Drone Software for a Smart Agriculture Project

A worldwide recognized leader in semiconductor industry tasked Auriga to develop a drone software utility to monitor the condition of trees in an apple garden as a part of a big smart agriculture project.

Projects Highlights
  • Dataflow programming using TensorFlow.
  • SSD-mobilenet neural network deployment.
  • Instant data calculations on video card algorithms developed.
  • Manual and semi-automated objects identification and labelling for machine learning.
  • Augmentation for deep learning implemented.
  • Data formatting for neural network.
  • GUI developed and tested.
Achieved Benefits
  • Data calculations on video card were boosted up to 4x times.
  • Enhanced machine learning and deep learning algorithms, increased objects DB significantly.
  • Developed all set of tools to be used by end-user and simplified GUI of the application.