Particle Image Reconstruction Method, Device, Computer Equipment and Storage Medium

Abstract

The present application relates to a particle image reconstruction method, device, computer equipment and storage medium. The method comprises: collecting original pulse image data, performing integral imaging processing on the original pulse image data using different integral times for multiple times, determining the target area including particles and the background area excluding particles in the original pulse image data; determining the first integral time corresponding to each target area according to the signal-to-noise ratio corresponding to different integral times of each target area, and determining the second integral time according to the brightness of the background area in each integral imaging result; performing integral imaging processing on each target area according to the first integral time, and performing integral imaging processing on the background area according to the second integral time, and reconstructing multiple frames of target particle images corresponding to the particle area according to the integral imaging results of each target area and background area. The use of this method can improve the accuracy of particle image reconstruction.

Description

Technical Field

[0001] This application relates to the field of image processing, and in particular, to a particle image reconstruction method, apparatus, computer device, storage medium, and computer program product. Background Art

[0002] Particle Image Velocimetry (PIV) technology plays a crucial role in fields such as fluid mechanics and aerodynamics. It can provide quantitative measurement of the velocity distribution of the flow field, which is essential for understanding the laws of fluid motion, optimizing the design of fluid machinery, and studying various fluid phenomena.

[0003] PIV technology usually uses a high-speed camera to capture particle images of the flow field area. However, in a flow field with a strong brightness contrast (such as a combustion flow field), the particle images obtained by the camera usually have overexposure or underexposure phenomena, resulting in oversaturation of the particle images in the bright area, and at the same time, weak particle signals in the dark area lead to insufficient particle signal-to-noise ratio, causing image information distortion or loss in some areas. Therefore, the traditional particle image acquisition technology has the problem of insufficient accuracy of the acquired particle images. Summary of the Invention

[0004] Based on this, it is necessary to provide a particle image reconstruction method, apparatus, computer device, and storage medium for the above technical problems.

[0005] In a first aspect, this application provides a particle image reconstruction method. The method includes:

[0006] Collect original pulse image data for the particle region within a preset sampling duration, perform integral imaging processing on the original pulse image data multiple times with different integration times, and determine the target region including particles and the background region not including particles in the original pulse image data according to the obtained integral imaging results;

[0007] Determine the signal-to-noise ratio of each target region corresponding to different integration times, determine the first integration time corresponding to each target region according to each signal-to-noise ratio, and determine the second integration time corresponding to the background region according to the brightness of the background region in each integral imaging result;

[0008] Perform integral imaging processing on each target region according to the first integration time respectively, and perform integral imaging processing on the background region according to the second integration time. Reconstruct multiple frames of target particle images corresponding to the particle region according to the integral imaging results of each target region and the background region.

[0009] In one embodiment, the original pulse image data is integrally imaged in multiple times with different integration times, and according to the obtained integral imaging results, a target region including particles and a background region not including particles in the original pulse image data are determined, including:

[0010] Taking the original pulse image data as target pulse image data, performing integral imaging processing on the target pulse image data with a target integration time to obtain an integral imaging result, determining a particle region including particles from the integral imaging result, and taking the part of the original pulse image data corresponding to the particle region as the target region;

[0011] Taking the part of the original pulse image data except each of the target regions as target pulse image data, increasing the target integration time, and jumping to the step of performing integral imaging processing on the target pulse image data with the target integration time to obtain an integral imaging result until all pixel points in the integral imaging result reach saturation;

[0012] Taking the part of the original pulse image data corresponding to the current target pulse image data as the background region.

[0013] In one embodiment, determining the signal-to-noise ratio of each of the target regions corresponding to different integration times, and determining a first integration time corresponding to each of the target regions according to the signal-to-noise ratios, includes:

[0014] Performing integral imaging processing on each of the target regions with different integration times to respectively obtain multiple integral imaging results of each of the target regions, and respectively determining the signal-to-noise ratio of each of the integral imaging results;

[0015] For any one of the target regions, based on the signal-to-noise ratios corresponding to the integral imaging results of the target region and the integration times corresponding to the integral imaging results of the target region, constructing a signal-to-noise ratio change curve of the signal-to-noise ratio of the target region with respect to the integration time, and taking the integration time corresponding to the maximum value in the signal-to-noise ratio change curve as the first integration time.

[0016] In one embodiment, determining a second integration time corresponding to the background region according to the brightness of the background region in each of the integral imaging results, includes:

[0017] According to each of the integral imaging results, for the pixel points located in the background region and with the brightness reaching saturation, respectively determining the saturation integration time when the brightness of each of the pixel points in the background region reaches saturation;

[0018] Determine the second integration time corresponding to each of the pixel points according to the overexposure suppression strategy and the saturation integration time.

[0019] In one embodiment, the reconstructing the multi-frame target particle images corresponding to the particle region according to the integral imaging results of the target regions and the background region includes:

[0020] Determine each imaging moment corresponding to the particle region within the preset sampling duration according to the pulse sampling time and the preset sampling duration;

[0021] Respectively determine the corresponding first integral imaging result in the integral imaging result of the background region at each of the imaging moments, and respectively determine the corresponding second integral imaging result in the integral imaging result of each of the target regions at each of the imaging moments;

[0022] Generate the target particle images corresponding to each of the imaging moments based on the first integral imaging result and the second integral imaging result.

[0023] In one embodiment, the above method further includes:

[0024] Perform denoising processing on each of the target particle images based on at least one of the pulse emission frequency corresponding to each pixel point in the original pulse image data, the pixel difference between every two adjacent target particle images, the first target filter, and the second target filter, to obtain the denoised target particle images;

[0025] Wherein, the first target filter is constructed based on the target shape feature of the particle, and the second target filter is constructed based on the target motion feature of the particle.

[0026] In one embodiment, performing denoising processing on each of the target particle images based on the pixel difference between every two adjacent initial particle images includes:

[0027] Perform a difference operation on any two adjacent target particle images. When the pixel difference corresponding to any pixel point in the two target particle images is greater than a first preset threshold and less than a second preset threshold, determine the pixel point as a noise pixel point, and perform denoising processing on each of the noise pixel points in the two target particle images.

[0028] In one embodiment, the target shape feature of the particle includes an area feature and an intensity feature. The area feature is that the number of pixel points corresponding to a single particle in the particle image is greater than 1, and the intensity feature is that among all the pixel points corresponding to a single particle in the particle image, the intensity of the central pixel point is greater than the intensity of the edge pixel points;

[0029] The target motion feature of the particle is that in the original pulse image data, the change values of the pulse emission frequencies of multiple adjacent pixel points have an associated relationship.

