A multi-target detection method based on entangled light quantum

Abstract

This invention proposes a multi-target detection method based on entangled photons. First, a pump light is generated using a continuous narrow-linewidth semiconductor laser, which, after polarization modulation and focusing, illuminates a PPKTP crystal. A spontaneous parametric down-conversion process generates polarization-entangled photon pairs, which are then separated into signal and reference light. These are received by a single-photon detector to obtain two time series. Second, coincidence measurements are performed on the time series in different scanning directions within the observation area to obtain a coincidence count matrix. Then, a two-dimensional constant false alarm rate (CFAR) detection algorithm is introduced, using ordered statistics to estimate background noise and adaptively adjust the detection threshold to screen candidate target points. Third, a density clustering algorithm based on correlation strength weighting is employed, combining the correlation strength and spatial distribution of entangled photon pairs to detect different targets and suppress isolated noise. Finally, the coincidence count peak points within each target cluster are selected as the final detection targets, thereby achieving multi-target detection with unknown numbers and randomly distributed locations.

Description

Technical Field

[0001] This invention relates to the field of quantum precision measurement, and in particular to a multi-target detection method based on entangled photons, which is applicable to the detection of multiple targets with unknown numbers and randomly distributed positions in non-cooperative scenarios. Background Technology

[0002] Optical target detection technology achieves accurate target detection by collecting and analyzing the echo signals reflected from targets, playing a core role in key fields such as autonomous driving, disaster relief, and environmental remote sensing. However, limited by the shot noise limit of classical physics, traditional optical detection technologies are easily overwhelmed by noise when facing distant, weak targets or strong background light interference, resulting in a sharp drop in detection sensitivity and making it difficult to meet the high-precision detection requirements in complex environments.

[0003] Quantum detection technology, based on the principles of quantum mechanics, utilizes the time correlation characteristics of entangled photon pairs to significantly suppress uncorrelated background noise, thereby achieving a signal-to-noise ratio gain that surpasses the classical limit. Theoretical studies have shown that quantum detection systems can effectively overcome the physical limitations of traditional detection methods in terms of power and signal-to-noise ratio, achieving high-sensitivity target detection even under low photon count conditions. Furthermore, thanks to the non-local correlation of entangled photon pairs, this technology possesses inherent anti-interference capabilities and stealth advantages, effectively resisting environmental noise interference and demonstrating enormous development potential in scenarios requiring high detection accuracy.

[0004] Current research on quantum target detection largely focuses on single-target detection, with relatively few studies on non-cooperative multi-target detection where the number and orientation of multiple targets are unknown. In multi-target scenarios with unknown numbers and randomly distributed locations, echo signals from neighboring targets easily interfere with each other, leading to reduced detection accuracy. Furthermore, environmental noise and signal broadening can easily be misidentified as real targets, increasing the false alarm rate and making it difficult to meet the requirements of high-precision detection. To address these issues, this paper proposes a multi-target detection method based on entangled photons, aiming to effectively suppress false alarms while improving the detection probability of multiple targets. Summary of the Invention

[0005] The purpose of this invention is to propose a multi-target detection method based on entangled photons. First, a two-dimensional constant false alarm rate (CFAR) detection model based on coincidence count is established, and the detection threshold is dynamically adjusted according to background noise to initially screen candidate target points. Then, to address the signal broadening problem caused by transmission medium scattering and detector jitter, a density clustering algorithm based on correlation strength weighting is proposed. This algorithm combines the correlation strength and spatial distribution of entangled photon pairs to accurately detect different targets and suppress isolated noise. Finally, the coincidence count peak points within each target cluster are selected as targets, achieving multi-target detection with unknown numbers and randomly distributed locations.

[0006] The technical solution adopted in this invention is: a multi-target detection method based on entangled photons, comprising the following specific steps:

[0007] Step 1: Use a continuous narrow-linewidth semiconductor laser with a bandwidth of 160MHz to generate pump light with a wavelength of 405nm, and adjust it into horizontally linearly polarized light through a half-wave plate, a quarter-wave plate and a polarization beam splitter.

