Plants - Books and Academic Papers

Recent Developments

• Real-time Realistic Rendering and Lighting of Forests, 2012
  • Eric Bruneton and Fabrice Neyret, Proceedings of Eurographics 2012
  • Good-looking realtime implementation of "sun and sky dome contributions, shadows between trees, inside trees, on the ground, and view-light masking correlations", without popping or aliasing.
  • Based on two new forest representations, z-fields and shader-maps, with a seamless transition between them.
  • 
  • Implemented in the open-source terrain library Proland.
• Texture-Lobes for Tree-Modelling, 2011
  • Livny, Pirk, Cheng, Yan, Deussen, Cohen-Or, Chen (China/Germany/Israel), ACM Trans. Graph. 2011
  • Abstracts the tree's foliage into canonical "lobe-texture" geometry, which can be stored and rendered very cheaply.
  • Can be used to reconstruct  laser scanned trees (from point clouds) or re-representing existing tree models (e.g. polygonal, from Xfrog)
  • 
  • It looks great, and the storage savings claimed are very attractive (e.g. 56MB Xfrog model vs. 65KB lobe model).  However, a complicated approach like this is more implementation than theory, so a public implementation would be needed to make it actually useful; the website says "application: available soon", so we can hope.
• Exploring 3D visualisation of vegetation, 2010
  • Joël Hempenius, Univ. Utrecht, thesis.
  • Gives an overview of the many issues and approaches for modeling and distributing plants on terrain: L-systems, AMAP, XFrog, GIS distribution, cellular automata, interpolation, etc.
  • A case study dataset was processed for the Ameland area in the Netherlands, using PlantGL and XFrog to make plant models and the VTP software (VTBuilder/Enviro) to visualize the results, with much evaluation and conclusions.
  • 
• Image-Error-Based Level of Detail for Landscape Visualization, 2010
  • Malte Clasen, Steffen Prohaska, VMV
  • LOD approach for fast realtime plant rendering, using a hierarchy of unconnected primitives (ellipsoids, lines)
  • Implemented in their Biosphere 3D software
    
• Tree detection from aerial imagery, 2009
  • Lin Yang (U. Alabama) & 3 others from Google, ACM GIS ’09
  • A pixel-level classifier is trained to assign a {tree, non-tree} label to each pixel in an aerial image, then refined by a partitioning algorithm to a clean image segmentation of tree and non-tree regions. Then, they use greedy selection to locate individual tree crowns.
  • This work was done at Google, so it probably relates to Google Earth's "3D tree" coverage, which was publicly released a year later (2010)
  • 
• NDVI-based Vegetation Rendering (PDF PPT)
  • Stefan Roettger, In Proc. CGIM '07, IASTED Press, February 2007
  • Describes using NDVI (Normalized Difference Vegetation Index) as a guide to distributing representative vegetation, derived from LandSat Channels 3 (red) and 4 (near infrared).  NDVI values are mapped to plant height, from low values for grass to higher values indicating tree size.
  • Example dataset has 10 million trees, reduced to 100k trees with culling, rendered at 30 fps.
• Rendering Grass Terrains in Real-Time with Dynamic Lighting
  • sketch and paper by Boulanger, Pattanaik, and Bouatouch, Siggraph 2006
  • uses per-pixel lighting, grass density maps, and three levels of detail: geometry, vertical, and horizontal slices, with smooth transitions between them
  • claims worst-case 20 fps, common-case 80 fps or better on a NVidia 7800 GTX
  • when highly anti-aliased, the results are beautiful
  • 
• Rendering Forest Scenes in Real-Time
  • Philippe Decaudin and Fabrice Neyret, Eurographics Symposium on Rendering, 2004
  • uses texture-slicing to produce dense forests corresponding to continuous non-repetitive fields made of thousands of trees with full parallax
  • the forest consists of a set of edge-compatible prisms containing forest samples which are aperiodically mapped onto the ground, with special treatment of silhouettes on ridges
  • relies on LODs and on a GPU-friendly structure to achieve real-time performance
  • there are more details on Decaudin's site Antisphere: Rendering Forest Scenes in Real-Time
  • follow-on work: GPU-Based Lighting and Shadowing of Complex Natural Scenes
• Interactive Visualization of Complex Plant Ecosystems
  • O. Deussen, C. Colditz, M. Stamminger, G. Drettakis, IEEE Visualization 2002, October 2002
  • describes an approach for drawing plant LODs using primitives (points, lines)
  • points are used mostly for leaves on trees, and flower petals
  • lines are used for the blades of grasses
  • octrees are used to efficiently choose which LOD  to use, and display lists are used for fast rendering
  • they can do massive scenes (>120 million triangles full detail) at several fps on a GeForce 3
• Generating spatial distributions for multilevel models of plant communities (pdf)
  • Brendan Lane and Dr. P, 2002 proceedings of Graphics Interface
  • describes two methods of distributing plants
    1. using an L-system-like approach (!) for simulating plant competition
    2. using a probability field which is deformed by each plant iteratively
• Modelling of ecosystems as a data source for real-time terrain rendering (pdf)
  • (Abstract) (Screenshots)
  • Johan Hammes, Proceedings of DEM2001 (Digital Earth Moving, First International Symposium)
  • presents a way in which well-known techniques for modeling ecosystems can be applied to generate the placement of plants on a terrain automatically at run-time
  • sufficiently fast to allow real-time computation, but also varied enough to allow for natural looking placement of plants and ecosystems while remaining deterministic

Papers of Bedřich Beneš et al.

