光谱成像仪快速光谱定标方法研究

光谱定标是确定光谱仪器各通道中心波长的过程，为了获取光谱辐亮度，通常需要对光谱仪器进行辐射定标，将光谱仪器输出的数值，映射为物理量——辐亮度。不同的光谱仪器的光谱响应不同，因此还需要在光谱定标过程中确定各个通道的光谱响应。光谱成像仪可以看成是多个光谱仪组成的，需要对所有点的中心波长和光谱响应进行定标。自第一台成像光谱仪诞生以来，其定标方法逐渐固定，通常需要采用光谱分辨率较光谱成像仪更高的单色仪输出准单色光进行光谱定标，其准单色光的光谱带宽远小于光谱成像仪的光谱响应带宽，可以将准单色光抽象为脉冲函数。根据脉冲函数的特性，改变准单色光的波长，扫描光谱成像仪的响应波长范围，是对光谱响应函数进行间隔采样的过程，通过光谱定标数据可以直接得到光谱成像仪的中心波长和光谱响应函数。随着技术的发展，探测器的灵敏度越来越高，光谱成像仪的分辨率也越来越高，为了完成光谱定标，对光谱定标需要的准单色光提出了更高的要求。然而准单色光的带宽越窄，其能量越低，获取满足信噪比要求的数据需要更长的时间，使定标的效率降低。从光谱定标的目的出发，结合准单色光和光谱成像仪光谱响应近似高斯函数的特点，通过理论分析，提出一种利用宽...

Fast Spectral Calibration Method of Spectral Imager

Spectral calibration is determining the central wavelength of each channel of a spectrometer. To obtain the spectral radiance, it is usually necessary to calibrate the spectrometer and map the output value of the spectrometer to a physical quantity radiance. Different spectrometers, The spectral response is different, so it is necessary to determine the spectral response of each channel during the spectral calibration process. The spectral imager can be regarded as a composition of multiple spectrometers, and the center wavelength and spectral response of all points need to be calibrated. Since the birth of the first imaging spectrometer, its calibration method has been gradually fixed. A monochromator with the higher spectral resolution is required, and its spectral bandwidth is much smaller than the spectral response bandwidth of the spectral imager so that quasi-monochromatic light can be considered a pulse function. According to the characteristics of the pulse function, changing the wavelength of the quasi-monochromatic light and scanning the response wavelength range of the spectral imager is a process of sampling the spectral response function at intervals.. Therefore, the spectral imager’s central wavelength and spectral response function can be directly obtained from the spectral calibration data. With the development of technology, the sensitivity of the detector is getting higher and higher, and the resolution of the spectral imager is getting higher and higher. Higher requirements are put forward for the quasi-monochromatic light required for the spectrum calibration. However, the narrower the bandwidth of the quasi-monochromatic light, the lower its energy, and it takes longer to obtain data that meets the signal-to-noise ratio, which reduces the efficiency of calibration. In this paper, we combined the characteristics of quasi-monochromatic light’s spectral line type and spectral response function approximating to Gaussian function. Through theoretical analysis, a method of spectral calibration using wide-band quasi-monochromatic light is proposed, which can effectively reduce the calibration step of spectral calibration improves the efficiency of calibration and is suitable for the rapid calibration of spectral imagers. This method is used for the spectral calibration of a space-borne hyperspectral imager. The spectral imager uses a prism to split light and has the characteristics of non-linear dispersion. The spectral resolution varies from 2 to 18 nm, and there is a large curve of spectral lines. As a result, the center wavelength of each pixel is different, and spectral calibration is required for each pixel. To avoid the discontinuity of the central wavelength of the adjacent field of view caused by the calibration of the separate field of view, the quasi-monochromatic light spot emitted by the monochromator illuminates the entire slit, and a cylindrical lens and ground glass are placed between the slit and the monochromator. The cylindrical lens is used to converge the light perpendicular to the slit direction to improve the energy utilization; the ground glass is used to homogenize the light, and the presence of ground glass greatly reduces the energy entering the spectral imager. Combining the method proposed in this paper increases of the accuracy the bandwidth of monochromatic light, and the increase of energy have finally completed the rapid calibration of the spectral imager. The mercury lamp verifies that the spectral calibration accuracy is 0.23 nm.