Method and system for preprocessing and compressing single-photon laser radar data

Abstract

The invention discloses a preprocessing and compression method and system for single-photon laser radar data, and belongs to the technical field of single-photon laser radar data processing. According to the method, firstly, a target distance image is collected by means of a single-photon laser radar imaging system, and original data is provided for subsequent processing; thirdly, constructing a frequency-time histogram for each pixel point of the acquired image so as to visually present the distribution condition of photons in the time dimension; then, Gaussian fitting and peak value screening operation is carried out on the histogram of each pixel point, effective photon signal peak values are accurately identified, false signals caused by noise and interference are eliminated, and the data quality is improved. And finally, inputting the photon data corresponding to the screened peak values into a multi-attention three-dimensional single-photon compression neural network, and realizing high-efficiency compression of the single-photon laser radar data by using the strong feature extraction, high-efficiency compression mechanism and flexible output control capability of the network, thereby providing powerful support for the application of the single-photon laser radar technology in multiple fields.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of single-photon lidar data processing, and relates to a single-photon lidar data preprocessing and compression method and system. BACKGROUND

[0002] With the continuous progress of lidar technology, single-photon lidar systems based on single-photon avalanche diode arrays (SPAD) have been widely used in many fields due to their excellent performance. In the field of high-precision three-dimensional imaging, it can accurately capture the three-dimensional shape of objects, providing key data support for industrial manufacturing, cultural heritage protection, etc.; in the automatic driving scene, it can perceive the three-dimensional information of the surrounding environment in real time, helping vehicles make accurate decisions and ensuring driving safety; in the field of topographic mapping, single-photon lidar can efficiently obtain large-area terrain data, providing high-precision basic information for geographic information systems (GIS); in the field of biomedical imaging, it provides a powerful tool for the study of microscopic structures and tissue morphology.

[0003] A significant advantage of such systems is that they can achieve sub-centimeter depth perception even in extremely low light conditions, with extremely high time resolution and detection sensitivity. This allows it to work stably in complex environments and obtain high-quality detection data, providing a solid guarantee for applications in various fields.

[0004] In practical applications, single-photon lidar usually outputs frequency-time histogram data. Specifically, for each pixel point in the field of view, the system will count the arrival time of photons over multiple laser emission periods, and then form a histogram on the time axis. This histogram clearly reflects the detection frequency in different time bins, providing an important basis for analyzing the distance, reflectivity, etc. of the target object.

[0005] However, due to the high timing accuracy of single photons, the time axis is often subdivided into thousands of time bins. For high-resolution fields of view, the original data size of each frame is extremely large, up to millions of data points. In continuous acquisition scenarios, the data volume grows exponentially. This massive amount of data puts a huge pressure on the system, greatly increasing the bandwidth, storage and computing costs. Especially in resource-constrained edge devices or real-time systems, the timeliness and accuracy of data processing are severely challenged, and efficient compression strategies are needed to alleviate data pressure.

[0006] In order to deal with the problem of large amount of single-photon lidar data, traditional compression algorithms such as histogram threshold screening, time domain downsampling, entropy encoding and principal component analysis are applied in data processing. These methods can reduce the data volume to a certain extent, and reduce the pressure of data storage and transmission.

[0007] However, they have many drawbacks in practical application. Especially in processing multi-peak signal, complex background noise or sparse data, these methods often cannot effectively preserve the detailed information of the target echo. Histogram threshold screening may mistakenly filter out some key but small amplitude signal peaks; time domain downsampling will cause the time resolution to be reduced, and part of the important information will be lost; entropy encoding and principal component analysis have limited compression effect when facing complex data structure, and may not be accurate enough for feature extraction of data. These problems make it difficult to recover high-precision target information when the compressed data is reconstructed, which seriously affects the performance and effect of single-photon lidar system in practical application. SUMMARY

[0008] The purpose of the present application is to solve the technical problem that the existing single-photon lidar generates a large amount of redundant data when obtaining high-precision depth information, and it is difficult to perform efficient real-time calculation, and to provide a single-photon lidar data preprocessing and compression method and system.

