Single-photon imaging method based on time information marking

Abstract

A single-photon imaging method based on time information marking comprises the steps that a single-frame binary image containing a timestamp is obtained by collecting sparse photons passing through a sample in a transmission mode, an estimated value of space coordinates of incident photons is obtained through PSF fitting and centroid averaging, an initial reconstructed image containing average photon number estimation is obtained through timestamp reconstruction, and a single-frame binary image containing the average photon number estimation is obtained. And obtaining a final reconstructed image through de-noising optimization processing. According to the method, the arrival time of a photon event is recorded, a probability distribution image of photons is obtained through reconstruction by using a timestamp algorithm, and the image is optimized through post-processing algorithms such as position calibration and filtering and denoising, so that a clear weak light imaging result is obtained.

Description

Technical Field

[0001] The present invention relates to a technology in the field of quantum imaging, in particular to a single-photon imaging method based on time information marking. Background Art The dynamic range and performance of traditional imaging solutions are greatly limited under ultra-weak light conditions, and it is often difficult to obtain clearer images, which brings huge restrictions to technical applications such as long-distance remote sensing and biological microscopy. Summary of the invention

[0002] In view of the problem that the prior art has difficulty in achieving clear imaging under ultra-weak light conditions and the shortcomings of small detection dynamic range and low detection sensitivity, the present invention proposes a single-photon imaging method based on time information marking. By recording the arrival time of the photon event and reconstructing the probability distribution image of the photon using a timestamp algorithm, the image is optimized through post-processing algorithms such as position calibration and filtering denoising to obtain clear weak-light imaging results.

[0003] The present invention is achieved through the following technical solutions:

[0004] The present invention relates to a single-photon imaging method based on time information tagging, comprising: obtaining a single-frame binary image containing a time stamp by collecting sparse photons transmitted through a sample, obtaining an estimated value of the spatial coordinates of the incident photons by PSF fitting and centroid averaging, obtaining an initial reconstructed image containing an average photon number estimate by time stamp reconstruction, and obtaining a final reconstructed image by denoising and optimization processing.

[0005] The sparse photons refer to photons that, after attenuation and loss during propagation, arrive at the detection end at only one photon in a single minimum time unit, excluding the arrival of multiple photons.

[0006] The timestamp, i.e., the time when the photon arrives at the camera, is derived from the camera time value or the number of frames per second and the position corresponding to a certain frame when it arrives.

[0007] The camera is preferably a device capable of detecting single photons, including single-photon detectors, large-area array single-photon cameras, and enhanced single-photon cameras.

[0008] The sample is a transparent sample, and light is emitted through the sample and received by the camera.

[0009] The timestamp reconstruction refers to: reconstructing the probability distribution information of the photon according to the time information of the photon.

[0010] The probability distribution information is the distribution information of photons in space.

[0011] The PSF fitting is to restore the distorted original data according to the point spread function theory.

[0012] The center-of-mass averaging refers to calculating the center-of-mass position of photons based on statistical averaging.

[0013] The present invention relates to a system for implementing the above method, comprising: a single-photon source generating unit, a single-photon camera and a data processing unit, wherein: the single-photon source generating unit generates sparse single-photon events, the single-photon camera collects transmitted light passing through a sample to be tested and extracts arrival time information of the photons, the data processing unit performs PSF fitting and centroid averaging according to the arrival time information to obtain an estimated value of the spatial coordinate of the incident photon, and obtains a final reconstructed image through timestamp reconstruction and denoising optimization processing. Technical Effects

[0014] The present invention can achieve a wider dynamic range and higher detection sensitivity under weak light detection conditions through marking of timestamp information based on a single-photon area array camera and a data processing algorithm. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is a flow chart of the present invention;

[0016] Figure 2 It is a schematic diagram of a single photon source generating unit;

[0017] In the figure: 1 object to be measured, 2 single-photon area array camera, 3 single-photon detector, 4 PPKTP crystal, 5 laser, 6 delay line, 7 wave plate, 8 filter;

[0018] Figure 3 It is a schematic diagram for fitting the spatial position of a photon event;

[0019] Figure 4 It is a schematic diagram of the effect of the embodiment. DETAILED DESCRIPTION

[0020] like Figure 1 As shown, this embodiment relates to a single-photon imaging method based on time information marking, including:

[0021] Step 1: Preparation of single photon source: Figure 1 As shown, a 25 mm long periodically poled KTP crystal (PPKTP) is pumped by a semiconductor laser with a wavelength of 405 nm to induce a spontaneous parametric down-conversion process to generate 810 nm correlated photon pairs.

