CN111854945A - A single-pixel ultraviolet polarization imaging method and system - Google Patents

Abstract

The invention discloses a single-pixel ultraviolet spectrum polarization imaging method and a single-pixel ultraviolet spectrum polarization imaging system, which comprise an ultraviolet light source, wherein the ultraviolet light source sequentially passes through a filtering unit, a first lens, an imaging target, a second lens, a spatial light modulator, a third lens, a polarization unit and a single-pixel ultraviolet detection module, the single-pixel ultraviolet detection module, a data acquisition module and an image restoration module are sequentially connected, a control module, the data acquisition module and the spatial light modulator are connected, and the control module is used for controlling the spatial light modulator and the data acquisition module. The invention applies the compression perception theory, carries out compression coding on the image or the spectrum information, can obtain the two-dimensional polarization image or the spectrum information of the target by using the ultraviolet detector with a single pixel, solves the data transmission problem caused by the high price and the large storage data amount of the existing ultraviolet imager, and provides a new scheme for realizing the ultraviolet polarization imaging.

Description

A single-pixel ultraviolet polarization imaging method and system

technical field

The invention relates to the field of optical imaging, in particular to a single-pixel ultraviolet spectral polarization imaging method and system.

Background technique

Ultraviolet imaging technology has important application requirements and research value. The ultraviolet array detection device used in the existing ultraviolet imaging system is difficult to manufacture, resulting in high overall system cost, and the used imaging principle has a large amount of stored data, and its application range is limited. Compared with the traditional ultraviolet imager, the single-pixel ultraviolet spectral imaging technology has relatively simple detection device process and low cost of the imaging system. At present, single-pixel imaging technology has been greatly developed in the visible and infrared bands and has achieved a certain degree of engineering applications. However, in the ultraviolet band, there is not much substantial progress in the research of high-performance ultraviolet detectors and single-pixel ultraviolet imaging optical systems. In addition, most of the research on polarization imaging technology focuses on the visible light and infrared bands. Polarization imaging can use the polarization vector information to enhance the ability to identify the imaging target, improve the imaging signal-to-noise ratio, and ultimately improve the imaging quality.

SUMMARY OF THE INVENTION

The technical problems to be solved by the present invention are that the existing ultraviolet imaging system is difficult to manufacture, the overall system cost is high, the amount of stored data of the imaging principle used is large, and its application range is limited, and the purpose is to provide a single-pixel ultraviolet spectrum. The polarization imaging method and system use compressed sensing theory to compress and encode image or spectral information, and only use a single-pixel UV detector to obtain a two-dimensional polarization image or spectral information of the target, which solves the problem of high cost of UV detection devices and imaging problems. The problem of slow speed and large amount of stored data.

The present invention is achieved through the following technical solutions:

A single-pixel ultraviolet spectrum polarization imaging system, comprising an ultraviolet light source, a filter unit, a first lens, an imaging target, a second lens, a spatial light modulator, a third lens, a polarization unit, a single-pixel ultraviolet detection module, a data acquisition module, An image restoration module and a control module, the light emitted by the ultraviolet light source passes through the filter unit to obtain narrow-band ultraviolet light, and the narrow-band ultraviolet light is focused and irradiated onto the imaging target through the first lens to obtain target imaging information, so The target imaging information is irradiated to the spatial light modulation module through the second lens for light modulation, and the modulated light containing the target imaging information is concentrated to the polarization unit through the third lens, and the polarization unit extracts the imaging target The ultraviolet polarization information of the imaging target is collected by the single-pixel ultraviolet detection module, and the data acquisition module converts the ultraviolet polarization information of the imaging target in the single-pixel ultraviolet detection module from optical signal information. is electrical signal information, the image restoration module is configured to process the electrical signal information to perform image restoration, and the control module controls the spatial light modulation module to perform light modulation.

In the present invention, the ultraviolet light source is firstly irradiated on the imaging target by the first lens after passing through the filtering unit, and then the light containing the image information of the imaging target is modulated by the spatial light modulator. According to the principle of compressed sensing, the spatial light modulator displays as The sensing matrix mask prepared in advance, and then the modulated light is passed through the polarization module to extract the polarization information of the imaging target, and the single-pixel UV detection module is used to detect the total light intensity of the modulated light to obtain the single-pixel measurement value, and the measurement value passes through the data. The acquisition module is transferred to the computer for image restoration using an algorithm.

The spatial light modulator coding and compression process is the core of the entire imaging system. Using the sparsity of the signal in a certain coding domain, the spatial signal is mapped to the coding domain for sampling, and the target image information can be recovered only by collecting a small number of measurement values, and the sampling frequency can be lower than the Nyquist frequency. , which reduces the requirements for the sampling rate and the amount of stored data at the same time.

Further, the filtering unit includes one or more filters, and the filters are narrow-band filters corresponding to the wavelength of the ultraviolet light source. The filtering unit is used to perform narrowband filtering on the ultraviolet light source, and can include multiple filters, which can perform filtering on different narrowband ultraviolet wavelength bands in time division, and can be used to obtain spectral information.

Further, the polarization unit includes one or more polarizers, and the polarizers are used to obtain polarization information of the imaging target. Preferably, the plurality of polarizers include polarizers with multiple polarization directions, so as to realize the selection of multiple polarization directions.

Further, the first lens, the first second lens and the third lens are all convex lenses.

Further, the single-pixel ultraviolet detection module includes a photoelectric sensor, and the photoelectric sensor is an ultraviolet avalanche photodiode.

Further, the spatial light modulator is a digital micromirror. Used to encode target image information projected to the spatial light modulator.

Further, based on the perceptual compression theory, the spatial light modulator and the data acquisition module are controlled by the control module, and the system includes the following control process: A1: The control module generates a perceptual matrix according to the image resolution of the target image, and the The perception matrix includes M rows; A2: Load the data of the Nth row in the perception matrix into the spatial light modulator, so as to control the spatial light modulator to deflect, and the data acquisition module completes one data acquisition, wherein N is a natural number greater than 1, and the initial value of N is 1; A3: if N is less than M, define N as N+1, and repeat step A2; A4: the data collection module collects the data after M times, The y matrix is obtained; A5: According to the y matrix, image reconstruction is performed by the compressed sensing restoration algorithm.

Another implementation manner of the present invention, a single-pixel ultraviolet spectral polarization imaging method, includes the following steps: B1: After the light of the ultraviolet light source is filtered, it is irradiated on the imaging target by the lens, and then collected to the spatial light modulator through the lens. ; B2: The light modulated by the spatial light modulator, through the lens and the polarization module, extracts the polarization information of the imaging target, B3: The single-pixel measurement value is then detected by the single-pixel UV detection module. Based on the compressed sensing theory, the The single-pixel measurement value is uploaded to the computer through data collection; B4: The target image is restored through the restoration algorithm.

The ultraviolet light source projects ultraviolet light of different wavelength bands onto the target through a spectroscopic device (ie filtering), and the spatial and spectral information carried by the reflected light of the target is encoded and compressed by the spatial light modulator, and then passes through the lens and the polarization module. It is coupled to the single-pixel UV detection module, and then the electrical signal obtained after photoelectric conversion is subjected to digital-to-analog conversion and data acquisition, and then decoded according to different light intensities and time-series coded signals, and the image is restored by hardware and software decoding of graphic compression and restoration algorithms. and extract spectral information.

Further, the restoration algorithm includes an orthogonal matching pursuit algorithm or a compressed sampling matching pursuit algorithm.

Compared with the prior art, the present invention has the following advantages and beneficial effects:

The invention combines imaging technology and spectrum technology, and on the basis of detecting two-dimensional space information, the ultraviolet spectrum information of the target can also be obtained, and the positioning, qualitative and quantitative analysis of the target with ultraviolet spectrum characteristics can be realized.

The invention uses the compressed sensing theory, compresses and encodes the image or spectral information, and only uses a single-pixel ultraviolet detector to obtain the two-dimensional polarization image or spectral information of the target, and solves the problem that the existing ultraviolet imager is expensive and storage. The problem of data transmission caused by the large amount of data, and provides a new solution for the realization of ultraviolet polarization imaging.

