City Engine

As an introduction: CityEngine® is a standalone software and a conceptual design and modelling tool for professional users in entertainment, architecture, urban planning, GIS and general 3D content production. It is primarily aimed at time-efficient creation of 3D cities and buildings. The main concept of CityEngine is the "procedural" approach towards modeling. The computer is given a code-based "procedure" which represents a series of commands - in this context geometric modeling commands - which then will be executed. Instead of the "classical" intervention of the user, who manually interacts with the model and models 3d geometries, the task is described "abstractly", in a rule file.

Coming across Bekerman's experiments with City Engine on his VenICE project, I was wondering whether Procedural Inc.'s approach for further development of their flagship product goes to city modelling or city planning. The example mentioned above is a modelling practice, while recent cooperation among Procedural Inc. and ESRI is a true evidence of their interest in stepping more deeply into the planning realm.

That is in fact the choice that many other practitioners such as Agency 9 have also made in recent years. Agency 9 AB which is a privately owned Stockholm-based visualization firm was founded in 2003. The firm started as a game design group and then shifted to production of online 3d maps and visualization tools. Their planning application, City Planner is a user-friendly and web-based tool for creating, sharing and communicating of future urban plans. The inputs to the program are 3d models of planned urban entities constructed in SketchUp, Maya or 3D Studio Max. Proposals are located in their surroundings and are made available to all stakeholders. This software provides the infrastructure required for adding geo-referenced feedback to a developing urban plan.The most dominant characteristic of CityPlanner is user-friendliness, simplicity and ease of communication through web. The tool has its focus on providing 3D visualization for planning procedures and does not include so many analytic functions or data-processing capabilities.

Sources: 

City Engine Help manual

http://www.youtube.com/watch?v=lJLvawS_5lA

Academia

In September, I had the opportunity to share outcomes of my recent studies with students of the master program "Sustainable Urban Planning and Design" in two occasions:

5th September, an overview of use of visualization in participatory planning was presented. This was in fact a revised version of my thesis work complemented by more background information, less focused on web-basedness and altered to better reflect the entire picture rather than my personal approach.

On the 14th of September, I gave a speech on 3d city models and 3d city modeling titled as: "An Overview of Three-Dimensional City Models; Historical Perspectives and a Contemporary Debate: Beauty vs. Brain". This was a summary on - among all - basic concepts in the field such as DEM(1), DSM(2); DTM(3), LOD(4), a history of creation of digital city models, different approaches, the vanishing borderline between digital maps and digital models and contemporary debates namely semantic city models and widespread use of CityGML.

In addition, an introductory lab session was held on use of the digital urban creation tool, City Engine, and implementation of procedural routines and grammar-based modeling. This was mainly focused on getting acquainted to the interface, experimenting basic functions of growing a street network from a scratch with help of shape files, creating urban lots, rendering those lots and streets into photo-realistic urban features based on the rules imposed by CGA files and customizing individual buildings by manipulating a set of parameters...

The above-mentioned were parts of the course AG2119 (Applied Analysis and International Planning Practice), year 2011, KTH.

1) Digital Elevation Model
2) Digital Surface Model
3) Digital Terrain Model
4) Level Of Detail

Villa Leipzig

Everything started with a simple three-dimensional model of the house we are living in to be located on Google Earth as a tribute to this lovely four-storied stately house. In fact, our landlord had previously lectured with a preacher's passion on the reputation of the edifice as one of the oldest buildings in the area. Nevertheless, it was just recently that I discovered that our house - which I had superficially called Örby House before - actually has a cordial historic name: Villa Leipzig!

Stockholm's Yellow Pages, 1922

According to Stockholm's Yellow Pages archive (above), the mansion dates back to latest 1922 and there was - among all - a bakery located in the house. Based on available photos, there has been a two-storied building called Villa Jennylund in the same location in 1910. This was consecutively replaced by current Leipzig House one decade later. It can be concluded that Villa Leipzig is at least 90 years old and has been almost in the same shape from the beginning.

Villa Jennylund, 1910

In 1945, there was a coffee and tobacco shop in the building which are now replaced by a residence and an office. Villa Leipzig undoubtedly deserved a 3d model on Google Earth! New features in Sketchup 8. have largely facilitated construction of new models by incorporating more accurate pictures from Google Map's Street View. Previously, Building Maker online application had been introduced by Google to construct digital models using enhanced oblique photos of the buildings from four different directions.

Villa Leipzig, 1945

Sources:
http://sv.wikipedia.org/wiki/%C3%96rby,_Stockholms_kommun
http://www.stockholmskallan.se/php/fupload/SMF/SD/SSMB_0023983_1922_12.pdf
http://www.hjle.se/

Google, Eniro, hitta and others

As part of an initial survey for a modelling project, I made a simple comparison over satellite photos of the terrain provided by prominent global actors such as google (Google Maps http://maps.google.com/) and two major Swedish providers, Eniro (http://kartor.eniro.se/) and Hitta (http://www.hitta.se/). Bing's images were nit detailed enough throughout this area.

Above you can see the result on a sample area. Hitta vividly offers the best quality while Eniro's images are of the lowest resolution. The objective here was to determine which source is more appropriate for us in modelling in cases where high-resolution aerial or satellite photos by e.g. Blom or C3 are not available.

Since there was also a need for oblique and facade images, I made an investigation into a wider range of online resources and checked which features were covered by each of them upon studied area. The features I chose for this purpose were 2-, 2.5- and 3-dimentional representations and panoramic views.

Again, Hitta proves to provide the best results. Specifically, the street view feature in Hitta (Gatubild) is distinguished from the other in the sense that it covers almost the entire city while that of Google Maps only has major roads represented and Eniro has no coverage in the area at all. Nevertheless, 2.5-dimentional views of Eniro are more useful than three dimensional map of Hitta for my purpose. The reason is that Hitta's 3d model is semi-automatically created by image sensering and thus does not fully recognize sides, edges and corners of the buildings, contains lots of distorted building volumes and buildings mixed with trees, whereas Eniro's oblique images (Utsikt) are precisely assembled outputs of high-resolution aerial photos.

In the following table, a summary of analysis is given.