Search

CN-122016047-A - Multimode resonance photon calculation spectrum imaging method and architecture

CN122016047ACN 122016047 ACN122016047 ACN 122016047ACN-122016047-A

Abstract

The invention provides a multimode resonance photon calculation spectrum imaging method and a multimode resonance photon calculation spectrum imaging framework, which comprise the steps of carrying out wavelength selection and intensity modulation on wide-spectrum incident light through a composite transmittance curve generated by interlayer light field coupling in a multilayer optical resonant cavity filter to obtain spatially heterogeneous coded filter light, receiving the spatially heterogeneous coded filter light by using a planar array image sensor, converting an optical signal of the filter light into original image data containing spectrum coding information, determining a composite transmittance matrix corresponding to the composite transmittance curve, and carrying out spectrum reconstruction based on the original image data and the composite transmittance matrix to obtain target hyperspectral image data. According to the method, the multi-layer optical resonant cavity filter is used for generating the compound transmittance curve with multiple peaks, non-period and narrow bandwidth, so that the number of spectrum channels is increased, a wide spectrum can be covered without switching the filter, the imaging efficiency is improved, meanwhile, the limited equidistant requirement is met, and the reconstruction quality and reliability of spectrum data are ensured.

Inventors

  • FANG LU
  • LI SHANGLONG

Assignees

  • 清华大学

Dates

Publication Date
20260512
Application Date
20260126

Claims (10)

  1. 1. A method of multimode resonant photon computational spectral imaging, comprising: performing wavelength selection and intensity modulation on wide-spectrum incident light through a composite transmittance curve generated by interlayer light field coupling in the multilayer optical resonant cavity filter to obtain spatially heterogeneous coded filter light; receiving the spatially heterogeneously encoded filtered light with an area array image sensor and converting an optical signal of the filtered light into raw image data comprising spectrally encoded information; Determining a composite transmittance matrix corresponding to the composite transmittance curve, and performing spectrum reconstruction based on the original image data and the composite transmittance matrix to obtain target hyperspectral image data.
  2. 2. The method of claim 1, wherein the multilayer optical resonant cavity filter comprises at least two layers of fabry-perot resonant cavities stacked along the optical axis direction, each layer of fabry-perot resonant cavity comprises an upper reflector and a lower reflector which are formed by multilayer dielectric films and a dielectric cavity layer therebetween, and a dielectric isolation layer is arranged between every two adjacent layers of fabry-perot resonant cavities and used for regulating and controlling interlayer coupling strength.
  3. 3. The method of claim 2, wherein the material of the dielectric cavity layer is determined based on a target spectral range.
  4. 4. The method of claim 3, wherein determining the material of the dielectric cavity layer according to the target spectral range comprises determining the material of the dielectric cavity layer to be SiO 2 or TiO 2 if the target spectral range is a visible light range and/or determining the material of the dielectric cavity layer to be Si 3 N 4 if the target spectral range is a near infrared range.
  5. 5. The method of claim 2, wherein the spatially heterogeneous distribution of the composite transmittance profile is achieved by introducing process tolerances during the fabrication of the multilayer optical resonator filter.
  6. 6. The method of claim 1, wherein determining the composite transmittance matrix corresponding to the composite transmittance curve comprises: Preprocessing the composite transmittance curve to obtain a standardized transmittance curve; for each spatial position index m and each discrete wavelength index n, obtaining a transmittance value at a corresponding wavelength from the normalized transmittance curve; And assigning the transmittance value to an nth element of an mth row in the composite transmittance matrix to obtain the composite transmittance matrix.
  7. 7. The method of claim 1, wherein performing spectral reconstruction based on the original image data and the composite transmittance matrix to obtain target hyperspectral image data comprises obtaining target hyperspectral image data by a spectral compressed sensing iterative solution algorithm based on the original image data and the composite transmittance matrix, wherein the spectral compressed sensing iterative solution algorithm is any one of an orthogonal matching pursuit OMP, a sparsity adaptive matching pursuit SAMP, an alternating direction multiplier method ADMM, or a gradient projection sparse reconstruction GPSR algorithm.
  8. 8. The method of claim 1, wherein performing spectral reconstruction based on the raw image data and the composite transmittance matrix to obtain target hyperspectral image data comprises inputting the raw image data and the composite transmittance matrix into a target hyperspectral neural network model to obtain target hyperspectral image data, wherein the target hyperspectral neural network model is an MST network or SPECAT network.
  9. 9. A multimode resonant photon computed spectral imaging device, the device comprising: the multi-layer optical resonant cavity filter module is used for carrying out wavelength selection and intensity modulation on wide-spectrum incident light through a composite transmittance curve generated by light field coupling among layers of the multi-layer optical resonant cavity filter to obtain spatially heterogeneous coded filter light; The imaging detection module is used for receiving the filtered light of the space heterogeneous coding by using an area array image sensor and converting an optical signal of the filtered light into original image data containing spectrum coding information; And the photon calculation processing module is used for determining a composite transmittance matrix corresponding to the composite transmittance curve, and carrying out spectrum reconstruction based on the original image data and the composite transmittance matrix to obtain target hyperspectral image data.
  10. 10. A multimode resonance photon computing spectral imaging architecture, wherein the multimode resonance photon computing spectral imaging architecture comprises a multimode resonance photon computing spectral imaging device.

