|| spectrometer || VIMS introduction || the 2 channels || VIMS-V || VIMS-IR || operations || performances ||
Instrument operation

VIMS-V and IR work in unison to provide a data set that appears as if it were made by a single device. Therefore in most of the possible VIMS operating modes, the operations described here are blind to the user. However it will be importantin some situations to understand the underlying operation of the instrument in order to properly command it. Due to their different detector configurations, recise synchronization of their mirror motion and data collection is critical.

Because the Visible Channel uses an area array Charge-Coupled Device (CCD) detector it acquires its data in "push-broom" mode; i.e. it views one row of a square scene at a time. This row is imaged as one row of pixels on the CCD, and the spectrometer spectrally disperses the image of this row so that each row of the CCD views the image in a different waveband contiguous with its neighboring rows. To acquire a square image, the CCD is read out after each row in the scene is acquired and the mirror moves to the next row. Each data readout from the CCD (which represents one row of the scene as seen in several wavelengths) is called "slice" and forms one layer in a data cube. When the instrument has completed recording a complete square scene, its image can be seen by viewing the face of the cube.

The IR detector, or Focal Plane Assembly (FPA), uses a linear array detector so it acquires its data in "whiskbroom" mode, where it views only a single spatial pixel per exposure. the spectrometer disperses the image of this pixel on the FPA so that each detector element views the pixel in a different contiguous waveband. To provide synchronous data with the VIMS-V, VIMS-IR must sweep its single pixel field of view along the identical row in the scene that the VIMS-V is observing within the same exposure time. Each readout of the IR channel FPA (which represents one pixel of the scene as seen in several wavelenghts) is called a spectrum.

During the imaging of one row in the scene, the Visible Channel remains stationary while the IR channel moves its mirror in 64 steps across the same row. The timing is such that the IR channel images the same row in the scene during the same time that the VIS channel is recording it.
The data slices from the two channels are combined (end to end) to form a single slice for the whole instrument. To create a two-dimensional image, the two channels begin at the top of the desired scene and acquire data row by row. This requires perfect synchronization and geometric alignement.

The pixel summing and image scanning process is further complicated because the two channels have different detector sizes. Each channel must synthesize a square nominal system pixel (0.5 x 0.5 mrad) by summing more than one exposure of each detector. The IR FPA has rectangular detectors, so during the time it is imaging a single square IFOV it is continuously integrating while its mirror moves over two of its actually rectangular IFOVs.

VIMS-V pixels are 0.167 mrad square, so a 3 x 3 grid of VCOH pixels must be summed to equal one nominal system pixel. To sum 3 IFOVs in one dimension the VCOH steps its mirror twice while continuously integrating during a single exposure time. This occurs simultaneously over a whole row in the scene, since it acquires a row at a time. When the CCD data is transferred to its horizontal register, the VCE commands the CCD to sum every three pixels in the other dimension as it is read out. This results in an equivalent 0.5 mrad square IFOV pixels. Five pixels are also summed in the spectral dimension to achieve the specified nominal visible spectral bandwidth of 7.3 nm.