[0008] Step 2: The horizontally linearly polarized light is sequentially modulated into 45° linearly polarized light by passing it through a half-wave plate and a quarter-wave plate, and then focused onto a PPKTP crystal by a lens. A pair of entangled photons with a wavelength of 810nm is generated through a spontaneous parametric downconversion process.

[0009] Step 3: Pass the entangled photon pair through a 50:50 dual-wavelength polarization beam splitter to separate the photon pair into a signal photon and a reference photon, resulting in signal light and reference light with equal intensity.

[0010] Step 4: Pass the reference light and signal light through photonic couplers respectively. The reference light is directly received locally by the single photon detector (SPD) 1 to generate a time pulse sequence. ; signal photons from to Scanning multiple targets, the signals reflected back to the local unit are received by the SPD2 to obtain another time pulse sequence. The time series obtained by the single-photon detectors in the reference optical path and the signal optical path are respectively expressed as follows:

[0011]

[0012] in, and These are the reference optical path and the signal optical path, respectively. and the The arrival time of each photon and This represents the total number of detected photons in that direction;

[0013] Step 5: Set the door width to be compatible. ,Will Delay Then, the time sequence of the reference optical path after the delay is calculated. and The coincidence count between the two values, when there exists a pair of photons that satisfy the condition... If the event is true, it is recorded as a valid event, and the validity count is incremented by 1; otherwise, it is not counted.

[0014] Step 6: Measure the coincidence count value in each scanning direction, and let... The total number of sampling points in the time delay dimension. Let be the total number of sampling points in the azimuth dimension, then the conformance counting matrix is... The definition is as follows:

[0015]

[0016] in, Indicates the first The delay unit and the first The coincidence count intensity at each azimuth unit;

[0017] Step 7: Convert the matrix As the input signal, the unit to be detected Construct a two-dimensional sliding window centered on the element, and set the size of the inner protective window to [value]. The outer reference window size is The annular region between the two windows serves as a reference cell for estimating background noise; the total number of cells in this region is [number missing]. ;

[0018] Step 8: Perform symmetric mirror filling on the matrix. Based on the reference window size, set the time-delay dimension fill half-width to [value missing]. The orientation dimension fill half-width is Construct an extended matrix Any position in this matrix element value ,in and These are the row and column indices mapped to the original matrix, respectively. For the row index... According to the current location With time delay dimension parameters and The relationship is calculated according to the following rules:

[0019]

[0020] Similarly, column indexes Based on azimuth dimension parameters and Calculated using the same rules;

[0021] Step 9: A two-dimensional sliding window is formed by the detection unit, protection unit, and reference unit, and then... (The sentence is incomplete and requires more context to translate accurately.) Iterate up through the data. Use the count value as the reference. As a test statistic According to the binary hypothesis testing theory, the goalless state can be... and having a goal Represented as:

[0022]

[0023] in, This is the coincidence count value between the noise photon and the reference photon. This is the coincidence count value between the signal photon and the reference photon;

[0024] Step 10: When the sliding window traverses to the unit to be detected... At that time, from Extract the coincidence count value of its reference cell to form a noise sample set. And sort them in ascending order. Select the first... Rank statistics As an estimate of background noise .according to The detection unit can be obtained separately. and Probability mass function under the assumption and ;

[0025] Step 11: Under the condition that the false alarm probability is satisfied... Under the constraints, the following optimization problem is constructed to maximize the detection probability:

[0026]

[0027] According to the Neyman-Pearson criterion, minimize the probability of missed detection. Equivalent to maximization The corresponding constant false alarm threshold The solution can be obtained using the following formula:

[0028]

[0029] Step 12: For calculation The maximum value is obtained by introducing Lagrange multipliers. Construct the objective function as follows:

[0030]

[0031] Therefore, it can be deduced that... The condition for minimization is And then sorted out The value at the critical point is:

[0032]

[0033] The likelihood ratio criterion can be deduced as follows:

[0034]

[0035] The above formula shows that when the calculated likelihood ratio Greater than the threshold If the threshold condition is met, the cell is considered to contain a target; otherwise, it is considered to contain only noise. After traversing the entire two-dimensional matrix, all target points that meet the threshold requirements are selected to form a candidate target point set. ;

[0036] Step 13: Set the spatial neighborhood radius Similarity threshold Calculate the weighted distance between any two points. If candidate point pairs satisfy Then determine belong The neighborhood of a cluster. Given the minimum number of neighboring nodes of a cluster. If any point Number of samples in the neighborhood Then determine The target cluster is defined as follows: A core point is identified, and all its density-reachable points are merged into a single cluster. Candidate points that are not core points within the cluster are designated as boundary points. After traversing all candidate points, a target cluster set is generated. The coincidence count peak points within each cluster are selected as the final detection targets. .

