The aim of this research project is to distinguish human skin from dead animal meat tissue contactless and reliably by means of spectral and temporal dynamic analysis. Based on the results a functional model will be developed, which measures narrowband reflection intensity at different wavelengths in the ultraviolet (UV) and visible spectral range. By means of elaborate analysis of stochastic signals, correlation changes are to be recognized and reliably detected.
Prior to the development of a robust detection method, a tissue and skin study will be conducted. For the detection of the different reflectance spectra, a calibrated measurement setup with a high-resolution, sensitive spectrometer is designed. The special challenge lies in the test setup itself, its calibration and the necessary optics for an experimental setup.
Due to the tissue and skin study significant spectral signatures in the near UV and in the visible wavelength range can be performed. Initial investigations have already identified three possible criteria to ensure that various types of tissue and skin types can be distinguished: distinctive spectral signatures in the ultraviolet spectral range, characteristic "W" structure of oxygenated hemoglobin and greater differences of various samples at certain wavelengths.
The functional model to be developed consists of one or more light sources and the corresponding detection unit including evaluation logic. The detection unit is composed of optics specially designed for the task and the actual optical detector, which will be made of a spectrometer or a system of photodiodes with narrowband optical filters. In addition to achieving its functionality, it is aimed to equip the device with a light source that is economical, low in maintenance and easy to implement.
Together with the optics and light source development, the appropriate electronics for controlling the light sources will be designed. With a microcontroller system, the measured electronically processed signals can be evaluated and classified accordingly.
Once the specified functional model is available, it is intended to provide both a lockable-sequent measurement campaign with different tissue and skin samples and an error analysis.
prevention, machine safetyDescription, key words:
Hand detection, distinguish hand - dead animal meat