This work studies the suitability of hyperspectral imaging microscopy for rapid and accurate atomic layer mapping of two-dimensional (2D) materials. A hyperspectral imaging system including a line-scan hyperspectral imaging microscope, system control, data acquisition, and data processing was custom built. A manual interpretation method based on hyperspectral library and spectral unmixing was developed for layer maps reconstruction with one-atomic-layer resolution. A deep fusion neural network based on the U-Net architecture was proposed for fully automated large-area atomic layer mapping of 2D materials.
«
This work studies the suitability of hyperspectral imaging microscopy for rapid and accurate atomic layer mapping of two-dimensional (2D) materials. A hyperspectral imaging system including a line-scan hyperspectral imaging microscope, system control, data acquisition, and data processing was custom built. A manual interpretation method based on hyperspectral library and spectral unmixing was developed for layer maps reconstruction with one-atomic-layer resolution. A deep fusion neural network b...
»