Description

Multimode resonance photon calculation spectrum imaging method and architecture Technical Field The disclosure relates to the technical field of spectrum imaging, in particular to a multimode resonance photon calculation spectrum imaging method and a multimode resonance photon calculation spectrum imaging architecture. Background The spectrum imaging technology realizes three-dimensional information characterization of space two-dimensional and spectrum one-dimensional by synchronously capturing the space image and continuous spectrum information of the target, and has irreplaceable application value in multiple fields. For example, in environmental monitoring, a broad spectrum range from ultraviolet (300-400 nm) to near infrared (760-1100 nm) needs to be covered to simultaneously identify trace Volatile Organic Compounds (VOCs) in the atmosphere and chlorophyll concentration in water, in the biomedical field, the technology can capture fine spectrum differences between tumor tissues and normal tissues and provide key information for unmarked pathological diagnosis, and in high-end industrial quality inspection, the composition and purity detection of semiconductor materials and precise optical elements also needs to be combined with a broad spectrum range and high resolution. In the spectrum imaging system, the filter element is a core component for realizing spectrum selection and resolution, and the performance of the filter element directly determines the spectrum resolution and luminous flux efficiency of the system, thereby affecting the imaging speed and information precision. Specifically, the Fabry-Perot Resonator Filter resonant cavity filter has the characteristics of relatively simple structure, easy tuning of the center wavelength of a passband, design of the bandwidth in a certain range, low manufacturing cost and the like, and becomes one of the commonly used filter devices in a computational spectrum imaging system. The fundamental principle of the Fabry-Perot resonant cavity filter is based on a resonant cavity formed by two parallel high-reflection mirrors, incident light generates multi-beam interference in the cavity, and a high-transmittance peak is formed only near a specific wavelength meeting resonance conditions, so that a spectrum screening function is realized. However, when the optical filter based on the single-layer Fabry-Perot resonant cavity in the related technology is applied to high-precision and high-efficiency calculation spectrum imaging, the single-layer Fabry-Perot cavity forms a single-peak or periodic transmittance curve, only 1-2 spectrum channels of information can be acquired in single imaging, based on the single-layer Fabry-Perot resonant cavity, the optical filter needs to be switched for multiple times to cover a wide spectrum range, so that the imaging efficiency is low and the dynamic scene adaptability is poor, the full width at half maximum (FWHM) of the single-layer Fabry-Perot cavity filter is limited by the fineness of a cavity, the transmission peak of the single-layer Fabry-Perot cavity filter is usually more than 15 nm, the target components with slight differences of spectrum characteristics (such as different types of biological tissues, similar pollutants of chemical components and the like) are difficult to accurately distinguish, the detection accuracy is limited, the single-layer Fabry-Perot cavity generates a single or periodic simple transmission curve, the column vector has relatively high coherence, the RIP condition is difficult to meet, the measurement matrix shows the pathological characteristics, the reconstruction quality and reliability of spectrum data are reduced, the transmission peaks with different wavelengths are easy to appear in the wide spectrum range, and the spectrum identification spectrum accuracy is reduced. Based on this, a multimode resonance photon calculation spectrum imaging method and architecture are needed to realize efficient and high-resolution spectrum imaging, and ensure the reconstruction quality and reliability of spectrum data. Disclosure of Invention The present disclosure aims to solve, at least to some extent, one of the technical problems in the related art. Therefore, a first object of the present disclosure is to provide a multi-mode resonance photon calculation spectrum imaging method, which generates a compound transmittance curve with multiple peaks, non-period and narrow bandwidth through a multi-layer optical resonant cavity filter, improves the number of spectrum channels, can cover a wide spectrum without switching the filter, improves imaging efficiency, determines a corresponding compound transmittance matrix for the transmittance of light waves with different wavelengths through a multi-layer optical resonant cavity represented by the compound transmittance curve, meets the requirement of limited equidistance, reduces reconstruction errors, and ensures reconstruction quality and reliabilit