A laser radar data quality supervision method and system

Abstract

The present invention relates to the field of data analysis technology, and provides a laser radar data quality supervision method and system. By acquiring a comprehensive set of environmental atmospheric detection data, designing an efficient positioning label system, constructing a multi-dimensional environmental atmospheric element vector, and determining a comprehensive data quality control detection and analysis viewpoint, comprehensive quality control and evaluation of laser radar detection data are achieved, providing strong support for research and application in related fields.

Description

Technical Field

[0001] The present invention relates to the field of data analysis technology, and in particular to a laser radar data quality supervision method and system. Background Art

[0002] With the continuous development of LiDAR technology, its application in environmental detection, terrain mapping, unmanned driving and other fields is becoming more and more extensive. LiDAR system can obtain high-precision three-dimensional information and atmospheric environment parameters of the target area by emitting laser beams and receiving reflected signals. However, in practical applications, the detection effect of LiDAR system is affected by many factors, such as environmental atmospheric conditions, the spatial position and time state of the system itself, etc.

[0003] In order to improve the detection accuracy and reliability of the LiDAR system, it is necessary to accurately detect and analyze the ambient atmospheric conditions in which it is located. Traditional LiDAR systems usually only focus on the detection data at a fixed position, ignoring the changes in the detection state of the system at different positions and at different times. This single detection method cannot fully reflect the actual working environment of the LiDAR system, thus limiting the improvement of its detection performance. Summary of the invention

[0004] In order to improve the above problems, the present invention provides a lidar data quality supervision method and system.

[0005] In a first aspect, an embodiment of the present invention provides a laser radar data quality supervision method, which is applied to a data quality supervision system, and the method includes:

[0006] Acquire an ambient atmosphere detection data set of a target laser radar system; wherein the ambient atmosphere detection data set includes at least one set of fixed ambient atmosphere detection data collected by the target laser radar system at a fixed position, and at least two sets of non-fixed ambient atmosphere detection data collected by the target laser radar system at a non-fixed position, wherein the fixed ambient atmosphere detection data and the non-fixed ambient atmosphere detection data are linkage ambient atmosphere index detection data obtained by collecting data from a selected space in a target detection area;

[0007] Parsing the fixed environment atmosphere detection data to obtain a first positioning tag, where the first positioning tag is used to characterize the distribution of the selected space corresponding to the fixed environment atmosphere detection data in the target detection area;

[0008] Based on the fixed environment atmosphere detection data, a fixed environment atmosphere element vector of the first positioning tag is obtained; based on the at least two sets of non-fixed environment atmosphere detection data, a non-fixed environment atmosphere element vector of the target laser radar system is obtained, and the non-fixed environment atmosphere element vector is used to characterize the spatial detection state and the temporal detection state of the target laser radar system;

[0009] Based on the fixed environment atmospheric element vector and the non-fixed environment atmospheric element vector, the data quality control detection and analysis viewpoint of the target lidar system is determined.

[0010] Preferably, obtaining the non-stationary ambient atmosphere element vector of the target laser radar system based on the at least two sets of non-stationary ambient atmosphere detection data includes:

[0011] Obtaining distribution of at least one ambient atmosphere data block in each group of the non-stationary ambient atmosphere detection data;

[0012] Based on the distribution of the at least one ambient atmosphere data block in each non-stationary ambient atmosphere detection data tuple, a trend feature of each ambient atmosphere data block in each non-stationary ambient atmosphere detection data tuple is obtained; wherein each non-stationary ambient atmosphere detection data tuple includes two consecutive groups of non-stationary ambient atmosphere detection data;

[0013] Based on the trend characteristics of each ambient atmosphere data block in each of the non-stationary ambient atmosphere detection data tuples, the non-stationary ambient atmosphere element vector of the target laser radar system is obtained.

[0014] Preferably, obtaining the non-stationary ambient atmosphere element vector of the target laser radar system based on the trend characteristics of each ambient atmosphere data block in each non-stationary ambient atmosphere detection data tuple includes:

[0015] For each of the non-stationary environment atmosphere detection data tuples, determine the tuple features of the trend features of at least one environment atmosphere data block in each of the non-stationary environment atmosphere detection data tuples, and obtain the averaged trend features of each of the non-stationary environment atmosphere detection data tuples; the feature value of the averaged trend feature is used to represent the time detection state of the target laser radar system in the non-stationary environment atmosphere detection data tuple, and the gradient of the averaged trend feature is used to represent the detection change of the target laser radar system in the non-stationary environment atmosphere detection data tuple;

[0016] Based on the average trend characteristics of each non-stationary environment atmosphere detection data binary group, the non-stationary environment atmosphere element vector of the target laser radar system is obtained.

[0017] Preferably, the obtaining of the distribution of at least one ambient atmosphere data block in each group of the non-stationary ambient atmosphere detection data includes:

[0018] Performing an embedding operation on the at least two sets of non-stationary ambient atmosphere detection data to obtain ambient atmosphere detection embedding data corresponding to each of the at least two sets of non-stationary ambient atmosphere detection data;

[0019] Based on the ambient atmosphere detection embedded data corresponding to the first non-stationary ambient atmosphere detection data in the at least two groups of non-stationary ambient atmosphere detection data, determine at least one ambient atmosphere data block in the ambient atmosphere detection embedded data corresponding to the first non-stationary ambient atmosphere detection data; wherein the first non-stationary ambient atmosphere detection data is the first group of non-stationary ambient atmosphere detection data in the at least two groups of non-stationary ambient atmosphere detection data;

[0020] A convolution strategy is adopted to determine the distribution of the at least one ambient atmosphere data block in the ambient atmosphere detection embedded data corresponding to each group of the non-fixed ambient atmosphere detection data based on the distribution of the at least one ambient atmosphere data block in the ambient atmosphere detection embedded data corresponding to the first non-fixed ambient atmosphere detection data.

[0021] Preferably, the method further comprises:

[0022] Using a data optimization algorithm to update the distribution of the at least one ambient atmosphere data block in the ambient atmosphere detection embedded data corresponding to each group of the non-stationary ambient atmosphere detection data, to obtain an updated distribution of the at least one ambient atmosphere data block in the ambient atmosphere detection embedded data corresponding to each group of the non-stationary ambient atmosphere detection data;

[0023] The obtaining of the trend characteristics of each ambient atmosphere data block in each of the non-stationary ambient atmosphere detection data tuples based on the distribution of the at least one ambient atmosphere data block in each of the non-stationary ambient atmosphere detection data tuples includes:

[0024] Based on the updated distribution of the at least one ambient atmosphere data block in each of the non-stationary ambient atmosphere detection data tuples, obtaining updated trend characteristics of each ambient atmosphere data block in each of the non-stationary ambient atmosphere detection data tuples;

[0025] The triggering period of the update operation includes at least one of the following:

[0026] After adopting the convolution strategy to determine the distribution of the at least one ambient atmosphere data block in the ambient atmosphere detection embedded data corresponding to each group of the non-stationary ambient atmosphere detection data, an update operation is performed on the distribution of the at least one ambient atmosphere data block in the ambient atmosphere detection embedded data corresponding to each group of the non-stationary ambient atmosphere detection data, and the convolution strategy is adopted based on the distribution of the at least one ambient atmosphere data block in the ambient atmosphere detection embedded data corresponding to the previous group of the non-stationary ambient atmosphere detection data to determine the distribution of the at least one ambient atmosphere data block in the ambient atmosphere detection embedded data corresponding to the next group of the non-stationary ambient atmosphere detection data;

[0027] Alternatively, after adopting the convolution strategy to determine the distribution of the at least one ambient atmosphere data block in the ambient atmosphere detection embedded data corresponding to each group of the non-fixed ambient atmosphere detection data, the distribution of the at least one ambient atmosphere data block in the ambient atmosphere detection embedded data corresponding to each group of the non-fixed ambient atmosphere detection data is updated.

[0028] Preferably, the updating operation of the distribution of the at least one ambient atmosphere data block in each group of the non-stationary ambient atmosphere detection data by using a data optimization algorithm to obtain the updated distribution of the at least one ambient atmosphere data block in each group of the non-stationary ambient atmosphere detection data includes:

[0029] Based on the updated distribution of the first ambient atmospheric data block at the q-1th period node, a distribution prediction result of the first ambient atmospheric data block at the qth period node is obtained, where q is an integer greater than 1; wherein the updated distribution at the 1st period node is the distribution of the first ambient atmospheric data block in the ambient atmospheric detection embedded data corresponding to the first non-fixed ambient atmospheric detection data; two consecutive groups of the non-fixed ambient atmospheric detection data in each of the non-fixed ambient atmospheric detection data tuples are separated by one period node;

[0030] Based on the distribution prediction result of the first ambient atmosphere data block at the qth period node, the detection distribution of the first ambient atmosphere data block at the qth period node, and the optimization coefficient corresponding to the data optimization algorithm, obtaining the updated distribution of the first ambient atmosphere data block at the qth period node;

[0031] Among them, the detection distribution of the first ambient atmospheric data block at the qth period node is the distribution obtained by adopting the convolution strategy; the optimization coefficient is determined based on the fluctuation weight corresponding to the distribution prediction result of the first ambient atmospheric data block at the qth period node, and the fluctuation weight corresponding to the distribution prediction result of the first ambient atmospheric data block at the qth period node is determined based on the fluctuation weight corresponding to the updated distribution of the first ambient atmospheric data block at the q-1th period node, and the fluctuation weight is used to represent the offset between the distribution prediction result and the detection distribution.

[0032] Preferably, the fixed environment atmospheric detection data includes meteorological element data collected by a ground observation instrument and atmospheric chemical composition data collected by an atmospheric composition analyzer, the meteorological element data is used to obtain the meteorological index vector of the target laser radar system, and the atmospheric chemical composition data is used to obtain the atmospheric chemical composition vector of the target laser radar system;

[0033] After acquiring the set of ambient atmosphere detection data of the target laser radar system, the method includes: adjusting the at least one set of fixed ambient atmosphere detection data to obtain at least one set of processed fixed ambient atmosphere detection data; the adjustment includes at least one of data matching, data deduplication, and data embedding; based on the at least one set of processed fixed ambient atmosphere detection data, determining first fixed ambient atmosphere detection data, wherein the first fixed ambient atmosphere detection data is the processed fixed ambient atmosphere detection data with the highest evaluation weight among the at least one set of processed fixed ambient atmosphere detection data;

[0034] The method of obtaining the fixed environment atmosphere element vector of the first positioning tag based on the fixed environment atmosphere detection data includes: obtaining the meteorological index vector of the first positioning tag based on the meteorological element data in the first fixed environment atmosphere detection data, and obtaining the atmospheric chemical composition vector of the first positioning tag based on the atmospheric chemical composition data in the first fixed environment atmosphere detection data; obtaining the fixed environment atmosphere element vector of the first positioning tag based on the meteorological index vector of the first positioning tag and the atmospheric chemical composition vector of the first positioning tag.

