According to a recent publication by the United States Patent and Trademark Office, Apple has just signed a series of patents on an Augmented Reality (AR) hardware device with enhanced computer vision capabilities, as well as software that has a new method for superimposing virtual information in a real environment.
Said solutions as a whole, could give us a new vision of the idea that Apple has in the world of AR and how the future of the company is projected in this potential technological branch.
Currently, the main handicap for the massive development of this type of technology is the great consumption of resources that it needs. However, the new patent of the Californian company goes further. An efficient environment monitoring method is detailed that provides information on detected objects.
Apple notes that the process has some phases that consume a greater amount of resources than others. For this reason, it discriminates between situations where there is a large power consumption (when downloading information from the environment, for example) and situations where this consumption may be lower (such as when the system is in environment scan mode).
Thus, Apple divides the process into the following three phases:
Feature detection: location in the environment of the objects with which we can interact.
Features description: definition of each.
Feature mapping: location in AR space
Instead of relying on existing computational display methods, Apple proposes a new streamlined technique that implements dedicated hardware and pre-learned data.
To make up for deficiencies in current computer vision systems, Apple proposes the use of geometry models, depth sensors, positioning data and other advanced technologies not commonly seen in today's AR devices.
For all this, Apple needs a system of two cameras to create the depth map with which to mesh all these ideas. This step has already been taken with the inclusion of the second camera in the new iPhone 7 Plus, so it is now possible to meet that requirement. With these patents, this depth map would be used to generate the geometric models on which the virtual data is superimposed.
There is still a lot of data to know about this technology. However, growing patent evidence on AR-related hardware and software suggests that the Cupertino-based company is putting great efforts in this regard.
