Translation of the Sample Slide
Coverage of the sample area is achieved by translating the sample slide in small increments along one direction with the motorized stage. This motion brings adjacent areas of the sample to the cross-hair position where a spectral image is recorded. Each recording covers a narrow line which is parallel to the horizontal cross-hair, and contains the spectrum of every point along the line. By repeating this motion, the spectra from a two-dimensional area of the sample are acquired. This process is known as “pushbroom hyperspectral scanning”, and allows both the sample image and the spectrum from every point in the image to be captured as a hyperspectral data cube. The dimensions of the cube include the X and Y dimensions of the spatial area and the spectral data associated with each pixel of the image.
Seen here (at left), in a view from the eyepiece, the microscope is focused on a group of cells that have been incubated with nanoparticles. Some of the particles are out of focus (large circles) and others appear focused inside the cell. The hyperspectral datacube is scanned from the area inside the red box. Motion of the slide causes the scan lines to move from the top to the bottom of the box. The blue and green scan lines shown here (marked 1 and 2 along the Y direction) intersect two nanoparticles which were inside the cell.
At each line the system takes a digital picture that contains the spectra of all points in the X direction. Here (at right) are spectral pictures taken at the blue and green lines. Spectra of the bright nanoparticle and fainter cell structure are visible. These two pictures belong to the full set of up to several hundred lines that will form the hyperspectral data cube.
Once the scan is completed, the CytoViva-ENVI software opens the datacube (seen at left). The datacube is an RGB (red, green, blue) recreation based on the spectral intensity measured at each pixel. At right, we are viewing plots of the spectra from the two nanoparticles, having clicked on pixels that overlapped their locations. Spectra of all points in the image are likewise found. These spectra can be exported for further analysis or saved as spectral libraries.