[0030] In a second aspect, the present application further provides a particle image reconstruction device. The device includes:

[0031] An integration module, configured to collect original pulse image data for a particle region within a preset sampling duration, perform integral imaging processing on the original pulse image data multiple times with different integration times, and determine a target region including the particle and a background region not including the particle in the original pulse image data according to the obtained integral imaging results;

[0032] A determination module, configured to determine the signal-to-noise ratio of each target region corresponding to different integration times, determine the first integration time corresponding to each target region according to the signal-to-noise ratios, and determine the second integration time corresponding to the background region according to the brightness of the background region in each integral imaging result;

[0033] A reconstruction module, configured to perform integral imaging processing on each target region according to the first integration time respectively, perform integral imaging processing on the background region according to the second integration time, and reconstruct multiple frames of target particle images corresponding to the particle region according to the integral imaging results of each target region and the background region.

[0034] In one embodiment, the integration module is further configured to:

[0035] Take the original pulse image data as target pulse image data, perform integral imaging processing on the target pulse image data with a target integration time to obtain an integral imaging result, determine a particle region including the particle from the integral imaging result, and take the part of the original pulse image data corresponding to the particle region as the target region;

[0036] Take the part of the original pulse image data except each target region as target pulse image data, increase the target integration time, and jump to the step of performing integral imaging processing on the target pulse image data with the target integration time to obtain an integral imaging result until all pixel points in the integral imaging result reach saturation;

[0037] Take the part of the original pulse image data corresponding to the current target pulse image data as the background region.

[0038] In one embodiment, the determination module is further configured to:

[0039] Integrate and image each of the target regions with different integration times to respectively obtain multiple integration imaging results for each of the target regions, and respectively determine the signal-to-noise ratios of each of the integration imaging results;

[0040] For any one of the target regions, based on the signal-to-noise ratios corresponding to each of the integration imaging results of the target region and the integration times corresponding to each of the integration imaging results of the target region, construct a signal-to-noise ratio change curve of the signal-to-noise ratio of the target region with respect to the integration time, and use the integration time corresponding to the maximum value in the signal-to-noise ratio change curve as the first integration time.

[0041] In one embodiment, the determination module is further configured to:

[0042] According to the pixel points located in the background region and having saturated brightness in each of the integration imaging results, respectively determine the saturation integration time when the brightness of each pixel point in the background region reaches saturation;

[0043] According to the overexposure suppression strategy and the saturation integration time, respectively determine the second integration time corresponding to each pixel point.

[0044] In one embodiment, the reconstruction module is further configured to:

[0045] According to the pulse sampling time and the preset sampling duration, determine each imaging moment corresponding to the particle region within the preset sampling duration;

[0046] Respectively determine the corresponding first integration imaging result in the integration imaging results of the background region for each of the imaging moments, and respectively determine the corresponding second integration imaging result in the integration imaging results of each of the target regions for each of the imaging moments;

[0047] Based on the first integration imaging result and the second integration imaging result, respectively generate target particle images corresponding to each of the imaging moments.

[0048] In one embodiment, the above device further includes:

[0049] A denoising module, configured to perform denoising processing on each of the target particle images based on at least one of the pulse emission frequencies corresponding to the pixel points in the original pulse image data, the pixel differences between every two adjacent target particle images, a first target filter, and a second target filter to obtain denoised target particle images;

[0050] Wherein, the first target filter is constructed based on the target shape characteristics of the particles, and the second target filter is constructed based on the target motion characteristics of the particles.

[0051] In one embodiment, the denoising module is further configured to:

[0052] Perform a difference operation on any two adjacent target particle images. When the pixel difference corresponding to any pixel point in the two target particle images is greater than a first preset threshold and less than a second preset threshold, determine the pixel point as a noise pixel point, and perform denoising processing on each of the noise pixel points in the two target particle images.

[0053] In one embodiment, the target shape features of the particles include an area feature and an intensity feature. The area feature is that the number of pixel points corresponding to a single particle in the particle image is greater than 1, and the intensity feature is that among all the pixel points corresponding to a single particle in the particle image, the intensity of the central pixel point is greater than the intensity of the edge pixel points;

[0054] The target motion feature of the particles is that in the original pulse image data, the change values of the pulse emission frequencies of adjacent multiple pixel points have an associated relationship.

[0055] In a third aspect, the present application further provides a computer device. The computer device includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the above-mentioned method of any one of the preceding items is implemented.

[0056] In a fourth aspect, the present application further provides a computer-readable storage medium. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the above-mentioned method of any one of the preceding items is implemented.

[0057] In a fifth aspect, the present application further provides a computer program product. The computer program product includes a computer program, and when the computer program is executed by a processor, the above-mentioned method of any one of the preceding items is implemented.

[0058] The above-mentioned particle image reconstruction method, device, computer device, and storage medium image the original pulse image data according to different integration times, identify the target area containing particles and the background area not containing particles in the original pulse image data according to the integration imaging results, and then determine the first integration time used for integrating and imaging the target area according to the signal-to-noise ratio of the target area corresponding to different integration times, and determine the second integration time corresponding to the background area according to the brightness of the background area in different integration imaging results. Furthermore, each area is respectively integrated and imaged using the determined integration time to obtain the target particle image corresponding to the particle area. In the embodiments of the present application, different integration times are determined for different areas, and each area is imaged using a better integration time. Therefore, when processing images with a large difference in brightness between different areas, it is also possible to take into account both bright and dark areas for image reconstruction, improving the reconstruction accuracy of the particle image. Description of the Drawings

[0059] Figure 1 It is a schematic flowchart of a particle image reconstruction method in an embodiment;

[0060] Figure 2 It is a schematic flowchart of step 102 in an embodiment;

[0061] Figure 3 It is a schematic diagram of iteratively determining a target region in an embodiment;

[0062] Figure 4 It is a schematic flowchart of determining a first integration time in step 104 in an embodiment;

[0063] Figure 5 It is a schematic diagram of a signal-to-noise ratio change curve in an embodiment;

[0064] Figure 6 It is a schematic flowchart of determining a second integration time in step 104 in an embodiment;

[0065] Figure 7 It is a schematic flowchart of step 106 in an embodiment;

[0066] Figure 8 It is a schematic diagram of reconstructing a target particle image in an embodiment;

[0067] Figure 9 It is a schematic diagram of denoising based on pulse firing frequency in an embodiment;

[0068] Figure 10 It is a schematic diagram of the external shape characteristics of a particle target in an embodiment;

[0069] Figure 11 It is a schematic diagram of a second target filter in an embodiment;

[0070] Figure 12 It is a structural block diagram of a particle image reconstruction device in an embodiment;

[0071] Figure 13 It is an internal structure diagram of a computer device in an embodiment. Detailed Description of the Embodiment

[0072] In order to make the objectives, technical solutions and advantages of the present application clearer and more understandable, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0073] In one embodiment, as Figure 1As shown, a particle image reconstruction method is provided. In this embodiment, the method is exemplified by being applied to a server. It can be understood that the method can also be applied to a terminal, or to a system including a terminal and a server, and is implemented through the interaction between the terminal and the server. In this embodiment, the method includes the following steps:

[0074] Step 102, collect the original pulse image data for the particle region within a preset sampling duration, perform integral imaging processing on the original pulse image data multiple times with different integration times, and determine the target region including particles and the background region not including particles in the original pulse image data according to the obtained integral imaging results.

[0075] In the embodiment of the present application, continuous original pulse image data is collected by a pulsed high-speed camera within a preset sampling duration, and the original pulse image data is integrated with different integration times to obtain imaging results corresponding to different integration times. The integration time refers to the time length for accumulating the original pulse image data when imaging one frame of the image. Specifically, which different integration times are used to integrate the original pulse image data can be determined according to actual needs. For example, starting from the shortest integration time of the pulsed high-speed camera, the integration time is gradually increased according to a preset integration time increase strategy (such as increasing the integration time by a certain step each time, or using a certain multiple of the previous integration time each time) to integrate the original pulse image data until all pixel points in the imaging result reach saturation when a certain integration time is used, so as to obtain integral imaging results corresponding to multiple different integration times.

[0076] After obtaining the imaging results corresponding to different integration times, the target region including particles and the background region not including particles in the original pulse image data can be divided according to each integral imaging result. For example, through any particle recognition algorithm (such as an edge detection algorithm or a trained neural network model), the parts with morphologies conforming to the particle contour morphology are identified from each integral imaging result, and then the particle regions containing particles in the integral imaging results are determined according to these parts, and the parts corresponding to the particle regions in the original pulse image data are used as the target region. After determining each target region, the part of the original pulse image data except the target region can be used as the background region.

[0077] It should be noted that a particle region may contain multiple particles or only 1 particle. For example, after detecting each particle in the integral imaging result, a region with a particle density greater than a preset threshold can be used as a particle region, or a region where a single particle is located can be used as a particle region. Since different first integration times need to be used for integral imaging of different target regions later, and the single-pixel dwell times of different particles usually have certain differences, when there are fewer particles in the particle region, the first integration time determined for the target region corresponding to the particle region is more accurate and the imaging accuracy is also higher, but at the same time, it will bring greater computing power consumption. Those skilled in the art can determine the division strategy of the particle region according to actual needs.

[0078] Step 104: Determine the signal-to-noise ratios of the respective target regions corresponding to different integration times, determine the first integration time corresponding to each target region according to the respective signal-to-noise ratios, and determine the second integration time corresponding to the background region according to the brightness of the background region in each integral imaging result.

[0079] In the embodiment of the present application, after obtaining each target region, the trend of the signal-to-noise ratio of each target region changing with the integration time can be determined, and based on this trend, the first integration time with better imaging effect can be determined when performing integral imaging on the target region. Here, integral imaging can be performed on each target region again with different integration times. For example, more integration times can be used for integral imaging compared to the previous step to obtain more integral imaging results of the target region, so as to obtain a more accurate trend of the signal-to-noise ratio of the target region changing with the integration time. Or alternatively, the parts corresponding to the target region can be determined from the respective integral imaging results obtained in the previous step, and the trend of the signal-to-noise ratio of the target region changing with the integration time can be determined according to the signal-to-noise ratios of these parts.

[0080] The signal-to-noise ratio of the target region can be determined based on the pixel value intensities of the particle-containing part and the non-particle-containing part in the target region in each integral imaging result. After determining the signal-to-noise ratios of the target region corresponding to different integration times, the integration time corresponding to the highest signal-to-noise ratio can be used as the first integration time, or alternatively, the relationship function between the signal-to-noise ratio and the integration time can be obtained by fitting, and the integration time corresponding to the maximum value of the relationship function can be used as the first integration time. The embodiment of the present application does not make specific limitations in this regard.

[0081] For the background region, in order to prevent overexposure of the background region during integral imaging and affect the imaging effect of the particles, an integration time that can make the brightness of the background region lower can be selected as the second integration time. Exemplarily, the brightness of the background region in each integral imaging result can be determined, and the integration time corresponding to the integral imaging result with the lowest brightness can be selected as the second integration time.

[0082] Step 106: Perform integral imaging processing on each target region according to the first integration time respectively, and perform integral imaging processing on the background region according to the second integration time. Based on the integral imaging results of each target region and the background region, reconstruct multiple frames of target particle images corresponding to the particle region.

[0083] In the embodiment of the present application, the first integration time determined for each target region is used to perform integral imaging processing on each target region, and the second integration time determined for the background region is used to perform integral imaging processing on the background region. After aligning the integral imaging results of each target region and the integral imaging results of the background region in the time domain, multiple frames of particle images reconstructed for the particle region can be obtained.

[0084] The particle image reconstruction method provided by the embodiment of the present application performs imaging on the original pulse image data according to different integration times, identifies the target regions containing particles and the background regions not containing particles in the original pulse image data based on the integral imaging results, and then determines the first integration time used for integral imaging of the target regions according to the signal-to-noise ratio of the target regions corresponding to different integration times, and determines the second integration time corresponding to the background region according to the brightness of the background region in different integral imaging results. Then, integral imaging is performed on each region using the determined integration time respectively to obtain the target particle images corresponding to the particle region. In the embodiment of the present application, different integration times are determined for different regions, and a better integration time is used for imaging each region. Therefore, when processing images with a large difference in brightness between different regions, it is also possible to take into account both bright and dark regions for image reconstruction, improving the reconstruction accuracy of particle images.

[0085] In one embodiment, as Figure 2 shown, in step 102, the original pulse image data is integrally imaged multiple times using different integration times, and based on the obtained integral imaging results, the target regions including particles and the background regions not including particles in the original pulse image data are determined, including:

[0086] Step 202: Take the original pulse image data as the target pulse image data, perform integral imaging processing on the target pulse image data using the target integration time to obtain an integral imaging result, determine the particle region including particles from the integral imaging result, and take the part of the original pulse image data corresponding to the particle region as the target region;

[0087] Step 204: Take the part of the original pulse image data except for each target region as the target pulse image data, increase the target integration time, and jump to the step of performing integral imaging processing on the target pulse image data using the target integration time to obtain an integral imaging result until all pixel points in the integral imaging result reach saturation;

[0088] Step 206: Use the part of the original pulse image data corresponding to the current target pulse image data as the background region.

