ADVANCED IMAGE MANAGEMENT AND CLOUD DETECTION ON-BOARD AERIAL VEHICLES
Reference number | |
Coordinator | SPACEMETRIC AB |
Funding from Vinnova | SEK 472 500 |
Project duration | June 2020 - June 2021 |
Status | Completed |
Venture | Innovair/SMF |
Important results from the project
In summary, Spacemetric is now able, using its Keystone software, to: - Simulate UAV and aerial images with Nvidia Jetson Tx2; - Perform image compression with simulated UAV imagery (in GPU) with Nvidia Jetson Tx2 and Nvidia Jetson NANO; and - Manage streaming data (video) using Nvidia Jetson Tx2. These functionalities are central for modern image data handling for UAVs, and SAAB perceive the project results as very interesting. Spacemetric perceives that the project´s goal, which i.a. was to increase the competitiveness of the company and Sweden, are well met.
Expected long term effects
In summary, Spacemetric are now able to, using its Keystone software: - Simulate UAV and aerial images with Nvidia Jetson Tx2; - Perform image compression (in GPU) with simulated UAV imagery using Nvidia Jetson Tx2 and Nvidia Jetson NANO; and - Use Nvidia Jetson Tx2 for streaming data (video). The expected effect is a significant strengthening of Spacemetric´s, and thus Sweden´s, competitiveness within this field. This is borne out by recent contacts with customers / partners. Spacemetric expect to turn the new capabilities into business possibilities.
Approach and implementation
After finalising the project´s detailed objectives, a brief method study was executed. Work on mini-computers (Nvidia Jetson Tx2 and Nvidia Jetson NANO, single-circuit computers that can be used on board UAVs, ensued. A prototype environment was set up with the mini-computers, software adapted for these computers was developed, and experiments with simulated UAV data begun. A GPU environment was used for image data compression, whilst image simulation and streaming video management were also main themes. The work was concluded with an online demo seminar with SAAB.