[0035] Preferably, the step of parsing the fixed environment atmospheric detection data to obtain a first positioning tag includes:

[0036] Using a debugged detection data analysis network to analyze the first fixed environment atmospheric detection data to obtain the first positioning tag;

[0037] The debugging process of the detection data analysis network is as follows:

[0038] Obtain a debugging sample set of the detection data analysis network, wherein the debugging sample set includes at least one debugging example binary, and each debugging example binary includes a detection data example and a training annotation corresponding to the detection data example; wherein the detection data example is linkage environmental atmospheric index detection data obtained by collecting data from a selected space of a target detection area, and the detection data example includes any one of local meteorological element data collected by a ground observation instrument and local atmospheric chemical composition data collected by an atmospheric composition analyzer, and the training annotation is used to characterize a positioning label of the selected space corresponding to the detection data example in the target detection area;

[0039] Performing interval mapping processing on each group of detection data examples in the debugging sample set to obtain processed detection data examples, wherein the interval mapping processing is used to process each group of detection data examples into data with consistent detection indicators;

[0040] Using the detection data parsing network to obtain a positioning prediction label corresponding to the detection data example based on the processed detection data example;

[0041] Based on the difference between the positioning prediction label and the training annotation, the detection data parsing network is debugged to obtain a debugged detection data parsing network.

[0042] Preferably, the determining of the data quality control detection analysis viewpoint of the target laser radar system based on the fixed environment atmospheric element vector and the non-fixed environment atmospheric element vector includes:

[0043] Based on the fixed environment atmosphere element vector and the non-fixed environment atmosphere element vector, a detection quality evaluation vector of the target laser radar system is obtained, where the detection quality evaluation vector is a knowledge vector obtained by combining the fixed environment atmosphere element vector and the non-fixed environment atmosphere element vector;

[0044] Determine the correlation weights between the detection quality evaluation vector and the detection quality evaluation vector of at least one past laser radar system tag in the cloud sharing server respectively;

[0045] If the first correlation weight corresponding to the first past lidar system label is the maximum correlation weight, and the maximum correlation weight is greater than the first threshold, then the data quality control detection and analysis point of view of the target lidar system is determined to be that the target lidar system matches the first past lidar system label.

[0046] Preferably, the determining of the data quality control detection analysis viewpoint of the target laser radar system based on the fixed environment atmospheric element vector and the non-fixed environment atmospheric element vector includes:

[0047] Determine a second correlation weight between the fixed environment atmospheric element vector and the fixed environment atmospheric element vector of at least one past laser radar system tag in the cloud sharing server, and determine a third correlation weight between the non-fixed environment atmospheric element vector and the non-fixed environment atmospheric element vector of at least one past laser radar system tag in the cloud sharing server;

[0048] Integrate the second correlation weight and the third correlation weight corresponding to the at least one past laser radar system label to obtain a global correlation weight corresponding to the at least one past laser radar system label;

[0049] If the global correlation weight corresponding to the second past lidar system label is the maximum correlation weight, and the maximum correlation weight is greater than the second threshold, then the data quality control detection and analysis point of view of the target lidar system is determined to be that the target lidar system matches the second past lidar system label.

[0050] Preferably, the determining of the data quality control detection analysis viewpoint of the target laser radar system based on the fixed environment atmospheric element vector and the non-fixed environment atmospheric element vector includes:

[0051] Using a quality control detection and analysis network to obtain a data quality control detection and analysis viewpoint of the target laser radar system based on the fixed environment atmospheric element vector and the non-fixed environment atmospheric element vector;

[0052] Among them, the quality control detection and analysis network is used to perform factor mining on the fixed environment atmospheric factor vector and the non-fixed environment atmospheric factor vector to obtain factor mining results; respectively determine the correlation weights between the factor mining results and the factor mining results of at least one past lidar system label in the cloud sharing server; based on at least one correlation weight, generate a data quality control detection and analysis viewpoint of the target lidar system, the fourth correlation weight corresponding to the third past lidar system label indicated by the data quality control detection and analysis viewpoint is the maximum correlation weight, and the maximum correlation weight is greater than the third threshold.

[0053] In a second aspect, an embodiment of the present invention provides a data quality supervision system, comprising at least one processor and a memory; the memory stores computer execution instructions; the at least one processor executes the computer execution instructions stored in the memory, so that the at least one processor executes the method described in the first aspect.

[0054] In a third aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program implements the method described in the first aspect when running.

[0055] The present invention proposes a laser radar data quality control detection and analysis method based on fixed ambient atmospheric element vectors and non-fixed ambient atmospheric element vectors. The method first obtains a set of ambient atmospheric detection data of the target laser radar system, including fixed ambient atmospheric detection data collected at fixed positions and non-fixed ambient atmospheric detection data collected at non-fixed positions. These data can fully reflect the detection status of the laser radar system under different working conditions.

[0056] By parsing the fixed environment atmospheric detection data, a first positioning tag representing its distribution in the target detection area can be obtained. Furthermore, based on the fixed environment atmospheric detection data, a fixed environment atmospheric element vector corresponding to the first positioning tag can be obtained. At the same time, based on at least two sets of non-fixed environment atmospheric detection data, a non-fixed environment atmospheric element vector representing the spatial detection state and temporal detection state of the target lidar system can be obtained.

[0057] Finally, by comprehensively considering the information of the fixed environment atmospheric element vector and the non-fixed environment atmospheric element vector, the data quality control detection analysis point of view of the target lidar system can be determined. This point of view can provide an important reference for subsequent data processing and application, and help improve the detection accuracy and reliability of the lidar system.

[0058] In summary, the purpose of the present invention is to provide a data quality control detection and analysis method that comprehensively considers the working environment and detection status of the laser radar system to solve the problems and shortcomings of the traditional method. Through the implementation of the present invention, the detection performance and application effect of the laser radar system can be significantly improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 A flowchart of a laser radar data quality supervision method provided in an embodiment of the present invention.

[0060] Figure 2 A structural diagram of a data quality monitoring system 200 provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0061] In order to better understand the above technical scheme, the technical scheme of the present invention is described in detail below through the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present invention and the specific features in the embodiments are detailed descriptions of the technical scheme of the present invention, rather than limitations on the technical scheme of the present invention. In the absence of conflict, the embodiments of the present invention and the technical features in the embodiments can be combined with each other.

[0062] Figure 1 A laser radar data quality supervision method is shown, which is applied to a data quality supervision system. The method includes the following steps 101 to 105.

[0063] Step 101: Obtain an ambient atmosphere detection data set of a target lidar system.

[0064] Among them, the environmental atmosphere detection data set includes at least one set of fixed environmental atmosphere detection data collected by the target lidar system in a fixed position, and at least two sets of non-fixed environmental atmosphere detection data collected by the target lidar system in a non-fixed position. The fixed environmental atmosphere detection data and the non-fixed environmental atmosphere detection data are linked environmental atmosphere indicator detection data collected from the selected space of the target detection area.

[0065] Step 102: parse the fixed environment atmospheric detection data to obtain a first positioning tag, where the first positioning tag is used to characterize the distribution of the selected space corresponding to the fixed environment atmospheric detection data in the target detection area.

[0066] Step 103: Based on the fixed environment atmosphere detection data, obtain the fixed environment atmosphere element vector of the first positioning tag.

[0067] Step 104: Based on the at least two sets of non-stationary ambient atmosphere detection data, obtain the non-stationary ambient atmosphere element vector of the target lidar system, and the non-stationary ambient atmosphere element vector is used to characterize the spatial detection state and the temporal detection state of the target lidar system.

[0068] Step 105: Determine the data quality control detection and analysis viewpoint of the target lidar system based on the fixed environment atmospheric element vector and the non-fixed environment atmospheric element vector.

[0069] In some application scenarios, the data quality supervision system first obtains the set of environmental atmospheric detection data of the target laser radar system. This set includes two parts of data: one is at least one set of fixed environmental atmospheric detection data collected when the laser radar system is in a fixed position; the other is at least two sets of non-fixed environmental atmospheric detection data collected when the laser radar system is in a mobile state. These data are collected from a specific space within the selected target detection area, and they reflect the linkage change indicators of the atmospheric environment in the space. Next, the system parses and processes the fixed environmental atmospheric detection data. The purpose of this step is to extract the first positioning tag implicit in the data. The first positioning tag is a key data element that can accurately characterize the specific distribution position of the selected space corresponding to the fixed environmental atmospheric detection data in the target detection area. After obtaining the first positioning tag, the system further processes the fixed environmental atmospheric detection data to generate a fixed environmental atmospheric element vector associated with the positioning tag. This vector is a set of numerical values ​​or symbols that comprehensively describe the characteristics and status of the atmospheric environmental elements detected by the laser radar at a fixed position. At the same time, the system also processes those non-fixed environmental atmospheric detection data. These data reflect the detection results of the atmospheric environment by the LiDAR in a mobile state. By processing these data, the system can generate a non-fixed environment atmospheric element vector. This vector not only contains spatial detection information, but also incorporates temporal detection information, so it can more comprehensively describe the detection state of the LiDAR in a non-fixed position. Finally, the system conducts a comprehensive analysis based on the fixed environment atmospheric element vector and the non-fixed environment atmospheric element vector. By comparing the numerical changes and trends of these two vectors, the system can evaluate the consistency and stability of the detection data of the target LiDAR system at different locations and times. Based on these evaluation results, the system finally determines a data quality control detection analysis point of view, which is a comprehensive and objective evaluation of the quality of the LiDAR system detection data. This evaluation not only points out the possible problems and deviations in the data, but also provides a strong basis for subsequent data correction and optimization.

[0070] In some application scenarios, first, a set of environmental atmospheric detection data collected by the target laser radar system under various working conditions is obtained. This includes fixed environmental atmospheric detection data collected at a fixed position, and non-fixed environmental atmospheric detection data collected in a mobile state. These data cover the atmospheric environmental indicators of the selected space in the target detection area, providing comprehensive and detailed environmental information. For the fixed environmental atmospheric detection data, in-depth analysis is performed to extract the key first positioning tag. This tag accurately identifies the distribution of the spatial position corresponding to the data in the target detection area, providing an accurate spatial reference for subsequent data processing. Based on these fixed environmental atmospheric detection data, a fixed environmental atmospheric element vector is further generated. This vector contains various key element information in the atmospheric environment, such as temperature, humidity, wind speed, etc., which intuitively show the detection results of the laser radar at a fixed position in a numerical form. At the same time, the non-fixed environmental atmospheric detection data is also processed to obtain a non-fixed environmental atmospheric element vector. This vector not only contains spatial detection information, but also incorporates time-varying factors, reflecting the dynamic detection results of the laser radar on the atmospheric environment in a mobile state. Finally, the fixed environment atmospheric element vector is combined with the non-fixed environment atmospheric element vector for comprehensive data quality control detection and analysis. By comparing the numerical changes and trends of these two vectors, the consistency and stability of the detection data of the lidar system at different locations and times can be evaluated. This analysis method helps to timely discover anomalies and deviations in the data, and provides strong support for subsequent data correction and optimization. In general, this technical solution obtains accurate data quality control results by fine-tuning and analyzing the lidar data. This not only improves the accuracy of temperature, humidity, wind and aerosol optical products, but also ensures that the temperature measurement accuracy, water vapor mixing ratio, wind speed and aerosol optical parameter detection errors meet the requirements of relevant guidelines. This processing method provides new ideas and solutions for the quality control and application of lidar data.

