hi everybody,
i just posted a new video demonstrating the implementation of some light scenarios into an augmented reality application. it was built on base of the shadow demo project from unity that can be downloaded here. for performance and usability reasons i removed the interaction possibilities with the model.
the model has about 6000 polygons and 8 light maps with 1024x2048px each. on my xperia x10 the app reaches about 17 fps while on a samsung galaxy s it's about 50 fps.
for any further questions or suggestions please leave a comment or send me a message.
Saturday, March 26, 2011
Wednesday, March 16, 2011
augmented reality to control your home
have a look at this very interesting project of a colleague. the idea of controlling your home with a mobile device like a mobile phone is not new but using augmented reality to do that is a great approach. although there might be some more steps to take to have such a system for everyday use there could be a lot of potential.
let me know what you think about it.
let me know what you think about it.
Sunday, March 13, 2011
fraunhofer multi touch table - 2d meets 3d
this is one of the projects i've seen at cebit. it's not really new but still could be an interesting way to discuss a project inside an office or with clients. still in the current implementation i see two main problems.
first it should be able to generate the necessary sections and plans from a model made with for example autodesk revit. all the necessary informations are already included in such a model and the extra step of programing the interactions just doesn't fit to make quick modifications to the model.
the second issue is related to the camera navigation. in my eyes it's extremely important to give the viewer also the possibility to look up and down and not just to rotate horizontally. this could maybe be done with another technology i've seen at the fair. this technology adds the possibility of measuring not only 2 coordinates of a finger but also the distance to the screen making a third dimension of input possible. that's just an idea but we will see what future brings.
first it should be able to generate the necessary sections and plans from a model made with for example autodesk revit. all the necessary informations are already included in such a model and the extra step of programing the interactions just doesn't fit to make quick modifications to the model.
the second issue is related to the camera navigation. in my eyes it's extremely important to give the viewer also the possibility to look up and down and not just to rotate horizontally. this could maybe be done with another technology i've seen at the fair. this technology adds the possibility of measuring not only 2 coordinates of a finger but also the distance to the screen making a third dimension of input possible. that's just an idea but we will see what future brings.
Thursday, March 3, 2011
CeBIT - off topic
hi guys,
as i'm visiting CeBIT in hannover at the moment, i decided to make some off-topic posts on interesting stuff i've seen there within the next few days. more infos will follow soon!
as i'm visiting CeBIT in hannover at the moment, i decided to make some off-topic posts on interesting stuff i've seen there within the next few days. more infos will follow soon!
Monday, February 21, 2011
indoor navigation
i just saw this video of a software called junaio allowing indoor navigation based on a combination of image recognition (marker based) and orientation. i'll try to get some more information, but for the moment have a look at the dome video from the mobile world congress in barcelona.
Tuesday, February 1, 2011
current methods of augmented reality - geo tags vs. trackables (part2)
hi,
following my previous post this time i will write about the second widely used method to generate augmented reality. in this case this is done using so called trackables. this is also the method i used in my sample videos [1, 2] so far.
the general idea behind is the following. you take a 2d image with a lot of contrast points and preferably sharp edges between the elements on the picture. you can take basically any picture but the results might be not so good if you don't follow these rules. next you have to establish a spatial correlation between this trackable and the 3d model you want to show in your augmented reality application.
as i'm no mathematician or computer scientist i can't provide you with in depth information on how the tracking is exactly done but i will try to update my link collection as soon as i find some time. just about the basic principle. the application with the predefined trackable, the 3d model and the spatial relation uses the camera to get information about the surrounding. within the camera view the app uses algorithms to search for patterns of contrast that can also be found on the trackable. this is done in order to generate the correct 2d representation of the 3d model and to place it at the right position of the screen. as this happens each frame the viewer gets the impression of real time motion according to the motion of the device around the trackable.
to get a better impression how it works just have a look at my previously mentioned videos. as i'm still working a lot on that topic be sure that further clips will follow.
current methods of augmented reality (part1)
the general idea behind is the following. you take a 2d image with a lot of contrast points and preferably sharp edges between the elements on the picture. you can take basically any picture but the results might be not so good if you don't follow these rules. next you have to establish a spatial correlation between this trackable and the 3d model you want to show in your augmented reality application.
as i'm no mathematician or computer scientist i can't provide you with in depth information on how the tracking is exactly done but i will try to update my link collection as soon as i find some time. just about the basic principle. the application with the predefined trackable, the 3d model and the spatial relation uses the camera to get information about the surrounding. within the camera view the app uses algorithms to search for patterns of contrast that can also be found on the trackable. this is done in order to generate the correct 2d representation of the 3d model and to place it at the right position of the screen. as this happens each frame the viewer gets the impression of real time motion according to the motion of the device around the trackable.
to get a better impression how it works just have a look at my previously mentioned videos. as i'm still working a lot on that topic be sure that further clips will follow.
current methods of augmented reality (part1)
Subscribe to:
Posts (Atom)