Font Size: a A A

Design And Application Of An LCTF-Based Hyperspectral Microscopic Imaging System For Biological Detection

Posted on:2016-04-30Degree:DoctorType:Dissertation
Country:ChinaCandidate:S Q ZhuFull Text:PDF
GTID:1108330479489554Subject:Information technology
Abstract/Summary:PDF Full Text Request
The hyperspectral microscopic imaging technology which can simultaneously obtain spatial and spectral information is one of the most important means to explore the microcosm. It is a noncontact and nondestructive detection method with high precision and replicability, and widely applied to the biological detection, especially in the area of composition identification. Several hyper-spectral technologies can be used to achieve different spectral images, such as absorption spectroscopy, fluorescence spectroscopy, and Raman spectroscopy. However, the application of spectral technologies are limited by the materials. For example, some materials that are almost transparent or opaque are not suitable for absorption spectral detection, and some materials cannot generate sufficient self-fluorescence for the fluorescence spectral detection. Therefore, none of these spectral technologies can be used to analyze all types of materials. Besides, samples extracted from different individuality often exhibit different spectral responses due to individual biological variations, even they are of the same kind. These are unfavorable for the application of hyperspectral technology in biological detection. To overcome these disadvantages and improve the detection ability of the system, an intensive study of hardware and software of the hyperspectral microscopic imaging system was made in this research. Based on the results of theoretical analysis and experiment, an LCTF-based hyperspectral microscopic imaging system that is capable of obtaining both transmission and fluorescence spectral images was set up. Moreover, a new spectral component extraction and identification method which can avoid the interference of individual biological variations was designed by analyzing the spectral information of the gastric cancerous tissues. What have done in this thesis are listed as follows:(1) Theoretical analysis and system design. First, based on the analyzing of the structure and function of different hyperspectral imaging systems, it was decided to combine a fluorescence spectral system with a transmission spectral system for obtaining double-type spectral images. Then the characteristics of the response of the system were analyzed. The response of the system is related to the light energy, exposure time, magnification, sensitivity of the CCD and kinds of losses of the optical signal, and a numerical model was set up to describe it by considering all of these factors. Severer methods about increasing response of the system were introduced after the study of the model.(2) Setting up light sources. As a double-type spectral imaging system, it needs two light sources. The lasers at the wavelength at 266 nm, 457 nm and 410 nm were set up and investigated. In accordance with the test results, the 410 nm laser diode was chosen for exciting the fluorescence in the system. Meanwhile, a white LED light source with a catadioptric optical structure that can collect more light energy was used to illuminate the sample. The signal-noise ratio can be increased by the improvement of utilization of the light.(3) Setting up the system. According to the theoretical analysis and the actual situation, the infinitycorrected imaging microscope objective, the zooming lenses, and the liquid crystal tunable filter, etc. were used to set up the hyperspectral microscopic imaging system. This system can obtain the transmission and fluorescence spectral images at the range of 450-700 nm with the highest spectral resolution of 1 nm and fastest image acquisition time of 100.02 ms. It has advantages of high signal-noise ratio, high imaging quality and good stability.(4) Experiment of bee pollen detection base on the hyperspectral microscopic imaging system. The bee pollen of the rose was used to investigate the ability of the double-type spectral detection of the system. With this system, we were able to determine that the center of the bee pollen is a thin film, rather than a tiny hole. Based on the fluorescence spectral images, we could confirm that the composition of the thin film is the same as the rest of the pollen. Hence, we were able to obtain the transmission spectrum at the thin film in the bee pollen; it is difficult to obtain the transmission spectrum at other locations. These experimental results were based on double-type spectral detection. Neither conventional microscopy nor a single-type hyper-spectral imaging system is capable of obtaining such data.(5) Experiment of gastric cancerous cells identification based on the hyperspectral microscopic imaging system. The samples of gastric tumor were used to study the methods of spectral extraction and identification. Samples from different patients often exhibit different spectral responses due to individual biological variations, and even in the same sample, the spectra of cancerous cells exhibit small differences. Thus, spectral imaging of such cells may not show any fixed characteristic peak or formation, which undermines the test results. In this research, a new method of identification based on the spectral zones division and zone features was proposed according to the comparative analysis of the normal cells and cancerous cells in the same sample. With the use of this method, the average accuracy of the identification of cancerous cells was 95%. In summary, the detectability and applicability of the hyperspectral microscopic imaging system can be improved effectively by using the double-type spectral imaging technology and the identification method based on the spectral zones division and zone features. The experiment results indicate that the system and the proposed methods were effective and valuable.
Keywords/Search Tags:hyperspectral imaging, biological detection, transmission spectrum, fluorescence spectrum, cancerous cells, bee pollen
PDF Full Text Request
Related items