Identification of Human Body Fluid Stains and Non-Biological Stains by Three-Dimensional Fluorescence Spectroscopy.

OBJECTIVES: To establish a rapid and nondestructive identification method for human body fluid stains and non-biological stains using three-dimensional fluorescence spectroscopy. METHODS: The collected three-dimensional fluorescence spectrum data of human saliva, 3% blood, coffee and Fanta® stains were processed with dimensionality reduction. After wavelet transform, spectral denoising and feature extraction, the classification formula was established. The Fisher discriminant was used for spectrum matching and recognition to establish the analysis method to distinguish stain types. RESULTS: According to the results of data training and comparison, all the recognition accuracies of Fanta®, coffee, saliva and blood were more than 91.39%. Among them, saliva reached 100% recognition accuracy. CONCLUSIONS: Three-dimensional fluorescence spectroscopy is a potential method for rapid and nondestructive identification of biological and non-biological stains.

Image reconstruction method based on CCD calibration in frequency domain.

We demonstrate a method to reconstruct CCD images by calculating out the pixel response function accurately with laser interference patterns. This method is proven theoretically to have the ability to improve image quality greatly and thus may find great application in high-quality imaging fields.

[Design of full-polarized and multi-spectral imaging system based on LCVR].

A new full-polarized multi-spectral imaging system is described, which uses electronically controlled LCVR (liquid crystal variable retarder) to modulate the full-polarized state of light in the visible to IR range. The system consisted of optical lenses, LCVRs, filters and CCD. Firstly, the system structure, working theory and optical design are introduced in the present paper. A polarization calibration method is provided and the calibration system was set up, which realized high-precision polarization calibration using a small polarized source. Then, a field experiment with the imaging system was carried out. Polarized spectral images with higher spectral and spatial resolution were collected. Finally, the data acquired were rough processed to get polarization degree image of the targets. It is concluded that the experiment has proved that the imaging system is effective in obtaining full-polarized and multi- spectral data. The image captured by the system can be applied to object identification and object classification.