[0009] In order to achieve the above purpose, the following technical scheme is adopted: The present application provides a single-photon lidar data preprocessing and compression method, comprising the following steps: Collecting target distance images through a single-photon lidar imaging system; Constructing a corresponding frequency-time histogram for each pixel point of the collected target distance image; Gaussian fitting is performed on the frequency-time histogram of each pixel point, and peak value screening is performed; The photon data corresponding to the screened peak values are input into a multi-attention three-dimensional single-photon compression neural network for data compression.

[0010] Further, the single-photon lidar imaging system adopts a paraxial single-photon lidar system, which includes a laser, a collimating mirror, a single-photon detector, a time-dependent single-photon counting module, a driver and a computer; The laser is electrically connected to the time-dependent single-photon counting module, and the laser emission time of the laser is used as the photon emission time; the laser emitted by the laser is collimated by the collimating mirror and directed to the target; the reflected light of the target enters the single-photon detector for signal detection; The single-photon detector is electrically connected to the time-correlated single-photon counting module, and the time when the single-photon detector receives the laser is taken as the photon reception time. A displacement stage is provided at the bottom of the single-photon detector; the driver is electrically connected to the displacement stage, and the driver controls the displacement stage to move the single-photon detector to adjust the target distance and achieve scanning of targets at different distances.

[0011] Furthermore, the step of performing Gaussian fitting on the frequency-time histogram of each pixel and filtering for peak values ​​specifically involves: The frequency-time histogram of each pixel is fitted with a Gaussian mixture function to obtain the photon distribution. One fitting function; The optimization is achieved using the least squares method. The parameters of the fitting function are obtained. The parameters of a fitting function; according to Peak values ​​are selected from the parameters of the fitted function to obtain the target peak value.

[0012] Furthermore, the photon distribution is fitted using a Gaussian mixture function as follows:

[0013] in, It is the first The amplitude of the Gaussian peak; It is the first The mean of the Gaussian peaks; It is the first The standard deviation of the Gaussian peak; Represents the location of Mount Gauss; The first image representing the target distance Line number The Gaussian fit result of the frequency-time histogram of the pixels in the column; exp() represents the natural exponential function; This indicates the total number of Gaussian peaks.

[0014] Furthermore, the optimization of the least squares method The parameters of the fitting function are as follows: Updated using least squares method Until the loss function is minimized; the expression for the loss function is:

[0015] in, The set of parameters to be optimized and iterated; The actual observed number in the frequency-time histogram One data point; This represents the fitting function based on the Gaussian mixture model distribution; Indicates the first The location of the Gaussian peak.

[0016] Furthermore, the aforementioned according to Peak values ​​are filtered from the parameters of the fitted function to obtain the target peak value. Specifically, when the amplitude of the fitted function is greater than a set threshold, the signal peak value is retained as the target peak value.

[0017] Furthermore, the step of inputting the photon data corresponding to the selected peaks into a multi-attention three-dimensional single-photon compression neural network for data compression specifically involves: Photon data within the time range of the target peak are input into a multi-attention three-dimensional single-photon compression neural network for data compression; the time range is:

[0018] in, It is the first The mean of the Gaussian peaks; It is the first The standard deviation of the Gaussian peak; Indicates the time range of the target peak.

[0019] Furthermore, the multi-attention three-dimensional single-photon compressed neural network includes a spatiotemporal attention feature extraction module, a deep compression extraction module, and a compression mapping module connected in sequence; The spatiotemporal attention feature extraction module includes several sequentially connected 3D convolution and activation units, attention mechanism module, regularization and normalization unit, and 3D max pooling downsampling unit; The deep compression extraction module includes a deep compression extraction stacked unit and a downsampling module connected in sequence; the deep compression extraction stacked unit includes a depth convolution unit, a point convolution unit, and a normalization and nonlinear activation unit; the output of the point convolution unit is connected to the output of the normalization and nonlinear activation unit; the deep compression extraction stacked unit is stacked repeatedly several times; The compression mapping module includes, in sequence, a flattening layer, a multi-layer fully connected transformation unit, and a Sigmoid mapping layer; the multi-layer fully connected transformation unit is repeatedly stacked several times.

[0020] A second aspect of the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the preprocessing and compression method for single-photon lidar data.

