Air detection analysis method and system based on laser radar

Abstract

The invention discloses an air detection analysis method and system based on a laser radar, and the method comprises the steps: carrying out the air detection through employing a laser radar technology, and collecting the multi-dimensional air quality big data of a preset region in a preset time period; multiple nodes are set based on different air parameters, and data are stored in a graph structure form. And traversing the data points through a breadth-first search algorithm, and marking primary abnormal data points. And then, continuous data extraction and serialization are carried out on each abnormal data point, prediction analysis and secondary evaluation are carried out by using an LSTM prediction model, and an abnormal evaluation result is generated. And based on the result, the air quality big data is processed, and a secondary monitoring scheme based on the laser radar is generated, so that more accurate and reliable air quality monitoring and management are realized.

Description

Technical Field

[0001] The present invention relates to the field of air monitoring, and more specifically, to an air detection and analysis method and system based on lidar. Background Art

[0002] With the acceleration of industrialization and urbanization, air quality problems have become increasingly serious. Traditional air quality monitoring methods rely on fixed-point sampling and laboratory analysis, which have problems such as long response times and limited coverage. As an emerging remote sensing technology, lidar can quickly obtain aerosol distribution information over a large area, thus enabling precise air detection.

[0003] However, in the face of a complex urban environment, lidar data may be interfered by various factors, such as building reflections, vehicle emissions, meteorological conditions, equipment usage errors, etc. These factors will affect the accuracy and reliability of the data. Traditional technologies often rely on simple data cleaning and comparative analysis to judge, without fully exploring the correlations between data, making it difficult to accurately judge abnormal data, and having low efficiency in screening and analyzing the large amounts of collected data. Without an efficient analysis process, the air monitoring and analysis effect of lidar data is not ideal and there are inaccuracies. Summary of the Invention

[0004] The present invention overcomes the defects of the prior art and provides an air detection and analysis method and system based on lidar.

[0005] In a first aspect of the present invention, there is provided an air detection and analysis method based on lidar, including:

[0006] Performing air detection through lidar technology, and collecting multi-dimensional large data of air quality within a preset area within a predetermined time period;

[0007] Based on different air parameters, setting multiple nodes for the large data of air quality, each node corresponding to an air parameter, and combining with the form of a graph structure, storing the large data of air quality in the form of a graph structure to obtain graph structure large data;

[0008] Using each node data in the graph structure large data as a data point, by introducing a breadth-first search algorithm, setting a preset abnormal judgment condition, traversing each data point and performing a primary evaluation of abnormal data points, and marking the abnormal data points to obtain primary abnormal data points;

[0009] Based on the primary abnormal data points, extracting continuous data within a preset time range for each data point, serializing the extracted data to form serialized data, and performing predictive analysis and secondary evaluation of abnormal data points on the serialized data through an LSTM prediction model to generate an abnormal evaluation result;

[0010] Based on the abnormal evaluation results, process the air quality big data and generate a secondary monitoring plan based on lidar.

[0011] In this solution, air detection is carried out through lidar technology, and multi-dimensional air quality big data in a preset area is collected within a predetermined time period. Specifically:

[0012] Set a predetermined time period and a preset area;

[0013] Conduct real-time air detection and data collection through lidar technology;

[0014] Collect multi-dimensional air quality big data in a preset area within a predetermined time period, and perform data cleaning and standardization preprocessing on the air quality big data.

[0015] In this solution, based on different air parameters, multiple nodes are set for the air quality big data, each node corresponds to an air parameter, and in combination with the graph structure form, the air quality big data is stored in the form of a graph structure to obtain graph structure big data. Specifically:

[0016] Based on the air quality big data, obtain different types of air parameters;

[0017] According to the graph structure storage mode, set multiple nodes for different types of air parameters, and the relationship between each node is represented by an edge. Store the air quality big data in the form of a graph structure to obtain graph structure big data.

[0018] In this solution, in the graph structure big data, using each node data as a data point, by introducing the breadth-first search algorithm, set preset abnormal judgment conditions, traverse each data point and perform a primary evaluation on the abnormal data points, and mark the abnormal data points to obtain the primary abnormal data points. Specifically:

[0019] In the graph structure big data, using each node data as a data point, set the distance between the starting node and the preset node, introduce the breadth-first search algorithm, and traverse each node;

[0020] During the traversal process, based on the current node, obtain multiple associated nodes, and the distance between the associated nodes and the current node is within the preset node distance;

[0021] Extract the data points corresponding to the current node and all associated nodes to form an intermediate data set. In the intermediate data set, use the Z-Score standard score method to determine whether the current node is an abnormal data point. After a primary evaluation, mark the abnormal data points to obtain the primary abnormal data points.

