Hi we are a small unit from Chennai having 3+ years of exp in Android programming. We have planned to move to Augmented Reality concept and we have decided to built the AR SDK from scratch! But we have no clue on where to start. We have already used KUDAN AND VUFORIA which are best at its sources but commercial. We know the principle behind the working concept of AR.
So what a SDK developer will plan before diving into coding? Kindly please dont close or downvote the question because as a 3+yrs experience we need some kind of more informations to achieve the task! Thanks for your support
First of all, you should know a bit more what this task require of you - you decided to do it from scratch without ANY effort estimation and cost estimation, because you do not know what is involved... This sounds like it might be a problem for you. I would strongly recommend making sure no AR open source SDK can assist you, since this is not a small project.
Now, if you insist... What I would suggest is to first set your goals - comparing to existing SDKs which you have examined, what do you want to achieve (performance-wise, memory-wise, API-wise, etc.).
Then, learn what feature-detection algorithms may be involved in order to achieve the required results, what type of system is required architecture-wise in order to meet the requirements. Design the architecture carefully, as the frame rate is the most critical issue usually on mobile AR apps. Also use the experience you got while trying to use other SDK-s - what was hard for you to use? What were you missing?
Once these are done, go and search for the technology to use - mainly, image processing library (such as OpenCV, and others).
Then, only then, you should start coding your SDK...
This is only the tip of the iceberg - you are up for a very big challenge.
Related
I have been looking for this for ever on the internet, but I can't seem to find it:
I want to create an augmented reality app for Android using Qualcom Vuforia and Eclipse.
Compiling and installing the Vuforia sample app worked, but I just don't get the logic of all behind it.
Do you know a thinner/ lighter tutorial for starting with Vuforia? (NO UNITY!!)
It is not very clear what exactly you are looking for - a tutorial for starting with it is one thing, and understanding the logic behind it is another thing...
Anyway, you can start here:
Getting started with Vuforia.
What exactly it is they are doing would not be revealed, of course, as it is not an open-source library, but you can learn some things from the Articles section. They have explanations on several topics, one of which could also assist you to better understand what's going on there:
Natural Features and Image Ratings
I want to develop an android augmented reality application, in which app should have a function to reconstruct a destructed objects(ex : buildings/statues) as shown in the following video link
https://www.youtube.com/watch?v=WOVjISxlhpU
I have gone through metaio, wikitude and vuforia sites each has some difficulties it self. and at Last i found vuforia has a feature call Smart Terrain where it is used for 3D animation and game development , the issue is only limited tutorials available to develop a customized application.
With that above link i found armedia.it and hyperspaces.inglobetechnologies.com those too have limited tutorials with the code.
Please let me know if any other SDK available their to fulfill my app feature share if there any useful tutorial to do such for the above sdk's
Thanks in advance
I do not think that there are any publicly available tools that can help you do this, except perhaps that armedia app. Reading through it, it seems like their approach is kind of laborious and fragile (align photos of user viewpoints with accurate 3D model). If you can't work through the tutorials these tools have, then posting here isn't going to get you what you want: SO is for asking specific technical questions (e.g., help fixing problems in code you have tried to build), not for general guidance and help.
FWIW, Vuforia's visual tracking will not work for something like this, I don't think, as is aimed at MUCH smaller scale things (for which you can build a target or object); typically things smaller than a person. Metaio is no longer available (Apple bought them).
Non-visual tracking (GPS + orientation) is not sufficient to attain this kind of tight registration.
I am looking to start writing apps for mobile devices. I know a little about this subjec, and I am proficient in both java and python. However, before jumping into creating apps I wanted to get the community's input on a question I have not yet found the answer to:
How/Where should I begin?
I understand this question is fairly general/basic for a community that focuses on solving/debugging complex programs, but if there is a recommended tutorial (or specific platform) that will help a beginner(such as myself), it would greatly reduce the frustration and amount of monotonous questions asked on this forum and others.
Regards
Coursera is offering 3 courses all using android development which started last week. That would be an excellent starting point for anyone:
https://www.coursera.org/specialization/mobilecloudcomputing/2
How/Where should I begin?
I don't know about you, but I often learn by example more than reading. You mentioned that you're proficient in java so thats a great start. After you understand the Android SDK it's really only the UI good practices you need to learn. So here's my suggestion.