• active in plant research 1996-2004
• A Stable Modeling of Large Plant Ecosystems, 2003
  • shows the importance of using environmental feedback (such as depletion of the soil) in addition to normal competition between plants
• A Real-Time Visual Models of Plants Interacting with Virtual Environment, and
• Virtual Climbing Plants Competing for Space, 2002
  • give a brief summary of work in progress which actually simulates both light and collision detection affecting plant growth

Research Projects

• LIAMA GreenLab
  • A French/Chinese academic cooperation, working on a stochastic, functional and interactive models for plant growth and architecture.
  •  INRIA is involved, so it may build on previous INRIA work (below).  Philippe de Reffye, one of the AMAP developers, is a guest researcher at LIAMA and contributed to the development of GreenLab.
  • GreenLab functional structural plant model + Scilab open source software for scientific computation → GreenScilab
  •  The libraries of GreenLab will also contain more Asian species which are still rare in 3D.
• Lenné3D
  • Was a years-long interdisciplinary project focusing on visualization of vegetation and GIS data, with a significant component having to do with plants  (see the Lenné3D entry on Government / Academic Research Projects for more info). In 2005, it spun off as a commercial services provider, Lenné3D GmbH
• Virtual Crops - Architecture and Process Oriented Modeling and Visualization of Crop Stands
  • a German project (2000-2005?) of several universities to support research into 3D modeling, simulation and visualization of crop plants
• Algorithmic Botany
  • site of the Biological Modeling and Visualization (BMV) research group at the University of Calgary.
  • Home of Dr. P (Przemyslaw Prusinkiewicz)
• Virtual Plants at INRIA
  • A multi-year academic project (in Montpellier, France) to analyze plant growth and structure, from a very low level, such as meristem functioning and development.
  • OpenAlea, an open-source platform provides a user-friendly environment for researchers to build plant models using a visual programming interface and provides a set of tools and models dedicated to plant modeling.  V-Plants is a set of packages in OpenAlea.
  • PlantGL, an open-source C++ graphic toolkit for the creation, simulation and analysis of 3D virtual plants embedded in the Python language.
  • There doesn't appear to be any realistic models of full plants, as the focus is more on abstract structure and biological growth at the microscopic level.

Books

• Trends in Real-Time Landscape Visualization and Participation, Proceedings at Anhalt University of Applied Sciences 2005
  • includes several Plant papers
• Computer generated plants (German/English)
  • a website for the book Digital design of Nature and related materials and software, including its German counterpart previously published:
• Computergenerierte Pflanzen. Techniken und Design digitaler Pflanzenwelten
  ("Computer-generated plants. Techniques and Design of Digital Plant Worlds"), 2002 [at amazon.com]
  • by Oliver Deussen, an overview of all CG plant and ecosystem efforts over the years
• The Algorithmic Beauty of Plants, 1990 [at bn.com] [online as a PDF]
  • Dr. Prusinkiewicz and Lindenmayer
  • a very complete introduction to L-systems and their applicability to plant modelling.
• Advances in Computational Life Sciences, Volume 1, 1997 [at amazon.com]
  • Plants to Ecosystem: Modelling of Biological Structures and Processes

Other Articles

• Digital Botany in Journal of Landscape Architecture (JoLA) 2, 2006 (pdf)
  •  A cover-story and visual essay by Jörg Rekittke and Philip Paar

Papers of Dr. P

• see BMV Publications for all the important plant L-systems papers

Other Important Papers

• Papers and Publications Related to Xfrog
• Plant Models Faithful to Botanical Structure and Development
  • P. de Reffye, C. Edelin, J. Francon, M. Jaeger and C. Puech, Siggraph '88
• Creation and Rendering of Realistic Trees
  • Jason Weber & Joseph Penn, Siggraph '95 [images]
• Interactive Modelling of Branching Structures
  • Lintermann, B., Deussen, O., Siggraph '96
• Realistic Modeling and Rendering of Plant Ecosystems [abstract] [pdf]
  • Deussen O., Hanrahan P., Lintermann B., Mech R., Pharr M., Prusinkiewicz P., Siggraph '98
• A Modelling Method And User Interface For Creating Plants [pdf]
  • overview of the motivation and implementation of the XFrog package, 1998