Research On Compressed Sensing Hyperspectral Biological Microscopic Imaging Based On LCVR

The merging of life sciences,optics,computers,and other fields has accelarated the development of life science detection methods and greatly improved the detection capabilities of humans.Since its inception,optical microscopy has played a major role in life science research.With the deeper exploration of biological targets,the demand for multidimensional optical microscopy imaging has grown.Hyperspectral microscopy imaging can measure both the two-dimensional spatial distribution of photons emitted by biological samples and their hyperspectral information,and has a widespread applications in life science research and applications such as target recognition,classification and medical diagnosis.A significant feature of hyperspectral microscopy is the huge amount of data,which poses challenges for transmission,storage,and rapid detection of data and is one of the main factors limiting its wide application.The combination of hyperspectral microscopy imaging technology and compressed sensing theory can recover the original image at a sampling rate far below that required by Nyquist criterion,significantly reducing the cost of signal storage,transmission,and processing.At both the hardware and software levels,it has further advanced the development of hyperspectral microscopy imaging technology.Under the paradigm of compressed sensing combined with the fast response characteristics of LCVR,this work investigates a new type of fast,compressed sensing hyperspectral microscopy imaging technology that has the advantages of small data acquisition,fast acquisition speed,no moving parts,and high throughput.This technology is suitable for the detection and identification of static biological targets.The research are as follow:1.The development of hyperspectral imaging technology in the field of microscopic imaging and the development of computational hyperspectral imaging are discussed.The spectral principle and scanning method of spectral imaging technology are examined in detail.The sparse representation of signals,linear coding and nonlinear decoding in compressed sensing are systematically studied.Under the compressed sensing framework,the discrete mathematical model and LCVR-based spectrum compression coding were drived.2.The LCVR mathematical model,which is the core component of spectral coding,is built and analyzed in detail.By constructing a nonlinear mathematical model of the phase retardation of LCVR as a function of voltage,angle of incidence,temperature,and wavelength,the effects of various factors on the phase retardation were characterized in detail,which provided a theoretical basis and guidance for the construction and engineering of the experimental system.3.An active illumination compression sensing hyperspectral microscopic imaging system based on LCVR is proposed.The active illumination-based hyperspectral microscopic imaging system is developed in the context of compressed sensing and uses a Liquid crystal modulation module to modulate and encode the incident light in the spectral domain.It can still accurately reconstruct the hyperspectral image even at a compression ratio of 11:1.This method of modulating the incident light does not require special modification of the optical microscope,so that the common optical microscopic imaging system also can perform hyperspectral imaging.Simulations of the system were performed using the data sets acquired in the laboratory,and the results show high reconstruction quality in both image and spectrum.Further,the experimental device of the system was built in the laboratory.After the system was calibrated,the feasibility of the system was tested by using slides.Finally,a hyperspectral microscopic imaging experiment of the cross section of herbal plant stem was carried out,and the hyperspectral reconstructed images in the wavelength range of 400-800 nm were obtained.