Start off with something super simple. Let's say, create your own version of a calculator. Something that can add, subtract, multiply, and divide. This should be dirty, ugly, but works correctly.
After doing this, take a look at some other android calculators on the play store. Check out Simple Loan Calculator. I use this example because it offers a lot of android UI components. ViewPager, ActionBar, etc. Download this app, take a look at it, and try to mimic it. This should provide an idea on how to work with the android UI components.
What's next? After you have done these two apps you should be relatively ready to tackle your own project.
HTML apps are easy to port across multiple mobile platforms. I have set up an Android HTML template as a starting point. You can just import this project into Eclipse and take a look at that.
https://github.com/jakewp11/HTML5_Android_Template
Here is my experience for ios,
There are three things that I consider as pillars
Objective C
Memory managment
Design patterns
The first one looks obvious but I'd stress on learning concepts like protocols, categories, extensions. As a beginner I thought that learning the syntax was enough, but time and again I had to jump back to the language concepts to understand what's going.
The memory managment , the most talked and confusing subject and now since we have automatic referenc counting to make things simpler and often new Dev skips learning memory concepts (I did it). So I suggest you to have an idea what actually happens under the hood of arc " the manual memory managment"
The design patters,
When I started off , to pass data and control , what I used to do was achieve every thing through the only design patten I knew (target action) . I lack the knowledge of design patters didn't how easily I could achive better results with less code.
Some resources:
Dev.apple.com
http://www.raywenderlich.com (one of my best tutorials are here, you can find one on design patterns)
http://rypress.com/tutorials/objective-c/
And yes , the sample projects at Dev.apple are the best tutors.
I have seen several posts here about cocos2d-android, so ambition to get more idea on it drag me at coco2ds-android-1 and a good example .
My analysis can not find any significant benefit of using coco2ds instead of usual 2d approach of surfaceView and SurfaceHolder.Callback .
I will be thankful if anybody have expertise over coco2ds-android will guide me about benefits to use it instead of usual gaming approach .
Just by clicking on links starting from the ones in OP, I have came across http://dan.clarke.name/2011/04/how-to-make-a-simple-android-game-with-cocos2d/ - which states the obvious answer you are looking for.
First of all, this is a 2D gaming engine. All of the physics and whatnot effects are just there for you. No need to re-implement from scratch
Secondly, this is actually a port of the iPhone gaming library with the same name - great news if you also plan on porting to iPhone. And thirdly it is open sourced, meaning you can tweak anything accordingly.
I have noticed however that this is a pure java library, so do not expect amazing performance. If performance is critical, google for something NDK based, not SDK based. I could not advise here as gaming is not my thing.
Just curious. I am already a Java android guy but am interested in the new app inventor beta for android.
I was wondering what "real" developers are thinking. A lot of people want to develop an android app but don't really have programming skills. Seems like a lot of apps could be written with app inventor - not sure since I haven't tried it yet.
Hopefully this clears up some things for people that are still just hearing about App Inventor. (Keep in mind App Inventor is extremely beta right now, so all of this is subject to change.)
If you're interested in playing around with some of the features it offers and haven't had the time or interest to get involved with the full Android SDK, then App Inventor is probably a good fit for you.
The interface of App inventor offers some unique features you won't find in the Android SDK and is pretty easy to get the hang of (make sure to look at the shortcuts for the Blocks editor, or it might start to feel tedious real fast). It has live debugging and live edits to both the interface and the supporting logic. That alone cuts down on time and frustration if you aren't quite sure about the proper way to do something yet and is probably my favorite feature of the whole system. New developers will especially find it rewarding to actually see their changes implemented on their phone without waiting for things to compile and run. It is a huge time saver and much easier to get used to than the traditional compile and wait process you face if you are trying to learn things through the SDK.
It is also quite handy if you have been playing around with the Android SDK and simply want to try something new out that App Inventor supports that you haven't had the time or aren't quite sure how to approach yet. So long as you keep things focused on features that App Inventor supports, this can provide valuable insight into understanding a problem by getting your phone doing something you can see and use in a matter of minutes. That's something even seasoned developers are likely to appreciate.