[0021] A third aspect of the present invention provides a preprocessing and compression system for single-photon lidar data, comprising: The data acquisition module collects target distance image through a single-photon laser radar imaging system; The histogram construction module constructs a corresponding frequency-time histogram for each pixel point of the collected target distance image; The peak value screening module performs Gaussian fitting on the frequency-time histogram of each pixel point and performs peak value screening; The data compression module inputs the screened peak value corresponding photon data into a multi-attention three-dimensional single-photon compression neural network for data compression.

[0022] Compared with the prior art, the present application has the following beneficial effects: The present application discloses a single-photon laser radar data preprocessing and compression method, which separately constructs a frequency-time histogram for each pixel point, converts single-photon data from "overall fuzzy distribution" to "pixel-level accurate timing characteristics", can clearly present the photon response law of each pixel point in different time dimensions, and effectively distinguishes the timing difference between target signal photons and background noise photons; Gaussian fitting can accurately model the effective signal trend in the histogram based on the natural distribution characteristics of the photon signal, reduce the interference of random noise on the signal characteristics; peak value screening further focuses on the effective signal peak after fitting, eliminates the photon data corresponding to low-amplitude noise, significantly improves the signal-to-noise ratio of the photon data, greatly reduces the effective data input into the compression network, reduces data redundancy, and reduces the subsequent compression cost; the multi-attention three-dimensional single-photon compression neural network can automatically identify the key feature area in the single-photon data through the attention mechanism, and allocate more compression resources to the key area, while efficiently compressing the background area with high redundancy, effectively improving the data compression ratio on the premise of ensuring the fidelity of the target data.

[0023] Further, the parameters of the fitting function are optimized by the least square method K , which can realize accurate solution of the fitting function parameters based on the error minimization target between the actual observation value and the predicted value of the fitting function of the photon count in the histogram, effectively reduce the interference of random noise on parameter estimation, and make the obtained K fitting function parameters have higher stability and reliability; based on the optimized K fitting function parameters, the peak value screening can accurately locate the peak position of each fitting function corresponding signal component, and then accurately separate the target signal corresponding peak value from the complex photon distribution, through the analysis of the fitting function parameters, the target signal peak value, background noise peak value and stray light interference peak value can be effectively distinguished, and the final screened target peak value can accurately reflect the distance information of the target, further reducing the influence of invalid noise signals on the subsequent compression link. BRIEF DESCRIPTION OF DRAWINGS

[0024] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiments. It should be understood that the following drawings only show some of the embodiments of the present application, and therefore should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can also be obtained without creative labor.

[0025] Figure 1 The flow chart of the preprocessing and compression method of the single-photon lidar data of the present application; Figure 2 The structure diagram of the paraxial single-photon lidar system for the embodiments of the present application; Figure 3 The photon data graph before coarse screening for the embodiments of the present application; Figure 4 The photon data graph after coarse screening for the embodiments of the present application; Figure 5 The structure diagram of the multi-attention three-dimensional single-photon compression neural network for the embodiments of the present application; Figure 6 The structure diagram of the space-time feature extraction module for the embodiments of the present application; Figure 7 The structure diagram of the deep compression extraction module for the embodiments of the present application; Figure 8 The structure diagram of the compression mapping module for the embodiments of the present application; Figure 9 The system block diagram of the preprocessing and compression system of the single-photon lidar data of the present application.

[0026] Among them, 1-time dependent single-photon counting module; 2-laser; 3-collimating mirror; 4-optical lens group; 5-filter; 6-single-photon detector; 7-displacement table; 8-driver; 9-computer; 901-data acquisition module; 902-histogram construction module; 903-peak screening module; 904-data compression module. DETAILED DESCRIPTION

[0027] In order to make the purpose, technical solutions and advantages of the embodiments of the present application more clear, the technical solutions in the embodiments of the present application will be described clearly and completely below in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are some of the embodiments of the present application, not all the embodiments. The components of the embodiments of the present application described and indicated in the drawings here can be arranged and designed in various different configurations.

[0028] Therefore, the following detailed description of the embodiments of the application provided in the accompanying drawings is not intended to limit the scope of the application claimed, but merely represents selected embodiments of the application. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the application without creative labor fall within the scope of the protection of the application.