[0071] The following is an adaptive explanation of the technical terms involved in step 101 to step 105.

[0072] Target LiDAR System: refers to the specific LiDAR system being studied or monitored, which is used to transmit lasers and receive reflected signals to detect and measure atmospheric characteristics in the target environment. For example, a certain type of meteorological LiDAR is used to collect and analyze information about temperature, humidity, wind speed, and particles such as aerosols in the atmosphere.

[0073] Ambient atmosphere detection data set: refers to a series of raw or processed data about the ambient atmosphere collected by the lidar system, which is used to analyze and understand the atmospheric conditions. A lidar scanning data set containing multiple days and multiple time periods reflects the changes in temperature, humidity, wind speed and aerosol distribution in the atmosphere of a certain area.

[0074] Fixed location: This refers to the LiDAR system’s geographic location remaining constant during data collection. The LiDAR was installed at a meteorological observatory on a mountaintop and collected data from this location for several days.

[0075] Non-fixed location: In contrast to fixed location, this refers to the LiDAR system being in motion during data collection, with its geographic location constantly changing. The LiDAR system is mounted on an aircraft and collects atmospheric data at different locations as the aircraft moves along its flight path.

[0076] Fixed-position ambient atmospheric detection data: refers to ambient atmospheric data collected when the LiDAR is working at a fixed position. The LiDAR at the mountaintop meteorological observatory collected data on temperature, humidity, and wind speed for three consecutive days.

[0077] Non-stationary ambient atmospheric detection data: refers to ambient atmospheric data collected by the LiDAR in a non-fixed position (mobile state). The LiDAR on the aircraft records the aerosol concentration and wind speed data at regular intervals during the flight.

[0078] Target detection area: refers to the specific geographical area where the LiDAR system conducts detection activities. In a LiDAR detection mission, the target area may be a city, a valley, or a sea area.

[0079] Selected Space: A specific three-dimensional spatial area within the target detection area that the LiDAR system focuses on detecting and analyzing. In an urban environment, the selected space may be a specific vertical air layer, such as an altitude range of 0 to 1000 meters.

[0080] Linked ambient atmospheric index detection data: refers to the data of multiple interrelated ambient atmospheric parameters collected simultaneously or continuously by the lidar system, which together describe the atmospheric conditions. In a lidar detection mission, the temperature, humidity, wind speed and aerosol concentration data collected simultaneously reflect the dynamic changes of the atmosphere in the selected space.

[0081] Parsing: Parsing refers to the process of detailed analysis and processing of data or information to extract key information or features. In lidar data processing, parsing may involve interpreting the raw signal and converting it into specific parameters that reflect atmospheric conditions, such as temperature, humidity, etc.

[0082] First Location Tag: A first location tag is an identifier or marker used to identify a specific location or spatial region. In a lidar dataset, the first location tag can be a geographic coordinate, altitude range, or other code or value that uniquely identifies a selected spatial location.

[0083] Distribution: Distribution describes the spread, configuration or arrangement of an object or phenomenon within a specific space or time. In lidar detection, distribution can refer to the variation of the concentration or density of a certain component (such as aerosol) in the atmosphere at different spatial locations.

[0084] Fixed environment atmospheric element vector: A fixed environment atmospheric element vector is a set of values ​​or parameters that describe the state of the atmospheric environment at a fixed location, usually including multiple dimensions such as temperature, humidity, and pressure. A LiDAR station records 24 hours of continuous temperature and humidity data at a fixed location. These data can form a two-dimensional fixed environment atmospheric element vector.

[0085] Non-stationary atmospheric element vector: The non-stationary atmospheric element vector is similar to the fixed atmospheric element vector, but it describes the atmospheric parameters collected by the LiDAR while in motion. The temperature, humidity, and wind speed data collected by an aircraft carrying a LiDAR during flight form a non-stationary atmospheric element vector as the aircraft position and time change.

[0086] Space detection status: Space detection status refers to the detection and recording of environmental conditions in a specific space area by the lidar system. The aerosol concentration distribution scanned and recorded by the lidar at a certain altitude reflects the space detection status of that altitude.

[0087] Time detection status: Time detection status refers to the detection and recording of environmental conditions by the lidar system at different time points, which reflects the changes in the environment over time. The lidar detects the same area for several consecutive hours, and the recorded temperature change curve over time shows the time detection status of the area.

[0088] Data quality control detection and analysis perspective: The data quality control detection and analysis perspective refers to a comprehensive perspective or method of quality control and analysis when processing and analyzing ambient atmospheric detection data collected by lidar. This perspective focuses on ensuring the accuracy, consistency and reliability of the data so that accurate temperature, humidity, wind and aerosol optical products can be derived based on these data in the future. In this process, the quality of the fixed and non-fixed ambient atmospheric element vectors will be comprehensively considered, outliers will be identified and excluded, deviations will be corrected, and necessary smoothing or interpolation of the data will be performed. When performing lidar data quality control, analysts may take a comprehensive detection and analysis perspective, first evaluating the stability of the fixed ambient atmospheric element vectors to ensure that they do not change abnormally over time; then, dynamically analyzing the non-fixed ambient atmospheric element vectors to check their continuity and rationality in space and time; finally, combining the two to form a comprehensive data quality control report, which indicates which data are reliable and which data may have quality problems and need to be further processed or eliminated. Such an analysis perspective helps ensure that the data set ultimately used for the calculation of temperature, humidity, wind and aerosol optical products is high quality and accurate.

[0089] For step 101, obtaining the environmental atmospheric detection data set of the target laser radar system is a key step in the application of laser radar technology, and plays a vital role in subsequent data analysis, product generation and decision support. This data set not only covers the environmental atmospheric information detected by the laser radar at fixed and non-fixed positions, but also ensures the comprehensiveness and diversity of the data, so as to more accurately reflect the atmospheric conditions in the target detection area. Specifically, when the target laser radar system is in a fixed position, it will continuously or regularly detect the surrounding environment and collect at least one set of fixed environmental atmospheric detection data. These data are obtained by long-term stable observation at a specific spatial point or area, so they have high spatial resolution and time continuity. Through detection at a fixed position, basic information of the atmospheric environment in the selected space, such as temperature, humidity, wind speed, etc., can be obtained. This information is the basis for understanding the state and behavior of the atmosphere. At the same time, in order to obtain more extensive and dynamic environmental atmospheric information, the target laser radar system also needs to detect at a non-fixed position. This usually involves installing the laser radar on a mobile platform, such as an aircraft, vehicle or ship, and collecting data at different locations and times as the platform moves. The collection of non-stationary environment atmospheric detection data requires at least two groups of detection to ensure the reliability and representativeness of the data. These data not only contain spatial change information, but also reflect the dynamic changes in time, providing a comprehensive view of the atmospheric environment. Whether it is fixed environment atmospheric detection data or non-stationary environment atmospheric detection data, they are collected from the atmospheric environment in the selected space of the target detection area. These data are not only interrelated, but also together form the basis of the linkage environment atmospheric indicator detection data. Through the comprehensive analysis of these data, we can have a deeper understanding of the complexity and dynamics of the atmospheric environment, and provide strong support for subsequent weather forecasting, climate research, air quality monitoring and other applications.

[0090] For step 102, parsing the fixed environment atmospheric detection data is one of the core links in the data processing process. This step aims to extract key information from the raw data and generate a first positioning tag. This tag is not only an identification of the data itself, but also a bridge connecting the data with the actual geographical location, laying the foundation for subsequent spatial analysis and environmental research. The parsing process involves a careful review and understanding of the data to determine its inherent structure and characteristics. For the fixed environment atmospheric detection data, the purpose of the analysis is to identify the spatial information implicit in the data. This information may include the specific coordinates, height, and relative position relationship of the detection point with the surrounding environment. By extracting and sorting this information, a first positioning tag can be formed. The generation of the first positioning tag is not arbitrary, but based on rigorous algorithms and scientific principles. The design of the tag needs to take into account the characteristics of the data, the purpose of the detection, and the needs of subsequent applications. An effective positioning tag should be able to accurately reflect the spatial distribution of the detection data in the target detection area, and should also be concise and easy to understand, so as to facilitate subsequent data processing and analysis. In the context of lidar detection, the importance of the first positioning tag is self-evident. It enables researchers to quickly locate the detection data of a specific spatial point and then analyze the atmospheric environment characteristics of that point. By comparing and analyzing the data of different spatial points, the spatial variability and distribution of the atmospheric environment in the target detection area can be revealed. These findings have important scientific value in understanding the processes and mechanisms of the atmospheric environment, predicting weather changes, and evaluating environmental impacts.

[0091] For step 103, obtaining the fixed environment atmospheric element vector of the first positioning tag based on the fixed environment atmospheric detection data is an important link in the field of environmental detection and analysis. This process is essentially to convert the original detection data into a more meaningful and operational information form. Fixed environment atmospheric detection data usually refers to the atmospheric environment data obtained by the laser radar system at a fixed position for a long time. These data contain rich information, such as temperature, humidity, wind speed, air pressure, and the concentration of various gases and particulate matter. These data not only reflect the static characteristics of the atmospheric environment, but also contain clues to the dynamic changes of the atmosphere. In order to use these data more effectively, they need to be processed and analyzed to extract key atmospheric element information. This process usually includes steps such as data cleaning, data conversion and data induction. Data cleaning is to remove outliers and noise to ensure the accuracy and reliability of data; data conversion is to convert the original data into a form that is more convenient for analysis, such as calculating derived variables or performing standardization processing; data induction is to organize and classify the data according to specific rules or models. In this process, the first positioning tag plays a key role. It is a tool for identifying and classifying fixed environment atmospheric detection data, so that the atmospheric environment information of a specific spatial location can be accurately identified and tracked. By associating the first positioning tag with the corresponding atmospheric element data, a fixed environment atmospheric element vector can be constructed. This vector is a multi-dimensional data set that contains information about various atmospheric elements at a selected spatial location. The acquisition of the fixed environment atmospheric element vector provides a powerful tool for further analysis of the atmospheric environment. By analyzing this vector, we can understand the overall condition of the atmospheric environment at a selected spatial location, identify the correlation and influence mechanism between different elements, and provide valuable scientific basis for fields such as environmental protection, climate change research, and weather forecasting.