[0037] Step ten includes the following steps:

[0038] Step 10 (a): Move the sliding window to any unit to be detected. Noise estimation is performed using the ordered statistics of its reference cell. The reference cell coincidence counts are extracted to form a noise sample set. And sorted in ascending order, we get:

[0039]

[0040] Step 10 (II): Select the sorted first... Statistical As an estimate of background noise quantiles , These are the preset quantile coefficients;

[0041] Step 10 (3): According to the Poisson distribution model, the unit to be detected is in and The probability mass functions under the assumptions are as follows:

[0042]

[0043] Step thirteen includes the following steps:

[0044] Step Thirteen (I): For the obtained set of candidate target points Each candidate point in the, where Indicates a time delay. Indicates angle, To satisfy the count value, define any two points and Spatial weighted distance between for:

[0045]

[0046] in, As a weighting factor, The radius of the spatial neighborhood. For intensity similarity threshold; spatial distance and relative intensity difference They are defined as follows:

[0047]

[0048] in, and These refer to the system's temporal resolution and angular resolution, respectively. The use of relative strength difference is to adapt to the dynamic changes in target strength.

[0049] Step 13 (II) If the candidate point pair satisfies Then determine belong The neighborhood of a cluster. Define the minimum number of neighboring nodes for a cluster. If any point The number of samples in the neighborhood satisfies Then determine The core point is identified, and all its density-reachable points are merged into a single cluster. Candidate points that are not core points within the cluster are designated as boundary points. This process is repeated for all candidate points to generate the target cluster set. Outliers that do not belong to any cluster are classified as noise;

[0050] Step Thirteen (III): For each target cluster The peak points of the target clusters are selected as the final detection targets. All target clusters are traversed to obtain the final set of detection targets. . Attached Figure Description

[0051] Figure 1 This is a schematic diagram of the overall system model of the present invention;

[0052] Figure 2 This is a schematic diagram of the counting principle of the present invention;

[0053] Figure 3 This is a schematic diagram of the two-dimensional constant false alarm rate (CFAR) detection principle of the present invention.

[0054] Figure 4 This is a schematic diagram of the density clustering algorithm based on association strength weighting of the present invention. Detailed Implementation Plan

[0055] The present invention will now be described in further detail with reference to the accompanying drawings:

[0056] Step 1: A continuous narrow-linewidth semiconductor laser with a bandwidth of 160MHz is used to generate pump light with a wavelength of 405nm. The pump light is then adjusted to be horizontally linearly polarized light by a half-wave plate, a quarter-wave plate, and a polarization beam splitter.

[0057] Step 2: The horizontally linearly polarized light is modulated into 45° linearly polarized light by passing it through a half-wave plate and a quarter-wave plate in sequence, and then focused onto the PPKTP crystal by a lens;

[0058] Step 3: After 45° linearly polarized light is irradiated onto the PPKTP crystal, two entangled photons with a wavelength of 810 nm are generated through a spontaneous parametric downconversion process. During the SPDC process, a pump photon with a wavelength of 405 nm splits into two photons with a wavelength of 810 nm after passing through the PPKTP crystal. These two photons are in a polarization entangled state after passing through the Sagnac interference ring structure. They are unpredictable and indistinguishable in terms of time and polarization state, and their energy and momentum are conserved with the pump light.

[0059] Step 4: Pass the entangled photon pair through a 50:50 dual-wavelength polarization beamsplitter. Utilizing the polarization state, the photon pair is separated into a reference photon and a signal photon, resulting in signal light and reference light with equal intensity I. The corresponding joint spectroscopic amplitude (JSA) can be expressed as:

[0060]

[0061] in, For particle number states, It is a constant. and These are the frequencies of the signal light and the reference light, respectively. and These are the generation operators for the signal photon and the reference photon, respectively. and These are the spectral envelope and phase-matching function of the pump light, respectively.

