At the momement we are limited in tracking 2D target images. Therefore recognizing the cannister as a whole won't be possible. However if there is an easily identifyable label on the canister, the user might scan this instead of the whole cannister and our SDK would be able to recognize it. Wikitude's Titanium Module works perfectly for such a usecase.
As next steps I suggest creating target images of the cannister's labels and quickly try it out in our SDK. This will give you a good impression on how well suited the cannister's label is for recognition.
over 5 years ago
I have a client that sells gas canisters for barbeques. They would like to create an app that allows a user to take a photo of a cannister to get details on it, ensuring that they purchase a correct type when their current one expires. Would this type of functionality work with the image recognition capabilites of Wikitude's Titanium Module? I'm assuming that it would from what I've read to-date; the cannisters in question come in two easily identifiable colours (blue and red) and an assortment of different sizes?