[0022] In this embodiment, one of the paths is selected as signal light to pass through the object and then be received by the single-photon area array camera, and the other path is selected as reference light to be collected by the single-photon detector and then used as an electrical signal to trigger the shutter of the area array camera.

[0023] Step 2: Target selection: The target is a transparent object. In the experiment, the photons that have passed through the object are collected and processed for analysis. In the experiment, a black and white letter image is selected as a sample, and the light can be received by the camera through the white part.

[0024] Step 3, signal detection and collection: The delay line is used to ensure that the optical paths of the two lights are the same, that is, the arrival of the reference light is used as an electrical signal to control the trigger switch of the camera intensifier to realize data collection. By adjusting the camera exposure time (50μs), the opening time of the image intensifier (10ns) and the intensity of the light source, the single image captured is controlled to contain only one photon per pixel at most, and a single-frame binary image of 0 or 1 with or without photons is obtained.

[0025] The image intensifier is a device placed at the front end of the camera, which can amplify the signal so that the camera can accurately capture single photon events.

[0026] Step 4: Data post-processing, including:

[0027] 4.1 Calibration of spatial position: Figure 3 As shown in FIG. 1 , after counting multiple frames of images of a light spot and calculating its average centroid position, that is, the position where the photon distribution is most concentrated, the light spot correction is achieved by fitting the PSF of the centroid.

[0028] 4.2 Estimate the average number of photons for each pixel in the calibrated image, and obtain the initial single-photon image after normalization. Specifically, when the acquisition rate is 100 frames per second, such as the 10th frame corresponding to the signal collected at 0.1s, by extracting the arrival time t1, t2, t3... of each photon, the time intervals t1, t2-t1, t3-t2... are obtained by subtracting them. The time intervals satisfy the relationship: Where: τ i is a time interval that satisfies:∑ i p i =1, p i is the Poisson distribution probability of the i-th photon, and the average photon time interval τ is obtained by statistics i , which satisfies τ i =1 / n, where n is the average number of photons, for τ i Taking the reciprocal gives an estimate of the average number of photons at that location.

[0029] 4.3 The initial single-photon image is filtered and convexly optimized to obtain the final reconstructed image.

[0030] The convex optimization process is implemented by, but not limited to, the technology described in "Sparse Poisson intensity reconstruction algorithms" (2009 IEEE / SP 15th Workshop on Statistical Signal Processing. 2009: 634-637.) by Harmany ZT, Marcia RF, Willett RM.

[0031] After specific actual experiments, in a dark laboratory environment, the camera exposure time was 50μs, the image intensifier was turned on (10ns), the camera cooling temperature was set to minus 30°, and the light path attenuation was adjusted to achieve an average of 500 photons detected per second per frame. The method was used to obtain a cumulative 50,000 frames of image results. The experimental results are shown in Figure 4 .

[0032] like Figure 4 The figures show the original image (a), the result of simply accumulating 5000 frames without using timestamp images (b), the result of timestamp processing but without optimization and enhancement, and the result of timestamp processing (c) and optimization and enhancement (d). It can be seen that adding timestamp processing can significantly remove most of the background noise and improve the signal-to-noise ratio of the image; adding optimization algorithms can improve the reconstruction of the image at the edge, and improve the smoothness and contrast of the image. Compared with the direct accumulation method, the image contrast of the optimized reconstruction result based on timestamps in the present invention is increased by 2.25 times.

[0033] In summary, the present invention can greatly improve the dynamic range and imaging quality of single-photon imaging under extremely weak light conditions. Traditional cumulative imaging under weak light conditions is often limited by the large differences in counts, resulting in large deviations in the reconstructed image. The timestamp scheme is based on the mining of the temporal degree of freedom. Lower photon intensities correspond to longer time intervals. When the system has a higher temporal resolution capability, the timestamp scheme can obtain a higher-precision photon number estimation result, thereby obtaining a clearer reconstructed image. The present invention has great application potential in the fields of biological microscopic imaging, long-distance remote sensing, etc.

