Seeking a better view: Guiding cameras for optimal imaging via reinforcement learning

In Biosciences, image capture is often used in microscopy, lab and field environments as the first step in obtaining reliable scientific measurements of plants. Good image capture is fundamental to many of these experiments, by measuring plants you can determine which are healthier, more robust, or producing more food. However, you can't measure what you can't see, and, capturing a better image could be crucial in measuring the subtle differences which indicate a higher yielding crop, or a resistance to pests.

This project will explore reinforcement and machine learning approaches to train robotic systems to capture better images with no human interaction. It will evaluate varied reward systems to train AI that can provide guidance on which view is likely to be best given what has been seen so far. Working across disciplines, the PhD will training automated image capture systems on a variety of plant subjects ranging in size and shape, with a view to improving the performance on a range of tasks including 3D reconstruction, feature detection and counting. The robotics will include linear 1D and 2D actuators, as well as a 6-dof robotic manipulator.