[0062] Step 5: Pass the reference light and signal light through photonic couplers respectively. The reference light is directly received locally by the single photon detector (SPD) 1 to generate a time pulse sequence. ; signal photons from to Scanning multiple targets, the signals reflected back to the local unit are received by the SPD2 to obtain another time pulse sequence. The time series obtained by the single-photon detectors in the reference optical path and the signal optical path are respectively expressed as follows:

[0063]

[0064] in, and These are the reference optical path and the signal optical path, respectively. and the The arrival time of each photon and This represents the total number of detected photons in that direction;

[0065] Step 6: Set the door width to be compatible. ,Will Delay Then, the time sequence of the reference optical path after the delay is calculated. and The coincidence count between the two values, when there exists a pair of photons that satisfy the condition... If the event is true, it is recorded as a valid event, and the validity count is incremented by 1; otherwise, it is not counted.

[0066] Step 7: Measure the coincidence count value in each scanning direction, and let... The total number of sampling points in the time delay dimension. Let be the total number of sampling points in the azimuth dimension, then the conformance counting matrix is... The definition is as follows:

[0067]

[0068] in, Indicates the first The delay unit and the first The coincidence count intensity at each azimuth unit;

[0069] Step 8: Convert the matrix As the input signal, the unit to be detected Construct a two-dimensional sliding window centered on the element, and set the size of the inner protective window to [value]. The outer reference window size is The annular region between the two windows serves as a reference cell for estimating background noise; the total number of cells in this region is [number missing]. ;

[0070] Step 9: Perform symmetric mirror filling on the matrix. Based on the reference window size, set the time-delay dimension fill half-width to [value missing]. The orientation dimension fill half-width is Construct an extended matrix Any position in this matrix element value ,in and These are the row and column indices mapped to the original matrix, respectively. For the row index... According to the current location With time delay dimension parameters and The relationship is calculated according to the following rules:

[0071]

[0072] Similarly, column indexes Based on azimuth dimension parameters and Calculated using the same rules;

[0073] Step 10: A two-dimensional sliding window is formed by the detection unit, protection unit, and reference unit, and then... (The sentence is incomplete and requires more context to translate accurately.) Iterate up through the data. Use the count value as the reference. As a test statistic According to the binary hypothesis testing theory, the goalless state can be... and having a goal Represented as:

[0074]

[0075] in, This is the coincidence count value between the noise photon and the reference photon. This is the coincidence count value between the signal photon and the reference photon;

[0076] Step 11: When the sliding window traverses to the unit to be detected... At that time, from Extract the coincidence count value of its reference cell to form a noise sample set. And sort them in ascending order. Select the first... Rank statistics As an estimate of background noise .according to The detection unit can be obtained separately. and Probability mass function under the assumption and ;

[0077] Step 12: Under the condition of satisfying the false alarm probability Under the constraints, the following optimization problem is constructed to maximize the detection probability:

[0078]

[0079] According to the Neyman-Pearson criterion, minimize the probability of missed detection. Equivalent to maximization The corresponding constant false alarm threshold The solution can be obtained using the following formula:

[0080]

[0081] Step Thirteen: For calculation The maximum value is obtained by introducing Lagrange multipliers. Construct the objective function as follows:

[0082]

[0083] Therefore, it can be deduced that... The condition for minimization is And then sorted out The value at the critical point is:

[0084]

[0085] The likelihood ratio criterion can be deduced as follows:

[0086]

[0087] The above formula shows that when the calculated likelihood ratio Greater than the threshold If the threshold condition is met, the cell is considered to contain a target; otherwise, it is considered to contain only noise. After traversing the entire two-dimensional matrix, all target points that meet the threshold requirements are selected to form a candidate target point set. ;

[0088] Step Fourteen: Set the spatial neighborhood radius Similarity threshold Calculate the weighted distance between any two points. If candidate point pairs satisfy Then determine belong The neighborhood of a cluster. Given the minimum number of neighboring nodes of a cluster. If any point Number of samples in the neighborhood Then determine The target cluster is defined as follows: A core point is identified, and all its density-reachable points are merged into a single cluster. Candidate points that are not core points within the cluster are designated as boundary points. After traversing all candidate points, a target cluster set is generated. The coincidence count peak points within each cluster are selected as the final detection targets. .