The real problem is more when you start hitting walls. Walls that you simply can't climb. And you quickly realize that the limits of this tool must be understood or you will probably spend a lot of time trying to do something that just isn't currently possible. Sooner or later your going find yourself in this position if you use App Inventor very much.
All coding is graphical.
Look at some of the demos or screen shots of what's available through the Blocks editor if you don't understand what that means. Setting values in those big colorful blocks is as close to code as you will get with App Inventor, for better or worse.
If you have any previous development experience, you are going to find yourself longing for the ability to do some simple coding to supplement what you are doing. Or at least the ability to export your project and pick up where you left off in Java. Sorry, this isn't that kind of tool.
Your only options when you get to this point are to pick up and start all over in code or forget about it, at least until the desired feature gets added to App Inventor (which may be never). That right there likely to irk a good number of people and turn them off to it.
If you are looking at getting into Android development, I see no problem with starting out with App Inventor. It might even save you hours of effort that might otherwise be spent trying to wrap your head around some concepts that are simple to understand when you do them, but a bit confusing if you are building them for the first time.
If you are a seasoned developer, or looking to make money or gain Android experience beyond just the concepts, this isn't your end destination by a long shot. If you are a student that is new to programming, this is probably a good environment with sufficient challenges and room to play around in to keep your interest for a decent amount of projects.
Despite it's current limits, I still see App Inventor as valuable resource to learning new concepts and as a playground for trying out new ideas. It is proving to be an awesome tool for introducing someone to Android development, and would probably do as good a job as you could ask for when introducing someone to programming in general.
I have been playing with it for a bit and my thoughts are that is is great for:
An introduction to programming. It is not something people can pick up and develop great programs with straight away, it does still require a lot of thought and logical thinking, but it doesn't require people to learn the correct code syntax
Prototyping. As it is relatively quick and easy to develop with (and has live debugging on the Android device), it allows you to prototype and see/interact with it in real time
Quick coding. If you want to develop a program quickly, this might be the simplest tool
As the tool develops, it will allow more feature-rich applications (it already allows integration with other back-end programs, etc.). The screen design right now is very basic, but will be improved soon.
Overall, I think that if you know Java for Android, then you will be able to pick up Google App Inventor within an hour and start making some functional apps. You might find it more suitable for building some apps, or you might prefer to stick with your current tools.
For people looking to start programming, I think it is an excellent introduction. They can concentrate on logic and program design, instead of setting up their IDE, code structure and syntax. Plus they will see results as they code.
To answer your question 'Should I learn it'. I would suggest that you play with it for an evening - that is all it will take you to get a good feel of it.
It probably won't make you a better developer, but all knowledge is good, so it can't hurt :)
I plan to play around with it anyway.
The generic answer to any "Should I learn ..." question is usually "Yes, it can't hurt to know things." From a practical standpoint, I would skip it. These sorts of things are usually nice ideas that never catch on. Sure this might be a neat way for non-programmers to build very simple applications and perhaps a real programmer could build a compelling application with it but it looks like if you are a programmer the benefits to using App Inventor don't seem that compelling.
I haven't checked it out, but I think it might be worth a look to help in UI development. Most software engineers are good at really solving problems with software or coming up with cool ideas for apps, but when it comes to UI design we find a lot of obstacles. At least myself, I've always been used to working with UI Designers.
I honestly don't like tools that generate code, (except WSDL to java or java to WSDL) but pretty much anything else I find myself spending more time trying to trouble shoot an issue when it comes up than the time I would've spent writing the app from scratch...
Appinventor is so easy to learn, I would suggest go for both options.
Concentrate most of your time on your main Dev environment and learn AppInventor, in the background, to give yourself a break.
The advantage of Appinventor is you can put together reasonably complex apps
in a relatively short period of time.
It's also pretty forgiving, for example, if you have a numeric variable, you can treat it either as a string, or a numeric value, meaning you can put together
prototypes pretty easily and quickly.
If you want an app putting together, quickly, Appinventor might be the better
option, if you need something more powerful, more control over you UI etc, then
a more sophisticated Dev environment might be called for.
With Appinventor being so easy to use and learn, you're giving yourself more options.
Considering Eclipse seems to introduce errors I can't find and can't remove--even after using Project-Clean, I'm looking forward to any alternative.