Description of drawings

The accompanying drawings described herein are used to provide further understanding of the embodiments of the present invention, and constitute a part of the present application, and do not constitute limitations to the embodiments of the present invention. In the attached image:

1 is a schematic diagram of the system structure of the present invention;

FIG. 2 is a flow chart of data acquisition of the present invention.

Detailed ways

In order to make the purpose, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the embodiments and the accompanying drawings. as a limitation of the present invention.

Example 1

As shown in FIG. 1 , Embodiment 1 is a single-pixel ultraviolet spectral polarization imaging system, which solves the problems of high cost, slow imaging speed, and large amount of stored data of current ultraviolet detection devices. The imaging system of the first embodiment includes, in sequence: an ultraviolet light source, a filter unit, a first lens, an imaging target, a second lens, a spatial light modulation module, a third lens, a polarization unit, an ultraviolet detection module, and an ultraviolet detection module, a data acquisition module The module and the image restoration module are connected in sequence. The control module is connected with the spatial light modulation module and the data acquisition module, and the control module controls the spatial light modulation module and the data acquisition module. Both the control module and the image restoration module are realized by computer.

First, a stable broadband ultraviolet light is emitted from the ultraviolet light source, and the broadband ultraviolet light passes through the filtering unit to obtain narrow-band ultraviolet light. The bandwidth is 10nm; the narrow-band ultraviolet light is focused and irradiated onto the imaging target through the lens to obtain the imaging information of the target, and then irradiated to the spatial light modulation module through the lens for light modulation; the spatial light modulation module (DMD) consists of tens of millions of It is composed of a small mirror with a size of 13.6um, which can be controlled by an external control signal to deflect ±12°. It is determined by the actual construction of the optical path. For example, the ultraviolet light irradiated on the -12° lens will not be irradiated to the subsequent lens for focusing, and the ultraviolet light irradiated on the +12° lens will pass through the lens, so that the light is radiated. Modulation; the modulated light containing the imaging information of the target is extracted by the polarization unit to extract the ultraviolet polarization information of the imaging target, and then collected by the ultraviolet detection module.

The first lens, the second lens and the third lens are all convex lenses.

The filtering unit is composed of filters, which are used for narrow-band filtering of the ultraviolet light source, and can include multiple filters, which can filter different narrow-band ultraviolet wavelength bands in a time-sharing manner, and can be used to obtain spectral information.

The polarizing unit is composed of a polarizing plate, and the polarizing plate is used to obtain the polarization information of the imaging target, and may include a polarizing plate with multiple polarization directions, so as to realize the selection of multiple polarization directions.

The spatial light modulation module is composed of a spatial light modulator, mainly a DMD (digital micromirror), which is used to encode the target image information projected to the DMD. The spatial light modulator coding and compression process is the core of the entire imaging system. Using the sparsity of the signal in a certain coding domain, the spatial signal is mapped to the coding domain for sampling, and the target image information can be recovered only by collecting a small number of measurement values, and the sampling frequency can be lower than the Nyquist frequency. , which reduces the requirements for the sampling rate and the amount of stored data at the same time.

In this embodiment 1, the ultraviolet light source is irradiated on the imaging target by the lens after passing through the filter, and then the light containing the image information of the imaging target is modulated by the spatial light modulator. The prepared perception matrix mask, and then the modulated light is passed through the polarization module to extract the polarization information of the imaging target, and the single-pixel UV detector module is used to detect the total light intensity of the modulated light to obtain the single-pixel measurement value, and the measurement value passes through the data. The acquisition module is transferred to the computer for image restoration using an algorithm.

This embodiment 1 provides a low-cost single-pixel ultraviolet spectral imaging system, which solves the problem of high cost of the current traditional ultraviolet imager, and provides a new solution for implementing an ultraviolet polarization imaging system. Combining imaging technology and spectral technology, on the basis of detecting two-dimensional spatial information, the ultraviolet spectral information of the target can also be obtained, and the positioning, qualitative and quantitative analysis of the target with ultraviolet spectral characteristics can be realized.

