In this talk we present our approach to automatic detection of critical failure states in pulsed Petawatt laser systems, used for investigations of exotic states of matter and medical applications. The beam shape is controlled to avoid high destructive energy densities. However, randomly occurring states threatening the device must be detected between pulses and trigger an interlock in the device firing at 10Hz.
Our automation approach, presented here, uses deep learning via the Caffe framework. The states we are aiming to detect are rare; thus, training data for this category is scarce. We address this by identifying regions of interest based on physical properties of the system.
Required audience experience: No specific knowledge required.
Objective of the talk: The audience will learn about feature classification using deep neural networks with Caffe and OpenCV. Other topics are the adjustment of the training process to a small number of interesting samples in the training set, training for the detection of rare events via region-of-interest identification as well as noisy and poor contrast intensity profiles.
Keywords: image classification, Caffe, smart laser operation, OpenCV
You can view Jeffrey’s slides via the link below:
You can watch Jeffrey’s presentation below: