A method and apparatus for wavelength calibration in a computational spectral video imaging system.

Abstract

This invention discloses a wavelength calibration method for a computational spectral video imaging system, comprising: controlling a spectral lamp to output M sets of standard spectral lines within the imaging spectral range of the computational spectral video imaging system to a narrowband filter group; M being a positive integer; filtering the light beam using the narrowband filter group to obtain a single-spectral-line beam output to the computational spectral video imaging system; imaging using the computational spectral video imaging system to obtain a wavelength calibration mask image corresponding to the center wavelength of the filters; and performing wavelength calibration using the wavelength calibration mask image and the spectral calibration mask image obtained from the computational spectral video imaging system. This invention solves the wavelength calibration problem of high spatiotemporal-spectral resolution computational spectral video imaging systems.

Description

Technical Field

[0001] This invention relates to a wavelength calibration method and apparatus for computational spectral video imaging systems, which can be applied to the wavelength calibration of computational spectral video imaging systems to achieve high-precision spectral video detection of highly maneuverable targets. It belongs to the field of computational imaging technology. Background Technology

[0002] This invention addresses the urgent need for rapid acquisition of hyperspectral video in the field of high-mobility target detection and recognition. It solves the wavelength calibration problem for high spatiotemporal-spectral resolution spectral video by designing a wavelength calibration method and apparatus for computational spectral video imaging systems. Traditional high-precision spectrometer wavelength calibration methods use spectral lamps for wavelength calibration. This primarily involves the spectrometer directly detecting the spectral lamps. Different wavelengths of spectral lamps are dispersed and focused at different spectral line positions on the focal plane, thus obtaining the spectral curve of the lamp. Characteristic peaks are selected from the generated spectral curves, and the difference between the wavelength of this peak and the characteristic peak of the spectral lamp is calculated. This difference represents the original spectral calibration accuracy of the spectrometer. Using this difference for wavelength correction can improve the spectral accuracy of the spectrometer. Traditional methods are insufficient to meet the wavelength calibration requirements of computational spectral video imaging systems, necessitating the exploration of novel wavelength calibration methods. Furthermore, traditional methods suffer from difficulties in independently controlling multiple spectral lines and spectral calibration image aliasing, making it impossible to perform wavelength calibration for the spectrum of computational spectral video imaging systems. Summary of the Invention

[0003] To address the challenges of traditional wavelength calibration methods, such as the difficulty in independently controlling multiple spectral lines and the aliasing of spectral calibration images when used in computational spectral video imaging systems, this invention proposes a wavelength calibration method and apparatus for computational spectral video imaging devices. The method establishes a wavelength calibration apparatus and constructs a mathematical model for wavelength calibration analysis. The wavelength calibration apparatus provided by this invention can perform time-domain separation of multiple spectral lines and then directly measure and image the computational spectral video imaging system to obtain a clear spectral-encoded calibration image. Combined with the established mathematical model, correlation analysis is performed between the spectral calibration image and the wavelength calibration image to obtain wavelength error data and perform calibration. Based on this achievement, wavelength calibration of computational spectral video imaging systems can be realized, enabling high-precision acquisition of spectral video information.

[0004] The technical solution of this invention is: A wavelength calibration method and apparatus for a computational spectral video imaging system, including: a spectral lamp, a narrowband filter group, a computational spectral video imaging system, and an image acquisition and reconstruction system.

[0005] The spectral lamp outputs multiple sets of characteristic spectral lines within the working spectrum band, which are then output to the narrowband filter group. The narrowband filter group receives light radiation from the spectral lamp, containing multiple sets of characteristic spectral line information. After passing through the narrowband filter, the output of the permissible characteristic spectral lines is provided to the computational spectral video imaging system. The computational spectral video imaging system receives light radiation after it has passed through a narrowband filter, encodes it, and then performs image acquisition by aliasing it onto the detector image plane before outputting it to the image acquisition system.

[0006] The image acquisition system receives aliased image information, performs similarity analysis with the calibration matrix of the computational spectral video imaging system, and performs wavelength calibration on the computational spectral video imaging system.

[0007] The spectral line lamp is used in combination with a narrowband filter group to output multiple sets of specific characteristic spectral lines, preventing the influence of stray light radiation outside the spectral lines.

[0008] A mathematical model for correlation analysis is established, and a similarity analysis is performed between the aliased images received by the image acquisition system and the calibration matrix of the computational spectral video imaging system to obtain the wavelength calibration accuracy of the computational spectral video imaging system.

[0009] The advantages of this invention compared to existing technologies are: 1) Compared with traditional wavelength calibration methods, this invention can independently control multiple spectral lines output by the spectral lamp, thereby avoiding spectral aliasing in the acquired image, providing single-band radiation for computational spectral video imaging systems, and acquiring single-band wavelength calibration images.

[0010] 2) Compared with traditional wavelength calibration methods, this invention establishes a mathematical model for correlation analysis of calibration images, and performs similarity analysis between the aliased images received by the image acquisition system and the calibration matrix of the computational spectral video imaging system, thereby performing high-precision wavelength calibration of the computational spectral video imaging system. Attached Figure Description

[0011] Figure 1 A schematic diagram of the wavelength calibration device setup; Figure 2 This is a schematic diagram of the computational spectral video imaging system of the present invention; Figure 3 A schematic diagram of the characteristic spectral lines of a spectral lamp (mercury-neon lamp). Figure 4 This is a mask image for spectral calibration and wavelength calibration of the present invention; Figure 5 This is a schematic diagram of the calibration accuracy analysis results using different spectral lines in this invention (normalized display). Detailed Implementation

[0012] Addressing the urgent need for rapid acquisition of hyperspectral video in the field of highly maneuverable target detection and recognition, and considering the difficulties of traditional wavelength calibration methods such as the inability to independently control multiple spectral lines and the aliasing of spectral calibration images when used in computational spectral video imaging systems, this invention proposes a wavelength calibration method and apparatus for computational spectral video imaging devices. The invention establishes a wavelength calibration apparatus and constructs a mathematical model for wavelength calibration analysis. The wavelength calibration apparatus provided by this invention can perform time-domain separation of multiple spectral lines and then directly measure and image the computational spectral video imaging system to obtain a clear spectral-coded calibration image. Combined with the established mathematical model, correlation analysis is performed between the spectral calibration image and the wavelength calibration image to obtain wavelength error data and perform calibration. Based on this achievement, wavelength calibration of computational spectral video imaging systems can be realized, achieving high-precision acquisition of spectral video information.

[0013] The working principle and process of the present invention will be further explained and described below with reference to the accompanying drawings.

[0014] The wavelength calibration device provided by this invention can perform time-domain separation of multiple spectral lines and then directly measure and image a computational spectral video imaging system to obtain a clear spectral coding calibration image. Combined with the established mathematical model, the correlation between the spectral calibration image and the wavelength calibration image is analyzed to obtain wavelength error data and perform calibration.

[0015] like Figure 1 As shown, a wavelength calibration method and apparatus for a computational spectral video imaging system are characterized by comprising: a spectral lamp, a narrowband filter group, a computational spectral video imaging system, and an image acquisition and reconstruction system.

[0016] A wavelength calibration system was built by combining a computational spectral video imaging device with a standard spectral line lamp. Standard spectral lines at 507.304 nm, 546.075 nm, 585.249 nm, 640.225 nm, and 650.653 nm were selected for wavelength calibration.

[0017] The spectral lamp outputs multiple sets of characteristic spectral lines within the working spectrum band, which are then output to the narrowband filter group. The narrowband filter group receives light radiation from the spectral lamp, containing multiple sets of characteristic spectral line information. After passing through the narrowband filter, the output of the permissible characteristic spectral lines is provided to the computational spectral video imaging system. The computational spectral video imaging system receives light radiation after it has passed through a narrowband filter, encodes it, and then performs image acquisition by aliasing it onto the detector image plane before outputting it to the image acquisition system.

[0018] The image acquisition system receives aliased image information, performs similarity analysis with the calibration matrix of the computational spectral video imaging system, and performs wavelength calibration on the computational spectral video imaging system.

[0019] Compared with traditional wavelength calibration methods, this invention can independently control multiple spectral lines output by the spectral lamp, thereby avoiding spectral aliasing in the acquired image and providing single-band radiation for computational spectral video imaging systems to acquire single-band wavelength calibration images.

[0020] Compared with traditional wavelength calibration methods, this invention establishes a mathematical model for correlation analysis of calibration images, and performs similarity analysis between the aliased images received by the image acquisition system and the calibration matrix of the computational spectral video imaging system, thereby performing high-precision wavelength calibration of the computational spectral video imaging system.

[0021] This invention expands the spectral range by selecting the appropriate main optical system and detector for the spectral range, and the technologies involved are applicable to ultraviolet, visible, infrared and other spectral ranges.

[0022] like Figure 2 As shown, the schematic diagram of the characteristic spectral lines emitted by the spectral line lamp of the present invention, taking the spectral line diagram of a mercury-neon lamp as an example, shows standard spectral lines such as 507.304nm, 546.075nm, 585.249nm, 640.225nm, and 650.653nm that can be used for wavelength calibration tests.

