A method for multi-spectral single-pixel imaging with pre-decoding dummy pattern and device thereof

Abstract

The present application belongs to the technical field of computational imaging, and particularly discloses a kind of multispectral single-pixel imaging methods of pre-decoding virtual mode, the imaging method includes the following steps: generating Hadamard matrix and spectral channel modulation function, according to Hadamard matrix and multiple spectral channel modulation function to generate mixed frequency illumination structured light;Using mixed frequency illumination structured light to irradiate target object, the intensity value of backscattering light or transmission light of target object is acquired;Fourier transform is carried out to Hadamard matrix, and the spectral information is acquired, and after the spectral information is rearranged, it is selected, and after selection, inverse Fourier transform is carried out to obtain the pre-decoding virtual mode spatial distribution of different spectral channels;According to the intensity value of backscattering light or transmission light, the multispectral image of target object is reconstructed using a step-by-step iterative algorithm according to the pre-decoding virtual mode spatial distribution of different spectral channels.The present application can improve the imaging efficiency of multispectral single-pixel and simplify the system operation process.

Description

Technical Field

[0001] This invention belongs to the field of computational imaging technology, and specifically relates to a multispectral single-pixel imaging method and device for pre-decoding virtual modes. Background Technology

[0002] Single-pixel imaging is an emerging computational imaging approach. Unlike traditional area array imaging, it uses a non-spatial resolution detector to acquire spatial information of the scene and reconstructs the image through correlation calculation algorithms. In recent years, single-pixel imaging has been widely studied in applications such as lidar, gas detection, and tracking of fast-moving objects.

[0003] Multispectral single-pixel imaging has attracted widespread attention in recent years due to its high-sensitivity imaging capabilities in the non-visible light band. Multispectral multiplexed structured illumination is a common modulation and coding method that synchronously acquires multispectral information of an object using a single single-pixel detector. Existing multispectral single-pixel imaging methods all feature post-processing decoding; that is, they require first acquiring aliased multispectral image information and then performing post-processing to obtain the object's multispectral information. This method cannot output the object's multispectral information in real time and is unsuitable for applications with high real-time requirements. Therefore, there is an urgent need for a multispectral single-pixel imaging method and its associated equipment to improve imaging efficiency and achieve rapid single-pixel multispectral imaging of objects. Summary of the Invention

[0004] To address the above problems, this invention discloses a multispectral single-pixel imaging method for pre-decoding virtual modes, the imaging method comprising the following steps:

[0005] Generate the Hadamard matrix and spectral channel modulation functions, and generate mixed illumination structured light based on the Hadamard matrix and multiple spectral channel modulation basis functions;

[0006] The target object is illuminated using mixed-frequency structured light to obtain the intensity value of the backscattered or transmitted light of the target object.

[0007] Perform a Fourier transform on the Hadamard matrix to obtain spectral information, rearrange the spectrum information and select the selected spectrum information, and then perform an inverse Fourier transform to obtain the spatial distribution of the pre-decoding virtual mode of different spectral channels.

[0008] The multispectral image of the target object is reconstructed using a step-by-step iterative algorithm based on the spatial distribution of the pre-decoded virtual mode of different spectral channels and the intensity values ​​of the backscattered or transmitted light.

[0009] Furthermore, the generation of the Hadamard matrix and spectral channel modulation functions, and the generation of mixed-frequency illumination structured light based on the Hadamard matrix and multiple spectral channel modulation basis functions, specifically includes the following steps:

[0010] Deterministic orthogonal Hadamard bases and spectral channel modulation base spatial distributions are generated using computers.

[0011] The spectral information of each spectral channel is spatially modulated by the spatial distribution of the modulation base of the spectral channels;

[0012] Multiple modulated spectral information are aliased.

[0013] Furthermore, the spectral channel modulation basis function is a plurality of N×N two-dimensional gray-level cosine structure coding matrices, which can be expressed as:

[0014]

[0015] Where Fa represents the spectral channel modulation function for a single color channel, f x f y For the frequencies of the encoding matrix, the numerical subscripts represent different frequency values. This is the initial phase.