[0092] For step 104, based on at least two sets of non-stationary environment atmospheric detection data, obtaining the non-stationary environment atmospheric element vector of the target laser radar system is a key link in the dynamic monitoring and analysis of the atmospheric environment. This vector not only integrates the detection information of the laser radar at different spatial positions, but also reflects the changing characteristics of the ambient atmosphere over time, thereby providing an important basis for a comprehensive understanding of the atmospheric environment state of the target area. The non-stationary environment atmospheric detection data is collected when the laser radar system is in a mobile state. As the laser radar moves, the spatial position of its detection is also constantly changing. These data record the atmospheric environment information at different locations, such as temperature, humidity, wind speed, aerosol concentration, etc., reflecting the spatial distribution characteristics and dynamic change process of the atmospheric environment. By processing and analyzing these non-stationary detection data, key atmospheric element information can be extracted and a non-stationary environment atmospheric element vector can be constructed. This vector is a multi-dimensional data set that contains information on various atmospheric elements at different spatial positions and time points. It can not only reflect the spatial change characteristics of the atmospheric environment, but also reveal the evolution process of the atmospheric environment over time. The acquisition of the non-stationary environment atmospheric element vector provides an important basis for evaluating the spatial detection state and temporal detection state of the target lidar system. The spatial detection state reflects the detection capability and effect of the lidar at different spatial positions, while the temporal detection state reveals the detection stability and continuity of the lidar at different time points. Through further analysis of this vector, the performance of the lidar system can be evaluated, providing strong support for subsequent detection mission optimization and system design. In addition, the non-stationary environment atmospheric element vector can also be used for dynamic monitoring and early warning of the atmospheric environment. Through real-time monitoring and analysis of this vector, abnormal changes and potential risks in the atmospheric environment can be discovered in a timely manner, providing valuable scientific basis for environmental protection, meteorological disaster early warning and other fields.

[0093] For step 105, determining the data quality control detection and analysis viewpoint of the target lidar system based on the fixed environment atmospheric element vector and the non-fixed environment atmospheric element vector is the core process of comprehensive quality control and evaluation of the lidar detection data. This process ensures the accuracy, consistency and reliability of the data, and provides a solid foundation for subsequent environmental analysis, model prediction and decision-making. The fixed environment atmospheric element vector reflects the state of the ambient atmosphere detected by the lidar at a fixed position, with high spatial resolution and temporal continuity. It provides an in-depth understanding of the basic information of the atmospheric environment in the selected space, such as the long-term trend and periodic fluctuation of key parameters such as temperature, humidity, and wind speed. These data are crucial for understanding the static characteristics and dynamic behavior of the atmospheric environment. The non-fixed environment atmospheric element vector captures the atmospheric environment information detected by the lidar in a mobile state, with extensive spatial coverage and temporal dynamics. It reveals the differences and correlations of the atmospheric environment at different locations and time points, and reflects the complex changes of the ambient atmosphere in spatial and temporal dimensions. These data provide valuable clues for evaluating the regional and global characteristics of the atmospheric environment. Combining the vectors of fixed environmental atmospheric elements with the vectors of non-fixed environmental atmospheric elements can form a comprehensive and integrated data quality control detection analysis perspective. This perspective not only takes into account the intrinsic quality and consistency of the data, but also takes into account the spatial representativeness and temporal dynamics of the data. It allows researchers to examine lidar detection data from multiple perspectives, identify and correct possible errors, outliers or deviations, and ensure the accuracy and reliability of the data. Under the guidance of this comprehensive analysis perspective, the performance of the lidar system can be comprehensively evaluated, including key indicators such as its detection accuracy, stability, and response speed. At the same time, the data set can also be preprocessed and calibrated as necessary to eliminate the impact of system errors and external environmental interference on the detection results. These works provide a more solid and reliable foundation for subsequent data analysis, model building, and prediction.

[0094] The beneficial effects of the present invention are mainly reflected in the following aspects: First, by obtaining the environmental atmosphere detection data set of the target laser radar system, including data collected at fixed positions and non-fixed positions, the comprehensiveness and diversity of the data are ensured. Such a data set not only reflects the detection capability of the laser radar at a fixed position, but also captures its detection information in a mobile state, providing rich materials for subsequent data analysis. Secondly, by parsing the fixed environmental atmosphere detection data, a first positioning tag is obtained, so that researchers can quickly locate the detection data of a specific spatial point, and then analyze the atmospheric environment characteristics of the point. The design of this positioning tag greatly improves the efficiency and accuracy of data processing, and lays the foundation for subsequent spatial analysis and environmental research. Thirdly, the fixed environmental atmosphere element vector of the first positioning tag obtained based on the fixed environmental atmosphere detection data, and the non-fixed environmental atmosphere element vector obtained based on at least two sets of non-fixed environmental atmosphere detection data, together constitute a comprehensive description of the detection capability of the target laser radar system. These vectors not only contain information on various atmospheric elements, but also reflect the spatial distribution and time variation characteristics of the ambient atmosphere, providing an important basis for evaluating the performance of the laser radar system. Finally, by comprehensively considering the fixed environment atmospheric element vectors and the non-fixed environment atmospheric element vectors, the data quality control detection and analysis viewpoint of the target lidar system was determined. This viewpoint not only ensures the accuracy and reliability of the data, but also provides methodological guidance for comprehensive evaluation of data quality. This comprehensive data quality control detection and analysis viewpoint helps to improve the application effect and value of lidar systems in the fields of environmental monitoring, weather forecasting, etc.

[0095] In summary, the present invention achieves comprehensive quality control and evaluation of lidar detection data by acquiring a comprehensive set of environmental atmospheric detection data, designing an efficient positioning label system, constructing a multi-dimensional environmental atmospheric element vector, and determining a comprehensive data quality control detection and analysis perspective, providing strong support for research and application in related fields.

[0096] In some preferred embodiments, the non-stationary ambient atmosphere element vector of the target lidar system is obtained based on the at least two groups of non-stationary ambient atmosphere detection data, including: obtaining the distribution of at least one ambient atmosphere data block in each group of the non-stationary ambient atmosphere detection data; based on the distribution of the at least one ambient atmosphere data block in each non-stationary ambient atmosphere detection data tuple, obtaining the trend characteristics of each ambient atmosphere data block in each non-stationary ambient atmosphere detection data tuple; wherein each non-stationary ambient atmosphere detection data tuple includes two consecutive groups of non-stationary ambient atmosphere detection data; based on the trend characteristics of each ambient atmosphere data block in each non-stationary ambient atmosphere detection data tuple, the non-stationary ambient atmosphere element vector of the target lidar system is obtained.

[0097] In some preferred embodiments, the process of acquiring the non-stationary ambient atmospheric element vector of the target lidar system is implemented by the following steps.

[0098] First, the system obtains the distribution of at least one ambient atmospheric data block in each group of non-stationary ambient atmospheric detection data. These data blocks are the basic units of ambient atmospheric detection data and may contain information on various atmospheric elements such as temperature, humidity, and wind speed. Non-stationary ambient atmospheric detection data are collected during the movement of the LiDAR system, so the distribution of these data blocks in different detection data groups will change with the change of spatial position.

[0099] Next, based on the distribution of these ambient atmospheric data blocks in each non-stationary ambient atmospheric detection data binary group, the system will further analyze the trend characteristics of each data block in each binary group. The non-stationary ambient atmospheric detection data binary group mentioned here refers to a data pair consisting of two consecutive sets of non-stationary ambient atmospheric detection data. By analyzing the distribution changes of these data blocks in the binary group, the changing trends and characteristics of atmospheric environmental elements in space and time can be revealed.

[0100] Finally, the system will comprehensively obtain the non-stationary ambient atmosphere element vector of the target lidar system based on the trend characteristics of each ambient atmosphere data block in each non-stationary ambient atmosphere detection data binary group. This vector is a multi-dimensional data set, which not only contains the information of various atmospheric elements, but also reflects the change trend and correlation of these elements in space and time. The acquisition of this vector provides an important data basis for evaluating the spatial detection status and temporal detection status of the lidar system.

[0101] It should be noted that in this process, the system may use various data processing and analysis methods, such as data interpolation, filtering, statistical analysis, etc., to improve the accuracy and reliability of the data. At the same time, the system also needs to consider various possible sources of error and interference factors, such as instrument errors, environmental noise, etc., to ensure that the obtained non-stationary environmental atmospheric element vector can truly reflect the actual state of the atmospheric environment.

[0102] Through the above steps, the system can effectively obtain the non-stationary environmental atmospheric element vector of the target lidar system, providing important support for subsequent data quality control detection and analysis. The acquisition of this vector not only improves the application effect and value of the lidar system in the field of atmospheric environment monitoring, but also provides strong data support for research and application in related fields.

[0103] In some other preferred embodiments, the non-stationary ambient atmosphere element vector of the target laser radar system is obtained based on the trend characteristics of each ambient atmosphere data block in each non-stationary ambient atmosphere detection data tuple, including: for each non-stationary ambient atmosphere detection data tuple, determining the tuple characteristics of the trend characteristics of at least one ambient atmosphere data block in each non-stationary ambient atmosphere detection data tuple, and obtaining the averaged trend characteristics of each non-stationary ambient atmosphere detection data tuple; the characteristic value of the averaged trend characteristic is used to represent the time detection state of the target laser radar system in the non-stationary ambient atmosphere detection data tuple, and the gradient of the averaged trend characteristic is used to represent the detection change of the target laser radar system in the non-stationary ambient atmosphere detection data tuple; based on the averaged trend characteristics of each non-stationary ambient atmosphere detection data tuple, the non-stationary ambient atmosphere element vector of the target laser radar system is obtained.

[0104] In some further preferred embodiments, the acquisition of the non-stationary ambient atmospheric element vector of the target lidar system is further refined into the following steps.

[0105] First, for each non-stationary ambient atmospheric detection data binary group, the system determines the binary features of the trend characteristics of each ambient atmospheric data block. The binary features here refer to the comparison results of the trend characteristics of the same data blocks in two consecutive non-stationary ambient atmospheric detection data groups, such as their difference, ratio or other statistics. By calculating these binary features, the changes and trends of atmospheric environmental elements during the movement of the lidar system can be revealed.

[0106] Next, the system will calculate the average trend feature of each binary group based on the binary features in each non-stationary environment atmospheric detection data binary group. The average trend feature is obtained by averaging or other statistical processing of the binary features of each data block in the binary group, which represents the overall trend change of the binary group. The characteristic value of this average trend feature can be used to represent the time detection state of the target lidar system in the binary group, that is, the average change level of the atmospheric environment elements in this time period.

[0107] At the same time, the system also calculates the gradient of the averaged trend feature to represent the detection changes of the target lidar system in the non-stationary environment atmospheric detection data binary. The gradient reflects the rate of change or slope of the averaged trend feature between different binary groups, which can reveal the speed and direction of change of atmospheric environmental elements in space and time.

[0108] Finally, the system will comprehensively derive the non-stationary environment atmospheric element vector of the target lidar system based on the average trend characteristics of each non-stationary environment atmospheric detection data binary group. This vector combines the average trend characteristics and gradient information of each binary group, providing a comprehensive and multi-dimensional description to characterize the lidar system's detection capability and change characteristics of the atmospheric environment in a mobile state.

[0109] Through the above steps, the system can obtain the non-stationary environmental atmospheric element vector of the target lidar system more precisely, providing richer information for subsequent data analysis and application. The acquisition of this vector not only helps to evaluate the performance status of the lidar system, but also provides strong data support for research and application in the fields of environmental monitoring and weather forecasting.

