| The Spectral-sensing Internet of Things (SSIOT) is a complementarily technological integration of near-infrared (NIR) spectroscopic analysis and traditional Internet of Things (IOT). A new-type spectroscopic sensor, monolithically integrated multichannel spectral sensor (MCSS), plays a significant role in this network. It is widely acknowledged that optical calibration is an important technical basis to employ the MCSS in practice, which ensures the measurement accuracy. Related work was studied including wavelength calibration, spectral responsivity calibration and impacts of temperature shifts. According to the relative spectral response and to principles of wavelength allocation, the wavelength of the MCSS was calibrated, neighboring pixels combined as one calibrated channel. Experimental results demonstrate good smoothness of measured spectra and wavelength calibration accuracy of 2.1 nm, and that the performance is as good as a commercial spectrometer. Apart from that, a low-cost method was used to calibrate the spectral responsivity from the relative value to the absolute one. Results demonstrate that the responsivity increases with the ordinal number of pixels growing. For example, the responsivity of NO.50 pixel is 6.47×108V/W, while that of NO.102pixel is 1.36×109V/W. For the variable-temperature spectral performance, it is also demonstrated by experiment results that within a temperature range of 50K (from 263.15K to 313.15K), the peak wavelength of each pixel slightly and linearly redshifts with the temperature rising, with the maximum wavelength deviation of less than 2.12 nm. Warming enables the responsivity to rise, whose variation is not linear and varies with different pixels. For instance, a temperature range of 50K enables the peak responsivity of NO.50 pixel and NO.98 one to grow by 18% and 30% respectively.38% is the maximum increment of the responsivity. The study on temperature shifts could pave the way for calibrating the MCSS out door in the near future. |
PDF Full Text Request