[0022] In this solution, based on the primary abnormal data points, continuous data within a preset time range is extracted for each data point, and the extracted data is serialized to form serialized data. Through the LSTM prediction model, the serialized data is predicted and analyzed, and a secondary evaluation of the abnormal data points is performed to generate an abnormal evaluation result. Specifically:

[0023] Based on the primary abnormal data points, in the air quality big data, continuous data within a preset time range is extracted for each data point to obtain multiple data segments;

[0024] Based on the time dimension, multiple data segments are serialized to form multiple serialized data;

[0025] Multiple serialized data are imported into the LSTM prediction model for data prediction to generate prediction sequence data corresponding to each abnormal data point;

[0026] The prediction sequence data is parsed and a secondary evaluation of the abnormal data points is performed to obtain the abnormal evaluation result.

[0027] In this solution, based on the abnormal evaluation result, the air quality big data is processed and a secondary monitoring scheme based on lidar is generated. Specifically:

[0028] According to the abnormal evaluation result, the primary abnormal data points are updated, the abnormal data points that do not meet the requirements are removed, and the secondary abnormal data points are retained;

[0029] According to the secondary abnormal data points, data retrieval is performed through graph-structured big data, and the corresponding data collection process is optimized and evaluated to generate a secondary monitoring scheme based on lidar.

[0030] In the second aspect of the present invention, an air detection and analysis system based on lidar is also provided. The system includes: a memory and a processor. The memory includes an air detection and analysis program based on lidar. When the air detection and analysis program based on lidar is executed by the processor, the following steps are implemented:

[0031] Air detection is performed through lidar technology, and multi-dimensional air quality big data within a preset area is collected within a predetermined time period;

[0032] Based on different air parameters, multiple nodes are set for the air quality big data, and each node corresponds to an air parameter. In combination with the graph structure form, the air quality big data is stored in the form of a graph structure to obtain graph-structured big data;

[0033] In the graph-structured big data, each node data is used as a data point. By introducing the breadth-first search algorithm and setting preset abnormal judgment conditions, each data point is traversed and evaluated for abnormal data points once, and the abnormal data points are marked to obtain the first-level abnormal data points.

[0034] Based on the first-level abnormal data points, continuous data within a preset time range is extracted for each data point, and the extracted data is serialized to form serialized data. Through the LSTM prediction model, the serialized data is predicted and analyzed and evaluated for abnormal data points a second time to generate an abnormal evaluation result.

[0035] Based on the abnormal evaluation result, the air quality big data is processed to generate a secondary monitoring plan based on lidar.

[0036] The third aspect of the present invention also provides a computer-readable storage medium, which includes an air detection and analysis program based on lidar. When the air detection and analysis program based on lidar is executed by a processor, the steps of the air detection and analysis method based on lidar as described in any one of the above are realized.

[0037] The present invention discloses an air detection and analysis method and system based on lidar. The lidar technology is used for air detection, and multi-dimensional air quality big data in a preset area is collected within a predetermined time period. Multiple nodes are set based on different air parameters, and the data is stored in a graph structure form. The data points are traversed through the breadth-first search algorithm, and the first-level abnormal data points are marked. Subsequently, continuous data is extracted and serialized for each abnormal data point, and the LSTM prediction model is used for prediction analysis and secondary evaluation to generate an abnormal evaluation result. Based on this result, the air quality big data is processed to generate a secondary monitoring plan based on lidar to achieve more accurate and reliable air quality monitoring and management. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 Shows a flowchart of an air detection and analysis method based on lidar according to the present invention;

[0039] Figure 2 Shows a flowchart for obtaining air quality big data in the present invention;

[0040] Figure 3 Shows a block diagram of an air detection and analysis system based on lidar according to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0041] To more clearly understand the above objects, features, and advantages of the present invention, the present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that, without conflict, the embodiments of the present application and the features in the embodiments can be combined with each other.

