From Scarlet Wiki
What are the AR development considerations?
Open Source or Proprietary
At the project outset, technical architecture and choice of software was considered integral to its level of success. As the project would not be developing any technical code or services, building on existing frameworks available as open source (e.g. AR browsers) it was imperative that the delivery solution would be intuitive,structurally sound and technical viable.
As a consequence of the cutting edge nature of Augmented Reality, a set of accepted standards has yet to be ratified by the W3C, although a working group discussing this subject has been formed to discuss a needs analysis. While most AR vendors are working on their own proprietary platforms, they are all relatively similar, so any future issues related to interoperability should be minimal. The majority use XML documents to hold the POI (or GLUE object) information with most nodes being generic (i.e. Longitude, Latitude, Name, Description, 3D model, Multimedia etc.).
It is hoped by the end of the project that the application of base level standards will be ubiquitous across most AR browsers enabling the outputs to be interoperable regardless of delivery choice.
Skills of Team
The production of media rich resources such as 3D or Video may be time intensive so it is worthwhile measuring achievable outputs to timeframes allocated by project managers. Many institutions have an AV department that can provide support and reference in terms of video and sound production. It is worth utilising their expertise if at all possible rather than learning complex new skills that are often time consuming and counterproductive when working with tight deadlines.
There are a number of tools primarily targeting non-technical users, assisting in the creation of AR location and visual recognition based channels. A few are listed below:
- Metaio Creator - Specific AR software to allow users to create a complete AR scenario and deploy it to Junaio in less than 5 minutes.
- Birdview - Birdsview is provided by www.birdsview.de and helps to create location-based POIs which are published in the Birdsview Channel in Junaio.
- Hoppala Augmentation - provides an easy-to-use graphical web interface to create location-based junaio channels, via a map interface, with just a few mouse clicks. Images and 3D models can be uploaded and kept in a personalized inventory. Hoppala Augmentation then publishes your channel to all major mobile Augmented Reality browsers.
- Layar Creator - Layar Creator allows a user to add digital content to static print media.
- Aurasma Studio - The Aurasma Studio is a web environment where partners can create content, apply for skinned versions of the Aurasma app or download the Aurasma kernel to embed into an existing app.
What are the AR tool considerations?
- There are a number of AR mobile specific tools that assist in the compatibility and delivery of secondary resources via an AR channel. The SCARLET project utilised the Junaio 3D Model Encrypter, taking existing 3D content and preparing it for use in Junaio.
- Another tool used to create tracking images from image files is available at http://www.junaio.com/develop/tools/
- Many AR browsers offer the option of hosting content on their servers. For the SCARLET project we took the decision to host on internal servers which allowed easier customisation. A server running Apache and the latest version of PHP was required to enable the SCARLET outputs to work effectively. It is worth having initial discussions with technical administrators so they can advise on server setup and back-end technologies required for AR outputs.
- If, as part of content creation you wish to include video, production and editing software is necessary. There is a range of software available ranging from basic low cost options such as iMovie or Windows Movie maker to advanced tools including special effects such as Adobe Premiere, Final Cut Pro or Cinema 4D.
- Free 3D modelling software is available for an entry level user, these include Blender and SketchUp (Pro version is available to educational institutions.)
What are the AR delivery considerations?
- To maximise student benefit, tablet devices were seen to “offer increased potential for AR applications since they feature larger screens (typically 7-10” compared to the 3-4” screens offered by most smartphones)“.
- The launch of the Apple iPad2 with rear facing camera meant that augmented user experiences could be enhanced immeasurably, particularly if these devices could be available in designated study areas for both casual and directed use. AR software could then be preloaded and bundled user support made available.
- It is important to maximise user adoption by providing access via different handset/tablet OS if possible. Presenting the AR experience on a cross platform App such as Junaio or Layar allows greater inclusivity than developing for a single platform.
- Many AR browsers require smartphones/tablets with fast processors and sufficiently high memory to operate effectively.
- MP3 players running Apple iOS such as later generation iPods can be used to deliver AR, provided they are connected to WiFi networks.
- Using AR apps and downloading media can be battery intensive so it is important that devices are charged on a regular basis
- Due to the environmental constraints of the technology being used inside, the traditional format of augmented delivery – POI’s (Points of Interest) mapped to GPS co-ordinates was problematic. Mobile devices struggle to detect accurate location-based data with their inbuilt GPS, or in some cases will not work at all due to compass interference.
- Remember to optimize content depending on the level of network or WiFi coverage available. Users become frustrated when content is not immediately displayed so carefully consider the types of media used.
- When creating image recognition based AR, it is important to consider environmental conditions such as lighting. Although many visual search algorithms are now very accurate, reflective surfaces, changeable lighting and shadows can have a detrimental impact on their effectiveness. Avoid reference images that are flat and blurry, too dark/bright. Instead use images with little symmetry and focus on high contrast areas that the device camera can uniquely identify.