[0016] Furthermore, the intensity value of the backscattered or transmitted light from the target object is expressed as:

[0017]

[0018] Among them, D i The backscattered or transmitted light intensity values ​​are collected after the target object is illuminated by mixed-frequency structured light; mr is the aliased multispectral image containing spectral information from multiple spectral channels; H i The deterministic orthogonal Hamadakji space distribution is generated by the computer, and the symbol * represents the product operation of corresponding elements of the matrix.

[0019] Furthermore, the selected spectral information is subjected to inverse Fourier transform to obtain the spatial distribution of the pre-decoded virtual mode for different spectral channels, which is then substituted into the solution formula as follows:

[0020]

[0021] H i，a =iFT{B*S a {FT{H i}}}

[0022] Where, m a D represents the spectral information corresponding to a single color in space. i This represents the intensity of backscattered or transmitted light collected after the target object is illuminated by mixed-frequency structured light. Parameter B represents spatial filtering in the Fourier frequency domain. The subscript a indicates a single color. FT stands for Fourier transform, iFT stands for inverse Fourier transform, and S... a This represents the spatial distribution H of the Hadamard.i Rearrangement of the Fourier spectrum, H i,a This indicates the spatial distribution of the pre-decoding virtual mode.

[0023] Furthermore, after performing a Fourier transform on the Hadamard matrix to obtain the spectral information, the process further includes:

[0024] Based on the characteristics of the modulation functions of multiple spectral channels, the frequency range of interest in the spectral information is selected, and the selected spectral information is moved to the center position;

[0025] A Gaussian center filter is used to extract the spectral information from multiple rearrangements.

[0026] Furthermore, the step-by-step iterative algorithm for reconstructing the multispectral image of the target object based on the spatial distribution of the pre-decoded virtual mode of different spectral channels and the intensity values ​​of the backscattered or transmitted light specifically includes the following steps:

[0027] The image information of a single spectral channel is obtained by real-time iterative calculation using the intensity values ​​of backscattered or transmitted light of the target object and the spatial distribution of multiple pre-decoded virtual modes.

[0028] By combining image information from individual spectral channels, a multispectral image of the target object can be reconstructed.

[0029] In another embodiment of the present invention, a multispectral single-pixel imaging device with a pre-decoded virtual mode is also disclosed, the imaging device comprising:

[0030] A lens group is used to collect light signals reflected from or transmitted through a target object onto a single-pixel detector.

[0031] Digital projectors are used to generate multispectral lighting patterns;

[0032] The first computing unit is used to generate the Hadamard matrix and spectral channel modulation function, and to acquire the intensity value of a single pixel detector.

[0033] The second calculation unit performs a Fourier transform on the Hadamard matrix to obtain spectral information, rearranges the spectral information, selects the selected information, and then performs an inverse Fourier transform to obtain the spatial distribution of the pre-decoding virtual mode for different spectral channels.

[0034] The third calculation unit reconstructs the multispectral image of the target object using a step-by-step iterative algorithm based on the spatial distribution of the pre-decoded virtual mode of different spectral channels and the intensity value of the backscattered or transmitted light.

[0035] In another embodiment of the present invention, a computer-readable storage medium is also disclosed, on which a computer program is stored, and the computer program, after being run, executes the multispectral single-pixel imaging method of the pre-decoding virtual mode as described in any of the above embodiments.

[0036] In another embodiment of the present invention, a computer device is also disclosed, including a processor and a storage medium, wherein a computer program is stored on the storage medium, and the processor reads from the storage medium and runs the computer program to perform the multispectral single-pixel imaging method of the pre-decoding virtual mode as described in any of the above embodiments.

[0037] Beneficial effects

[0038] 1) This invention improves the imaging efficiency of multispectral single pixels and simplifies the system operation process by introducing a pre-decoding virtual mode.

[0039] 2) Gaussian center filters can emphasize or suppress signal components at specific frequencies, reducing interference with subsequent spectral processing and imaging.

[0040] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures pointed out in the description, claims and drawings. Attached Figure Description

[0041] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0042] Figure 1 A flowchart of a multispectral single-pixel imaging method for pre-decoding virtual modes according to an embodiment of the present invention is shown;

[0043] Figure 2 A schematic diagram of a multispectral single-pixel imaging method for pre-decoding virtual modes according to an embodiment of the present invention is shown.