[0042] In the following description, many specific details are set forth in order to fully understand the present invention. However, the present invention can also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited by the specific embodiments disclosed below.

[0043] Figure 1 The flowchart of a lidar-based air detection and analysis method of the present invention is shown.

[0044] As Figure 1 shown, the first aspect of the present invention provides a lidar-based air detection and analysis method, including:

[0045] S102, performing air detection through lidar technology, and collecting multi-dimensional air quality big data within a preset area within a predetermined time period;

[0046] S104, based on different air parameters, setting multiple nodes for the air quality big data, each node corresponding to an air parameter, and combining with the graph structure form, storing the air quality big data in the form of a graph structure to obtain graph structure big data;

[0047] S106, taking each node data as a data point in the graph structure big data, by introducing the breadth-first search algorithm, setting a preset abnormal judgment condition, traversing each data point and performing a primary evaluation on the abnormal data points, and marking the abnormal data points to obtain primary abnormal data points;

[0048] S108, based on the primary abnormal data points, extracting continuous data within a preset time range for each data point, serializing the extracted data to form serialized data, and performing predictive analysis and secondary evaluation of abnormal data points on the serialized data through an LSTM prediction model to generate an abnormal evaluation result;

[0049] S110, based on the abnormal evaluation result, performing data processing on the air quality big data and generating a secondary monitoring scheme based on lidar.

[0050] It should be noted that the lidar device has the characteristics of high resolution and long-distance measurement, and can realize real-time monitoring and data feedback of the target area. The air quality data packet includes data such as PM2.5, PM10, concentrations of various gaseous pollutants, wind speed, and wind direction.

[0051] Figure 2The flowchart for obtaining air quality big data in the present invention is shown;

[0052] According to an embodiment of the present invention, for air detection by lidar technology, multi-dimensional air quality big data within a preset area is collected within a predetermined time period, specifically:

[0053] S202, set a predetermined time period and a preset area;

[0054] S204, perform real-time air detection and data collection through lidar technology;

[0055] S206, within a predetermined time period, collect multi-dimensional air quality big data within a preset area, and perform data cleaning and standardization preprocessing on the air quality big data.

[0056] It should be noted that the predetermined time period is a relatively long time period, so as to collect and form big data.

[0057] According to an embodiment of the present invention, based on different air parameters, multiple nodes are set for the air quality big data, each node corresponds to an air parameter, and in combination with the graph structure form, the air quality big data is stored in the form of a graph structure to obtain graph structure big data, specifically:

[0058] Based on the air quality big data, different types of air parameters are obtained;

[0059] According to the graph structure storage mode, multiple nodes are set for different types of air parameters, the relationship between each node is represented by an edge, and the air quality big data is stored in the form of a graph structure to obtain graph structure big data.

[0060] It should be noted that the relationship between nodes is the relationship between different air parameters, such as the collection time relationship, the collection data area relationship, the data attribute relationship, etc.

[0061] According to an embodiment of the present invention, in the graph structure big data, each node data is used as a data point, by introducing a breadth-first search algorithm, a preset abnormal judgment condition is set, each data point is traversed and an abnormal data point is evaluated once, and the abnormal data point is marked to obtain a first abnormal data point, specifically:

[0062] In the graph structure big data, each node data is used as a data point, the distance between the starting node and the preset node is set, and a breadth-first search algorithm is introduced to traverse each node;

[0063] During the traversal process, based on the current node, multiple associated nodes are obtained, and the distance between the associated node and the current node is within the preset node distance;

[0064] Extract the data points corresponding to the current node and all associated nodes to form an intermediate data set. In the intermediate data set, use the Z-Score standard score method to determine whether the current node is an outlier data point. After one evaluation, mark the outlier data points to obtain the first-round outlier data points.

[0065] It should be noted that the calculation of the node distance is the shortest distance between two nodes based on the connecting edge in the graph structure.

[0066] According to an embodiment of the present invention, based on the first-round outlier data points, continuously extract the data within a preset time range for each data point, and serialize the extracted data to form serialized data. Through the LSTM prediction model, perform prediction analysis and secondary evaluation of outlier data points on the serialized data to generate an anomaly evaluation result. Specifically:

[0067] Based on the first-round outlier data points, in the air quality big data, continuously extract the data within a preset time range for each data point to obtain multiple data segments;

[0068] Based on the time dimension, serialize multiple data segments to form multiple serialized data;

[0069] Import multiple serialized data into the LSTM prediction model for data prediction to generate the predicted sequence data corresponding to each outlier data point;

[0070] Perform data parsing and secondary evaluation of outlier data points on the predicted sequence data to obtain the anomaly evaluation result.

