We conduct a range of work around robotic and computer vision that enables robots and autonomous systems to “see” so that they can work in challenging agricultural environments!
To do this we frequently do research into the use of deep learning techniques so that we can perform fast and accurate detection and segmentation. Some example areas of research include:
- instance-based semantic segmentation,
- spatio-temporal models,
- efficient models, and
- multi-modal fusion.
The data that we use include 2D (colour imagery) as well as depth data so that we can perform instance-based semantic segmentation. A recent example can be seen in the videos below which are examples “Crop Agnostic Monitoring Driven by Deep Learning”.
Below you can see some other examples of the data and imagery that we often deal with. If you’re looking for some of the data sets that we’ve captured and made available go here.