[0029] It should be noted that similar reference numbers and letters represent similar items in the following drawings, and therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings.

[0030] The application will be further described in detail below in conjunction with the accompanying drawings: Referring to Figure 1 The application discloses a preprocessing and compression method for single-photon lidar data, comprising the following steps: S1, single-photon lidar imaging system is built: Referring to Figure 2 First, an efficient and low-noise paraxial single-photon lidar system is constructed. The paraxial single-photon lidar system comprises a laser 2, a collimating mirror 3, an optical lens group 4, a filter 5, a single-photon detector 6, a time-correlated single photon counting (TCSPC) module 1 and a displacement table 7.

[0031] The light emitted by the laser 2 is directed to the target through the collimating mirror 3, and the target reflected light is transmitted to the single-photon detector 6 through the optical lens group 4 and the filter 5 in turn. The displacement table 7 carries the single-photon detector 6 and controls the movement of the single-photon detector 6 by the driver 8. The "start" end of the time-correlated single photon counting module 1 is electrically connected with the laser 2; the "stop" end of the time-correlated single photon counting module 1 is electrically connected with the single-photon detector 6; the "enable" end of the time-correlated single photon counting module 1 is electrically connected with the computer 9; and the computer 9 is electrically connected with the driver 8.

[0032] The computer 9 triggers the time-correlated single photon counting module 1 to start and makes the laser 2 emit laser pulses. The laser is collimated by the collimating mirror 3 and irradiates the target. The target reflected laser is transmitted to the single-photon detector 6 through the optical lens group 4 and the filter 5 in turn and is detected by the single-photon detector 6. The laser emission time is recorded as "start" as the photon emission time, and the time when the single-photon detector 6 receives the laser is recorded as "stop" as the photon receiving time. The time-correlated single photon counting module 1 records the time difference between "start" and "stop" to calculate the target distance. At the same time, the single-photon detector 6 on the displacement table 7 is moved by the driver 8 to realize scanning of targets at different distances.

[0033] The laser emitted by the laser 2 is irradiated to the target object surface through the collimating mirror 3, and is received by the single-photon detector 6 after being reflected by the target. To realize the multi-distance imaging capability, a set of displacement tables 7 are introduced to accurately move the position of the single-photon detector 6, so as to realize the adjustment of the echo focal plane and the expansion of the imaging distance range, under the premise of ensuring that the position of the front imaging lens remains fixed.

[0034] S2, single-photon lidar data collection: Referring to Figure 3 The single-photon lidar imaging system is aligned with the target to be measured, and data is collected, and the target distance image obtained has a size of , wherein is the number of frames collected.

[0035] S3, data coarse screening: S301, data preprocessing: For each pixel point of the collected target distance image , it is expressed as a time series in the time dimension , wherein is the length of the time dimension, respectively represent the time when the photons arrive when the length of the time dimension is ; and and are the coordinates of the pixel point in the target distance image, wherein represents the row number, represents the column number.

[0036] The size of the collected target distance image is: , wherein is the size of the field of view of the single-photon detector, is the length of the time series collected.

[0037] S302, frequency-time histogram construction: For the data of each pixel point , first, an appropriate time bin width is set to construct the frequency-time histogram of the pixel point. The histogram represents the number of photon arrivals in different time bins, and the expression is:

[0038] , wherein the axis of the histogram represents time, represents the photon count in the time bin ; and represents the counting function.

[0039] S303, Gaussian mixture function fitting: Frequency-time histogram of each pixel obtained through Gaussian mixture function To perform fitting and obtain the main signal peaks, the specific steps are as follows: S303-1, Frequency-Time Histogram for Each Pixel Gaussian fitting is performed using a Gaussian mixture function, the specific expression of which is:

[0040] in, It is the first The amplitude of each Gaussian peak represents its peak intensity; It is the first The mean of the Gaussian peaks This represents the location of the Gaussian peak, i.e., the time it takes for the photon to arrive. It is the first The standard deviation of a Gaussian peak represents the width of the peak; The total number of Gaussian peaks; exp() represents the natural exponential function; The first image representing the target distance Line number Gaussian fitting results of the frequency-time histogram of the pixels in the column.

