Greenhouses are getting more popular and the existing tend to become modernized. This includes climate control, and crop monitoring which currently is done in most cases with the use of multiple sensors and actuators, but they still remain very labour intensive. To optimally control all these complex aspects there is a need for more advanced automation.
Drones can add significant value to greenhouses since they can automate many tasks in the production process. For example, they can cover large areas in short periods of time while collecting plant phenotypes. The limited operational space is the main and most impactful constrain. For flying autonomously drones should primarily be able to know its exact position in space, hence localization and mapping are essential. Finally, collision avoidance and path planning are necessary mechanisms for the drone to enable the drone navigation in this environment.
The aim of this project is to automatize the UAV data acquisition, processing and analysis workflow. This will be achieved within the main steps:
- Review of the state-of-the art
- Implement state-of-art algorithms
- Practical deployments on embedded platforms and integration with drone hardware
- Writing report
- Taketomi, T., Uchiyama, H., & Ikeda, S. (2017). Visual SLAM algorithms: a survey from 2010 to 2016. IPSJ Transactions on Computer Vision and Applications, 9(1).
- Cvišić, I., Ćesić, J., Marković, I., & Petrović, I. (2018). SOFT-SLAM: Computationally efficient stereo visual simultaneous localization and mapping for autonomous unmanned aerial vehicles. Journal of Field Robotics, 35(4), 578–595.
- Attada, V., & Katta, S. (2019). A Methodology for Automatic Detection and Classification of Pests using Optimized SVM in Greenhouse Crops. International Journal of Engineering and Advanced Technology, 8(6), 1485–1491.
- Enthusiast about aerial robots
- Willing to see a pilotless drone
- Excited to learn more about SLAM
Theme(s): Sensing & measuring