[0110] Under some other exemplary design ideas, the obtaining of the distribution of at least one ambient atmospheric data block in each group of the non-fixed ambient atmospheric detection data includes: performing an embedding operation on the at least two groups of non-fixed ambient atmospheric detection data to obtain the ambient atmospheric detection embedded data corresponding to each of the at least two groups of non-fixed ambient atmospheric detection data; based on the ambient atmospheric detection embedded data corresponding to the first non-fixed ambient atmospheric detection data in the at least two groups of non-fixed ambient atmospheric detection data, determining at least one ambient atmospheric data block in the ambient atmospheric detection embedded data corresponding to the first non-fixed ambient atmospheric detection data; wherein the first non-fixed ambient atmospheric detection data is the first group of non-fixed ambient atmospheric detection data in the at least two groups of non-fixed ambient atmospheric detection data; and using a convolution strategy to determine the distribution of the at least one ambient atmospheric data block in the ambient atmospheric detection embedded data corresponding to each of the groups of the non-fixed ambient atmospheric detection data based on the distribution of the at least one ambient atmospheric data block in the ambient atmospheric detection embedded data corresponding to the first non-fixed ambient atmospheric detection data.

[0111] Under some other exemplary design ideas, the process of the system obtaining the distribution of at least one ambient atmosphere data block in each group of non-stationary ambient atmosphere detection data is achieved through the following steps.

[0112] First, the system will perform embedding operations on at least two sets of non-stationary ambient atmospheric detection data to obtain their corresponding ambient atmospheric detection embedded data. Embedding operation is a technique that maps raw data into a high-dimensional space. Through this mapping, more features and patterns in the raw data can be extracted. Here, embedding operation can help the system better capture and represent the distribution and characteristics of ambient atmospheric data blocks in non-stationary ambient atmospheric detection data.

[0113] Next, the system will determine at least one ambient atmospheric data block based on the ambient atmospheric detection embedded data corresponding to the first set of non-stationary ambient atmospheric detection data (i.e., one of the data in the first non-stationary ambient atmospheric detection data tuple) in the at least two sets of non-stationary ambient atmospheric detection data. These data blocks are parts of the embedded data that represent specific atmospheric environmental elements, and they can be determined by various methods, such as clustering, segmentation, etc.

[0114] Then, the system will adopt a convolution strategy to determine the distribution of these data blocks in the ambient atmosphere detection embedded data corresponding to each group of non-stationary ambient atmosphere detection data based on the distribution of at least one ambient atmosphere data block in the ambient atmosphere detection embedded data corresponding to the first non-stationary ambient atmosphere detection data binary group. The convolution strategy is a commonly used technology in image processing, which can perform feature extraction and pattern recognition on local areas by means of sliding windows. Here, the convolution strategy can help the system effectively capture and compare the distribution and characteristics of ambient atmosphere data blocks in different non-stationary ambient atmosphere detection data groups.

[0115] Through the above steps, the system can accurately obtain the distribution information of at least one ambient atmospheric data block in each group of non-stationary ambient atmospheric detection data. This information provides an important basis for subsequent data processing and analysis, and helps the system to more comprehensively understand and evaluate the ambient atmospheric detection capabilities and performance of the target lidar system in a mobile state. At the same time, this method based on embedding operations and convolution strategies also has certain versatility and flexibility, and can adapt to the processing needs of non-stationary ambient atmospheric detection data of different types and scales.

[0116] In some other possible embodiments, the method further includes: using a data optimization algorithm to update the distribution of the at least one ambient atmosphere data block in the ambient atmosphere detection embedded data corresponding to each group of the non-fixed ambient atmosphere detection data, to obtain the updated distribution of the at least one ambient atmosphere data block in the ambient atmosphere detection embedded data corresponding to each group of the non-fixed ambient atmosphere detection data.

[0117] Then, based on the distribution of the at least one ambient atmosphere data block in each non-fixed ambient atmosphere detection data tuple, the trend characteristics of each ambient atmosphere data block in each non-fixed ambient atmosphere detection data tuple are obtained, including: based on the updated distribution of the at least one ambient atmosphere data block in each non-fixed ambient atmosphere detection data tuple, the updated trend characteristics of each ambient atmosphere data block in each non-fixed ambient atmosphere detection data tuple are obtained.

[0118] Among them, the triggering cycle of the update operation includes at least one of the following: after adopting the convolution strategy to determine the distribution of the at least one ambient atmospheric data block in the ambient atmospheric detection embedded data corresponding to each group of the non-fixed ambient atmospheric detection data, the distribution of the at least one ambient atmospheric data block in the ambient atmospheric detection embedded data corresponding to each group of the non-fixed ambient atmospheric detection data is updated, and the convolution strategy is adopted based on the distribution of the at least one ambient atmospheric data block in the ambient atmospheric detection embedded data corresponding to the previous group of the non-fixed ambient atmospheric detection data to determine the distribution of the at least one ambient atmospheric data block in the ambient atmospheric detection embedded data corresponding to the next group of the non-fixed ambient atmospheric detection data; or, after adopting the convolution strategy to determine the distribution of the at least one ambient atmospheric data block in the ambient atmospheric detection embedded data corresponding to each group of the non-fixed ambient atmospheric detection data, the distribution of the at least one ambient atmospheric data block in the ambient atmospheric detection embedded data corresponding to each group of the non-fixed ambient atmospheric detection data is updated.

[0119] In some other possible embodiments, the method further introduces a data optimization algorithm to improve the distribution of ambient atmosphere data blocks in the non-stationary ambient atmosphere detection data. The following are detailed implementation steps.

[0120] First, the system will use a data optimization algorithm to update the distribution of at least one ambient atmospheric data block in the ambient atmospheric detection embedded data corresponding to each group of non-stationary ambient atmospheric detection data binary. The data optimization algorithm can be various machine learning algorithms, statistical methods or optimization techniques, such as gradient descent, genetic algorithm, particle swarm optimization, etc. These algorithms can adjust and optimize the distribution of ambient atmospheric data blocks according to certain criteria or objective functions to improve their accuracy and reliability.

[0121] Through the update operation, the system can obtain the updated distribution of at least one ambient atmosphere data block in the ambient atmosphere detection embedded data corresponding to each set of non-stationary ambient atmosphere detection data tuples. These updated distributions more accurately reflect the changes and characteristics of the ambient atmosphere data blocks in the actual detection process.

[0122] Then, the system will recalculate the trend characteristics of each ambient atmospheric data block in each non-stationary ambient atmospheric detection data binary group based on these updated distributions. Since the distribution has been optimized and updated, the calculated trend characteristics will be more accurate and reliable. These updated trend characteristics can better reveal the changing trends and laws of atmospheric environmental elements during the movement of the lidar system.

[0123] Regarding the triggering period of the update operation, there are two possible implementations.

[0124] One is to immediately perform an update operation after using a convolution strategy to determine the distribution of at least one ambient atmospheric data block in the ambient atmospheric detection embedded data corresponding to each set of non-stationary ambient atmospheric detection data binary. Then, the system will use a convolution strategy to determine the distribution of these data blocks in the ambient atmospheric detection embedded data corresponding to the next set of non-stationary ambient atmospheric detection data binary based on the distribution of at least one ambient atmospheric data block in the ambient atmospheric detection embedded data corresponding to the previous set of non-stationary ambient atmospheric detection data binary. This method can achieve real-time data update and optimization, and is suitable for application scenarios with high real-time requirements.

[0125] The other is to perform a one-time update operation after using a convolution strategy to determine the distribution of at least one ambient atmospheric data block in the ambient atmospheric detection embedded data corresponding to each group of non-stationary ambient atmospheric detection data. This method can uniformly process and optimize all non-stationary ambient atmospheric detection data, and is suitable for application scenarios with high requirements for data consistency and accuracy.

[0126] Through the above steps and implementation methods, the system can more accurately acquire and analyze the distribution and trend characteristics of ambient atmospheric data blocks in non-stationary ambient atmospheric detection data, providing more powerful support for subsequent data quality control detection and analysis.

[0127] In some other replaceable embodiments, the data optimization algorithm is used to update the distribution of the at least one ambient atmosphere data block in each group of the non-fixed ambient atmosphere detection data to obtain the updated distribution of the at least one ambient atmosphere data block in each group of the non-fixed ambient atmosphere detection data, including: based on the updated distribution of the first ambient atmosphere data block at the q-1th period node, obtaining the distribution prediction result of the first ambient atmosphere data block at the qth period node, q is an integer greater than 1; wherein the updated distribution of the 1st period node is the distribution of the first ambient atmosphere data block in the ambient atmosphere detection embedded data corresponding to the first non-fixed ambient atmosphere detection data; two consecutive groups of the non-fixed ambient atmosphere detection data in each non-fixed ambient atmosphere detection data tuple are separated by one period node; based on the The updated distribution of the first ambient atmospheric data block at the qth period node is obtained based on the distribution prediction result of an ambient atmospheric data block at the qth period node, the detection distribution of the first ambient atmospheric data block at the qth period node and the optimization coefficient corresponding to the data optimization algorithm; wherein the detection distribution of the first ambient atmospheric data block at the qth period node is the distribution obtained by adopting the convolution strategy; the optimization coefficient is determined based on the fluctuation weight corresponding to the distribution prediction result of the first ambient atmospheric data block at the qth period node, the fluctuation weight corresponding to the distribution prediction result of the first ambient atmospheric data block at the qth period node is determined based on the fluctuation weight corresponding to the updated distribution of the first ambient atmospheric data block at the q-1th period node, and the fluctuation weight is used to represent the offset between the distribution prediction result and the detection distribution.

[0128] In some other alternative embodiments, the system adopts a data optimization method based on periodic nodes to update the distribution of ambient atmosphere data blocks in the non-stationary ambient atmosphere detection data. The following are detailed implementation steps.

[0129] First, the system predicts the distribution of the first ambient atmosphere data block at the qth period node based on the updated distribution at the q-1th period node. Here, q is an integer greater than 1, representing the sequence number of the period node. The updated distribution of the first period node is known, which corresponds to the distribution of the first ambient atmosphere data block in the ambient atmosphere detection embedded data in the first non-fixed ambient atmosphere detection data tuple. Two consecutive groups of non-fixed ambient atmosphere detection data in each non-fixed ambient atmosphere detection data tuple are separated by one period node.

[0130] Predictive distribution can be achieved through various time series prediction methods, such as autoregressive models, moving average models, neural networks, etc. These methods can use historical data (i.e. the updated distribution of the nodes in the q-1th period) to predict future data (i.e. the distribution of the nodes in the qth period).

[0131] Next, the system will obtain the detection distribution of the first ambient atmospheric data block at the qth period node. This detection distribution is obtained through the convolution strategy, which reflects the actual detection situation of the data block at the current period node.

[0132] Then, the system will combine the predicted distribution, the detected distribution and the optimization coefficient corresponding to the data optimization algorithm to calculate the updated distribution of the first ambient atmospheric data block at the qth period node. The optimization coefficient is determined based on the fluctuation weight corresponding to the predicted distribution. The fluctuation weight represents the degree of deviation or uncertainty between the predicted distribution and the detected distribution. If the predicted distribution differs greatly from the detected distribution, the fluctuation weight will be larger, otherwise it will be smaller.