[0023] like Figure 3 As shown, the characteristic spectral bands emitted by the spectral lamp are controlled by narrowband filters, resulting in a single spectral line output. By placing narrowband filters with different center wavelengths (such as 505±10nm, 545±10nm, 585±10nm, 640±10nm, and 650±10nm) in front of the spectral lamp, the combination of the spectral lamp and the filters suppresses the influence of out-of-band spectral lines, resulting in only a single spectral line output. like Figure 4 As shown, a standard spectral lamp, after controlling the output of a single spectral line, illuminates the computational spectral video imaging system. The image acquisition and reconstruction system then captures an image at that wavelength. A similarity comparison analysis is performed between this image and the spectral calibration mask image. The spectral calibration mask image is obtained by the system through wavelength-by-wavelength calibration using a monochromator. When the monochromator illuminates the computational spectral video imaging system, the position of the image on the encoder plate on the focal plane shifts in a certain direction as the wavelength moves. Only at a specific wavelength will the image on the encoder plate reproduce the image illuminated by the spectral lamp.

[0024] like Figure 5 The diagram shows the results of wavelength calibration analysis using different spectral lines. The spectral calibration mask image obtained during monochromator calibration is compared with the wavelength calibration mask image obtained using a standard spectral lamp. The spectral distribution of their similarity is analyzed, and the wavelength positions of the similarity peaks are recorded. Wavelength calibration is performed based on the difference between the peak wavelength positions of the calibration matrix and the spectral line positions of the standard spectral lamp.

[0025] By registering images and determining their correlation, the system automatically selects resolvable characteristic spectral images of single pixels within the spectral band (images with the highest contrast, which are beneficial for improving the accuracy of spectral reconstruction), and finally forms a complete and realistic calibration database for high-resolution spectral image reconstruction in ground scene imaging experiments.

[0026] Based on this principle, the coded template calibration images at different wavelength positions obtained from the monochromator calibration are calculated and analyzed through image comparison. (j ranges from 1 to N, where N is the spectral width in nm) and images of characteristic peaks illuminated by spectral lamps. Differences Minimum difference value The image with the highest similarity is identified as the same image, and the wavelength of this image reflects the corresponding wavelength image under the monochromator wavelength calibration.

[0027]

[0028] In the formula, m and n are the number of detector pixels; Let be the j-th image; i is the cyclic factor for the number of pixels traversed by the detector.

[0029] The contents not described in detail in this specification are common knowledge to those skilled in the art.

Claims

1. A wavelength calibration method for a computational spectral video imaging system, characterized in that, include: The control spectral lamp outputs M sets of standard spectral lines within the imaging spectral range of the calculated spectral video imaging system to the narrowband filter group; M is a positive integer; The beam is filtered using a narrowband filter array to obtain a beam with a single spectral line, which is then output to the computational spectral video imaging system. Imaging was performed using a computational spectral video imaging system to obtain a wavelength calibration mask image corresponding to the center wavelength of the filter. Wavelength calibration is performed using wavelength calibration mask images and spectral calibration mask images obtained from a computational spectral video imaging system.

2. The wavelength calibration method for a computational spectral video imaging system according to claim 1, characterized in that, Each set of standard spectral lines is equipped with a narrowband filter corresponding to the center wavelength.

3. The wavelength calibration method for a computational spectral video imaging system according to claim 2, characterized in that, M is 5; Narrowband filters include: narrowband filters with a center wavelength of 505±10nm, narrowband filters with a center wavelength of 545±10nm, narrowband filters with a center wavelength of 585±10nm, narrowband filters with a center wavelength of 640±10nm, and narrowband filters with a center wavelength of 650±10nm.

4. A wavelength calibration method for a computational spectral video imaging system according to claim 2 or 3, characterized in that, The spectral calibration mask image is obtained by wavelength-by-wavelength calibration using a computational spectral video imaging system.

5. The wavelength calibration method for a computational spectral video imaging system according to claim 1, characterized in that, The method for wavelength calibration is as follows: Based on the wavelength calibration mask image and the spectral calibration mask image obtained from the spectral video imaging system, determine the difference value. When the wavelength is minimized, the corresponding j-th wavelength calibration mask image is obtained; Obtain the standard spectral wavelength corresponding to the j-th wavelength calibration mask image, determine the difference between the monochromator wavelength and the standard spectral wavelength as the wavelength calibration error, take the standard spectral wavelength as the true value, and use the wavelength calibration error to correct the wavelength of the monochromator calibration image.

6. The wavelength calibration method for a computational spectral video imaging system according to claim 5, characterized in that, Difference value The method for determining this is as follows: Where m and n are the number of pixel rows and columns of the detector in the computational spectral video imaging system, respectively; The wavelength calibration mask image obtained when calculating the spectral video imaging system for standard spectral line incidence of a spectral lamp. When calibrating the monochromator, calculate the pixel value at pixel position i in the j-th wavelength calibration mask image acquired by the spectral video imaging system; j∈[1,N], N is the spectral width, which is the spectral range of ±3nm centered on the standard spectral line; i corresponds to the detector pixel position, and its value ranges from 1 to m×n.

Images