Decoding method for the spectral mixing pixels in Hadamard transform spectral imager

Jun Xu,Bingliang Hu,Dazheng Feng,Xiaohui Fan,Qingming Qian
DOI: https://doi.org/10.3969/j.issn.1007-2276.2012.06.024
2012-01-01
Abstract:A sort of engineering Hadamard transform spectral imager based on DMD was introduced. Because the size of individual micromirrors did not match the pixel size of the detector, or for other unavoidable errors in the optical design and the system assembling, a sort of spectral mixing phenomenon appeared on some pixels of the encoded images acquired from the detector. Their encoding principle was described and analyzed. The analysis demonstrated that these spectral mixing pixels could not be decoded by inverse Hadamard transform directly. To decode these spectral mixing pixels accurately and improve the quality of the decoded spectral images, a decoding method for these spectral mixing pixels of the encoded images was proposed. Firstly, a decoding coefficient for the spectral mixing pixel was gotten by directing a beam of laser into the spectral imager to fill the field of view. Then the coefficient was brought into the encoding equation of the spectral mixing pixel to decode all the spectral elements. The experimental results prove the feasibility of the decoding method.
What problem does this paper attempt to address?