Skip to content

First run

This walkthrough takes you from an empty app to a processed spectral cube. It uses the simulated instrument, so you don't need hardware. Before you start, make sure the backend is running.

1. Check the connection

Open QCLM. The status bar at the bottom shows Server latency in milliseconds when the app can reach the backend.

If it shows no latency:

  1. Click Customize in the left sidebar.
  2. Under Backend, click Test Connection.
  3. If the backend is on another machine, enter its Server URL, for example http://lab-server:8080/api/v1, and click OK.

See Troubleshooting if the test fails.

2. Create a run

  1. Click + New Run.
  2. Fill in Configure run:
    • Run name, for example First_Test
    • Specimen ID (optional)
    • System type: Infrared (IR) or Visible Light (Optical)
    • Objective lens and Storage location
  3. Optionally, pick a Run template: Standard Scan, Rapid Screening or Maintenance Run.
  4. Click Create Run ▶.

The backend records the run as an experiment, and the acquisition window opens.

3. Acquire a hyperspectral scan

The acquisition window has instrument controls on the left, the live image in the centre and the analysis tabs on the right.

  1. In the Instrument tab, under Acquisition, choose a scan mode:

    Tab Use it for
    Single One wavelength, for focusing and live viewing
    Hyper A full sweep: set Start, End and Step
    Discrete A hand-picked list of wavelengths (a sparse scan)
  2. Open Hyper and set a small range for a quick test, such as 10 steps.

  3. Click Record.

Frames appear in the viewer as the simulated detector captures them, and the Frames tab lists them. Hold Shift and move the mouse over the image to see the spectral profile at the cursor. Click Stop to end the scan early.

4. Build the spectral cube

When the scan finishes, the processing service assembles the frames into a spectral cube: one image per wavelength, stacked into a 3-D array.

  • If Auto-process pipeline is enabled in the Processing tab, the cube is built automatically.
  • Otherwise, use Cube Build in the Processing tab, or Rebuild Cube from selected in the Frames tab.

The Processing tab also holds the calibration options: reference correction, destriping, normalization, band cropping, downsampling and ROI bounds. Opening data and processing describes each one.

5. Look at the results

  • Bands: browse individual wavelength images and build an RGB composite from three bands.
  • Plotting: plot mean spectra over your ROIs.
  • Indices: compute spectral index maps such as RABD and band ratios.
  • Inference: run a classification model and view the class map.
  • Analysis: run MNF/PCA, Pure Pixel Index and end-member extraction.

6. Save your work

Open the I/O tab to:

  • export images (TIFF, PNG), plots and CSV data
  • use Export Project to save the run's parameters, so you can reopen the same setup later with Import Project

Next: the user guide covers each step in more depth.