I am a very newbie to Augmented Reality software. I want to design a simple app. As a part of this app, There will be a series of uniquely designed tags. These tags will be on some assets. In the application, I want to store some metadata for each asset. Imagine a DB table with fields like :(asset Id, name, var1, var2...) holding the asset meta-data.
So, when The augmented reality app detects a unique image then it will show its meta-data information, over the marker. It is that simple. In summary, I want to know how can I use image markers to differentiate assets? Sorry If I am asking a very basic question.
Regards,
Ferda
First of all your question is too broad. How are you planning to implement this application? First, you have to decide whether you will be using an Augmented Reality SDK or computer vision techniques?
My suggestions would be based on amount of devices you want to use this application or target platform, choosing 1 SDK from ARCore, Vuforia or ARKit. I am not familiar with ARKit but in ARCore and Vuforia, either augmented images or image targets are hold in an image database. So you can get the image id or name of any target you detected using your device. In conclusion you can visualize specific assets for specific images.
Below you can see ARCore Augmented Image database. As you can see every image has a name. In your code you can differentiate images using image.Name, then visualize corresponding meta data over the marker.
Also in both SDKs, you can define your own database but your images should not have repetitive features and have high contrast sections.
Vuforia has a similar concept as well. The difference between ARCore and Vuforia depends on what which devices you target and quality of image tracking. Vuforia can detect images better in my opinion. It is really fast in terms of detecting images.
Related
I'm using vuforia scanner to detect and recognize a 3D object. It works well with one object but now i want to recognize multiple objects at once. I have gone through so many links but they only speak of multiple image targets and not 3d objects. If not using vuforia, is there any other sdk to do so?
I messed with object recognition once but I'm pretty sure the databases are basically the "same" as 2D image target databases. That is, you can tell Vuforia to load more than one of them and they'll run simultaneously. I don't have Vuforia installed at the moment, but I know the setting is in the main script attached to the camera (you have to fiddle with it when creating your project in the first place to get it to use something other than the sample targets).
There is, however, the limit on how many different targets Vuforia will recognize at once (IIRC is something really small, like 2 or 3). So be aware of this when planning your project.
Im trying to create a AR Game in Unity for educational project.
I want to create something like pokemon go: when the camera open the object will be fixed somewhere on the real world and you will have to search for it with the camera.
My problem is that ARCore and vuforia groundDetection (I dont want to use targets) are only limited for few types of phone and i tried to use kudan sdk but it didnt work.
Any one can give me a tool or a tutorial on how to do this? I just need ideas or someone to tell me where to start?
Thanks in advance.
The reason why plane detection is limited to only some phones at this time is partially because older/less powerful phones cannot handle the required computing power.
If you want to make an app that has the largest reach, Vuforia is probably the way to go. Personally, I am not a fan of Vuforia, and I would suggest you use ARCore (and/or ARKit for iOS).
Since this is an educational tool and not a game, are you sure Unity is the way to go? I am sure you may be able to do it in Unity, but choosing the right platform for a project is important - just keep that in mind. You could make a native app instead.
If you want to work with ARCore and Unity (which is a great choice in general), here is the first in a series of tutorials that can get you started as a total beginner.
Let me know if you have other questions :)
You can use GPS data from phone to display object when the user arrived specific place you can show the object. You can search GPS based Augmented Reality on google. You can check this video : https://www.youtube.com/watch?v=X6djed8e4n0
I'm already make an augmented reality app that can read images marker, but I wonder if i can make augmented reality without marker in unity.
Can anyone tell me how?
Maybe What you need is SLAM, I think. Simultaneous Localization And Mapping, and is markerless, just recognite and track the environment.
These are 2 videos about slam:
https://www.youtube.com/watch?v=HbaEw5-YvA0
https://www.youtube.com/watch?v=_YLzcWX-gWU
One is from kudan, one is from wikitude. If this feature is what you want. Then I am sure that what you need is SLAM.
You can get more from:
wikitude slam
kudan
And, vuforia is not adviced. In fact, Smart Terrain does recoginate the environment, it still need a marker.
You can use Google ARCore, Vuforia or 8thWall SDKs. All of them have motion tracking or extended tracking. So you do not have to use markers. You can take a look at ARCore HelloAR example:
https://developers.google.com/ar/develop/unity/tutorials/hello-ar-sample
you can using ARUnity. ARUnity is the Unity plugin for ARToolKit. (Well the marker would be some kind of image which is used for tracking. ARToolKit calls it NFT.)
You can download it here:
http://www.artoolkit.org/download-artoolkit-sdk
(Scroll down for the ARUnity download link)
Documentation is available here:
http://www.artoolkit.org/documentation/doku.php?id=6_Unity:unity_getting_started
Best is, it is free and open-source :).
I am thinking of building a camera application - with the ability to do image processing (adjust contrast, apply different image filters) while you are taking picture or after the pictures has taken.
The app will also have the ability of drag and drop icons.
At the end you are able to export the edited images either to the camera roll or app memory.
There is already many apps out there like this. (Line Camera) etc...
Just wondering what is the best way to build such app.
Can I build the app purely with Objective C ios sdk? or do i need to build it with C++/cocos2d, etc...
Thanks for your help!
Your question is very broad, so here is a broad answer...
Accessing the camera/photo library
First you'll need to access the camera using UIImagePickerController to either take a new photo or grab one from your photo library. You can read up on how to accomplish this here: Camera Programming Topics for iOS
Image Manipulation
AviarySDK has much of this already built for you. Very easy to set up and use in your apps. You can download their sample app for free in the app store if you want to see what it can do. Check it out here: http://aviary.com/
Alternatively, read up on Core Image if you'd like to avoid third-party libraries. See Core Image Programming Guide for more information.
There is absolutely no need for cocos2d which is a game engine.
You can accomplish everything you mentioned using only Objective-C.
If you want real-time effects you will need to dive into OpenGL. you can use GLKit if you target iOS 5 and above.
I need some help for finding the person with Camera.
is it possible to fetch person from the groups if he has install applicator in that device from the camera?
the basic concept is when user rotate camera to any person camera will fetch image of that user and will search it and compare it with application database for if that user is register in application or not and display his information.
if its possible and you have some reference then please share with me.
Thank you for read question.
Core Image has a new CIFaceFeature to recognize faces in real time; you can start with these examples to take an overview:
SquareCam (from Apple)
iOS Facial Recognition
Easy Face detection with Core Image
Then you have to design the logic to compare and store images.