Single-photon image enhancement method and system based on time-stamping of area array photons

Abstract

A single-photon image enhancement method and system with a planar array photon time tagging, which uses a planar array detector to obtain the arrival time and spatial position of a single photon, and converts it into the position and transmittance of a target object image; then, after Gaussian fitting each photon to obtain an accurate signal position, a convex optimization method is used to obtain an ultra-high resolution reconstruction result map of the target object to achieve image enhancement. The present invention uses a planar array single-photon detection device to record the flight time of a single photon, and uses an algorithm to reconstruct the intensity of an object, thereby achieving a method for low-photon ultra-weak light space-time imaging.

Description

Technical Field

[0001] The present invention relates to a technology in the field of image processing, in particular to a single-photon image enhancement method and system for planar array photon time marking. Background Art

[0002] Generally, each pixel in a high-quality image contains more than 10 3 However, in many practical scenarios, such as fluorescence microscopy and biomedical imaging, samples are very sensitive to light intensity and measurements may cause damage, which requires innovative imaging schemes with the lowest possible light flux. The shortcoming of existing photon counting strategies is that more photons need to be collected to achieve higher image contrast, ignoring the time parameter that may improve image quality. Summary of the invention

[0003] In view of the defect of the prior art that the fewer the number of finite photons under low-light conditions, the smaller the number of photons received by the detector, and the greater the error in judging the number of photons, the present invention proposes a single-photon image enhancement method and system with planar array photon time marking, which utilizes planar array single-photon detection devices to record the flight time of single photons, and uses an algorithm to reconstruct the intensity of objects, thereby realizing a method for low-photon ultra-weak light space-time imaging.

[0004] The present invention is achieved through the following technical solutions:

[0005] The present invention relates to a single-photon image enhancement method with planar array photon time tagging. A planar array detector is used to obtain the arrival time and spatial position of a single photon, and convert it into the position and transmittance of a target object image. Then, after Gaussian fitting is performed on each photon to obtain an accurate signal position, a convex optimization method is used to obtain an ultra-high resolution reconstruction result image of the target object to achieve image enhancement.

[0006] The planar array detector includes: a direct detection device and an indirect detection device, wherein: the direct detection device directly detects the single photon signal and accurately records the flight time of the photon and analyzes the Poisson process of the photon; the indirect detection device is pre-coupled with an image intensifier and indirectly detects the spatial and temporal information of the target object, thereby providing raw data for realizing accurate reconstruction of the target object.

[0007] The conversion is specifically: g n =t n+1 -t n , where: g n is the time interval between the nth photon and the n+1th photon, t n is the arrival time of the nth photon at position (x, y), ω is the noise, λ (x，y) is the estimated amount of photons at position (x, y), i.e., the transmittance.

[0008] The Gaussian fitting is specifically as follows: a single photon is sensed and amplified by the image intensifier, and an area of ​​N×N pixels is obtained on the camera. The light intensity distribution of this area can be fitted by a Gaussian function:

[0009] Where: A is the amplitude, x, y represents the pixel position, σ x , σ y is the variance, and the position of the central maximum point (x0, y0) is obtained by fitting with the pixel intensity as the estimated position of the photon arrival.

[0010] The convex optimization method is to obtain a convex function by taking the negative logarithm of the time equation satisfied by the arrival of a single photon, and then solve the minimum value of the convex function to obtain the optimal solution for the corresponding pixel.

[0011] The reconstruction is performed by first balancing the time equation satisfied by the photon arrival time obtained for each pixel of the entire image with the term for preventing overfitting; then numerical iteration is performed based on the initial value, and the pixel value corresponding to the iteration result is the reconstruction result of the pixel.

[0012] The balance is specifically: Where: β is the balance term, b is the parameter, f is the operation function, and A is the linear space operator.

[0013] The numerical iteration is specifically: k+1 =f(f k+1 ), the obtained result is used as a variable, and it is iterated repeatedly until a certain error range is met and the optimal value is reached.

[0014] The present invention relates to a system for implementing the above method, comprising: a light source, an attenuation module, a sparse photon time recording module and a data processing module, wherein: the attenuation module performs light intensity attenuation processing according to photon statistical distribution information, and obtains a result that only a single photon is present in the process of interaction with a target object; the sparse photon time recording module performs nanosecond or even picosecond precision recording processing according to the arrival time information of a single photon, and obtains an accurate recording result of the time of each photon in each pixel; the data processing module performs convex optimization processing according to the time information recorded for each pixel, and obtains the optimal reflectivity or transmittance result corresponding to each pixel.

[0015] Technical Effects

[0016] The present invention reconstructs image intensity by utilizing the time information of the light reaching the detector after interacting with the target object, rather than reconstructing by accumulating the number of photons. This enables restoration of higher image contrast under the same sparse number of photons, and further reconstructs high-quality images in low-light environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a schematic diagram of the present invention;

[0018] Figure 2 Schematic diagram of the temporal changes of a discrete single photon before and after it passes through an object;

[0019] Figure 3 This is a diagram of the single photon identification and Gaussian fitting centroiding process;

[0020] Figure 4 Reconstruct results for the sparse photon algorithm. DETAILED DESCRIPTION

[0021] like Figure 1 As shown, this embodiment relates to a single-photon image enhancement method for planar array photon time marking, comprising:

[0022] Step 1. Determine the measurement variable: When reducing the photon radiation to 2.5 photons / 50μs, select an area of ​​300 pixels × 300 pixels to record each pixel of the target object imaging and the arrival time of the first 3-5 photons, that is, the time information as the data to be processed.