[0089] In the embodiments of the present application, after each integral imaging process, the target regions containing particles are identified from the integral imaging results, and the integral time of the non-target region part in the original pulse image data is gradually increased, and the integral imaging is iteratively performed to determine all the target regions. For the schematic diagram of this process and the determined target regions, see Figure 3 as shown. Figure 3 The regions selected after each iteration in

[0090] The initially determined target integral time can be a relatively short time. Using a relatively short integral time for integral imaging can avoid the situation that the high-speed particles in the imaging result cause motion blur and overexposure of the region, which affects particle recognition. After integral imaging is performed according to the target integral time, based on the particle recognition algorithm, particles are identified from the obtained integral imaging results and the particle regions are divided. The division method of the particle regions can be seen in the description of the foregoing embodiments, and the embodiments of the present application will not be elaborated herein.

[0091] After using the part of the original pulse image data corresponding to the particle region as the target region, the part other than the target region can be used as the target pulse image data in the next iteration, so as to avoid the interference of the particles with stronger energy determined in the previous rounds on the imaging result in the next iteration. At the same time, the target integral time can be increased according to the preset integral time increase strategy, and integral imaging processing is performed on the target pulse image data until all the pixel points in the integral imaging result reach saturation. At this time, it can be considered that the weakest energy particles that the pulse high-speed camera can theoretically detect have been detected, and the remaining regions where particles have not been detected (that is, the regions corresponding to the current target pulse image data) are background regions without particles.

[0092] In one embodiment, as Figure 4 shown, in step 104, determining the signal-to-noise ratios of the target regions corresponding to different integral times, and determining the first integral time corresponding to each target region according to the signal-to-noise ratios, includes:

[0093] Step 402: Perform integral imaging processing on each target region using different integral times, respectively obtain multiple integral imaging results of each target region, and respectively determine the signal-to-noise ratios of the integral imaging results;

[0094] Step 404: For any target region, based on the signal-to-noise ratio (SNR) corresponding to each integral imaging result of the target region and the integration time corresponding to each integral imaging result of the target region, construct an SNR variation curve of the SNR of the target region with respect to the integration time, and use the integration time corresponding to the maximum value in the SNR variation curve as the first integration time.

[0095] In the embodiments of the present application, after determining each target region, different integration times can be re - used to perform integral imaging processing on each target region. The number of integration times used in this process can be more than the number of integration times used in step 102, and the difference between each integration time used can also be shorter than that in step 102, so as to make the relationship between the determined SNR and the integration time more accurate.

[0096] Since the particles are in a high - speed motion state and the single - pixel dwell time is limited, when the integration time is lower than the single - pixel dwell time of the particles, the SNR of the target region corresponding to the particles is proportional to the integration time. When the integration time exceeds the single - pixel dwell time of the particles, due to the displacement of the particles resulting in imaging blur and trailing, the SNR will be inversely proportional to the integration time. The theoretical formula for the relationship between the SNR and the integration time can be seen in formula (1):

[0097] Formula (1)

[0098] Where, is the SNR at the integration time t, is the particle intensity per unit time, is the background intensity per unit time, is the single - pixel dwell time of the particles.

[0099] Therefore, it can be determined that when the integration time is the single - pixel dwell time of the particles, the SNR of the target region is the largest and the imaging effect of the target region is also the best. After determining the SNR of the target region for different integration times, the SNR variation curve of the SNR with respect to the integration time can be obtained by function fitting or interpolation, and then the integration time corresponding to the maximum value (i.e., the maximum point of the SNR) in the SNR variation curve is used as the first integration time.

[0100] Refer to Figure 5 as shown in the figure, which is a schematic diagram of the SNR variation curve. In Figure 5It can be determined that the integration time when the signal-to-noise ratio change curve of the target area where particle 1 is located reaches the maximum value is 1 ms, the integration time when the signal-to-noise ratio change curve of the target area where particle 2 is located reaches the maximum value is 3.5 ms, and the integration time when the signal-to-noise ratio change curve of the target area where particle 3 is located reaches the maximum value is 6 ms. Therefore, the first integration time of the target area where particle 1 is located can be determined to be 1 ms, the first integration time of the target area where particle 2 is located can be determined to be 3.5 ms, and the first integration time of the target area where particle 3 is located can be determined to be 6 ms.

[0101] The particle image reconstruction method provided by the embodiment of the present application constructs a signal-to-noise ratio change curve of the target area with respect to the integration time, and selects the integration time corresponding to the maximum value in the signal-to-noise ratio change curve as the first integration time. Therefore, for each target area, the integration imaging process can be performed using the first integration time that can maximize its signal-to-noise ratio, improving the accuracy of the integration imaging result.

[0102] In one embodiment, as Figure 6 shown, in step 104, determining the second integration time corresponding to the background area according to the brightness of the background area in each integration imaging result includes:

[0103] Step 602, respectively determining the saturation integration time when the brightness of each pixel point in the background area reaches saturation according to the pixel points located in the background area and having saturated brightness in each integration imaging result;

[0104] Step 604, respectively determining the second integration time corresponding to each pixel point according to the overexposure suppression strategy and the saturation integration time.

[0105] In the embodiment of the present application, when the same second integration time is used for the integration imaging of the background area, the brightness of each pixel point may not be consistent, and there may be a problem of uneven background area. Therefore, in order to make the brightness of the background area in the finally obtained particle image relatively uniform, the saturation integration time that can make each pixel point reach saturation can be determined for each pixel point in the background area, and then the second integration time can be obtained by processing the saturation integration time based on the overexposure suppression strategy, so that the integration imaging process can be performed for each pixel point according to different second integration times, and finally the integration imaging results with similar brightness can be obtained for each pixel point.

[0106] After performing integral imaging processing on the original pulse image data with different integration times and determining the background region in the original pulse image data, the part corresponding to the background region can be determined from each integral imaging result, and it can be determined which integral imaging result each pixel point in the background region reaches saturation at, so as to obtain the saturation integration time of each pixel point. Alternatively, according to the brightness of each pixel point in each integral imaging result, a relationship curve between the brightness of the pixel point and the integration time can be fitted, and the integration time corresponding to when the brightness of the pixel point reaches the saturation value in the relationship curve is used as the saturation integration time. The embodiments of the present application do not make specific limitations on this.

[0107] After processing the saturation integration time based on the overexposure suppression strategy, the second integration time can be obtained. The embodiments of the present application do not make specific limitations on the overexposure suppression strategy, as long as after performing integral imaging processing on the pixel points in the background region with the second integration time obtained according to the overexposure suppression strategy, the brightness differences of each pixel point in the obtained integral imaging result are within a certain range, and the brightness of each pixel point is lower than the preset brightness threshold. Exemplarily, the overexposure suppression strategy can be multiplying the saturation integration time by an overexposure suppression coefficient (such as 0.2), or the overexposure suppression strategy can also be subtracting a preset value from the saturation integration time, etc. The embodiments of the present application do not make specific limitations on this.

