Adaptive Air Mission - Acitve Deep Learning applied to Insulator Inspection

  • 182 views

  • 1 comments

  • 3 favorites

  • hochgeladen 8. Februar 2022

Drone technologies offer new possibilities for complex flight applications as object inspection, area maintenance or agricultural support. Autonomous unmanned air-operations require high sophisticated methods for image processing and control. We develop methods for image based inspections, e.g. inspections of power lines. All inspection missions are implemented by autonomous flight control algorithms on our new adaptive research multicopter platform AREIOM. The AREIOM platform introduces several levels of control, provides high computation power and allows automated  interaction between flight-control and mission-control. 

In general the flight control is based on a predefined flight path. Therefore points of the flight path are defined by their GPS coordinates and the flight altitude is set. In addition a minimum distance to the object of interest and fundamental restrictions have to be considered.

The data of mission sensors and intelligent onboard data processing based on algorithms and powerful hardware are used for adaptive minimum distance control. In addition legal requirements for example No-Fly regions are included in the drone mission. So the flight path will be adapted based on this information.

Tags:
Bildungseinrichtungen: TU Chemnitz
Berechtigung: 1

1 Kommentare

Awesome work. From biomedical, agriculture and genetic engineering, deep learning is proving to be game changer everywhere.

10. Februar 2022 09:14:18 CET