[0041] S303-2, using the least squares method to optimize the fitting function based on the Gaussian mixture function. ,in Let:

[0042] but:

[0043] During the optimization process, the partial derivative of each parameter is calculated and iteratively updated. This continues until the loss function converges. The loss function is expressed as:

[0044] in, This represents the number of times the frequency-time histogram was actually observed. One data point; The actual observed number in the frequency-time histogram The fitted function values ​​corresponding to each data point; Indicates the first The location of the Gaussian peak.

[0045] Then the optimal parameter set is obtained by fitting. is expressed as:

[0046]

[0047] Finally, the parameters of the fitting function are obtained, and each parameter of the fitting function is expressed as ; wherein argmin represents the value of the independent variable that makes the objective function take the minimum value.

[0048] S304, peak selection and screening: According to the fitting result of the Gaussian mixture function, i.e. the parameters of the fitting function, the peak screening is performed, and the significant peak is selected according to the amplitude Only when is greater than the set threshold value , the peak is retained as the target peak, and the expression is:

[0049] In this way, noise peaks and invalid peaks can be avoided.

[0050] The time range of the target peak is determined by its mean value and standard deviation :

[0051] The time range contains the main part of the target peak, and the photon data in the time range is used as the input in the subsequent compression process, and the time range can be recorded as .

[0052] S305, data output: S304 is fitted by the Gaussian mixture model and the peak screening, and the main peak region of each pixel point, i.e. the time range is obtained. The photon data in the time range will be retained for subsequent data compression processing. Specifically: For each pixel point , the photon data in the time range is retained; the time bin data not in the range is removed, thereby reducing the data amount and removing the background noise. Finally, the output data will be a compressed time sequence containing the peak region, and the data size is , is expressed as a compressed time sequence, see Figure 4 , and the photon counting cube is finally screened.

[0053] S4, data compression processing:​ The data size after coarse screening by Gaussian fitting is , represents the number of effective bins extracted in the time dimension. The input data is sent to the neural network in the form of a 5-dimensional tensor [Batch, Height, Width, Depth, Channel], with a default batch size of 1 and a channel number of 1. The input tensor size is: ; Batch represents the sample size; Height represents the image height; Width represents the image width; Depth represents the image depth; Channel represents the number of image channels.

[0054] Referring to Figures 5-8 , the multi-attention three-dimensional single-photon compression neural network includes a space-time attention feature extraction module, a deep compression extraction module, and a compression mapping module. The photon data corresponding to the screened peak values are sequentially processed by the space-time attention feature extraction module, the deep compression extraction module, and the compression mapping module to output the final compressed data.

[0055] The core purpose of the space-time attention feature extraction module is to extract discriminative space-time fusion features from the input high-dimensional single-photon three-dimensional tensor to enhance the network's expression ability for weak signals in single-photon events. The space-time attention feature extraction module receives input data with a dimension of , and outputs intermediate feature data with a dimension of , providing multi-dimensional feature support for subsequent deep compression processing. The space-time attention feature extraction module mainly includes the following structural components, which are repeatedly stacked in a layered form times: (1) Three-dimensional convolution and activation unit (Conv3D+ReLU): The input screened peak value corresponding photon data is first processed by a set of three-dimensional convolution operations (Conv3D, 3D Convolutional Layer), and the convolution kernel covers the space and time dimensions, thereby realizing space-time feature perception in the local region. The convolution output is then activated by the ReLU (Rectified Linear Unit, Rectified Linear Unit) activation function, enhancing the network's non-linear modeling ability and accelerating the convergence speed.

[0056] (2) Attention mechanism module (CBAM, Convolutional Block Attention Module): The feature tensor processed by the three-dimensional convolution and activation unit is input into the CBAM module. The CBAM module combines channel attention and spatial attention mechanisms to model the importance relationship between channels and the saliency of spatial positions, respectively, to achieve adaptive enhancement of key region features and effectively suppress background noise and redundant signals.

[0057] (3) Regularization and normalization unit (Dropout + LayerNorm): Dropout operation is introduced after CBAM output to reduce the risk of overfitting, and then layer normalization (LayerNorm) processing is performed to improve network stability and training efficiency, especially suitable for processing variable-length time series and weak signal scenarios.

