A quantum imaging method based on Gaussian filtering optimization

Abstract

The present invention proposes a quantum imaging method based on Gaussian filtering optimization, and Gaussian filtering is performed on the coincidence count value to improve the quality of quantum imaging. First, a laser is used to generate pump light, and the pump light is irradiated to a periodically polarized potassium titanyl phosphate (PPKTP) crystal through lens beam expansion and half-wave plate and quarter-wave plate modulation, and a reference photon and a signal photon with entangled characteristics are generated through the spontaneous parametric down-conversion process of PPKTP; then, a surface detector is used to detect the reference photon, and the signal photon is irradiated to the target to be imaged, and a bucket detector is used to detect the signal photon; secondly, the time series recorded by the two detectors are counted to obtain the coincidence count value at the corresponding image pixel position; thirdly, the coincidence count value is filtered and optimized through Gaussian filtering and wavelet transform; finally, a quantum image is obtained according to the processed coincidence count value.

Description

Technical Field

[0001] The present invention belongs to the field of quantum information technology, and specifically in the process of quantum imaging based on entangled light source, a Gaussian filtering method is used to optimize the coincidence count value to improve the quantum imaging quality. Background Art

[0002] In recent years, the combination of quantum mechanics, information science and computer science has led to a boom in the research of quantum information technology. Based on the principles of quantum superposition and quantum entanglement, quantum information technology is expected to break through the bottleneck of traditional technology, solve problems that cannot be solved by classical information technology and conventional computers, massively improve computer operation speed, ensure absolute security of information communication, and increase measurement accuracy by orders of magnitude. Quantum entanglement is a core concept in quantum information technology. It reveals the non-locality of state functions of two or more particles at a distance, and is one of the basic characteristics that distinguish quantum mechanics from classical mechanics.

[0003] Traditional geometric optical imaging is based on the point-to-point conjugate relationship between the object and the image, while quantum imaging uses the spatial correlation characteristics of entangled photons to perform imaging, which has non-local characteristics. The imaging optical path consists of a signal optical path and a reference optical path. There is only a barrel-shaped detector without imaging function in the signal optical path. Although there is a planar array detector for imaging in the reference optical path, it does not contain any object information and cannot obtain imaging results based on only one signal. Imaging can only be performed by counting the coincidences of the two signals. It is precisely because this imaging method goes against conventional thinking habits that it was called "ghost imaging" when it was discovered.

[0004] In addition to the characteristic of non-locality, quantum imaging has several other important characteristics: first, the resolution of quantum imaging can break through the diffraction limit of classical imaging; second, quantum imaging is not affected by optical path disturbances and has the ability to resist atmospheric turbulence; third, incoherent light sources can be used for coherent imaging. Quantum imaging based on entangled light counts the photons in the signal optical path and the reference optical path, and obtains imaging results based on the coincidence count values. Factors such as ambient light will introduce abnormal coincidence counts, which will in turn cause image quality to deteriorate. Therefore, improving imaging quality by filtering out noise has important research value in the field of quantum imaging based on entangled light. With the development of science and technology, quantum imaging technology has broad application prospects in many aspects such as all-day and all-time target recognition, underwater detection, terrain mapping, satellite and ground remote sensing, etc. Summary of the invention

[0005] The purpose of the present invention is to provide a quantum imaging method based on Gaussian filtering optimization. The present invention uses the Gaussian filtering method to perform smoothing filtering on the coincidence count value generated during the imaging process, and then performs wavelet transformation on it, performs Gaussian smoothing filtering on the high-frequency wavelet coefficients to filter out noise, and finally performs inverse transformation on the wavelet coefficients to obtain the optimized coincidence count value, and performs quantum imaging based on the coincidence count value.

[0006] The specific steps are as follows:

[0007] Step 1: Generate pump light using a semiconductor laser with a wavelength of 405nm;

[0008] Step 2: Use lenses with focal lengths of 100 mm and 50 mm to form a beam reduction combination to reduce the pump light and make the energy more concentrated. At the same time, use a half-wave plate (HWP) and a quarter-wave plate (QWP) with a transmission wavelength of 405 nm to modulate the pump light into linear polarization.