[0071] It should be noted that the continuous data extraction within the preset time range is to collect the data object according to the collection time corresponding to the outlier data point, and obtain a segment of data before and after the collection time point to form a data segment. Each data segment corresponds to a first-round outlier data point. The secondary evaluation is based on the analysis of the predicted data and the actual air monitoring situation for evaluation, perform secondary outlier judgment and screening on the first-round outlier points, and generate the result after the secondary judgment, thereby reducing the misjudgment probability.

[0072] It is worth mentioning that in the face of a complex urban environment, lidar data may be interfered by various factors, such as building reflections, vehicle emissions, meteorological conditions, equipment use errors, etc. These factors will affect the accuracy and reliability of the data. Traditional technologies often rely on simple data cleaning and comparative analysis for judgment, without fully exploring the correlation between data, making it difficult to achieve accurate judgment of outlier data, and the efficiency of screening and analyzing the collected big data is low, without an efficient analysis process, resulting in an unsatisfactory air monitoring analysis effect of lidar data and inaccurate situations.

[0073] Based on this, the present invention analyzes the air quality big data based on lidar technology, sets graph structure nodes based on air parameters, stores big data in a graph structure, and further introduces a breadth-first search algorithm to traverse the nodes, that is, quickly traverse the big data and analyze the abnormal points once. Subsequently, a secondary abnormality analysis is performed on the sequence data from which the abnormal points are extracted, thereby ensuring the accuracy of the abnormal data analysis, improving the efficiency and accuracy of air quality monitoring, and at the same time improving the analysis accuracy of the lidar data, providing strong data support for subsequent air quality control.

[0074] According to an embodiment of the present invention, based on the abnormal evaluation results, the air quality big data is processed and a secondary monitoring scheme based on a laser radar is generated, specifically:

[0075] According to the abnormal evaluation results, the primary abnormal data points are updated, the abnormal data points that do not meet the requirements are eliminated, and the secondary abnormal data points are retained;

[0076] According to the secondary abnormal data points, data retrieval is performed through graph structure big data, the corresponding data collection process is optimized and evaluated, and a secondary monitoring plan based on lidar is generated.

[0077] It should be noted that the secondary evaluation of abnormal data points is to evaluate the primary abnormal data points and screen out the secondary abnormal data points. The secondary monitoring scheme includes secondary monitoring, deletion or data correction of abnormal data to reduce the impact of abnormal data.

[0078] According to an embodiment of the present invention, it also includes:

[0079] In graph structured big data, obtain the number of nodes corresponding to an abnormal data point;

[0080] Calculate the proportion of abnormal data points based on the number of nodes and the total number of nodes;

[0081] Determine whether the proportion of abnormal data points is greater than the expected value. If so, obtain multiple data segments corresponding to the abnormal data points;

[0082] Sort and serialize a data segment in forward and reverse directions based on the time dimension to obtain sequence data and reverse sequence data;

[0083] The sequence data and the reverse sequence data are respectively imported into the LSTM prediction model for prediction to obtain the first prediction data and the second prediction data;

[0084] The complete prediction data is generated by combining the first prediction data with the second prediction data. The isolation forest model is used to evaluate whether the abnormal data point is an abnormal data point in the complete prediction data. If so, it is marked as a secondary abnormal point, otherwise, the abnormal data point is removed.

[0085] It should be noted that the expected value is a dynamically set value by the user. It is worth mentioning that during the actual abnormal data evaluation process, due to the influence of building reflections, vehicle emissions, meteorological condition changes, equipment errors of lidar, etc., a large number of abnormal points may appear in the data. At this time, it is necessary to make dynamic analysis and adjustment of the abnormal points to ensure the reasonable evaluation of the abnormal points. Therefore, in this solution, based on the actual primary abnormal evaluation, the proportion of abnormal nodes in the graph structure can be introduced for dynamic serialization prediction analysis. When the proportion is greater than the expected value, the abnormal evaluation process is dynamically adjusted, and forward prediction and reverse prediction are combined. Here, the forward sequence data is to predict the future data trend according to the time sequence, and the reverse is to sort the data in reverse based on the time dimension to realize the prediction of past data. Combining the two sets of data to form a complete data set, and judging whether a primary abnormal point is a real abnormality through the Isolation Forest to improve the discrimination accuracy. If the proportion is within the expected value, the abnormal evaluation is carried out based on methods such as the Z-Score standard score method to realize the dynamic regulation evaluation mode and ensure the accurate evaluation efficiency and analysis efficiency of the data collected by the lidar.

