Fig. 2: Localization of the microcavities.

a Schematics of the pushbroom hyperspectral imaging. b Principle of the spectral localization. c Intensity images acquired through a phantom (1.7l^*) at three wavelengths, corresponding to WGM peaks from different microcavities. The circles indicate the reconstructed locations. Scale bar, 50 μm. d Intensity map summed over all wavelengths equivalent to a regular image. Circles mark the reconstructed locations obtained as the average over all spectral peaks from one microcavity, and crosses mark the actual microcavity locations measured without the phantom. Scale bar, 50 μm. e Localization and size measurement of a large number (45) of microcavities below a phantom (1.1l^*). Circle sizes are proportional to the reconstructed sizes of the microcavities. (Inset) Area marked by the red square imaged without the phantom. Scale bar, 500 μm. f The depth calibration curve obtained by measurement of the light distribution width of the transmitted signal at the phantom surface as a function of the microcavity depth. Each data point is the mean of the distribution width measured on 4 microcavities. For each microcavity the width was measured in four different directions. The error bars represent the standard error of the mean. The simulated curve was obtained by Monte Carlo simulation. (Insets) Images of the corresponding light distributions. Scale bar, 100 μm. g 3D reconstruction of the locations for the microcavities dispersed inside the phantom. The color of the spheres indicates the microcavity size and the image at the top shows the fluorescence intensity distribution at the phantom surface.