Object Recognition Demo

For humans, recognizing objects in their field of vision requires minimal effort. Understanding an object from various distances and angles also presents us little difficulty. Digital cameras allow us to emulate vision through a computer, but they do not have the natural ability of discerning objects from one another in their field of view. Using Object Recognition (OR) technology, we can help computers "see" more effectively by searching the area to identify objects in a real environment.

There are multiple ways to visualize objects for scanning. We have found an overlay to be a popular solution for our clients. First, the AR application will scan the environment to locate the item being scanned. Then, once it recognizes the target object's geometry, the application casts a digital overlay which can provide information specific to areas, components, or the whole piece of equipment. A second visualization method is through single component selection, where one specific object is being searched for in the area. Once the component is found, it will confirm that the item selected is correct, and proceed to the next component as shown in the image above.

The use cases for OR are vast. Use it for troubleshooting or identifying components of complex machinery, or to show the internal workings of a system with an X-Ray view utilizing real-time data. Even expedite equipment audits, where a technician can take images and audio recordings to document the status and condition of the equipment. This information can then be saved and stored for liability purposes and referenced at any time.

Before starting work on a project for a client, we performed a series of tests with early object recognition technology. We can confirm the technology has come a very long way since this video was captured nearly three years ago, however, this demo shows how a device’s camera can locate and identify specific shapes. We can’t wait to share how far this technology has come in the very near future!