[0086] Figure 3 The block diagram of an air detection and analysis system based on lidar according to the present invention is shown.

[0087] The second aspect of the present invention also provides an air detection and analysis system 3 based on lidar. The system includes: a memory 31 and a processor 32. The memory 31 includes an air detection and analysis program based on lidar. When the air detection and analysis program based on lidar is executed by the processor 32, the following steps are realized:

[0088] Perform air detection through lidar technology, and collect multi-dimensional air quality big data in a preset area within a predetermined time period;

[0089] Based on different air parameters, set a variety of nodes for the air quality big data, each node corresponding to an air parameter. Combining with the graph structure form, store the air quality big data in the form of a graph structure to obtain graph structure big data;

[0090] Using each node data in the graph structure big data as a data point, by introducing the breadth-first search algorithm, setting preset abnormal judgment conditions, traversing each data point and performing a primary evaluation of abnormal data points, and marking the abnormal data points to obtain primary abnormal data points;

[0091] Based on the primary abnormal data points, extract continuous data within a preset time range for each data point, and serialize the extracted data to form serialized data. Through the LSTM prediction model, perform prediction analysis and secondary evaluation of abnormal data points on the serialized data to generate an abnormal evaluation result;

[0092] Based on the abnormal evaluation results, process the air quality big data and generate a secondary monitoring plan based on lidar.

[0093] It should be noted that the lidar device has the characteristics of high resolution and long-distance measurement, and can realize real-time monitoring and data feedback of the target area. The air quality data packet includes data such as PM2.5, PM10, concentrations of various gaseous pollutants, wind speed, and wind direction.

[0094] According to an embodiment of the present invention, for air detection by lidar technology, multi-dimensional air quality big data in a preset area is collected within a predetermined time period, specifically:

[0095] Set a predetermined time period and a preset area;

[0096] Perform real-time air detection and data collection by lidar technology;

[0097] Within a predetermined time period, collect multi-dimensional air quality big data in a preset area, and perform data cleaning and standardization preprocessing on the air quality big data.

[0098] It should be noted that the predetermined time period is a relatively long time period to collect and form big data.

[0099] According to an embodiment of the present invention, based on different air parameters, multiple nodes are set for the air quality big data, each node corresponds to an air parameter, and in combination with the graph structure form, the air quality big data is stored in the form of a graph structure to obtain graph structure big data, specifically:

[0100] Based on the air quality big data, obtain different types of air parameters;

[0101] According to the graph structure storage mode, set multiple nodes for different types of air parameters, represent the relationship between each node with an edge, and store the air quality big data in the form of a graph structure to obtain graph structure big data.

[0102] It should be noted that the relationship between nodes is the relationship between different air parameters, such as the time relationship of collection, the regional relationship of collected data, the attribute relationship between data, etc.

[0103] According to an embodiment of the present invention, in the graph structure big data, taking each node data as a data point, by introducing a breadth-first search algorithm, setting preset abnormal judgment conditions, traversing each data point and performing a primary evaluation on abnormal data points, and marking the abnormal data points to obtain primary abnormal data points, specifically:

[0104] In graph-structured big data, each node's data is used as a data point. The starting node and the preset node distance are set, and the breadth-first search algorithm is introduced to traverse each node.

[0105] During the traversal process, based on the current node, multiple associated nodes are obtained, and the distance between the associated nodes and the current node is within the preset node distance.

[0106] The data points corresponding to the current node and all associated nodes are extracted to form an intermediate data set. In the intermediate data set, through the Z-Score standard score method, it is judged whether the current node is an abnormal data point. After one evaluation, the abnormal data points are marked to obtain the first-time abnormal data points.

[0107] It should be noted that the calculation of the node distance is the shortest distance between two nodes based on the connecting edge in the graph structure.