[0058] (4) Three-dimensional maximum pooling downsampling unit (MaxPooling3D): The module output finally realizes spatial and temporal dimension downsampling through three-dimensional maximum pooling, effectively compresses the feature volume and retains the main information, providing more compact feature representation for subsequent deep compression.

[0059] Deep compression extraction module, which adopts a deep separable three-dimensional convolution mechanism, has spatial-time joint modeling capability and high parameter efficiency, and the input data dimension is , and the output data dimension is . The deep compression extraction module includes a deep compression extraction stacking unit and a downsampling module; the deep compression extraction stacking unit is mainly formed by stacking a deep convolution unit, a point convolution unit, and a normalization and nonlinear activation unit in sequence, specifically: (1) Deep convolution unit (Depthwise Conv3D): The input tensor is first subjected to a deep convolution operation, i.e., performing a three-dimensional convolution operation on each input channel independently, to extract local spatio-temporal features without introducing cross-channel information interaction, significantly reducing the computational burden and parameter size.

[0060] (2) Point convolution unit (Pointwise Conv3D): Then, point convolution (i.e., 1×1\times1\times1\times1 three-dimensional convolution) is performed to realize cross-channel feature fusion and improve expression ability. This module allows the channel dimension to be reconstructed while maintaining the size of the input feature map unchanged.

[0061] (3) Normalization and nonlinear activation unit (BatchNorm + ReLU): The point convolution output is processed by batch normalization (Batch Normalization) to stabilize the training process, and then the ReLU activation function is introduced to introduce nonlinear mapping capability.

[0062] (4) Multiple stacking and residual connection (repeat × n and jump connection): The above three unit modules are repeatedly stacked The depth and representation ability of the model are enhanced, and a short circuit connection structure is adopted to directly connect the input of a nonlinear activation unit to the final output of a stacked module to form a residual path, thereby improving training stability and gradient flow.

[0063] (5) Down-sampling module (MaxPooling3D): The output of the residual connection is sent to a three-dimensional maximum pooling layer for down-sampling in the spatial and temporal dimensions, further compressing the feature size and enhancing the modeling ability in the temporal dimension.

[0064] The compression mapping module is located at the end of the entire compression neural network and is used to further map the extracted and compressed multi-dimensional space-time features to a one-dimensional compressed vector to meet the needs of low-bit coding or input reconstruction. The compression mapping module can dynamically adjust the parameter configuration of the fully connected network structure according to the target compression code length , thereby realizing flexible output dimension control and compression ratio adjustment, and balancing compression accuracy and reconstruction fidelity. Specifically, the structure of the compression mapping module is as follows: (1) Flatten layer: First, the five-dimensional tensor input is unfolded into a one-dimensional vector, breaking down the structural barriers between space, time, and channels, and preparing for subsequent fully connected mapping.

[0065] (2) Multi-layer fully connected transformation unit (Fully Connected Layer + ReLU): The unfolded high-dimensional vector is input into multiple fully connected layers (Fully Connected, FC) in turn, each followed by a ReLU activation function to introduce non-linear representation ability. This part can be repeatedly stacked to further compress and refine high-dimensional features.

[0066] (3) Sigmoid mapping layer: The last layer is a Sigmoid activation function, which is used to compress the output range to interval, adapting to the normalized output form required by binary coding, photon intensity probability modeling or downstream objective functions.

[0067] (4) Dynamic dimension regulation mechanism: The final output dimension of the multi-layer fully connected transformation layer can be dynamically set according to the compression target parameters , thereby realizing an output compressed vector size of . Through this mechanism, the system can flexibly adapt to different transmission bandwidth, storage resources or accuracy requirements.