[0034] The above-mentioned specific implementation can be partially adjusted in different ways by those skilled in the art without departing from the principle and purpose of the present invention. The protection scope of the present invention shall be based on the claims and shall not be limited by the above-mentioned specific implementation. Each implementation scheme within its scope shall be subject to the constraints of the present invention.

Claims

1. A single-photon imaging method based on time information labeling, characterized in that: include: A single-frame binary image containing a timestamp is obtained by collecting sparse photons transmitted through the sample, and an estimate of the spatial coordinates of the incident photons is obtained by PSF fitting and centroid averaging. An initial reconstructed image containing an estimate of the average number of photons is obtained by timestamp reconstruction, and a final reconstructed image is obtained by denoising and optimization processing. The sparse photons refer to the photons that arrive at the detection end only one at a single minimum time unit after attenuation and loss during propagation, excluding the arrival of multiple photons. The timestamp, i.e., the time when the photon arrives at the camera, is derived from the camera time value or the number of frames per second and the position corresponding to a certain frame when it arrives.

2. The single-photon imaging method based on time information labeling according to claim 1 is characterized in that: The camera is a single-photon detector, a large-area array single-photon camera or an enhanced single-photon camera.

3. The single-photon imaging method based on time information labeling according to claim 1 is characterized in that: The timestamp reconstruction refers to: reconstructing the probability distribution information of the photon, that is, the distribution information of the photon in space, according to the time information of the photon.

4. The single-photon imaging method based on time information labeling according to claim 1 is characterized in that: The PSF fitting is to restore the distorted original data according to the point spread function theory.

5. The single-photon imaging method based on time information labeling according to claim 1 is characterized in that: The center-of-mass averaging refers to calculating the center-of-mass position of photons based on statistical averaging.

6. The single-photon imaging method based on time information labeling according to any one of claims 1 to 5, characterized in that: include: Step 1, preparation of single-photon source: a 25 mm long periodically polarized KTP crystal is pumped by a semiconductor laser with a wavelength of 405 nm to induce a spontaneous parametric down-conversion process to generate 810 nm correlated photon pairs, one of which is selected as signal light to pass through the object and then be received by the single-photon array camera, and the other is collected by the single-photon detector as reference light and then used as an electrical signal to trigger the shutter of the array camera; Step 2: Selection of target: The target is a transparent object. In the experiment, photons transmitted through the object are collected and processed for analysis. Step 3, signal detection and collection: The delay line is used to ensure that the optical path of the two beams is the same, that is, the arrival of the reference light is used as an electrical signal to control the trigger switch of the camera intensifier to realize data collection. By adjusting the exposure time of the camera, the opening time of the image intensifier and the intensity of the light source, the single image captured is controlled to contain only one photon per pixel at most, and a single-frame binary image with or without photons is obtained; Step 4: Data post-processing, including: 4.1 Calibration of spatial position: Count multiple frames of images of a light spot, calculate its average centroid position, that is, the position where the photon distribution is most concentrated, and then calibrate the light spot by fitting the PSF of the centroid; 4.2 Estimate the average number of photons for each pixel in the calibrated image and obtain the initial single-photon image after normalization. Specifically, extract the arrival time t1, t2, t3… of each photon and subtract them to obtain the time interval t1, t2-t1, t3-t2…. The time interval satisfies the relationship: Where: τ i is a time interval that satisfies:∑ i p i =1, p i is the Poisson distribution probability of the i-th photon, and the average photon time interval τ is obtained by statistics i , which satisfies τ i =1 / n, where n is the average number of photons, for τ i Taking the reciprocal gives an estimate of the average number of photons at that location; 4.3 The initial single-photon image is filtered and convexly optimized to obtain the final reconstructed image.

7. A single-photon imaging system based on time information marking for implementing any of the methods described in claims 1-6, characterized in that: include: A single-photon source generating unit, a single-photon camera and a data processing unit, wherein: the single-photon source generating unit generates sparse single-photon events, the single-photon camera collects the transmitted light passing through the sample to be tested and extracts the arrival time information of the photons, the data processing unit performs PSF fitting and centroid averaging according to the arrival time information to obtain the estimated value of the spatial coordinate of the incident photon, and obtains the final reconstructed image through timestamp reconstruction and denoising optimization processing.

Images