Hugon-spectrometer

The automation of vicarious ground calibration would help increasing the frequency of data collection, improving the quality of the sensor monitoring.

The Remote Sensing Group has worked on LED-based radiometers to develop the me-thodology necessary to such an automation. Unfortunately, the LED radiometers have a very limited spectral resolution, providing very limited information on the spectral structure of the ground reflectance. It is important to retrieve sufficient spectral information about the ground reflectance to achieve good calibrations, especially in the visible and near infra-red were the reflectance of RSG test sites displays the most variability.

To solve this issue, efforts have been carried out to develop a low-cost sturdy spectro-meter.

This instrument is based on a very simple and common design. It collects light through a slit. A plano-convex lens collimates the light getting through the slit. The collimated beam goes through a high-dispersion prism which allows to disperse the spectrum. The light is eventually focused by another plano-convex lens after the prism. This forms a different image of the slit for each wavelength. A CMOS linear detectors array is placed in the focal plane to carry out the measurement. The pixel position can be related to wavelength and the flux collected by a pixel can be related to radiance, assuming the geometry of the instrument is well-known. This allows to perform a hyperspectral measurement of the spectral radiance reflected by the ground.

The instrument is piloted by an electronic board used for our Automated Sun Radiometers. The board contains a programmable micro-controller, allowing to generate and transmit the synchronization signals to the detector. The board drives also a digitizer that collects the video signal generated by the detector and converts it to 16 bits numbers. This electronic board can be wired to a computer through a serial port. A Labview interface has been created to retrieve the measurements.

The first prototype should span 435nm to 700nm with a spectral resolution between 3nm and 10nm. This prototype is still under fabrication and is expected to undergo its first field tests in April 2007.