[0133] Specifically, the optimization coefficient can be a function or transformation form of the fluctuation weight. For example, it can be the inverse, square root, logarithm, etc. of the fluctuation weight. By selecting a suitable function or transformation form, the system can adjust the size of the optimization coefficient according to the size of the fluctuation weight, thereby achieving a reasonable balance and compromise between the predicted distribution and the detected distribution.

[0134] Finally, the system will use the updated distribution as the final distribution of the first ambient atmospheric data block at the node of the qth period, and can be used for subsequent data processing and analysis. At the same time, this process can be repeated continuously to update and optimize the distribution of the ambient atmospheric data block in the non-stationary ambient atmospheric detection data period by period.

[0135] Through this data optimization method based on periodic nodes, the system can more accurately capture and track the changing rules and trend characteristics of the ambient atmospheric data blocks in the non-stationary ambient atmospheric detection data. This is of great significance for improving the performance evaluation of the lidar system and the accuracy of environmental monitoring.

[0136] In some optional technical solutions, the fixed environment atmospheric detection data includes meteorological element data collected by ground observation instruments and atmospheric chemical composition data collected by atmospheric composition analyzers. The meteorological element data is used to obtain the meteorological index vector of the target lidar system, and the atmospheric chemical composition data is used to obtain the atmospheric chemical composition vector of the target lidar system.

[0137] After obtaining the set of ambient atmosphere detection data of the target lidar system, the method includes: adjusting the at least one set of fixed ambient atmosphere detection data to obtain at least one set of processed fixed ambient atmosphere detection data; the adjustment includes at least one of data matching, data deduplication, and data embedding; based on the at least one set of processed fixed ambient atmosphere detection data, determining first fixed ambient atmosphere detection data, wherein the first fixed ambient atmosphere detection data is the processed fixed ambient atmosphere detection data with the highest evaluation weight among the at least one set of processed fixed ambient atmosphere detection data.

[0138] Furthermore, the fixed environment atmospheric element vector of the first positioning tag is obtained based on the fixed environment atmospheric detection data, including: obtaining the meteorological index vector of the first positioning tag based on the meteorological element data in the first fixed environment atmospheric detection data, and obtaining the atmospheric chemical composition vector of the first positioning tag based on the atmospheric chemical composition data in the first fixed environment atmospheric detection data; obtaining the fixed environment atmospheric element vector of the first positioning tag based on the meteorological index vector of the first positioning tag and the atmospheric chemical composition vector of the first positioning tag.

[0139] In some optional technical solutions, the system processes the atmospheric detection data of the fixed environment in a more detailed and accurate way. The following are the detailed implementation steps.

[0140] First of all, it should be made clear that the fixed environment atmospheric detection data consists of two parts: meteorological element data collected by ground observation instruments and atmospheric chemical composition data collected by atmospheric composition analyzers. Meteorological element data are mainly used to obtain meteorological index vectors of the target lidar system, which may include temperature, humidity, wind speed, wind direction, etc.; while atmospheric chemical composition data are used to obtain atmospheric chemical composition vectors of the target lidar system, such as the content of carbon dioxide, oxygen, nitrogen and other possible pollutants.

[0141] After acquiring the ambient atmospheric detection data set of the target LiDAR system, the system will perform a series of adjustments on these fixed ambient atmospheric detection data. These adjustments may include steps such as data matching, data deduplication, and data embedding. Data matching is to ensure that data from different sources or formats can be unified in standards and units; data deduplication is to eliminate duplicate or redundant data to ensure the uniqueness and accuracy of the data; data embedding is to map the data into a high-dimensional space to extract more features and patterns.

[0142] After adjustment and processing, the system will obtain at least one set of processed fixed environment atmospheric detection data. Then, the system will determine the first fixed environment atmospheric detection data according to the evaluation weight of these data. The evaluation weight here can be set according to factors such as the source, quality, and timeliness of the data. The processed fixed environment atmospheric detection data with the highest evaluation weight will be selected as the first fixed environment atmospheric detection data.

[0143] Next, the system will obtain the fixed environment atmospheric element vector of the first positioning tag based on the first fixed environment atmospheric detection data. Specifically, the system will obtain the meteorological index vector and the atmospheric chemical composition vector based on the meteorological element data and the atmospheric chemical composition data respectively. These two vectors reflect the meteorological conditions and atmospheric chemical composition conditions at the location of the target lidar system, respectively.

[0144] Finally, the system will merge or fuse the meteorological index vector and the atmospheric chemical composition vector to obtain the fixed ambient atmospheric element vector of the first positioning tag. This vector combines information from both meteorology and chemical composition, and can more comprehensively describe the ambient atmospheric conditions at the location of the target lidar system.

[0145] Through this processing method, the system can more effectively utilize the fixed ambient atmosphere detection data to obtain the ambient atmosphere element vector of the target lidar system, providing a more reliable foundation for subsequent data analysis and application.

[0146] In some further optional embodiments, parsing the fixed environment atmospheric detection data to obtain a first positioning tag includes: using a debugged detection data parsing network to parse the first fixed environment atmospheric detection data to obtain the first positioning tag; wherein the debugging process of the detection data parsing network includes steps 201-204.

[0147] Step 201, obtaining a debugging sample set of the detection data analysis network, wherein the debugging sample set includes at least one debugging example tuple, each debugging example tuple includes a detection data example and a training annotation corresponding to the detection data example; wherein the detection data example is the linked environmental atmospheric index detection data obtained by collecting data from a selected space in a target detection area, and the detection data example includes any one of local meteorological element data collected by a ground observation instrument and local atmospheric chemical composition data collected by an atmospheric composition analyzer, and the training annotation is used to characterize the positioning label of the selected space corresponding to the detection data example in the target detection area.

[0148] Step 202: perform interval mapping processing on each group of detection data examples in the debugging sample set to obtain processed detection data examples, wherein the interval mapping processing is used to process each group of detection data examples into data with consistent detection indicators.

[0149] Step 203: Using the detection data parsing network to obtain a positioning prediction label corresponding to the detection data example based on the processed detection data example.

[0150] Step 204: Debug the detection data parsing network based on the difference between the positioning prediction label and the training annotation to obtain a debugged detection data parsing network.

[0151] In some other specific embodiments, the process of the system processing the fixed environment atmospheric detection data and generating the first positioning tag is more detailed. The following are detailed operation steps and explanations.

[0152] First, the system will obtain the first fixed environment atmospheric detection data, which has been previously processed to ensure its accuracy and completeness. These data may include meteorological element data collected by ground observation instruments, such as temperature, humidity, wind speed, etc., and atmospheric chemical composition data collected by atmospheric composition analyzers, such as carbon dioxide concentration, oxygen content, etc.

[0153] Next, the system uses a specially tuned detection data parsing network to parse these fixed-position ambient atmospheric detection data. This network acts like a translator, converting complex ambient atmospheric data into easy-to-understand positioning tags.

[0154] But before that, in order to ensure that the translator can work accurately, a series of debugging is required. This process includes four main steps: the system first needs to obtain the debugging sample set of the detection data parsing network. This set contains multiple debugging example pairs, each of which contains a detection data example and a corresponding training annotation. The detection data example is the environmental atmospheric data collected from a specific space in the target detection area, and the training annotation is the geographical location label corresponding to these data; since different detection data may have different units and ranges, direct comparison and analysis may be misleading. Therefore, the system needs to perform interval mapping on the data in these debugging sample sets. This process is like converting all data to the same metric for subsequent comparison and analysis; next, the system will use the detection data parsing network to process these processed detection data examples and generate corresponding positioning prediction labels. These prediction labels are based on the network's learning and experience, and they represent the system's preliminary judgment of the geographical location represented by these data; finally, the system will compare the differences between these positioning prediction labels and training annotations, and debug the detection data parsing network based on these differences. This process is like a feedback loop, where the system continuously adjusts and optimizes the parameters and structure of the network based on errors and deficiencies to improve the accuracy and efficiency of its parsing.

[0155] After this series of complex processing, the system can finally use the debugged detection data analysis network to analyze the first fixed environment atmospheric detection data and generate an accurate first positioning tag. This tag not only represents the precise information of the target lidar system's location, but also provides an important basis for subsequent data analysis and application.

[0156] Through such a technical solution, we can not only make full use of the rich information in the ambient atmospheric data, but also improve the accuracy and efficiency of positioning, providing strong support for various practical application scenarios.

[0157] In some exemplary embodiments, the data quality control detection and analysis point of view of the target lidar system based on the fixed environment atmospheric element vector and the non-fixed environment atmospheric element vector is determined, including: obtaining a detection quality evaluation vector of the target lidar system based on the fixed environment atmospheric element vector and the non-fixed environment atmospheric element vector, the detection quality evaluation vector being a knowledge vector obtained by combining the fixed environment atmospheric element vector and the non-fixed environment atmospheric element vector; respectively determining the correlation weight between the detection quality evaluation vector and the detection quality evaluation vector of at least one past lidar system label in the cloud sharing server; if the first correlation weight corresponding to the first past lidar system label is the maximum correlation weight, and the maximum correlation weight is greater than a first threshold, then determining the data quality control detection and analysis point of view of the target lidar system is that the target lidar system matches the first past lidar system label.

[0158] In some exemplary embodiments, the system performs a series of operations to determine the data quality control detection analysis view of the target lidar system. The following are detailed steps and explanations.

[0159] First, the system generates a detection quality evaluation vector of the target lidar system based on the fixed environment atmospheric element vector and the non-fixed environment atmospheric element vector. This detection quality evaluation vector is a comprehensive knowledge vector obtained by combining the fixed and non-fixed environment atmospheric element vectors, which reflects the impact of various environmental factors on the detection quality of the target lidar system during the detection process.

[0160] Next, the system needs to evaluate the validity and accuracy of this detection quality evaluation vector. To do this, the system accesses a cloud shared server that stores the labels of past lidar systems and their corresponding detection quality evaluation vectors. The system then calculates the correlation weight between the detection quality evaluation vector of the target lidar system and the detection quality evaluation vectors of these past labels.

[0161] The correlation weight is an indicator that measures the similarity between two vectors. The larger its value is, the more similar the two vectors are, which means that the detection environment of the target lidar system is closer to the detection environment of a previous lidar system.

[0162] After calculating all the correlation weights, the system will find the largest one, which is called the maximum correlation weight. The past lidar system label corresponding to this maximum correlation weight is called the first past lidar system label.

[0163] Finally, the system determines whether the maximum correlation weight exceeds a preset first threshold. If so, the system determines that the target LiDAR system's data quality control detection analysis viewpoint matches the target LiDAR system with the first past LiDAR system label. In other words, the system considers the target LiDAR system's detection quality to be equivalent to the detection quality represented by the first past LiDAR system label.

[0164] This technical solution not only takes into account the current detection environment of the target LiDAR system, but also draws on the experience data of past LiDAR systems, thereby improving the accuracy and reliability of data quality control detection and analysis. At the same time, through the use of cloud shared servers, data sharing and reuse are achieved, improving the efficiency and scalability of the system.

