Evaluation Criteria For Computational Spectral Imaging Quality

Computational spectral imaging integrates calculations into the spectral imaging process to achieve the purpose of improving the signal-to-noise ratio, speeding up imaging, and reducing the size of the spectrometer. At present, the commonly used evaluation methods for calculating the spectrum are mostly directly borrowed from the evaluation criteria of conventional spectral imaging, or only the spatial information evaluation method of the scene is used to evaluate the quality of the space spectrum restoration of the image. From the perspective of spectral imaging applications, the evaluation criteria for computing the quality of spectral imaging is proposed, that is spectral imaging stability is used as the first evaluation criterion. Considering that the main application field of spectral imaging is the identification of the target species and the determination of the content, the accuracy of the center position of the spectral peak of the target test set is used as the evaluation method for qualitative identification, and the trend line of the spectral curve of multiple measurements is used as the quantitative evaluation index for content determination. Two types of calculated spectral image reconstruction results are displayed. One type of mean square error is twice that of the other type. Under the evaluation criterion that the smaller the mean square error, the better the evaluation criteria, the better reconstruction results can not satisfy the spectral qualitative application. However, the method proposed in this article can quickly and effectively judge whether the restored spectrum can meet the needs of spectral analysis applications.