[0089] Step eleven includes the following steps:

[0090] Step 11 (a): Move the sliding window to any unit to be detected. Noise estimation is performed using the ordered statistics of its reference cell. The reference cell coincidence counts are extracted to form a noise sample set. And sorted in ascending order, we get:

[0091]

[0092] Step 11 (II): Select the sorted number... Statistical As an estimate of background noise quantiles , These are the preset quantile coefficients;

[0093] Step 11 (III): According to the Poisson distribution model, the unit to be detected is in and The probability mass functions under the assumptions are as follows:

[0094]

[0095] Step fourteen includes the following steps:

[0096] Step Fourteen (I): For the obtained set of candidate target points Each candidate point in the, where Indicates a time delay. Indicates angle, To satisfy the count value, define any two points and Spatial weighted distance between for:

[0097]

[0098] in, As a weighting factor, The radius of the spatial neighborhood. For intensity similarity threshold; spatial distance and relative intensity difference They are defined as follows:

[0099]

[0100] in, and These refer to the system's temporal resolution and angular resolution, respectively. The use of relative strength difference is to adapt to the dynamic changes in target strength.

[0101] Step Fourteen (II): If the candidate point pairs satisfy... Then determine belong The neighborhood of a cluster. Define the minimum number of neighboring nodes for a cluster. If any point The number of samples in the neighborhood satisfies Then determine The core point is identified, and all its density-reachable points are merged into a single cluster. Candidate points that are not core points within the cluster are designated as boundary points. This process is repeated for all candidate points to generate the target cluster set. Outliers that do not belong to any cluster are classified as noise;

[0102] Step Fourteen (III): For each target cluster The peak points of the target clusters are selected as the final detection targets. All target clusters are traversed to obtain the final set of detection targets. .