[0009] Step 3: The modulated pump light is irradiated on a periodically polarized potassium titanyl phosphate (PPKTP) crystal, and the photons in the beam undergo parametric down-conversion with a certain probability to obtain entangled two-photon pairs;

[0010] Step 4: Use a lens with a focal length of 25mm and a lens with a focal length of 35mm to expand the entangled light beam, and use a polarization beam splitter with a wavelength of 810nm to separate the entangled photon pairs. The ones that pass through the polarization beam splitter are signal photons, and the ones that are reflected are reference photons. The signal photons and reference photons have the same light intensity I and satisfy:

[0011]

[0012] The symbol “∝” indicates proportional to, d is the length of the PPKTP crystal, c is the speed of light, and Δk is the phase mismatch, satisfying:

[0013]

[0014] Among them, k p represents the pump light wave vector, k i represents the reference light wave vector, k s represents the signal light wave vector, Λ is the polarization period, and 2π / Λ is the grating wave vector;

[0015] Step 5: The signal photons are irradiated to the target to be imaged through a lens with a focal length of 175 mm, then pass through a lens with a focal length of 125 mm, and then pass through a narrow-band filter with a wavelength of 810 nm before being collected by the barrel detector;

[0016] Step 6: After the reference photon passes through the reflection silver mirror, it is incident on the digital micromirror device, then passes through a lens with a focal length of 125 mm, and then passes through a narrow-band filter with a wavelength of 810 nm before being collected by the surface detector;

[0017] Step 7: Collect two level pulse signals with a certain delay to obtain two time series, and mark the reference optical path CH1 and the signal optical path CH2 according to different flag positions;

[0018] Step 8: Set the coincidence gate width W, perform coincidence counting on the two sequences of CH1 and CH2, and obtain the coincidence count value C(x, y) at the corresponding image pixel position;

[0019] Step 9: Use the Gaussian filter method to perform smoothing filtering on the coincidence count value C(x,y), where the Gaussian filter template size is set to 3×3, and the coincidence count value after processing is:

[0020]

[0021] Where, σ is the standard deviation (assuming σ = 1.6);

[0022] Step 10: Decompose the processed coincidence count value using a two-dimensional discrete wavelet transform method to obtain wavelet coefficients, and perform smoothing filtering on the high-frequency wavelet coefficients where the noise is located using the Gaussian filtering method in step 9;

[0023] Step 11: Perform wavelet reconstruction on the low-frequency wavelet coefficients and the processed high-frequency wavelet coefficients to obtain the denoised coincidence count value, select the maximum coincidence count value as the grayscale value 255 and the minimum count value as the grayscale value 0, and use the linear mapping method to map the coincidence count value to the grayscale value to obtain the final quantum image. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure (1) is a schematic diagram of the quantum imaging optical path structure of the present invention;

[0025] Figure (2) is a schematic diagram of the coincidence counting principle of the present invention;

[0026] Figure (3) is a schematic diagram of the decomposition and reconstruction process of the two-dimensional discrete wavelet transform of the present invention. Specific implementation plan

[0027] The present invention is further described in detail below in conjunction with the accompanying drawings:

[0028] Step 1: Generate pump light using a semiconductor laser with a wavelength of 405nm;

[0029] Step 2: Use lenses with focal lengths of 100 mm and 50 mm to form a beam reduction combination to reduce the pump light and make the energy more concentrated. At the same time, use a half-wave plate (HWP) and a quarter-wave plate (QWP) with a transmission wavelength of 405 nm to modulate the pump light into linear polarization.

[0030] Step 3: The modulated pump light is irradiated on a periodically polarized potassium titanyl phosphate (PPKTP) crystal, and the photons in the beam undergo parametric down-conversion with a certain probability to obtain entangled two-photon pairs;

[0031] Step 4: Use a lens with a focal length of 25mm and a lens with a focal length of 35mm to expand the entangled light beam, and use a polarization beam splitter with a wavelength of 810nm to separate the entangled photon pairs. The ones that pass through the polarization beam splitter are signal photons, and the ones that are reflected are reference photons. The signal photons and reference photons have the same light intensity I and satisfy:

[0032]

[0033] The symbol “∝” indicates proportional to, d is the length of the PPKTP crystal, c is the speed of light, and Δk is the phase mismatch, satisfying:

[0034]

[0035] Among them, k p represents the pump light wave vector, k i represents the reference light wave vector, k s represents the signal light wave vector, Λ is the polarization period, and 2π / Λ is the grating wave vector;

[0036] Step 5: The signal photons are irradiated to the target to be imaged through a lens with a focal length of 175 mm, then pass through a lens with a focal length of 125 mm, and then pass through a narrow-band filter with a wavelength of 810 nm before being collected by the barrel detector;