[0108] Taking the overexposure suppression strategy of multiplying the saturation integration time by an overexposure suppression coefficient as an example, for any pixel point in the original pulse image data, the integration time formula of this pixel point is as shown in formula (two):

[0109] Formula (two)

[0110] Wherein, is the integration time of the pixel point located at position , is the first integration time of the target region to which this pixel point belongs when this pixel point belongs to the target region, is the saturation integration time determined for this pixel point when this pixel point belongs to the background region, is the overexposure suppression coefficient.

[0111] The particle image reconstruction method provided by the embodiments of the present application determines the saturation integration time when each pixel point in the background region reaches saturation, processes the saturation integration time according to the overexposure suppression strategy to obtain the corresponding second integration time for each pixel point, and performs integral imaging processing on each pixel point in the background region with different second integration times. Therefore, the optimal integration time that can suppress overexposure can be determined for each pixel point in the background region, avoiding the phenomenon of overexposure in some regions of the background region and improving the imaging accuracy of the particle image.

[0112] In one embodiment, as Figure 7 shown, in step 106, according to the integral imaging results of each target region and the background region, multiple frames of target particle images corresponding to the particle region are reconstructed, including:

[0113] Step 702, determining each imaging moment corresponding to the particle region within the preset sampling duration according to the pulse sampling time and the preset sampling duration;

[0114] Step 704, respectively determining the corresponding first integral imaging result in the integral imaging result of the background region at each imaging moment, and respectively determining the corresponding second integral imaging result in the integral imaging result of each target region at each imaging moment;

[0115] Step 706, generating target particle images corresponding to each imaging moment based on the first integral imaging result and the second integral imaging result.

[0116] In the embodiment of the present application, the pulse sampling time is also the shortest duration required for the pulse high-speed camera to sample one frame of original pulse image data. Therefore, the integration time is a multiple of the pulse sampling time of the pulse high-speed camera, and the pulse sampling time can be used as the basic time unit for particle image reconstruction. Theoretically, in the preset sampling duration, a total of preset sampling duration divided by the pulse sampling time times of sampling can be performed, so that preset sampling duration divided by the pulse sampling time number of target particle images can also be obtained.

[0117] The imaging moment of each frame of target particle image in the preset sampling duration can be determined according to the pulse sampling time, and then the corresponding first and second integral imaging results at this imaging moment can be determined from the integral imaging results of each target region and the background region, and these integral imaging results are combined into the target particle image corresponding to this imaging moment.

[0118] Taking the pulse sampling time as T s , and the integration time of a certain target region is T z as an example, an integral imaging result can be obtained for this target region every n = T z / T s pulse sampling times. Further, referring to the description in the foregoing embodiment, since the integration time of the target region is the single-pixel residence time of the particles in the target region, it means that the particles will not move within one integration time. Therefore, if the target region obtains an integral imaging result for the time period from T m to T m+n , then the first integral imaging results corresponding to these n imaging moments from T to T m to T m+n can all be .

[0119] By a similar method, the corresponding first integral imaging result in the integral imaging results of other target regions and the corresponding second integral imaging result in the integral imaging results of the background region can be determined for each imaging moment. After obtaining the first integral imaging results corresponding to all target regions and the second integral imaging results corresponding to the background region for a specific imaging moment, these integral imaging results can be combined to obtain the target particle image.

[0120] See Figure 8 as shown, which is a schematic diagram of the above process. Figure 8 where the imaging moments are from t 0 to t 5 . The original pulsed image data collected from the moment of t 0 to the moment of t 5 can be seen in the upper half of Figure 8 . Through the signal-to-noise ratio analysis of the foregoing embodiments, it can be obtained that for all target regions (hereinafter referred to as top target regions) where the particles in the top region have moved to, the first integration time is 3T s . For all target regions (hereinafter referred to as middle target regions) where the particles in the middle region have moved to, the first integration time is 2T s . For all target regions (hereinafter referred to as bottom target regions) where the particles in the lower region have moved to, the first integration time is T s .

[0121] By using the first integration time to perform integral imaging on each target region respectively, the integral imaging result of the 1st bottom target region can be obtained at t 0 . The integral imaging results of the 2nd bottom target region and the 1st middle target region can be obtained at t 1 . The integral imaging results of the 3rd bottom target region and the 1st top target region can be obtained at t 2 and so on. At the same time, since the particles in the middle region have been in the 1st middle target region from t 0 to t 1 , the position of the particles in the middle region reflected by the integral imaging result of the 1st middle target region at t 1 is also the position of the particles in the middle region at t 0 . Therefore, the integral imaging result of the 1st middle target region obtained at t 1 can be copied to t 0 . Similarly, similar operations can be performed on other middle target regions and other top target regions. The target particle image at each imaging moment finally obtained is as shown in the lower half of Figure 8 .

[0122] The particle image reconstruction method provided by the embodiments of the present application can perform time-frame alignment on the integral imaging results of each target region after obtaining the integral imaging results, and can generate target particle images with higher clarity for each imaging moment in the preset sampling duration, improving the generation accuracy of the target particle images.

[0123] In one embodiment, the above method further includes:

[0124] Performing denoising processing on each target particle image based on at least one of the pixel difference between every two adjacent target particle images, the first target filter, and the second target filter to obtain the denoised target particle image;

[0125] Wherein, the first target filter is constructed based on the target shape characteristics of the particles, and the second target filter is constructed based on the target motion characteristics of the particles.

[0126] In the embodiments of the present application, since the pulse emission sensitivities of the pixels in the pulsed high-speed camera will inevitably be inconsistent, and the camera circuit will generate dark current during actual operation, spatial and temporal noises may appear in the reconstructed target particle images, resulting in a decrease in the signal-to-noise ratio and affecting the subsequent particle image velocimetry (PIV) effect. Therefore, after reconstructing the target particle images, denoising processing can also be performed on the target particle images.

[0127] The denoising method can adopt pulse emission frequency analysis, the pixel difference between every two adjacent target particle images (i.e., the frame difference method), or a filter, where the filter can be constructed based on the target shape characteristics and target motion characteristics that particles generally have in the particle images taken for the particles. The target shape characteristics can include the intensity characteristics of the pixel points containing the particles, the shape characteristics of the region where the particles are located, etc., and the target motion characteristics can include the change law of the pulse emission frequencies of adjacent pixel points when the particles move.