Example 2

The second embodiment is a single-pixel ultraviolet polarization imaging system based on the first embodiment. The ultraviolet light source projects the ultraviolet light of different wavelength bands onto the target object through a spectroscopic device (ie, a filter), and the space carried by the reflected light from the target object After the spectral information is compressed by DMD encoding, the target information is coupled to the single-pixel UV detector through lens aggregation and ultraviolet polarization, and then the electrical signal obtained after photoelectric conversion is subjected to digital-to-analog conversion and data acquisition. It decodes the strong and time-series coded signals, uses the graphics compression and restoration algorithm software and hardware to decode the restored image and extracts spectral information.

The DMD model sampled in this Example 2 is DLP7000UV with a resolution of 1024*768, which is suitable for the application of ultraviolet light sources with wavelengths from 200 nm to 400 nm. The photoelectric sensor used in the UV detection module is the UV avalanche photodiode APD, which has two working modes: linear mode and Geiger mode. In Geiger mode, it works above the reverse-biased breakdown voltage, the gain can reach more than 1M, and it has single-photon sensitivity, which can be used for spectral imaging in weak scenes. In the linear mode, although the sensitivity cannot reach the single-photon level, it works relatively faster and can be used for high-speed spectral imaging under strong light conditions.

In this embodiment 2, an ultraviolet single-pixel detector without spatial resolution is used as a sensing device to convert the optical signal into an electrical signal, and then obtain the information of the imaging target. Compressed sensing is its theoretical basis. If the signal is compressible or can be sparsely represented in a certain frequency domain, the signal can be measured using a matrix uncorrelated with the sparse basis in the frequency domain. Project a signal from a high-dimensional space to a low-dimensional space. The unknown element is the high-dimensional space signal, and the measured value is the low-dimensional space projection information.

The mathematical model of compressed sensing is: y=ΦX. The sparsity or compressibility of the signal is an important premise of compressed sensing. The real signal existing in nature is generally not absolutely sparse, but can be approximately sparse in a certain transform domain, that is, a compressible signal. For example, the transform domain can be discrete cosine transform, or wavelet transform. Image reconstruction is to first obtain the y matrix by measurement, and then because the general image signal X itself is not sparse, it needs to be sparsely represented on a certain sparse basis, that is, X=Ψθ, Ψ is a known sparse basis matrix, θ is a sparse vector (theta has only K non-zero values (K<<N), then the matrix A=ΦΨ can be obtained; the measured y and matrix A can be used to represent θ; since the sparse base matrix Ψ is known, compression can be used Another mathematical premise for perceptual image reconstruction is that the observation matrix Φ and the sparse basis matrix Ψ are irrelevant. Since the sparse basis matrix Ψ is known, the image X can be reconstructed through the restoration algorithm.

The restoration algorithm may use an Orthogonal Matching Pursuit (OMP) algorithm, a Compressive Sampling MP (Compressive Sampling MP) algorithm, and the like. Regarding the orthogonal matching pursuit algorithm, the core idea is to realize the orthogonalization of the column vector through Schmidt orthogonalization. In the iterative process, the orthogonalized atom and the signal are projected and compared to obtain the residual, so the residual sum The previously selected atoms are always orthogonal. Then each iteration will update the selected set of atoms to form a new set of atoms, which will further improve the computational efficiency of the algorithm and avoid repeated searches for atoms. The process is as follows:

(1) Initialize the sparse vector x0=0, the residual r0=y, the number of iterations k=1, and let the candidate set E0 be empty;

其中绝对值内为计算向量内积；(2) select the atom that best matches the residual from the previous iteration,

Among them, the absolute value is the inner product of the calculation vector;

(3) Add atoms to the candidate set, that is, Ek=Ek-1∪Ik;

(4) Update the sparse vector: xk=(ФEk)*y is obtained according to the least squares method, wherein ()* represents the pseudo-inverse matrix;

(5) Update residual rk=y-ФEkxk;

(6) Let k=k+1, judge whether the iteration condition is satisfied, if so, stop, otherwise jump to the second step.

As shown in FIG. 2 , the data collection process based on compressed sensing in Embodiment 2 is shown.