[0037] Step 6: After the reference photon passes through the reflection silver mirror, it is incident on the digital micromirror device, then passes through a lens with a focal length of 125 mm, and then passes through a narrow-band filter with a wavelength of 810 nm before being collected by the surface detector;

[0038] Step 7: Collect two level pulse signals with a certain delay to obtain two time series, and mark the reference optical path CH1 and the signal optical path CH2 according to different flag positions;

[0039] Step 8: Set the coincidence gate width W, perform coincidence counting on the two sequences of CH1 and CH2, and obtain the coincidence count value C(x, y) at the corresponding image pixel position;

[0040] Step 9: Use the Gaussian filter method to perform smoothing filtering on the coincidence count value C(x,y), where the Gaussian filter template size is set to 3×3, and the coincidence count value after processing is:

[0041]

[0042] Where, σ is the standard deviation (assuming σ = 1.6);

[0043] Step 10: Decompose the processed coincidence count value using a two-dimensional discrete wavelet transform method to obtain wavelet coefficients, and perform smoothing filtering on the high-frequency wavelet coefficients where the noise is located using the Gaussian filtering method in step 9;

[0044] Step 11: Perform wavelet reconstruction on the low-frequency wavelet coefficients and the processed high-frequency wavelet coefficients to obtain the denoised coincidence count value, select the maximum coincidence count value as the grayscale value 255 and the minimum count value as the grayscale value 0, and use the linear mapping method to map the coincidence count value to the grayscale value to obtain the final quantum image.

Claims

1. A quantum imaging method based on Gaussian filtering optimization, characterized in that The following steps are involved: Step 1: Generate pump light using a semiconductor laser with a wavelength of 405nm; Step 2: Use lenses with focal lengths of 100 mm and 50 mm to form a beam reduction combination to reduce the pump light and make the energy more concentrated. At the same time, use a half-wave plate (HWP) and a quarter-wave plate (QWP) with a transmission wavelength of 405 nm to modulate the pump light into linear polarization. Step 3: The modulated pump light is irradiated on a periodically polarized potassium titanyl phosphate (PPKTP) crystal, and the photons in the beam undergo parametric down-conversion with a certain probability to obtain entangled two-photon pairs; Step 4: Use a lens with a focal length of 25 mm and a lens with a focal length of 35 mm to expand the entangled light beam, and use a polarization beam splitter with a wavelength of 810 nm to separate the entangled photon pairs. The photons passing through the polarization beam splitter are signal photons, and the photons reflected are reference photons. The signal photons and the reference photons have the same light intensity I and satisfy: The symbol "∝" indicates proportional to, d is the length of the PPKTP crystal, c is the speed of light, and Δk is the phase mismatch, satisfying: Among them, k p represents the pump light wave vector, k i represents the reference light wave vector, k s represents the signal light wave vector, Λ is the polarization period, and 2π / Λ is the grating wave vector; Step 5: The signal photons are irradiated to the target to be imaged through a lens with a focal length of 175 mm, then pass through a lens with a focal length of 125 mm, and then pass through a narrow-band filter with a wavelength of 810 nm before being collected by the barrel detector; Step 6: After the reference photon passes through the reflection silver mirror, it is incident on the digital micromirror device, then passes through a lens with a focal length of 125 mm, and then passes through a narrow-band filter with a wavelength of 810 nm before being collected by the surface detector; Step 7: Collect two level pulse signals with a certain delay to obtain two time series, and mark the reference optical path CH1 and the signal optical path CH2 according to different flag positions; Step 8: Set the coincidence gate width W, perform coincidence counting on the two sequences of CH1 and CH2, and obtain the coincidence count value C(x, y) at the corresponding image pixel position; Step 9: Use the Gaussian filter method to perform smoothing filtering on the coincidence count value C(x,y), where the Gaussian filter template size is set to 3×3, and the coincidence count value after processing is: Where, σ is the standard deviation (assuming σ = 1.6); Step 10: Decompose the processed coincidence count value using a two-dimensional discrete wavelet transform method to obtain wavelet coefficients, and perform smoothing filtering on the high-frequency wavelet coefficients where the noise is located using the Gaussian filtering method in step 9; Step 11: Perform wavelet reconstruction on the low-frequency wavelet coefficients and the processed high-frequency wavelet coefficients to obtain the denoised coincidence count value, select the maximum coincidence count value as the grayscale value 255 and the minimum count value as the grayscale value 0, and use the linear mapping method to map the coincidence count value to the grayscale value to obtain the final quantum image.

Images