[0023] Step 2: Build the imaging system: Figure 2 As shown in the figure, according to the characteristics of single-photon imaging, a planar array imaging acquisition device with nanosecond time resolution is selected to amplify and record single photons. 50,000 frames are collected, and the position of a single photon and the time information of each photon are recorded in each frame. The signal output time is removed to obtain data containing a time series with 0 or 1 single photon per pixel, and the time interval information between photons is extracted.

[0024] Step 3: Preliminary data processing to optimize low-resolution images: Figure 3 As shown in the figure, due to the point spread function of the optical system, a single photon occupies 3*3 pixels on the camera, which is used as a criterion to distinguish it from noise. At the same time, a two-dimensional Gaussian function is used to fit the light intensity distribution, and the position of the maximum point in the center is obtained as an estimate of the position where the photon arrives. This operation is performed on all collected data frames to obtain the precise position and time information of all photons.

[0025] Step 4: Data image super-resolution reconstruction: Figure 4 As shown in the figure, the joint distribution probability of the flight arrival time of sparse photons satisfies the convex function characteristics. After convex optimization of the entire image, the intensity information optimization value of the corresponding pixel, that is, the corresponding grayscale value, is obtained. With the goal of obtaining the optimal solution for the entire image, the image with the minimum error value is obtained through iteration as the final optimization result.

[0026] The joint distribution probability of arrival time means that the time series of photon occurrence corresponding to each pixel can be described by a Poisson process, and the time of all photons occurring in this process can be represented by a joint conditional occurrence probability.

[0027] After specific practical experiments, in this embodiment, when there are about 3-5 photons per pixel, the contrast obtained by reconstructing the cumulative number of photons is 0.4, while the contrast obtained by reconstructing the time information can reach 0.9, and the contrast is improved by 2.25 times.

[0028] Compared with the existing technology, this method analyzes and records the time information of single-photon flight events in the case of few photons, and obtains a 2.25-fold performance improvement in contrast when there are only 3-5 photons per pixel. The result is better than the imaging scheme of photon number accumulation.

[0029] 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 image enhancement method for area array photon time marking, characterized in that: A planar array detector is used to obtain the arrival time and spatial position of a single photon, and convert it into the position and transmittance of the target object image. Then, after Gaussian fitting is performed on each photon to obtain the precise signal position, a convex optimization method is used to obtain an ultra-high resolution reconstruction result image of the target object to achieve image enhancement. The transformation is specifically: g n =t n+1 -t n , where: g n is the time interval between the nth photon and the n+1th photon, t n is the arrival time of the nth photon at position (x, y), ω is the noise, λ (x，y) is the estimated amount of photons at position (x, y), i.e., transmittance; The Gaussian fitting is specifically as follows: a single photon is sensed and amplified by the image intensifier, and an area of ​​N×N pixels is obtained on the camera. The light intensity distribution of this area can be fitted by a Gaussian function: Where: A is the amplitude, x, y represents the pixel position, σ x , σ y is the variance, and the position of the maximum point in the center (x0, y0) is obtained by fitting with the pixel intensity as the estimated position of the photon arrival; The convex optimization method is as follows: after taking the negative logarithm of the time equation satisfied by the arrival of a single photon to obtain a convex function, solving the minimum value of the convex function, that is, obtaining the optimal solution for the corresponding pixel; The reconstruction is performed by first balancing the time equation satisfied by the photon arrival time obtained for each pixel of the entire image with the term for preventing overfitting; then numerical iteration is performed based on the initial value, and the pixel value corresponding to the iteration result is the reconstruction result of the pixel.

2. The single-photon image enhancement method of area array photon time marking according to claim 1 is characterized in that: The planar array detector includes: a direct detection device and an indirect detection device, wherein: the direct detection device directly detects the single photon signal and accurately records the flight time of the photon and analyzes the Poisson process of the photon; the indirect detection device is pre-coupled with an image intensifier and indirectly detects the spatial and temporal information of the target object, thereby providing raw data for realizing accurate reconstruction of the target object.

3. The single-photon image enhancement method of area array photon time marking according to claim 1 is characterized in that: The balance is specifically: Where: β is the balance term, b is the parameter, f is the operation function, and A is the linear space operator.

4. The single-photon image enhancement method of area array photon time marking according to claim 1 is characterized in that: The numerical iteration is specifically: f k+1 =ff k+1 , the obtained result is used as a variable, and it is iterated repeatedly until a certain error range is met and the optimal value is reached.

5. A system for implementing the single-photon image enhancement method of planar photon time marking as described in any one of claims 1 to 4, characterized in that: include: A light source, an attenuation module, a sparse photon time recording module and a data processing module, wherein: the attenuation module performs light intensity attenuation processing according to the photon statistical distribution information, and obtains the result that only a single photon interacts with the target object; the sparse photon time recording module performs nanosecond or even picosecond precision recording processing according to the arrival time information of a single photon, and obtains the accurate recording result of each photon time for each pixel; the data processing module performs convex optimization processing according to the time information recorded for each pixel, and obtains the optimal reflectivity or transmittance result corresponding to each pixel.

Images