[0044] Figure 3 A schematic diagram of spectral rearrangement according to an embodiment of the present invention is shown. Detailed Implementation

[0045] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0046] like Figure 1 and Figure 2 As shown, this invention discloses a multispectral single-pixel imaging method for pre-decoding virtual modes, the imaging method comprising the following steps:

[0047] Generate the Hadamard matrix and spectral channel modulation functions, and generate mixed illumination structured light based on the Hadamard matrix and multiple spectral channel modulation basis functions;

[0048] The target object is illuminated using mixed-frequency structured light to obtain the intensity value of the backscattered or transmitted light of the target object.

[0049] Perform a Fourier transform on the Hadamard matrix to obtain spectral information, rearrange the spectrum information and select the selected spectrum information, and then perform an inverse Fourier transform to obtain the spatial distribution of the pre-decoding virtual mode of different spectral channels.

[0050] The multispectral image of the target object is reconstructed using a step-by-step iterative algorithm based on the spatial distribution of the pre-decoded virtual mode of different spectral channels and the intensity values ​​of the backscattered or transmitted light.

[0051] For example, in this embodiment, three modulation basis functions for the red, green, and blue spectral channels are generated, denoted as Fi. red F green and F blue The three spectral channels are used to modulate the basis functions and the generated Hadamard matrix H. i Interactions generate mixed-frequency illumination structured light P i The Hadamard matrix H is among them. i The mixed-frequency illumination structured light P is a set of hadamar patterns. i It is formed by superimposing light signals of different frequencies. In this embodiment, the mixed-frequency illumination structured light is formed by superimposing structured light of three frequencies: red, green, and blue. The mixed-frequency illumination structured light P... i The target object is illuminated, and then the reflected or transmitted light from the illuminated object is collected. The detection intensity of the reflected or transmitted light is obtained and stored, denoted as D. i .

[0052] For the Hadamard matrix H iPerform a Fourier transform, denoted as FT, to obtain its spectral information FT{H i}. The Hadamard matrix H is obtained through Fourier transform. i Transform from the spatial domain to the frequency domain. Based on the characteristics of the modulation functions of the red, green, and blue spectral channels, analyze the spectral information FT{H... i The spectrum is rearranged, selected, and truncated sequentially to obtain the premodulation spectrum information of the red, green, and blue spectral channels, denoted as B*S. red {FT{H i}}、B*S green {FT{H i}} and B*S blue {FT{H i Next, an inverse Fourier transform (iFT) is performed on the premodulated spectral information to obtain the spatial distribution of the pre-decoding virtual mode for the three spectral channels, denoted as H. i,red H i,green and H i,blue .

[0053] Using the detection intensity D i The image information of the three spectral channels is obtained by real-time iterative calculation of the pre-decoded virtual mode spatial distribution of the red, green, and blue spectral channels, denoted as m. red m green and m blue After obtaining the image information from the three spectral channels, the image information from the three spectral channels is synthesized to obtain a pseudo-color image.

[0054] It is worth noting that increasing the number of spectral channels will result in a richer color image, better displaying the details and features of the target object. This embodiment uses only three spectral channels as an example. This embodiment utilizes the dissimilarity between mixing image decoding and modulation mode decoding to perform pre-decoding of the iterative modulation mode, improving multispectral imaging efficiency and simplifying the system operation process.

[0055] Rearrangement refers to dividing the Fourier spectrum into several regions and then rearranging the relative positions of these regions. For example... Figure 3 As shown, the left column represents the Fourier spectra of red, green, and blue respectively. The red Fourier spectrum is divided into two regions, upper and lower, and then the two regions are rearranged. The green Fourier spectrum is divided into two regions, left and right, and then the two regions are rearranged. The blue Fourier spectrum is divided into four regions, and then the positions of the four regions are rearranged.

[0056] Preferably, after performing a Fourier transform on the Hadamard matrix to obtain spectral information in different spectral channels, the method further includes:

[0057] Based on the characteristics of the modulation functions of multiple spectral channels, the frequency range of interest in the spectral information is selected, and the selected spectral information is moved to the center position;

[0058] A Gaussian center filter is used to extract the spectral information from multiple rearrangements.

