I want to understand the various methods used for 3D city modelling and their advantage, disadvantage.
Dear Anuj Tiwari,
I would say that terrestrial LiDAR would be the best answer to your question. Hovever 3d-Laser scanners are quite expensive and in many case not affordable for Master thesis or small projects.
Another cheaper solution could be using 3d-reconstruction based on optical devices and stereo pairs like for instance those used in the following software: Agisoft (http://www.agisoft.ru/); Image 3D (http://images-3d.software.informer.com/); StereoPhoto Maker (http://stereo.jpn.org/eng/stphmkr/)...etc.
More details on how to make 3D Images are available online (e.g. http://www.3ddigitalphoto.com/phantograms.pdf).
Best regards
Hello Maurizio, thanks .....
i have gone through some literature and found single satellite image based, panormic image based, video based solution for 3D modeling. thats great if you can provide some details about the use of these solution with implimentation condition ].. thanks in advance...
Hello tiwari,
please go through the attached paper,
towards automatic validation and healing of citygml models for
geometric and semantic consistency
@Asmat Ali: Thanks for your reply Sir, this is really an important point of concern.. We are using Autodesk 3D, Ravit softwares with high end processing hardware... but software like CityEngine are also pioneer in real time 3D modeling with photogrammetry and grammer oriented 3D modeling.
There are several methods to create 3D city models. It depends on the Level of Detail you would like to obtain. There are semi-automatic methods and manual methods as well as different data source. You may find this article interesting: http://www.int-arch-photogramm-remote-sens-spatial-inf-sci.net/XL-1-W1/213/2013/isprsarchives-XL-1-W1-213-2013.pdf
Ideally, you need airborne and ground lidar as well as airbore photogrammetric image as well as close-range images in the form of mobile mapping. This data set can be relaxed to a smaller set but at the expenses of more manual data collection. On a different vein, you can use mobile mapping with planimetric maps that can obtained from a GIS. In general, you need sources of 3D information for the sides of the buildings, which can be complemented with sources of 2D information to provide the top of the buildings.
For small areas you can use Unmanned Aerial Vehicles (UAV) to capture top view as well as oblique images.
You can get very nice 3D models from oblique aerial images and image matching. Acute3D does that. The problem with this technique is that the model you get is "dumb". It is only a triangulated mesh of unclassified points.
Points from lidar can be classified easily and so it is easier to get semantically rich 3D models with classified objects: buildings with walls, roof, trees... If you have the vectors of building footprints you can get fully authomatic LoD2 models . If you don't have footprints you should do some editing to get acceptable 3D models. You can do that with Terrasolid software. The final building model is a boundary representation of roof planes and walls. You can attach textures form oblique images to walls and roofs.
You can get building models from point clouds from images and/or lidar also with CityGML. In this case, the buildings are selected from a library of basic building templates.
In general, point clouds from lidar have lower density than point clouds from image matching. The accuracy of elevation and slope is usually better with lidar but to place accurately the roof edges you need images.
The best 3D model depends on the application.
I know this response comes late, but another possibility is to use X3D geometric primitives (e.g. boxes) and then apply textures on top. X3D is an international Web-based standard. In the past we've done some X3D city modeling you can see at:
http://projects.felixlup.net/s3d/S3D.x3db
You have to use an X3D player like Bitcontact Management X3D plugin or InstantReality player.
About oblique models, there is some GIS software that have solutions for integration of 3d skin models and regular feature data. Unfortunately is propietary software, look at www.supermap.com there are some ideas.
In different circumstances, according to the project criteria and the purpose of our project, it is possible that the methods of the lidar, drone or aerial images are suitable for the production of the model.
- Badges
- Science topic
- Similar topics
- Geoscience
- Geoinformatics
- Geographic Information System