Claims

1. A multi-target detection method based on entangled photons, characterized in that... Includes the following steps: Step 1: Use a continuous narrow-linewidth semiconductor laser with a bandwidth of 160MHz to generate pump light with a wavelength of 405nm, and adjust it into horizontally linearly polarized light through a half-wave plate, a quarter-wave plate and a polarization beam splitter. Step 2: The horizontally linearly polarized light is sequentially modulated into 45° linearly polarized light by passing it through a half-wave plate and a quarter-wave plate, and then focused onto a PPKTP crystal by a lens. A pair of entangled photons with a wavelength of 810nm is generated through a spontaneous parametric downconversion process. Step 3: Pass the entangled photon pair through a 50:50 dual-wavelength polarization beam splitter to separate the photon pair into a signal photon and a reference photon, resulting in signal light and reference light with equal intensity. Step 4: Pass the reference light and signal light through photonic couplers respectively. The reference light is directly received locally by the single photon detector (SPD) 1 to generate a time pulse sequence. ; signal photons from to Scanning multiple targets, the signals reflected back to the local unit are received by the SPD2 to obtain another time pulse sequence. The time series obtained by the single-photon detectors in the reference optical path and the signal optical path are respectively expressed as follows: in, and These are the reference optical path and the signal optical path, respectively. and the The arrival time of each photon and This represents the total number of detected photons in that direction; Step 5: Set the door width to be compatible. ,Will Delay Then, the time sequence of the reference optical path after the delay is calculated. and The coincidence count between the two values, when there exists a pair of photons that satisfy the condition... If the event is true, it is recorded as a valid event, and the validity count is incremented by 1; otherwise, it is not counted. Step 6: Measure the coincidence count value in each scanning direction, and let... The total number of sampling points in the time delay dimension. Let be the total number of sampling points in the azimuth dimension, then the conformance counting matrix is... The definition is as follows: in, Indicates the first The delay unit and the first Synchronous count intensity at each azimuth unit; Step 7: Convert the matrix As the input signal, the unit to be detected Construct a two-dimensional sliding window centered on the element, and set the size of the inner protective window to [value]. The outer reference window size is The annular region between the two windows serves as a reference cell for estimating background noise; the total number of cells in this region is [number missing]. ; Step 8: Perform symmetric mirror filling on the matrix. Based on the reference window size, set the half-width of the time-delay dimension fill to [value missing]. The orientation dimension fill half-width is Construct an extended matrix Any position in this matrix element value ,in and These are the row and column indices mapped to the original matrix, respectively. For the row index... According to the current location With time delay dimension parameters and The relationship is calculated according to the following rules: Similarly, column indexes Based on azimuth dimension parameters and Calculated using the same rules; Step 9: A two-dimensional sliding window is formed by the detection unit, protection unit, and reference unit, and then... (The sentence is incomplete and requires more context to translate accurately.) Iterate up through the data. Use the count value as the reference. As a test statistic According to the binary hypothesis testing theory, the goalless state can be... and having a goal Represented as: in, This is the coincidence count value between the noise photon and the reference photon. This is the coincidence count value between the signal photon and the reference photon; Step 10: When the sliding window traverses to the unit to be detected... At that time, from Extract the coincidence count value of its reference cell to form a noise sample set. And sort them in ascending order. Select the first... Rank statistics As an estimate of background noise .according to The detection unit can be obtained separately. and Probability mass function under the assumption and ; Step 11: Under the condition that the false alarm probability is satisfied... Under the constraints, the following optimization problem is constructed to maximize the detection probability: According to the Neyman-Pearson criterion, minimize the probability of missed detection. Equivalent to maximization The corresponding constant false alarm threshold The solution can be obtained using the following formula: Step 12: For calculation The maximum value is obtained by introducing Lagrange multipliers. Construct the objective function as follows: Therefore, it can be deduced that... The condition for minimization is And then sorted out The value at the critical point is: The likelihood ratio criterion can be deduced as follows: The above formula shows that when the calculated likelihood ratio Greater than the threshold If the threshold condition is met, the cell is considered to contain a target; otherwise, it is considered to contain only noise. After traversing the entire two-dimensional matrix, all target points that meet the threshold requirements are selected to form a candidate target point set. ; Step 13: Set the spatial neighborhood radius Similarity threshold Calculate the weighted distance between any two points. If candidate point pairs satisfy Then determine belong The neighborhood of a cluster. Given the minimum number of neighborhood points of a cluster. If any point Number of samples in the neighborhood Then determine The target cluster is defined as follows: A core point is identified, and all its density-reachable points are merged into a single cluster. Candidate points that are not core points within the cluster are designated as boundary points. After traversing all candidate points, a target cluster set is generated. The coincidence count peak points within each cluster are selected as the final detection targets. .

2. The noise estimation method in a multi-target detection method based on entangled photons according to claim 1, characterized in that: Step ten includes the following steps: Step 10 (a): Move the sliding window to any unit to be detected. Noise estimation is performed using the ordered statistics of its reference cell. The reference cell coincidence counts are extracted to form a noise sample set. And sorted in ascending order, we get: Step 10 (II): Select the sorted first... Statistical As an estimate of background noise quantiles , These are the preset quantile coefficients; Step 10 (3): According to the Poisson distribution model, the unit to be detected is in and The probability mass functions under the assumptions are as follows:

3. The target clustering method in the multi-target detection method based on entangled photons according to claim 1, characterized in that: Step thirteen includes the following steps: Step Thirteen (I): For the obtained set of candidate target points Each candidate point in the, where Indicates time delay. Indicates angle, To satisfy the count value, define any two points and Spatial weighted distance between for: in, As a weighting factor, The radius of the spatial neighborhood. For intensity similarity threshold; spatial distance and relative intensity difference They are defined as follows: in, and These refer to the system's temporal resolution and angular resolution, respectively. The use of relative strength difference is to adapt to the dynamic changes in target strength. Step 13 (II) If the candidate point pairs satisfy Then determine belong The neighborhood of a cluster. Define the minimum number of neighboring nodes for a cluster. If any point The number of samples in the neighborhood satisfies Then determine The core point is identified, and all its density-reachable points are merged into a single cluster. Candidate points that are not core points within the cluster are designated as boundary points. This process is repeated for all candidate points to generate the target cluster set. Outliers that do not belong to any cluster are classified as noise; Step Thirteen (III): For each target cluster The peak points of the target clusters are selected as the final detection targets. All target clusters are traversed to obtain the final set of detection targets. .

Images