First, load the deflection pattern of the DMD, that is, one row of the generated perception matrix Φ into the DMD and control it to deflect according to the deflection pattern. At the same time, the data acquisition module is controlled to collect the signal passing through the ultraviolet detector module, that is, a measurement is completed. After repeatedly loading and measuring M times (the number of rows of the perception matrix Φ), the y matrix is obtained, and then the image can be restored through the compressed sensing restoration algorithm. The global loading rate of DMD can reach 20kHz. If the image resolution is 128*128, the required mask images are 8192, and it takes less than one second to load all the mask images.

The ultraviolet light source in the above embodiment 1 or 2 may be a common light source or a broad-spectrum light source.

The specific embodiments described above further describe the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above descriptions are only specific embodiments of the present invention and are not intended to limit the scope of the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims (10)

1. a single-pixel ultraviolet spectrum polarization imaging system, is characterized in that, comprises ultraviolet light source, filter unit, first lens, imaging target, second lens, spatial light modulator, third lens, polarization unit, single-pixel ultraviolet detection module, data acquisition module, image restoration module and control module, the light emitted by the ultraviolet light source passes through the filtering unit to obtain narrow-band ultraviolet light, and the narrow-band ultraviolet light is focused and irradiated onto the imaging target through the first lens, Obtaining target imaging information, the target imaging information is irradiated to the spatial light modulation module through the second lens for light modulation, and the modulated light containing the target imaging information is condensed to the polarization unit through the third lens, and the The polarization unit extracts the ultraviolet polarization information of the imaging target, and then collects the ultraviolet polarization information of the imaging target by the single-pixel ultraviolet detection module, and the data acquisition module collects the ultraviolet polarization information of the imaging target in the single-pixel ultraviolet detection module. The information is converted from optical signal information to electrical signal information, the image restoration module is used for processing the electrical signal information to perform image restoration, and the control module controls the spatial light modulation module to perform light modulation. 2 . The single-pixel ultraviolet spectral polarization imaging system according to claim 1 , wherein the filtering unit comprises one or more filters, and the filters are narrow-band filters corresponding to the wavelength of the ultraviolet light source. 3 . . 3 . The single-pixel ultraviolet spectral polarization imaging system according to claim 1 , wherein the polarization unit comprises one or more polarizers. 4 . 4 . The single-pixel ultraviolet spectral polarization imaging system according to claim 3 , wherein the plurality of polarizers comprise polarizers with a plurality of polarization directions. 5 . 5 . The single-pixel ultraviolet spectral polarization imaging system according to claim 1 , wherein the first lens, the first second lens and the third lens are all convex lenses. 6 . 6 . The single-pixel ultraviolet spectral polarization imaging system according to claim 1 , wherein the single-pixel ultraviolet detection module comprises a photoelectric sensor, and the photoelectric sensor is an ultraviolet avalanche photodiode. 7 . 7 . The single-pixel ultraviolet spectral polarization imaging system according to claim 1 , wherein the spatial light modulator is a digital micromirror. 8 . 8. The single-pixel ultraviolet spectral polarization imaging system according to claim 1, wherein the system comprises the following control process: A1: The control module generates a perception matrix according to the image resolution of the target image, and the perception matrix includes M rows; A2: Load the data of the Nth row in the perception matrix into the spatial light modulator to control the spatial light modulator to deflect, and the data acquisition module completes one data acquisition, where N is a natural number greater than 1 , and the initial value of N is 1; A3: If N is less than M, define N as N+1, and repeat step A2; A4: The data acquisition module collects the data after M times to obtain a y matrix; A5: According to the y matrix, image reconstruction is performed by the compressed sensing restoration algorithm. 9. A single-pixel ultraviolet spectral polarization imaging method, characterized in that, comprising the following steps: B1: After the light of the ultraviolet light source is filtered, it is irradiated on the imaging target by the lens, and then concentrated to the spatial light modulator through the lens; B2: The light modulated by the spatial light modulator, through the lens and the polarization module, extracts the polarization information of the imaging target, B3: The single-pixel measurement value is then detected by the single-pixel UV detection module, and based on the compressed sensing theory, the single-pixel measurement value is uploaded to the computer through data collection; B4: The target image is restored through the restoration algorithm. 10 . The single-pixel ultraviolet spectral polarization imaging method according to claim 9 , wherein the restoration algorithm comprises an orthogonal matching pursuit algorithm or a compressive sampling matching pursuit algorithm. 11 .

Images