[0128] The principle of denoising based on the pulse emission frequency is that the brightness of the particles in the particle region is generally within a fixed range. Therefore, when using a pulsed high-speed camera to capture the particle region, the pulse emission frequency corresponding to each pixel point should also be within a fixed range. This fixed range can be obtained through statistical methods. As Figure 9 shown, it is a histogram of the number of pixels with a certain pulse emission frequency obtained after statistically analyzing the pulse emission frequencies corresponding to each pixel point. Through relevant algorithms in statistics, the outliers in the histogram can be determined, and then the pixel points corresponding to these outliers can be determined as the pixel points containing temporal noise, and through any image denoising algorithm, denoising processing can be performed on these pixel points in each target particle image.

[0129] In one embodiment, denoising each target particle image based on the pixel difference between every two adjacent initial particle images includes:

[0130] Performing a difference operation on any two adjacent target particle images. When the pixel difference corresponding to any pixel point in the two target particle images is greater than a first preset threshold and less than a second preset threshold, the pixel point is determined as a noise pixel point, and denoising is performed on each noise pixel point in the two target particle images.

[0131] In the embodiments of the present application, the frame difference method is used to denoise the target particle images. Since in the particle image velocimetry (PIV) scenario, the shooting position of the pulsed high-speed camera is fixed, the background area remains stationary during the shooting process. Reflected in the image, the pixel values of the background area in different target particle images generally do not differ much. Only the movement of the particles and the temporal noise generated by the accumulation of dark current will cause changes in the pixel values of the target particle images. At the same time, compared with the pixel value changes caused by temporal noise, the pixel value changes caused by the movement of the particles are generally more intense. Therefore, based on this characteristic, the first preset threshold and the second preset threshold can be statistically obtained through simulation experiments and other means. If after performing a difference operation on two target particle images, the pixel difference corresponding to a certain pixel point is less than or equal to the first preset threshold, it can be considered that this pixel point belongs to the background area. If the pixel difference corresponding to a certain pixel point is greater than or equal to the second preset threshold, it can be considered that this pixel point belongs to the particle movement area. When the pixel difference corresponding to a pixel point is greater than the first preset threshold and less than the second preset threshold, it can be considered that the change in the pixel value of this pixel point is caused by noise. At this time, through any image denoising algorithm, denoising can be performed on this pixel point in these two target particle images respectively.

[0132] In one embodiment, the target shape features of the particles include an area feature and an intensity feature. The area feature is that the number of pixel points corresponding to a single particle in the particle image is greater than 1, and the intensity feature is that among all the pixel points corresponding to a single particle in the particle image, the intensity of the central pixel point is greater than the intensity of the edge pixel points;

[0133] The target motion feature of the particles is that in the original pulsed image data, the change values of the pulse emission frequencies of adjacent multiple pixel points have an associated relationship.

[0134] In the embodiments of the present application, particles generally occupy multiple pixel points in the particle image, and the occupied pixel point array presents a dot-like shape feature (the intensity of the central pixel point is the largest, and the intensity gradually decreases outward), while noise usually occupies a single pixel point, or the occupied pixel point array has an irregular shape, such as Figure 10As shown in the figure. Therefore, based on this feature, the target shape feature of the particle can be determined as that the number of pixel points corresponding to a single particle in the particle image is greater than 1, and among all the pixel points corresponding to a single particle in the particle image, the intensity of the central pixel point is greater than that of the edge pixel points. Based on the above target shape feature, a high-pass filter can be designed as the first target filter, and each target particle image is convolved to obtain a denoised image. The specific formula of the high-pass filter is shown in Formula (III):

[0135] Formula (III)

[0136] wherein, is the target particle image, is the convolution kernel of the high-pass filter, is the value of the target particle image at the coordinate value , is the value of the convolution kernel relative to deviating from . By setting the threshold value of the high-pass filter, the target conforming to the target shape feature of the particle can be obtained, thereby removing part of the time-domain noise.

[0137] The principle of the target motion feature of the particle is that since the particle is in a high-speed motion state, there is a certain correlation relationship between the pulse emission frequencies of adjacent pixel points in the original pulse image data. Specifically, when the pulse emission frequency feature of the current pixel changes, the pulse emission frequency features of adjacent pixels change correlatively, and the two change values are in a corresponding relationship. Therefore, based on the above particle motion characteristics, a multi-scale time-domain filter as shown in Figure 11 can be designed as the second target filter. A time window is superimposed on each current-frame target particle image, and the target number of pixel points whose pulse emission frequency features change as above detected in a time window in each region is calculated respectively through convolution, and the target numbers in two scale windows and the size of the target number in the window and the set prior threshold are compared, so as to screen and filter out the time-domain noise that does not conform to the motion feature, and finally restore and reconstruct a clear high signal-to-noise ratio image.

[0138] The above particle image reconstruction method has the following remarkable advantages compared with the traditional particle image velocimetry (PIV) method of high-speed cameras and the existing pulse image reconstruction technology:

[0139] Pulse image reconstruction under a wide dynamic range: It can reconstruct a clear image for particle image velocimetry (PIV) in high dynamic range imaging scenarios such as combustion flow fields or strong backlight flow fields, overcoming the problems of overexposure in high-brightness regions and underexposure in low-brightness regions of traditional cameras, as well as the inability of traditional pulse imaging technology to take into account both bright and dark regions and the blurring and trailing of high-speed particles.

[0140] Denoising method based on particle spatio-temporal characteristics: A denoising method is proposed for the particle image velocimetry (PIV) scenario, which combines the characteristics of particles in the spatial domain (intensity and shape features) and the temporal domain (motion features). It effectively suppresses and removes the spatial noise caused by inconsistent pixel pulse emission sensitivities and the temporal noise generated by dark current accumulation, achieving high signal-to-noise ratio imaging.

[0141] Image reconstruction algorithm based on region adaptive integration: An image reconstruction algorithm based on region adaptive integration is proposed, which is specifically designed and optimized for the characteristics of pulse image data and moving particles. Based on the characteristic of independent integration of each pixel in the pulse image data, the target region and background region containing particles are iteratively and finely divided. According to the signal-to-noise ratio change curve of the target region, the pixel residence time of each particle is determined to obtain the optimal integration time that takes into account both bright and dark regions. At the same time, overexposure suppression is performed according to the saturation integration time of the background region to prevent image overexposure, realizing the conversion of the original pulse data of the pulsed high-speed camera into high-dynamic particle images.

[0142] It should be understood that although the steps in the flowcharts involved in the above-described embodiments are shown in sequence according to the arrows, these steps do not necessarily have to be executed in the order indicated by the arrows. Unless there is a clear indication in this article, the execution of these steps does not have a strict order limit, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-described embodiments may include multiple steps or multiple stages. These steps or stages do not necessarily have to be executed at the same time, but can be executed at different times. The execution order of these steps or stages does not necessarily have to be sequential, but can be executed alternately or in turn with at least a part of other steps or steps or stages in other steps.

