MACHINE VISION

AREIAL VIEW - DRONE - PIPE

2019.01

With the very fast development of aircrafts, areial view becomes more and more available and useful, like in Maps & Geospatial, Military, and Moives & Art.. So our group focus on the areial view in machince vision, and decide to use drone as our toolkit. Pipes are the objects we want to shoot because they are everywhere in cities and they are linear and continuous.

Pipe Recognition - Workflow

This workflow shows how we use different technology toolkits(thermal camera, depth camera and roborealm) to get different informations(temperature, volume and color) to help us realize the recognition of different types of pipes. Moreover it shows how we apply these machine visions in people's daily life.

Recognition Logic

Application

1.city of pipes

The first application is based on a future 'pipe city'. We imagine in the future, all the transportation including pedestrian, autonomous vehicles, metros and parcels, will happen in pipes to get more speedy, direct, and safer. And the drone can recognize all of the diiferent pipes for different functions and go there directly to offer the service like maintaining and delivery.

Play this Video ▷

2. drone charge on infrastructures in streets

The second application of our workflow is more practical and more related to the use of drone. When we use drone as our toolkit, we find a problem that most drones in the market have -- very shot duration of flight because the limited weight and volume of the battery. So we imagine that every infrustructure in streets can help charge the drones. And the lampposts, traffic lights and electric poles which are kind of pipes can be recognized and used conviniently by drones. So we collect datas of these pipe-infrustructres in machine visions and make drone recognize them.

recognition logic:

Play this Video ▷