Applications of object detection networks at high-power laser systems and experiments

• DOI: 10.48550/arxiv.2210.02539
• Publication Date: 2022-01-01
• Authors: Jinpu Lin; Florian Haberstroh; S. Karsch; A. Döpp
• Keywords: Plasma Physics (physics.plasm-ph); Accelerator Physics (physics.acc-ph); Physics - Data Analysis, Statistics and Probability; 0103 physical sciences; FOS: Physical sciences; Physics - Accelerator Physics; 01 natural sciences; Physics - Plasma Physics; Data Analysis, Statistics and Probability (physics.data-an); Physics - Optics; Optics (physics.optics)

The recent advent of deep artificial neural networks has resulted in a dramatic increase performance for object classification and detection. While pre-trained with everyday objects, we find that state-of-the-art detection architecture can very efficiently be fine-tuned to work on variety tasks high-power laser laboratory. In this manuscript, three exemplary applications are presented. We show the plasma waves laser-plasma accelerator detected located optical shadowgrams. wavelength density estimated accordingly. Furthermore, present all peaks an electron energy spectrum accelerated beam, beam charge each peak is Lastly, demonstrate damage system. reliability detector demonstrated over one thousand shots application. Our study shows suitable assist online offline experiment analysis, even small training sets. believe presented methodology adaptable yet robust, encourage further facilities regarding control diagnostic tools, especially those involving image data.