| Internship
                work for Harris E.S.S., 1998RealSite Viewing
                application annotationsI was requested
                to add annotations and other objects to the
                viewing application for the RealSite project.
                RealSite extrudes urban scenes from multiple air
                and or satellite images. Once ground geometry
                (buildings) are realized into three dimensions
                the images used to acquire the geometry are
                applied as textures. The result is a model which
                closely represents the urban scene.  My work was more
                or less very general, a great deal of freedom was
                permitted in what I was to create. Here I will
                display some samples, but first let me introduce
                you to the viewer. Below is a full screen shot of
                the viewing application with just about
                everything 'turned on'. There are no
                'annotations' in this scene, however, the rose
                compass and bottom menu are my work. 
 At the top of the
                screen is a heading indicator, to the right a map
                inset, and in the lower left corner a needle
                compass. These were implemented before my
                arrival. The view is of a prototype geometry
                model (not actually created by RealSite) and
                shows Harris buildings.  A rose compass is
                located in the upper left, which is the last
                object I created. The heading of the user is
                indicated by highlighting the petals with red
                (currently a South East heading). An attitude
                ball displays pitch and roll (indicating where
                the horizon should be). A small black stud
                indicates altitude. This object took ~16 hours to
                implement using Performer (in which all my work was
                done).  A menubar is on
                the bottom of the screen, it is a prototype for a
                metadata console. Certain types of information
                may not be displayed graphically in the world
                model, when that data seems relevant the data
                console indicates data availability by flashing
                LEDs. The user may then browse through available
                data and optionally display, save, &tc. (This
                is a prototype, only the appearance was
                implemented) Below arrows and
                a volume container are displayed. The view is of
                the Harris wickham site, where I worked. The
                large dark blob in the center is a lake =) 
 The five 2D
                arrows in the scene (Two yellow and black, white
                and black, cyan, and pink) are samples of a wide
                range of arrows which dynamically rotate about
                their directional axis to maintain maximum
                visibility for the user. Some arrows have color
                animation for motion and blinking. There is also
                an arrow which simply receives a texture with
                alpha channel so that any drawn arrow may be
                used.  In the center is
                a transparent pyramid, with label. It, and it's
                many variations, are intended to enclose large
                volumes of space. Here a small one is used to
                mark an atrium which might have been overlooked
                otherwise. The label on top (unimaginatively the
                label reads 'PIN_PYRAMID') hovers around the
                volume and uses many techniques to maintain high
                visibility, including circular translation,
                rotation, and scaling with attention to display
                resolution. Variations also include narrow
                formations which mark a single position well.  Below is the
                entrance to one of the wickham buildings.  
 To help clarify
                scale to the user, silhouettes of a man, car, and
                trees are provided. They have a 'billboarding
                action' which allows them to rotate for
                visibility, and the man 'walks' a simple path.
                Any texture may easily be used for a silhouette.  To the right is a
                sign capable of holding any amount of text.
                Normally it is lowered so that only the large
                label is visible. A heuristic determines when the
                sign should raise its self to be visible, or
                perhaps the user may simply click the label to
                raise the sign (not implemented).  
 Above are four
                simple icons intended to represent
                metatdata(text, video, photo, geo/spatial) in the
                scene. They have many deforming behaviors which
                attempt to make them look 'happy' (that's what
                the man asked for..). These include a wobbling,
                bouncing, and spinning approach. The icons
                position relative to the user determines action
                type and amount. For example, when far away the
                icons spin but when approached they rotate to
                face the user.  These objects,
                and additional variations, were
                described in a final report.  Thanks go to
                Harris, and Mr. Ellery Chan, the man(boss). |