same information. This can be seen as something akin to the bird’s-eye views shown in some early map-like artefacts that were produced to show cityscapes. Since photorealism has become achievable, little research has been undertaken into alternative methods of information display, and the benefits of photorealism to user understanding currently remain unclear (Schumann et al., 1996; Markosian et al., 1997; Ferwerda, 2003; Gooch and Gooch, 2001). Non-photorealism is predominantly focused on user understanding. It effectively communicates subtle information while also highlighting important details. Research suggests that the human mind is able to complete abstract information, and that realism does not directly influence human image interpretation (Duke et al., 2003). Non-photorealism represents a form of functional realism, whereby knowledge relating to the properties of objects are provided to the user, allowing them to make reliable visual judgements (Gooch and Willemsen, 2002). Photorealism provides the same visual response or stimulation as the original scene. In contrast, functional realism provides users with the same visual information, allowing them to use imagery to help them complete real-world tasks (Ferwerda, 2003). 8.3 An NPR technique for mobile city models Döllner and Walther (2003) presented a non-photorealistic rendering technique for 3D city models. Principles derived from cartography, cognition and non-photorealism form the basis of this technique, which may hold a number of advantages over photorealistic representations. Less graphical and geometric detail is required for the construction of these models in order to produce favourable results. The potential of non-photorealistic models for small-screen delivery was © 2007 by Taylor & Francis Group, LLC 8. Evaluation of Effectiveness of Non-Realistic 3D Graphics on Small-Screen Devices 143 outlined, and it was found that these models had a number of theoretical advantages over photorealism in this context. Most studies concerning 3D map design for mobile devices focus on the development and use of photorealistic imagery and do not attempt to explore alternative methods of display. It is argued that more focus needs to be directed at determining the most effective technique to display this information on handheld devices, as current research does not appear to address this issue. Contemporary cartography has followed a similar path to that of computer graphics, but as computer graphics realises the need for NPR in some applications, 3D cartography is still primarily interested in realism. Döllner and Walther (2003) have identified a possible need for non-photorealism in city models and have devised an NPR technique. While Döllner and Walther’s (2003) technique is theoretically feasible, there was an interest to test its effectiveness in a real-world situation. This research involved the development of a prototype that simulated their technique. The prototype was then used to compare the technique with a realistic representation before testing in the field with real users. By doing this, it was hoped that a clearer understanding relating to effective 3D map design for small-screen devices could be established. 8.4 Developing the prototypes A prototype was built to demonstrate how Döllner and Walther’s (2003) expressive rendering of city models could be applied to a map for display on a small-screen, and to compare this with the use of a photorealistic map. Expressive rendering involved manipulating an initial image produced with a CAD package so that it begins to resemble a hand-drawn sketch. Lines defining the edges of buildings are made to change thickness along their length, colours are chosen from basic hues and the general impression of the image appears more uneven that a typical CAD output. Expressive rendering has as its aim to produce simple images that impart a more general image of a building, rather than a hard-edged precise CAD drawing. This prototype contained a 2.5D map of an urban area, encompassing both pan and zoom functionality. It was designed to be delivered on a small-screen device for evaluation. The development of the prototype involved four stages: 1. Data collection; 2. Base map production; 3. Map design; and 4. Preparation for delivery. The map coverage of the prototype included the southern part of the city of Melbourne’s CBD and the eastern portion of the Southgate Arts and Leisure Precinct. This area was selected because of its popularity as an entertainment destination for both tourists and Melbournians. The area features a variety of land cover types, including parkland, varying building densities and forms, different © 2007 by Taylor & Francis Group, LLC ... - tailieumienphi.vn