[0165] For some possible embodiments, determining the data quality control detection and analysis viewpoint of the target lidar system based on the fixed environment atmospheric element vector and the non-fixed environment atmospheric element vector includes: respectively determining the second correlation weight between the fixed environment atmospheric element vector and the fixed environment atmospheric element vector of at least one past lidar system label in the cloud sharing server, and respectively determining the third correlation weight between the non-fixed environment atmospheric element vector and the non-fixed environment atmospheric element vector of at least one past lidar system label in the cloud sharing server; integrating the second correlation weight and the third correlation weight corresponding to the at least one past lidar system label to obtain the global correlation weight corresponding to the at least one past lidar system label; if the global correlation weight corresponding to the second past lidar system label is the maximum correlation weight, and the maximum correlation weight is greater than the second threshold, then determining the data quality control detection and analysis viewpoint of the target lidar system is that the target lidar system matches the second past lidar system label.

[0166] In some possible embodiments, the system determines the data quality control detection analysis viewpoint of the target lidar system through the following steps.

[0167] First, the system calculates the second correlation weight between the fixed environment atmospheric element vector and the fixed environment atmospheric element vector of at least one past lidar system tag stored in the cloud sharing server. The purpose of this step is to measure the similarity between the fixed environment atmospheric element of the current target lidar system and the fixed environment atmospheric element of the past lidar system tag.

[0168] At the same time, the system also determines the third correlation weight between the non-stationary ambient atmospheric element vector and the non-stationary ambient atmospheric element vector of at least one past lidar system tag stored in the cloud sharing server. The purpose of this step is to evaluate the similarity between the non-stationary ambient atmospheric element of the current target lidar system and the non-stationary ambient atmospheric element of the past lidar system tag.

[0169] After calculating the second and third relevance weights, the system integrates these relevance weights to obtain the global relevance weight corresponding to each past lidar system label. The global relevance weight comprehensively considers the similarity between the fixed and non-fixed ambient atmospheric elements and the past lidar system labels, and is a more comprehensive evaluation indicator.

[0170] Next, the system will find the maximum value in the global correlation weight, and the past lidar system label corresponding to this maximum value is called the second past lidar system label. At the same time, the system will determine whether this maximum global correlation weight exceeds a preset second threshold.

[0171] If the maximum global correlation weight exceeds the second threshold, the system determines that the data quality control detection analysis view of the target lidar system is that the target lidar system matches the second past lidar system label. This means that the detection environment of the current target lidar system is very similar to the detection environment of a past lidar system (i.e., the system represented by the second past lidar system label), so the experience data of the past lidar system can be used to optimize the data quality control strategy of the current system.

[0172] Through such a technical solution, the system can more accurately evaluate the detection environment of the target LiDAR system and make effective data quality control recommendations based on the experience data of past LiDAR systems. This not only improves the accuracy and efficiency of data quality control, but also helps to improve the overall performance of the target LiDAR system.

[0173] For some other possible embodiments, the determining of the data quality control detection and analysis viewpoint of the target lidar system based on the fixed environment atmospheric element vector and the non-fixed environment atmospheric element vector includes: using a quality control detection and analysis network to obtain the data quality control detection and analysis viewpoint of the target lidar system based on the fixed environment atmospheric element vector and the non-fixed environment atmospheric element vector; wherein the quality control detection and analysis network is used to perform element mining on the fixed environment atmospheric element vector and the non-fixed environment atmospheric element vector to obtain element mining results; respectively determining the correlation weights between the element mining results and the element mining results of at least one past lidar system label in the cloud sharing server; based on at least one correlation weight, generating the data quality control detection and analysis viewpoint of the target lidar system, the fourth correlation weight corresponding to the third past lidar system label indicated by the data quality control detection and analysis viewpoint being the maximum correlation weight, and the maximum correlation weight being greater than a third threshold.

[0174] In some other possible embodiments, the system uses a quality control detection and analysis network to determine the data quality control detection and analysis viewpoint of the target laser radar system. The following are detailed steps and explanations.

[0175] First, the system uses the quality control detection and analysis network to process the fixed ambient atmospheric element vector and the non-fixed ambient atmospheric element vector. This network is designed to deeply mine the key elements in these two vectors to obtain element mining results. This process can be regarded as a deep analysis of the ambient atmospheric data by the system, with the aim of extracting the most relevant feature information for the target lidar system data quality control.

[0176] Next, the system compares these feature mining results with the feature mining results of past lidar system labels stored in the cloud sharing server. Specifically, the system calculates the correlation weights between the current feature mining results and the feature mining results of each past label. These correlation weights reflect the similarity between the data quality control requirements of the current target lidar system and the past lidar systems.

[0177] The system then generates a data quality control detection analysis view of the target LiDAR system based on these relevance weights. This view is based on one or more past LiDAR system labels, which represents the system's understanding and judgment of the current data quality control status of the target LiDAR system. In this process, the system pays special attention to past labels with large relevance weights, because they may provide the most valuable reference information for the current system.

[0178] Finally, the system selects a past LiDAR system label that is most relevant to the current target LiDAR system as the third past LiDAR system label. The fourth correlation weight corresponding to this label is the maximum value among all correlation weights, and this maximum value needs to exceed a preset third threshold. If this condition is met, the system determines that the data quality control detection and analysis point of view of the target LiDAR system matches the third past LiDAR system label. This means that the system believes that the data quality control requirements of the current target LiDAR system are very similar to those of the third past LiDAR system, so its experience can be used to optimize the data quality control strategy of the current system.

[0179] Through such a technical solution, the system can make full use of historical data and advanced analysis networks to evaluate and improve the data quality control effect of the target lidar system. This not only improves the accuracy and efficiency of data quality control, but also helps to improve the overall performance and reliability of the system.

[0180] Figure 2 FIG. 2 is a schematic diagram of a data quality monitoring system 200 provided in an embodiment of the present invention. Figure 2 The data quality supervision system 200 shown includes a processor 210. The processor 210 can call and run a computer program from a memory to implement the method in the embodiment of the present invention.

[0181] Alternatively, if Figure 2 As shown, the data quality supervision system 200 may further include a memory 230. The processor 210 may call and run a computer program from the memory 230 to implement the method in the embodiment of the present invention.

[0182] The memory 230 may be a separate device independent of the processor 210 , or may be integrated into the processor 210 .

[0183] Alternatively, if Figure 2 As shown, the data quality monitoring system 200 may further include a transceiver 220. The processor 210 may control the transceiver 220 to interact with other devices. Specifically, the transceiver 220 may send information or data to other devices, or receive information or data sent by other devices.

[0184] Optionally, the data quality monitoring system 200 may implement the corresponding processes corresponding to the storage engine or components in the storage engine (such as a processing module) or a device deployed with a storage engine in each method of the embodiments of the present invention, which will not be described here for the sake of brevity.

[0185] It should be understood that the processor of the embodiment of the present invention may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method embodiment can be completed by the hardware integrated logic circuit in the processor or the instruction in the form of software. The above processor can be a general processor, a digital signal processor (Digital Signal Processor, DSP), an application specific integrated circuit (Application Specific Integrated Circuit, ASIC), a field programmable gate array (Field Programmable Gate Array, FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present invention can be implemented or executed. The general processor can be a microprocessor or the processor can also be any conventional processor. The steps of the method disclosed in the embodiment of the present invention can be directly embodied as a hardware decoding processor to perform, or the hardware and software modules in the decoding processor can be combined to perform. The software module can be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory, and the processor reads the information in the memory and completes the steps of the above method in combination with its hardware.

[0186] It can be understood that the memory in the embodiment of the present invention can be a volatile memory or a non-volatile memory, or can include both volatile and non-volatile memories. Among them, the non-volatile memory can be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory can be a random access memory (RAM), which is used as an external cache. By way of example and not limitation, many forms of RAM are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM), and direct RAM bus random access memory (DR RAM). It should be noted that the memory of the systems and methods described herein is intended to include, but is not limited to, these and any other suitable types of memory.

[0187] It should be understood that the above-mentioned memory is exemplary but not restrictive. For example, the memory in the embodiment of the present invention may also be static random access memory (static RAM, SRAM), dynamic random access memory (dynamic RAM, DRAM), synchronous dynamic random access memory (synchronous DRAM, SDRAM), double data rate synchronous dynamic random access memory (double data rate SDRAM, DDR SDRAM), enhanced synchronous dynamic random access memory (enhanced SDRAM, ESDRAM), synchronous link dynamic random access memory (synch link DRAM, SLDRAM) and direct memory bus random access memory (Direct Rambus RAM, DR RAM), etc. That is to say, the memory in the embodiment of the present invention is intended to include but not limited to these and any other suitable types of memory.

[0188] Based on the above, a computer-readable storage medium is provided, on which a computer program is stored, and the computer program implements the above method when running.

[0189] The above are only embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations.

Claims

1. A laser radar data quality supervision method, characterized in that: Applied to a data quality supervision system, the method comprises: Acquire an ambient atmosphere detection data set of a target laser radar system; wherein the ambient atmosphere detection data set includes at least one set of fixed ambient atmosphere detection data collected by the target laser radar system at a fixed position, and at least two sets of non-fixed ambient atmosphere detection data collected by the target laser radar system at a non-fixed position, wherein the fixed ambient atmosphere detection data and the non-fixed ambient atmosphere detection data are linkage ambient atmosphere index detection data obtained by collecting data from a selected space in a target detection area; Parsing the fixed environment atmosphere detection data to obtain a first positioning tag, where the first positioning tag is used to characterize the distribution of the selected space corresponding to the fixed environment atmosphere detection data in the target detection area; Based on the fixed environment atmosphere detection data, a fixed environment atmosphere element vector of the first positioning tag is obtained; based on the at least two sets of non-fixed environment atmosphere detection data, a non-fixed environment atmosphere element vector of the target laser radar system is obtained, and the non-fixed environment atmosphere element vector is used to characterize the spatial detection state and the temporal detection state of the target laser radar system; Determining a data quality control detection analysis viewpoint of the target laser radar system based on the fixed environment atmospheric element vector and the non-fixed environment atmospheric element vector; The step of determining the data quality control detection and analysis viewpoint of the target laser radar system based on the fixed environment atmospheric element vector and the non-fixed environment atmospheric element vector comprises: Based on the fixed environment atmosphere element vector and the non-fixed environment atmosphere element vector, a detection quality evaluation vector of the target laser radar system is obtained, wherein the detection quality evaluation vector is a knowledge vector obtained by combining the fixed environment atmosphere element vector and the non-fixed environment atmosphere element vector; Determine the correlation weights between the detection quality evaluation vector and the detection quality evaluation vector of at least one past laser radar system tag in the cloud sharing server respectively; If the first correlation weight corresponding to the first past lidar system label is the maximum correlation weight, and the maximum correlation weight is greater than the first threshold, then the data quality control detection and analysis point of view of the target lidar system is determined to be that the target lidar system matches the first past lidar system label.