[0143] Based on the same inventive concept, an embodiment of the present application further provides a particle image reconstruction device for implementing the above-mentioned particle image reconstruction method. The solution provided by this device to solve the problem is similar to the solution described in the above method. Therefore, the specific limitations in one or more embodiments of the particle image reconstruction device provided below can refer to the limitations on the particle image reconstruction method in the above text and will not be repeated here.

[0144] In one embodiment, as Figure 12 shown, a particle image reconstruction device 1200 is provided, including: an integration module 1202, a determination module 1204, and a reconstruction module 1206, where:

[0145] An integration module 1202 is configured to collect original pulse image data for a particle region within a preset sampling duration, perform integration imaging processing on the original pulse image data multiple times with different integration times, and determine a target region including particles and a background region not including particles in the original pulse image data according to the obtained integration imaging results;

[0146] A determination module 1204 is configured to determine the signal-to-noise ratios of the respective target regions corresponding to different integration times, determine a first integration time corresponding to each target region according to the respective signal-to-noise ratios, and determine a second integration time corresponding to the background region according to the brightness of the background region in each integration imaging result;

[0147] A reconstruction module 1206 is configured to perform integration imaging processing on each target region according to the first integration time respectively, perform integration imaging processing on the background region according to the second integration time, and reconstruct multiple frames of target particle images corresponding to the particle region according to the integration imaging results of each target region and the background region.

[0148] The particle image reconstruction device provided by the embodiment of the present application performs imaging on the original pulse image data according to different integration times, identifies a target region including particles and a background region not including particles in the original pulse image data according to the integration imaging results, then determines a first integration time used for performing integration imaging on the target region according to the signal-to-noise ratios of the target regions corresponding to different integration times, and determines a second integration time corresponding to the background region according to the brightness of the background region in different integration imaging results, and then performs integration imaging on each region respectively using the determined integration time, so as to obtain target particle images corresponding to the particle region. The embodiment of the present application determines different integration times for different regions, and performs imaging on each region using a better integration time. Therefore, when processing images with a large difference in brightness between different regions, it is also possible to take both bright and dark regions into account for image reconstruction, improving the reconstruction accuracy of particle images.

[0149] In one embodiment, the integration module 1202 is further configured to:

[0150] Use the original pulse image data as target pulse image data, perform integration imaging processing on the target pulse image data using a target integration time to obtain an integration imaging result, determine a particle region including particles from the integration imaging result, and use the part of the original pulse image data corresponding to the particle region as the target region;

[0151] Take the part of the original pulse image data except for each target region as the target pulse image data, increase the target integration time, and jump to the step of performing integral imaging processing on the target pulse image data with the target integration time to obtain an integral imaging result until all pixel points in the integral imaging result reach saturation;

[0152] Take the part of the original pulse image data corresponding to the current target pulse image data as the background region.

[0153] In one embodiment, the determining module 1204 is further configured to:

[0154] Perform integral imaging processing on each target region with different integration times, respectively obtain multiple integral imaging results of each target region, and respectively determine the signal-to-noise ratios of each integral imaging result;

[0155] For any one of the target regions, based on the signal-to-noise ratios corresponding to the integral imaging results of the target region and the integration times corresponding to the integral imaging results of the target region, construct a signal-to-noise ratio change curve of the signal-to-noise ratio of the target region with respect to the integration time, and take the integration time corresponding to the maximum value in the signal-to-noise ratio change curve as the first integration time.

[0156] In one embodiment, the determining module 1204 is further configured to:

[0157] According to the pixel points located in the background region and having saturated brightness in each integral imaging result, respectively determine the saturation integration time when the brightness of each pixel point in the background region reaches saturation;

[0158] According to the overexposure suppression strategy and the saturation integration time, respectively determine the second integration time corresponding to each pixel point.

[0159] In one embodiment, the reconstruction module 1206 is further configured to:

[0160] According to the pulse sampling time and the preset sampling duration, determine each imaging moment corresponding to the particle region within the preset sampling duration;

[0161] Respectively determine the corresponding first integral imaging result in the integral imaging result of the background region at each imaging moment, and respectively determine the corresponding second integral imaging result in the integral imaging result of each target region at each imaging moment;

[0162] Based on the first integral imaging result and the second integral imaging result, respectively generate the target particle image corresponding to each imaging moment.

[0163] In one embodiment, the above-mentioned device further includes:

[0164] A denoising module, configured to perform denoising processing on each of the target particle images based on at least one of the pulse firing frequencies corresponding to the pixel points in the original pulse image data, the pixel difference between every two adjacent target particle images, a first target filter, and a second target filter, to obtain denoised target particle images;

[0165] Wherein, the first target filter is constructed based on the target shape characteristics of the particles, and the second target filter is constructed based on the target motion characteristics of the particles.

[0166] In one embodiment, the denoising module is further configured to:

[0167] Perform a difference operation on any two adjacent target particle images. When the pixel difference corresponding to any pixel point in the two target particle images is greater than a first preset threshold and less than a second preset threshold, determine the pixel point as a noise pixel point, and perform denoising processing on each of the noise pixel points in the two target particle images.

[0168] In one embodiment, the target shape characteristics of the particles include an area characteristic and an intensity characteristic. The area characteristic is that the number of pixel points corresponding to a single particle in the particle image is greater than 1, and the intensity characteristic is that among all the pixel points corresponding to a single particle in the particle image, the intensity of the central pixel point is greater than the intensity of the edge pixel points;

[0169] The target motion characteristic of the particles is that in the original pulse image data, there is a correlation relationship between the change values of the pulse firing frequencies of adjacent multiple pixel points.

[0170] Each module in the above-mentioned device can be implemented in whole or in part by software, hardware, and their combination. The above-mentioned modules can be embedded in the processor in the computer device in hardware form or be independent of it, or can be stored in the memory in the computer device in software form, so that the processor can call and execute the operations corresponding to the above-mentioned modules.

[0171] In one embodiment, a computer device is provided. The computer device can be a server, and its internal structure diagram can be as Figure 13As shown. The computer device includes a processor, a memory, and a network interface connected via a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, it implements a particle image reconstruction method.

[0172] Those skilled in the art can understand that Figure 13 the structure shown in is only a block diagram of some structures related to the solution of this application, and does not constitute a limitation on the computer device to which the solution of this application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements.

[0173] In one embodiment, a computer device is provided, including a memory and a processor. A computer program is stored in the memory. When the processor executes the computer program, the steps in the above method embodiments are implemented.