[0059] For example, the red spectral channel modulation function selects the red frequency range from the spectral information, the green spectral channel modulation function selects the green frequency range from the spectral information, and the blue spectral channel modulation function selects the blue frequency range from the spectral information. The three spectral information values ​​selected from the three spectral channels are then moved to their respective center positions. A Gaussian center filter is then used to truncate the selected spectral information, with the center position of the Gaussian center filter corresponding to the center position of the moved spectral information. The Gaussian center filter can emphasize or suppress signal components at specific frequencies, reducing interference with subsequent spectral processing and imaging.

[0060] Furthermore, the generation of the Hadamard matrix and spectral channel modulation functions, and the generation of mixed-frequency illumination structured light based on the Hadamard matrix and multiple spectral channel modulation basis functions, specifically includes the following steps:

[0061] Deterministic orthogonal Hadamard bases and spectral channel modulation base spatial distributions are generated using computers.

[0062] The spectral information of each spectral channel is spatially modulated by the spatial distribution of the modulation base of the spectral channels;

[0063] Multiple modulated spectral information are aliased.

[0064] For example, computer-generated deterministic orthogonal Hadamardi spatial distribution H i Through the spatial distribution H of Hadamaki i The cosine structure encoding matrix spatially modulates each spectrum, and then the matrices are superimposed to construct a multispectral illumination pattern P. i In this embodiment, the red, green, and blue spectral channels are used as an example. The multispectral illumination pattern Pi can be represented as:

[0065] P i =F red *H i +F green *H i +F blue *H i

[0066] Among them, F red *H i F green *H i and F blue *Hi These are illumination patterns corresponding to the red, green, and blue spectral channels, respectively. The symbol * represents the product operation of the corresponding elements of the matrix.

[0067] Furthermore, the spectral channel modulation basis function is a plurality of N×N two-dimensional gray-level cosine structure coding matrices, which can be expressed as:

[0068]

[0069] Among them, F a The spectral channel modulation function representing a single color channel, f x f y For the frequencies of the encoding matrix, the numerical subscripts represent different frequency values. This is the initial phase.

[0070] In this embodiment, the spectral channel modulation function is three N×N two-dimensional gray-level cosine structure coding matrices, which can be represented as:

[0071]

[0072]

[0073]

[0074] Among them, f x f y For the frequency of the encoding matrix, This is the initial phase.

[0075] Furthermore, the intensity value of the backscattered or transmitted light from the target object is expressed as:

[0076]

[0077] Among them, D i The backscattered or transmitted light intensity values ​​are collected after the target object is illuminated by mixed-frequency structured light; mr is the aliased multispectral image containing spectral information from multiple spectral channels; H i The deterministic orthogonal Hamadakji space distribution is generated by the computer, and the symbol * represents the product operation of corresponding elements of the matrix.

[0078] Specifically, the transmitted or backscattered light from the target object is collected by a lens group, then captured and stored for analysis using a single-pixel detector. In this embodiment, the intensity values ​​of the transmitted or backscattered light are calculated using a generated deterministic orthogonal Hamadakji spatial distribution.

[0079] Furthermore, the rearranged spectral information is subjected to inverse Fourier transform to obtain the spatial distribution H of the pre-decoded virtual mode obtained by iterative solution. i,aSubstituting this into the solution formula, it can be expressed as:

[0080]

[0081] H i，a =iFT{B*S a {FT{H i}}}

[0082] Where, m a D represents the spectral information corresponding to a single color in space. i This represents the intensity of backscattered or transmitted light collected after the target object is illuminated by mixed-frequency structured light. Parameter B represents spatial filtering in the Fourier frequency domain. The subscript a indicates a single color. FT represents Fourier transform, iFT represents inverse Fourier transform, and Sa represents the rearrangement of the Fourier spectrum of the Hadamard spatial distribution Hi. i,a This indicates the spatial distribution of the pre-decoded data.