[0068] In this step, the input of the spatiotemporal attention feature extraction module receives the filtered photon counting cube, which has been coarsely filtered by the Gaussian fitting module. The deep compression extraction module is used to filter the photon counting cube through three-dimensional convolution and attention mechanisms. Compressed into a photon counting matrix and the photon counting matrix The input is sent to the compression mapping module; the compression mapping module is used to process the photon counting matrix through a flattening layer and a fully connected layer. Perform compression mapping to compressed data This refers to compressed single-photon data. The compressed single-photon data... Compared with the filtered single-photon data The following quantitative relationships exist between them:

[0069] To enable flexible adjustment of the compression ratio, the compression mapping module includes: A fully connected layer, A positive integer used to boost the compression target parameter in the output dimension. Given the feature mapping capability, to ensure the integrity of the representation of input single-photon data under high compression ratio, the compression ratio is... for: .

[0070] See Figure 9 One embodiment of the present invention provides a preprocessing and compression system for single-photon lidar data, comprising: Data acquisition module 901 acquires target distance images through a single-photon lidar imaging system; Histogram construction module 902 constructs a frequency-time histogram for each pixel of the acquired target distance image; Peak filtering module 903 performs Gaussian fitting on the frequency-time histogram of each pixel and then performs peak filtering. The data compression module 904 inputs the photon data corresponding to the selected peaks into a multi-attention three-dimensional single-photon compression neural network for data compression.

[0071] One embodiment of the present application provides a storage medium, specifically a computer readable storage medium, which is a memory device in a terminal device, used for storing programs and data. It can be understood that the computer readable storage medium herein can include a built-in storage medium in the terminal device, and of course can include an expansion storage medium supported by the terminal device, and can be any tangible medium containing or storing programs, which can be used by or in combination with an instruction execution system, device or apparatus. The computer readable storage medium provides a storage space, which stores an operating system of the terminal. Moreover, one or more instructions suitable for being loaded and executed by the processor are also stored in the storage space, and the instructions can be one or more computer programs (including program codes). It should be noted that more specific examples (non-exhaustive list) of the computer readable storage medium include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory, a read-only memory, an erasable programmable read-only memory, an optical fiber, a portable compact disk read-only memory, an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0072] The computer readable storage medium also includes a data signal carried in baseband or propagated as a carrier wave in a propagated data signal, in which the readable program code is carried. Such a propagated data signal can take various forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination of the above. The readable storage medium can also be any readable medium other than the readable storage medium, which can send, propagate or transmit programs for use by or in combination with an instruction execution system, device or apparatus. The program code contained on the readable storage medium can be transmitted by any suitable medium, including but not limited to wireless, wired, optical cable, etc., or any suitable combination of the above.

[0073] The program code for performing the operations of the present application can be written in any combination of one or more programming languages, including an object oriented programming language such as Java, C++, etc., and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code can execute entirely on the user's computing device, partly on the user's device, as a stand-alone software package, partly on the user's device and partly on a remote computing device or entirely on the remote computing device or server. In the latter scenario, the remote computing device can be connected to the user's device through any kind of network, including a local area network or a wide area network, or the connection can be made to an external computing device (for example, through the Internet using an Internet Service Provider).

[0074] The one or more instructions stored in the computer-readable storage medium can be loaded and executed by the processor to implement the corresponding steps of the preprocessing and compression method of the single-photon lidar data in the above embodiments.

[0075] The above only is the preferred embodiment of the present application, and is not used to limit the present application, for the person skilled in the art, the present application can have various changes and changes. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for preprocessing and compressing single-photon lidar data, characterized in that, The method comprises the following steps: target distance image acquisition is performed by a single-photon laser radar imaging system; a corresponding frequency-time histogram is constructed for each pixel point of the acquired target distance image; gaussian fitting is performed on the frequency-time histogram of each pixel point, and peak value screening is performed; the photon data corresponding to the screened peak value is input into a multi-attention three-dimensional single-photon compression neural network for data compression.

2. The preprocessing and compression method of single-photon lidar data according to claim 1, characterized in that, The single-photon laser radar imaging system adopts a paraxial single-photon laser radar system, which comprises a laser, a collimating mirror, a single-photon detector, a time-dependent single-photon counting module, and a driver; the laser is electrically connected to the time-dependent single-photon counting module, so that the laser emission time of the laser is used as the photon emission time; the laser emitted by the laser is collimated by the collimating mirror and then directed to the target; the reflected light of the target enters the single-photon detector for signal detection; the single-photon detector is electrically connected to the time-dependent single-photon counting module, so that the time when the single-photon detector receives the laser is used as the photon reception time; a displacement table is arranged at the bottom of the single-photon detector; the driver is electrically connected to the displacement table, and the driver controls the displacement table to drive the single-photon detector to move, so as to adjust the target distance and realize scanning of targets at different distances.