[0174] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by the processor, the steps in the above method embodiments are implemented.

[0175] In one embodiment, a computer program product is provided, including a computer program. When the computer program is executed by the processor, the steps in the above method embodiments are implemented.

[0176] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0177] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memories. Non-volatile memories can include read-only memory (ROM), magnetic tapes, floppy disks, flash memories, optical memories, high-density embedded non-volatile memories, resistive random access memories (ReRAM), magnetoresistive random access memories (MRAM), ferroelectric random access memories (FRAM), phase change memories (PCM), graphene memories, etc. Volatile memories can include random access memory (RAM) or external cache memories, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The databases involved in the embodiments provided in the present application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., without limitation. The processors involved in the embodiments provided in the present application can be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, data processing logics based on quantum computing, etc., without limitation.

[0178] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope recorded in this specification.

[0179] The above-described embodiments merely represent several implementation manners of the present application. The description is relatively specific and detailed, but it should not be construed as a limitation on the patent scope of the present application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the appended claims.

Claims

1. A particle image reconstruction method, characterized in that: The method comprises: Collecting raw pulse image data for a particle region within a preset sampling time, performing integral imaging processing on the raw pulse image data for multiple times using different integral times, and determining a target region including particles and a background region excluding particles in the raw pulse image data according to each integral imaging result obtained; Determine the signal-to-noise ratio of each target area corresponding to different integration times, determine the first integration time corresponding to each target area according to each signal-to-noise ratio, and determine the second integration time corresponding to the background area according to the brightness of the background area in each integral imaging result; Integral imaging processing is performed on each of the target areas according to the first integration time, and integral imaging processing is performed on the background area according to the second integration time. Based on the integral imaging results of each of the target areas and the background area, a multi-frame target particle image corresponding to the particle area is reconstructed.

2. The method according to claim 1, characterized in that The step of performing integral imaging processing on the raw pulse image data for multiple times using different integral times, and determining a target area including particles and a background area not including particles in the raw pulse image data according to each integral imaging result obtained, comprises: The original pulse image data is used as target pulse image data, and the target pulse image data is subjected to integral imaging processing using a target integration time to obtain an integral imaging result, and a particle region including particles is determined from the integral imaging result, and a portion of the original pulse image data corresponding to the particle region is used as a target region; The steps of using the portion of the original pulse image data other than the target regions as target pulse image data, increasing the target integration time, jumping to the step of performing integral imaging processing on the target pulse image data using the target integration time to obtain an integral imaging result, determining a particle region including particles from the integral imaging result, and using the portion of the original pulse image data corresponding to the particle region as the target region, until all pixel points in the integral imaging result are saturated; The portion of the original pulse image data corresponding to the current target pulse image data is used as the background area.

3. The method according to claim 1, characterized in that The step of determining the signal-to-noise ratios of the target areas corresponding to different integration times, and determining the first integration time corresponding to the target areas according to the signal-to-noise ratios, comprises: Performing integral imaging processing on each of the target areas using different integral times to obtain a plurality of integral imaging results for each of the target areas, and determining a signal-to-noise ratio of each of the integral imaging results; For any of the target areas, based on the signal-to-noise ratio corresponding to each of the integrated imaging results of the target area and the integration time corresponding to each of the integrated imaging results of the target area, a signal-to-noise ratio variation curve of the signal-to-noise ratio of the target area with respect to the integration time is constructed, and the integration time corresponding to the maximum value in the signal-to-noise ratio variation curve is used as the first integration time.

4. The method according to claim 1, characterized in that: The determining, according to the brightness of the background area in each of the integral imaging results, a second integration time corresponding to the background area comprises: According to the pixel points in the background area and whose brightness reaches saturation in each of the integral imaging results, respectively determining the saturation integration time when the brightness of each of the pixel points in the background area reaches saturation; According to the overexposure suppression strategy and the saturation integration time, the second integration time corresponding to each of the pixel points is determined respectively.

5. The method according to claim 1, characterized in that The step of reconstructing a plurality of target particle images corresponding to the particle region according to the integrated imaging results of each of the target region and the background region comprises: Determine, according to the pulse sampling time and the preset sampling duration, each imaging moment corresponding to the particle region within the preset sampling duration; Respectively determining a first integral imaging result corresponding to the integral imaging result of the background area at each of the imaging moments, and respectively determining a second integral imaging result corresponding to the integral imaging result of the target area at each of the imaging moments; Based on the first integral imaging result and the second integral imaging result, target particle images corresponding to each imaging moment are generated respectively.

6. The method according to claim 1, characterized in that The above method further includes: Based on the pulse emission frequency corresponding to each pixel point in the original pulse image data, the pixel difference between each two adjacent target particle images, and at least one of the first target filter and the second target filter, denoising each target particle image to obtain a denoised target particle image; The first target filter is constructed based on the target shape characteristics of the particles, and the second target filter is constructed based on the target motion characteristics of the particles.

7. The method according to claim 6, characterized in that The method further comprises: performing denoising processing on each of the target particle images based on the pixel difference between each two adjacent initial particle images, comprising: A difference processing is performed on any two adjacent frames of target particle images. When the pixel difference corresponding to any pixel point in the two frames of target particle images is greater than a first preset threshold and less than a second preset threshold, the pixel point is determined as a noise pixel point, and denoising is performed on each of the noise pixel points in the two frames of target particle images.

8. The method according to claim 6, characterized in that The target shape feature of the particle includes an area feature and an intensity feature, wherein the area feature is that the number of pixel points corresponding to a single particle in the particle image is greater than 1, and the intensity feature is that among all the pixel points corresponding to a single particle in the particle image, the intensity of the central pixel point is greater than the intensity of the edge pixel points; The target motion feature of the particle is that in the original pulse image data, the change values ​​of the pulse emission frequencies of a plurality of adjacent pixel points have a correlation relationship.

9. A particle image reconstruction device, characterized in that: The device comprises: An integration module is used to collect raw pulse image data for a particle region within a preset sampling time, perform integral imaging processing on the raw pulse image data using different integral times for multiple times, and determine a target region including particles and a background region excluding particles in the raw pulse image data according to each integral imaging result obtained; A determination module, used to determine the signal-to-noise ratio of each target area corresponding to different integration times, determine the first integration time corresponding to each target area according to each signal-to-noise ratio, and determine the second integration time corresponding to the background area according to the brightness of the background area in each integral imaging result; A reconstruction module is used to perform integral imaging processing on each of the target areas according to the first integration time, and to perform integral imaging processing on the background area according to the second integration time, and to reconstruct a multi-frame target particle image corresponding to the particle area based on the integral imaging results of each of the target areas and the background area.

10. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 8 are implemented.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

Images