[0083] In this embodiment, the reciprocity between aliased image decoding and modulation mode decoding is utilized to achieve near real-time synchronous iterative acquisition of multispectral images by pre-decoding the modulation mode. i，a It is a pre-decoded spatial distribution that did not participate in the actual detection experiment; it is called a pre-decoded virtual mode.

[0084] Taking the acquisition of an image of the red channel of a target object as an example:

[0085]

[0086] H i，red =iFT{B*S red {FT{H i}}}

[0087] Parameter B represents spatial filtering in the Fourier frequency domain. The center position of the filter corresponds to the middle position of the spectrum after frequency shifting, which can effectively obtain the spectrum of the red channel.

[0088] Furthermore, the step-by-step iterative algorithm for reconstructing the multispectral image of the target object based on the spatial distribution of the pre-decoded virtual mode of different spectral channels and the intensity values ​​of the backscattered or transmitted light specifically includes the following steps:

[0089] The image information of a single spectral channel is obtained by real-time iterative calculation using the intensity value of the backscattered or transmitted light of the target object and multiple spectral spatial distributions.

[0090] By combining image information from individual spectral channels, a multispectral image of the target object can be reconstructed.

[0091] For example, the intensity value Di of the backscattered or transmitted light of the target object and the three spectral spatial distribution functions H (red, green, and blue) can be used. i,red H i,blue and H i,green Real-time iterative calculations are performed separately to obtain image information for the red, green, and blue spectral channels, denoted as m. red m blue and m green Finally, by superimposing the image information from the red, green, and blue spectral channels, a pseudo-color image can be obtained.

[0092] In another embodiment of the present invention, a multispectral single-pixel imaging device is also disclosed, the imaging device comprising:

[0093] A lens group is used to collect light signals reflected from or transmitted through a target object onto a single-pixel detector.

[0094] Digital projectors are used to generate multispectral lighting patterns;

[0095] The first computing unit is used to generate the Hadamard matrix and spectral channel modulation function, and to acquire the intensity value of a single pixel detector.

[0096] The second calculation unit performs a Fourier transform on the Hadamard matrix to obtain spectral information in different spectral channels, rearranges the spectral information, selects the selected information, and then performs an inverse Fourier transform to obtain the spatial distribution of different spectral channels.

[0097] The third calculation unit reconstructs the multispectral image of the target object using a step-by-step iterative algorithm based on the spatial distribution of different spectral channels and the intensity values ​​of the backscattered or transmitted light.

[0098] In another embodiment of the present invention, a computer-readable storage medium is also disclosed, on which a computer program is stored, and the computer program executes the multispectral single-pixel imaging method as described in any of the above embodiments.

[0099] In another embodiment of the present invention, a computer device is also disclosed, including a processor and a storage medium, wherein a computer program is stored on the storage medium, and the processor reads from the storage medium and runs the computer program to perform the multispectral single-pixel imaging method as described in any of the above embodiments.

[0100] Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A multispectral single-pixel imaging method for pre-decoding virtual modes, characterized in that, The imaging method includes the following steps: Generate the Hadamard matrix and spectral channel modulation functions, and generate mixed illumination structured light based on the Hadamard matrix and multiple spectral channel modulation basis functions; The target object is illuminated using mixed-frequency structured light to obtain the intensity value of the backscattered or transmitted light of the target object. Perform a Fourier transform on the Hadamard matrix to obtain spectral information, rearrange the spectrum information and select the selected spectrum information, and then perform an inverse Fourier transform to obtain the spatial distribution of the pre-decoding virtual mode of different spectral channels. The multispectral image of the target object is reconstructed using a step-by-step iterative algorithm based on the spatial distribution of the pre-decoded virtual mode of different spectral channels and the intensity values ​​of the backscattered or transmitted light.

2. The multispectral single-pixel imaging method for pre-decoding virtual modes according to claim 1, characterized in that, The process of generating the Hadamard matrix and spectral channel modulation functions, and generating mixed illumination structured light based on the Hadamard matrix and multiple spectral channel modulation basis functions, specifically includes the following steps: Deterministic orthogonal Hadamard bases and spectral channel modulation base spatial distributions are generated using computers. The spectral information of each spectral channel is spatially modulated by the spatial distribution of the modulation base of the spectral channels; Multiple modulated spectral information are aliased.