3. The preprocessing and compression method of single-photon lidar data according to claim 1, characterized in that, The gaussian fitting and peak value screening of the frequency-time histogram of each pixel point are specifically as follows: The frequency-time histogram of each pixel point is fitted with a Gaussian mixture function to obtain a fitting function ; parameters of the fitting function by least squares optimization parameters of the fitting function parameters of the fitting function According to The peak screening is performed according to the parameters of the fitting function, and a target peak value is obtained.

4. The preprocessing and compression method of single-photon ladar data according to claim 3, wherein, The photon distribution is fitted by a gaussian mixture function, and the gaussian mixture function is specifically as follows: wherein, is the amplitude of the th Gaussian peak; is the mean of the th Gaussian peak; is the standard deviation of the th Gaussian peak; represents the position of the Gaussian peak; represents the frequency-time histogram of the pixel point of the th row and the th column of the target range image; exp() represents the natural exponential function; represents the total number of Gaussian peaks.

5. The preprocessing and compression method of single-photon ladar data according to claim 4, wherein, the parameters of the fitting function are optimized by least squares method, specifically: the parameters of the fitting function are optimized by least squares method, specifically: updating by least squares method until the loss function is minimized; the expression of the loss function is: in, The set of parameters to be optimized and iterated; The actual observed number in the frequency-time histogram One data point; This represents the fitting function based on the Gaussian mixture model distribution; Indicates the first The location of the Gaussian peak.

6. The preprocessing and compression method of single-photon ladar data according to claim 1, wherein, The according The parameters of the fitting function are used to screen the peak value, and a target peak value is obtained, specifically: when the amplitude of the fitting function is greater than a set threshold, the signal peak value is retained as the target peak value.

7. The preprocessing and compression method of single-photon ladar data according to claim 6, wherein, The photon data corresponding to the screened peak value is input into a multi-attention three-dimensional single-photon compression neural network for data compression, and the specific process is as follows: The photon data in the time range of the target peak value is input into the multi-attention three-dimensional single-photon compression neural network for data compression; the time range is: wherein, is the mean value of the th Gaussian peak; is the standard deviation of the th Gaussian peak; denotes the time range of the target peak.

8. The preprocessing and compression method of single-photon ladar data according to claim 7, wherein, The multi-attention three-dimensional single-photon compression neural network comprises a space-time attention feature extraction module, a deep compression extraction module, and a compression mapping module connected in sequence; The space-time attention feature extraction module comprises a plurality of three-dimensional convolution and activation units, an attention mechanism module, a regularization and normalization unit, and a three-dimensional maximum pooling downsampling unit connected in sequence; The deep compression extraction module comprises a deep compression extraction stacking unit and a downsampling module connected in sequence; the deep compression extraction stacking unit comprises a depth convolution unit, a point convolution unit, and a normalization and nonlinear activation unit; the output of the point convolution unit is connected to the output of the normalization and nonlinear activation unit; the deep compression extraction stacking unit is repeatedly stacked for several times; The compression mapping module comprises a flattening layer, a multi-layer fully connected transformation unit, and a sigmoid mapping layer connected in sequence; the multi-layer fully connected transformation unit is repeatedly stacked for several times.

9. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a computer program, and the computer program is executed by a processor to realize the single-photon laser radar data preprocessing and compression method of any one of claims 1-8.

10. A system for pre-processing and compression of single-photon ladar data, based on the method for pre-processing and compression of single-photon ladar data according to any one of claims 1 to 8, characterized in that It comprises: a data acquisition module, which acquires target distance images by a single-photon laser radar imaging system; a histogram construction module, which constructs a corresponding frequency-time histogram for each pixel point of the acquired target distance image; A peak screening module performs Gaussian fitting on the frequency-time histogram of each pixel point and performs peak screening; A data compression module inputs the screened peak value corresponding photon data into a multi-attention three-dimensional single-photon compression neural network for data compression.

Images