Algorithms Of Image Pretreatment Study Based On Coated Video Spectrometer

Multi-spectral image technique which combines the geometrical feature of image with the rich spectral characteristics has been widely used in many fields,such as hyperspectral remote sensing,biomedicine,artificial intelligence.There is dozens or even hundreds of the spectrum information in multispectral images.So the data and transmission rate of multispectral image is large and slow.In order to improve the transmission rate of multispectral image and reduce the demand space of data storage,this paper introduces a real-time transmission technology of multispectral image data,and studies the demosaicking algorithm for multispectral image based on the film-coating video spectrometer.The research is mainly to restore the mosaic multispectral image collected by the film-coating video spectrometer to completement multispectral image.The recovery process mainly includes: 1.Extract the single spectral segment image according to the mosaic template;2.Enhance the spatial information of all the single spectral segment images closed to the resolution of the detector;3.Reconstruct the spectral information of pixel missing,that each pixel has a complete spectral curve.This paper analyzes the Taylor series interpolation algorithm for two-dimensional image restoration and applies it to the demosaicking algorithm for multispectral image.Another improved Taylor series interpolation method is proposed,in which the first and second derivative of the image after wavelet transformed are used to guide the subsequent Taylor series interpolation to recover all single spectral images.In addition,the effect of spatial resolution enhancement and the accuracy of image element reconstruction spectra,as well as the factors that affect the quality of intact multispectral images,are analyzed in the course of the research,which ensures the reliability of the research method.