2. The method according to claim 1, characterized in that The step of obtaining the non-stationary ambient atmosphere element vector of the target laser radar system based on the at least two sets of non-stationary ambient atmosphere detection data includes: Obtaining distribution of at least one ambient atmosphere data block in each group of the non-stationary ambient atmosphere detection data; Based on the distribution of the at least one ambient atmosphere data block in each non-stationary ambient atmosphere detection data tuple, a trend feature of each ambient atmosphere data block in each non-stationary ambient atmosphere detection data tuple is obtained; wherein each non-stationary ambient atmosphere detection data tuple includes two consecutive groups of non-stationary ambient atmosphere detection data; Based on the trend characteristics of each ambient atmosphere data block in each of the non-stationary ambient atmosphere detection data tuples, the non-stationary ambient atmosphere element vector of the target laser radar system is obtained.

3. The method according to claim 2, characterized in that The method of obtaining the non-stationary ambient atmosphere element vector of the target laser radar system based on the trend characteristics of each ambient atmosphere data block in each non-stationary ambient atmosphere detection data tuple includes: For each of the non-stationary environment atmosphere detection data tuples, determine the tuple features of the trend features of at least one environment atmosphere data block in each of the non-stationary environment atmosphere detection data tuples, and obtain the averaged trend features of each of the non-stationary environment atmosphere detection data tuples; the feature value of the averaged trend feature is used to represent the time detection state of the target laser radar system in the non-stationary environment atmosphere detection data tuple, and the gradient of the averaged trend feature is used to represent the detection change of the target laser radar system in the non-stationary environment atmosphere detection data tuple; Based on the average trend characteristics of each non-stationary environment atmosphere detection data binary group, the non-stationary environment atmosphere element vector of the target laser radar system is obtained.

4. The method according to claim 2, characterized in that The obtaining of the distribution of at least one ambient atmosphere data block in each group of the non-stationary ambient atmosphere detection data includes: Performing an embedding operation on the at least two sets of non-stationary ambient atmosphere detection data to obtain ambient atmosphere detection embedded data corresponding to each of the at least two sets of non-stationary ambient atmosphere detection data; wherein the embedding operation is a technique for mapping original data into a high-dimensional space, and the embedding operation is used to extract features and patterns from the original data; Based on the ambient atmosphere detection embedded data corresponding to the first non-stationary ambient atmosphere detection data in the at least two groups of non-stationary ambient atmosphere detection data, determine at least one ambient atmosphere data block in the ambient atmosphere detection embedded data corresponding to the first non-stationary ambient atmosphere detection data; wherein the first non-stationary ambient atmosphere detection data is the first group of non-stationary ambient atmosphere detection data in the at least two groups of non-stationary ambient atmosphere detection data; A convolution strategy is adopted to determine the distribution of the at least one ambient atmosphere data block in the ambient atmosphere detection embedded data corresponding to each group of the non-fixed ambient atmosphere detection data based on the distribution of the at least one ambient atmosphere data block in the ambient atmosphere detection embedded data corresponding to the first non-fixed ambient atmosphere detection data.

5. The method according to claim 4, characterized in that The method further comprises: Using a data optimization algorithm to update the distribution of the at least one ambient atmosphere data block in the ambient atmosphere detection embedded data corresponding to each group of the non-stationary ambient atmosphere detection data, to obtain an updated distribution of the at least one ambient atmosphere data block in the ambient atmosphere detection embedded data corresponding to each group of the non-stationary ambient atmosphere detection data; The obtaining of the trend characteristics of each ambient atmosphere data block in each of the non-stationary ambient atmosphere detection data tuples based on the distribution of the at least one ambient atmosphere data block in each of the non-stationary ambient atmosphere detection data tuples includes: Based on the updated distribution of the at least one ambient atmosphere data block in each of the non-stationary ambient atmosphere detection data tuples, obtaining updated trend characteristics of each ambient atmosphere data block in each of the non-stationary ambient atmosphere detection data tuples; Wherein, the triggering cycle of the update operation includes at least one of the following: after adopting the convolution strategy to determine the distribution of the at least one ambient atmospheric data block in the ambient atmospheric detection embedded data corresponding to each group of the non-stationary ambient atmospheric detection data, the distribution of the at least one ambient atmospheric data block in the ambient atmospheric detection embedded data corresponding to each group of the non-stationary ambient atmospheric detection data is updated, and the distribution of the at least one ambient atmospheric data block in the ambient atmospheric detection embedded data corresponding to the next group of the non-stationary ambient atmospheric detection data is determined based on the distribution of the at least one ambient atmospheric data block in the ambient atmospheric detection embedded data corresponding to the previous group of the non-stationary ambient atmospheric detection data; or, after adopting the convolution strategy to determine the distribution of the at least one ambient atmospheric data block in the ambient atmospheric detection embedded data corresponding to each group of the non-stationary ambient atmospheric detection data, the distribution of the at least one ambient atmospheric data block in the ambient atmospheric detection embedded data corresponding to each group of the non-stationary ambient atmospheric detection data is updated; Among them, a data optimization algorithm is used to update the distribution of the at least one ambient atmospheric data block in each group of the non-fixed ambient atmospheric detection data to obtain the updated distribution of the at least one ambient atmospheric data block in each group of the non-fixed ambient atmospheric detection data, including: based on the updated distribution of the first ambient atmospheric data block at the q-1th period node, obtaining the distribution prediction result of the first ambient atmospheric data block at the qth period node, q is an integer greater than 1; wherein the updated distribution of the 1st period node is the distribution of the first ambient atmospheric data block in the ambient atmospheric detection embedded data corresponding to the first non-fixed ambient atmospheric detection data; two consecutive groups of the non-fixed ambient atmospheric detection data in each tuple of the non-fixed ambient atmospheric detection data are separated by one period node; based on the first ambient atmospheric data The updated distribution of the first ambient atmospheric data block at the qth period node is obtained by combining the distribution prediction result of the first ambient atmospheric data block at the qth period node, the detection distribution of the first ambient atmospheric data block at the qth period node and the optimization coefficient corresponding to the data optimization algorithm; wherein, the detection distribution of the first ambient atmospheric data block at the qth period node is the distribution obtained by adopting the convolution strategy; the optimization coefficient is determined based on the fluctuation weight corresponding to the distribution prediction result of the first ambient atmospheric data block at the qth period node, the fluctuation weight corresponding to the distribution prediction result of the first ambient atmospheric data block at the qth period node is determined based on the fluctuation weight corresponding to the updated distribution of the first ambient atmospheric data block at the q-1th period node, and the fluctuation weight is used to represent the offset between the distribution prediction result and the detection distribution.

6. The method according to claim 1, characterized in that The fixed environment atmospheric detection data includes meteorological element data collected by a ground observation instrument and atmospheric chemical composition data collected by an atmospheric composition analyzer, wherein the meteorological element data is used to obtain a meteorological index vector of the target laser radar system, and the atmospheric chemical composition data is used to obtain an atmospheric chemical composition vector of the target laser radar system; After acquiring the set of ambient atmosphere detection data of the target laser radar system, the method includes: adjusting the at least one set of fixed ambient atmosphere detection data to obtain at least one set of processed fixed ambient atmosphere detection data; the adjustment includes at least one of data matching, data deduplication, and data embedding; based on the at least one set of processed fixed ambient atmosphere detection data, determining first fixed ambient atmosphere detection data, wherein the first fixed ambient atmosphere detection data is the processed fixed ambient atmosphere detection data with the highest evaluation weight among the at least one set of processed fixed ambient atmosphere detection data; The step of obtaining the fixed environment atmospheric element vector of the first positioning tag based on the fixed environment atmospheric detection data comprises: obtaining the meteorological index vector of the first positioning tag based on the meteorological element data in the first fixed environment atmospheric detection data, and obtaining the atmospheric chemical composition vector of the first positioning tag based on the atmospheric chemical composition data in the first fixed environment atmospheric detection data; obtaining the fixed environment atmospheric element vector of the first positioning tag based on the meteorological index vector of the first positioning tag and the atmospheric chemical composition vector of the first positioning tag; The step of parsing the atmospheric detection data of the fixed environment to obtain the first positioning tag includes: parsing the atmospheric detection data of the fixed environment using a debugged detection data parsing network to obtain the first positioning tag; The debugging process of the detection data analysis network is as follows: obtaining a debugging sample set of the detection data analysis network, the debugging sample set contains at least one debugging example binary group, each debugging example binary group contains a detection data example and a training annotation corresponding to the detection data example; wherein the detection data example is the linked environmental atmospheric index detection data obtained by collecting the selected space of the target detection area, the detection data example includes any one of the local meteorological element data collected by the ground observation instrument and the local atmospheric chemical composition data collected by the atmospheric composition analyzer, and the training annotation is used to characterize the positioning label of the selected space corresponding to the detection data example in the target detection area; performing interval mapping processing on each group of detection data examples in the debugging sample set to obtain the processed detection data examples, and the interval mapping processing is used to process each group of detection data examples into data with consistent detection indicators; using the detection data analysis network to obtain the positioning prediction label corresponding to the detection data example based on the processed detection data example; debugging the detection data analysis network based on the difference between the positioning prediction label and the training annotation to obtain the debugged detection data analysis network.

7. The method according to claim 1, characterized in that The step of determining the data quality control detection and analysis viewpoint of the target laser radar system based on the fixed environment atmospheric element vector and the non-fixed environment atmospheric element vector comprises: Determine a second correlation weight between the fixed environment atmospheric element vector and the fixed environment atmospheric element vector of at least one past laser radar system tag in the cloud sharing server, and determine a third correlation weight between the non-fixed environment atmospheric element vector and the non-fixed environment atmospheric element vector of at least one past laser radar system tag in the cloud sharing server; Integrate the second correlation weight and the third correlation weight corresponding to the at least one past laser radar system label to obtain a global correlation weight corresponding to the at least one past laser radar system label; If the global correlation weight corresponding to the second past lidar system label is the maximum correlation weight, and the maximum correlation weight is greater than the second threshold, then the data quality control detection and analysis point of view of the target lidar system is determined to be that the target lidar system matches the second past lidar system label.

8. The method according to claim 1, characterized in that The step of determining the data quality control detection and analysis viewpoint of the target laser radar system based on the fixed environment atmospheric element vector and the non-fixed environment atmospheric element vector comprises: Using a quality control detection and analysis network to obtain a data quality control detection and analysis viewpoint of the target laser radar system based on the fixed environment atmospheric element vector and the non-fixed environment atmospheric element vector; Among them, the quality control detection and analysis network is used to perform factor mining on the fixed environment atmospheric element vector and the non-fixed environment atmospheric element vector to obtain factor mining results; respectively determine the correlation weights between the factor mining results and the factor mining results of at least one past lidar system label in the cloud sharing server; based on at least one correlation weight, generate a data quality control detection and analysis viewpoint of the target lidar system, the fourth correlation weight corresponding to the third past lidar system label indicated by the data quality control detection and analysis viewpoint is the maximum correlation weight, and the maximum correlation weight is greater than the third threshold; and, the data quality control detection and analysis viewpoint of the target lidar system is matched with the third past lidar system label, and the data quality control requirements of the current target lidar system are similar to those of the third past lidar system.

9. A data quality supervision system, characterized in that: The method comprises at least one processor and a memory; the memory stores computer-executable instructions; the at least one processor executes the computer-executable instructions stored in the memory, so that the at least one processor executes the method according to any one of claims 1 to 8.

Images