3. The multispectral single-pixel imaging method for pre-decoding virtual modes according to claim 1, characterized in that, The modulation basis function of the spectral channel is a plurality of N×N two-dimensional gray-level cosine structure coding matrices, which can be represented as: Among them, F a The spectral channel modulation function representing a single color channel, f x f y For the frequencies of the encoding matrix, the numerical subscripts represent different frequency values. This is the initial phase.

4. The multispectral single-pixel imaging method for pre-decoding virtual modes according to claim 3, characterized in that, The intensity value of the backscattered or transmitted light from the target object is expressed as: Among them, D i The backscattered or transmitted light intensity values ​​are collected after the target object is illuminated by mixed-frequency structured light; mr is the aliased multispectral image containing spectral information from multiple spectral channels; H i The deterministic orthogonal Hamadakji space distribution is generated by the computer, and the symbol * represents the product operation of corresponding elements of the matrix.

5. The multispectral single-pixel imaging method for pre-decoding virtual modes according to claim 3, characterized in that, Performing an inverse Fourier transform on the selected spectral information yields the spatial distribution of the pre-decoded virtual modes for different spectral channels, which, when substituted into the solution formula, is expressed as: H i，a =iFT{B*S a {FT{H i }}} Where, m a D represents the spectral information corresponding to a single color in space. i This represents the intensity of backscattered or transmitted light collected after the target object is illuminated by mixed-frequency structured light. Parameter B represents spatial filtering in the Fourier frequency domain. The subscript a indicates a single color. FT stands for Fourier transform, iFT stands for inverse Fourier transform, and S... a This represents the spatial distribution H of the Hadamard. i Rearrangement of the Fourier spectrum, H i,a This indicates the spatial distribution of the pre-decoding virtual mode.

6. The multispectral single-pixel imaging method for pre-decoding virtual modes according to claim 1, characterized in that, After performing a Fourier transform on the Hadamard matrix to obtain the spectral information, the process further includes: Based on the characteristics of the modulation functions of multiple spectral channels, the frequency range of interest in the spectral information is selected, and the selected spectral information is moved to the center position; A Gaussian center filter is used to extract the spectral information from multiple rearrangements.

7. The multispectral single-pixel imaging method for pre-decoding virtual modes according to claim 1, characterized in that, The process of reconstructing the multispectral image of the target object using a step-by-step iterative algorithm based on the spatial distribution of the pre-decoded virtual mode of different spectral channels and the intensity values ​​of the backscattered or transmitted light specifically includes the following steps: The image information of a single spectral channel is obtained by real-time iterative calculation using the intensity values ​​of backscattered or transmitted light of the target object and the spatial distribution of multiple pre-decoded virtual modes. By combining image information from individual spectral channels, a multispectral image of the target object can be reconstructed.

8. A multispectral single-pixel imaging device with pre-decoded virtual mode, characterized in that, The imaging device includes: A lens group is used to collect light signals reflected from or transmitted through a target object onto a single-pixel detector. Digital projectors are used to generate multispectral lighting patterns; The first computing unit is used to generate the Hadamard matrix and spectral channel modulation function, and to acquire the intensity value of a single pixel detector. The second calculation unit performs a Fourier transform on the Hadamard matrix to obtain spectral information, rearranges the spectral information, selects the selected information, and then performs an inverse Fourier transform to obtain the spatial distribution of the pre-decoding virtual mode for different spectral channels. The third calculation unit reconstructs the multispectral image of the target object using a step-by-step iterative algorithm based on the spatial distribution of the pre-decoded virtual mode of different spectral channels and the intensity value of the backscattered or transmitted light.

9. A computer-readable storage medium, characterized in that, The medium contains a computer program, which, when run, executes the multispectral single-pixel imaging method of the pre-decoded virtual mode as described in any one of claims 1 to 7.

10. A computer device, characterized in that, It includes a processor and a storage medium, on which a computer program is stored, and the processor reads from the storage medium and runs the computer program to perform the multispectral single-pixel imaging method of the pre-decoded virtual mode as described in any one of claims 1 to 7.

Images