[0108] According to the embodiments of the present invention, based on the first-time abnormal data points, continuous data within a preset time range is extracted for each data point, and the extracted data is serialized to form serialized data. Through the LSTM prediction model, the serialized data is predicted and analyzed and the abnormal data points are evaluated for the second time to generate an abnormal evaluation result. Specifically:

[0109] Based on the first-time abnormal data points, in the air quality big data, continuous data within a preset time range is extracted for each data point to obtain multiple data segments.

[0110] Based on the time dimension, multiple data segments are serialized to form multiple serialized data.

[0111] Multiple serialized data are imported into the LSTM prediction model for data prediction to generate prediction sequence data corresponding to each abnormal data point.

[0112] The prediction sequence data is parsed and the abnormal data points are evaluated for the second time to obtain the abnormal evaluation result.

[0113] It should be noted that the extraction of continuous data within a preset time range means that according to the collection time corresponding to the abnormal data point and the data collection object, a period of data before and after the collection time point is collected to form a data segment. Each data segment corresponds to a first-time abnormal data point. The second evaluation is based on the analysis of the predicted data and the actual air monitoring situation for evaluation, and the first-time abnormal points are judged and screened for the second time to generate the result after the second judgment, thereby reducing the misjudgment probability.

[0114] It is worth mentioning that when facing a complex urban environment, lidar data may be interfered by various factors, such as building reflections, vehicle emissions, meteorological conditions, equipment usage errors, etc. These factors will affect the accuracy and reliability of the data. Traditional technologies often rely on simple data cleaning and comparative analysis judgments, without fully exploring the correlations between data, making it difficult to accurately judge abnormal data, and having low efficiency in screening and analyzing the large amounts of data collected. Without an efficient analysis process, the air monitoring analysis effect of lidar data is not ideal and there are inaccuracies.

[0115] Based on this, the present invention analyzes air quality big data based on lidar technology, sets graph structure nodes based on air parameters, stores the big data in a graph structure, and further traverses the nodes by introducing a breadth-first search algorithm, that is, quickly traverses the big data and performs a one-time analysis of abnormal points. Subsequently, a secondary abnormal analysis is performed on the sequence data of the extracted abnormal points, ensuring the accuracy of abnormal data analysis, improving the efficiency and accuracy of air quality monitoring, and at the same time improving the analysis accuracy of lidar data, providing strong data support for subsequent air quality regulation.

[0116] According to an embodiment of the present invention, based on the abnormal evaluation result, data processing is performed on the air quality big data and a secondary monitoring scheme based on lidar is generated, specifically:

[0117] According to the abnormal evaluation result, the first abnormal data points are updated, the abnormal data points that do not meet the requirements are removed, and the secondary abnormal data points are retained;

[0118] According to the secondary abnormal data points, data retrieval is performed through the graph structure big data, the corresponding data acquisition process is optimized and evaluated, and a secondary monitoring scheme based on lidar is generated.

[0119] It should be noted that the secondary evaluation of the abnormal data points is performed among the first abnormal data points to screen out the secondary abnormal data points. The secondary monitoring scheme includes schemes such as secondary monitoring, deletion, or data correction of abnormal data to reduce the impact of abnormal data.

[0120] According to an embodiment of the present invention, it further includes:

[0121] In the graph structure big data, obtain the number of nodes corresponding to the first abnormal data points;

[0122] Based on the number of nodes and the total number of nodes, calculate the proportion of abnormal data points;

[0123] Judge whether the proportion of abnormal data points is greater than the expected value. If so, obtain multiple data segments corresponding to the first abnormal data points;

[0124] Sort and serialize a data segment in the forward and reverse directions based on the time dimension to obtain sequence data and reverse sequence data;

[0125] Import the sequence data and the reverse sequence data into the LSTM prediction model for prediction respectively to obtain first prediction data and second prediction data;

[0126] Generate complete prediction data by combining the first prediction data and the second prediction data, and use the Isolation Forest model to evaluate whether the abnormal data points are abnormal data points in the complete prediction data. If so, mark them as secondary abnormal points; otherwise, eliminate the abnormal data points.

[0127] It should be noted that the expected value is a user-dynamic setting value. It is worth mentioning that in the actual abnormal data evaluation process, due to the influence of building reflections, vehicle emissions, meteorological condition changes, equipment errors of lidar, etc., a large number of abnormal points may appear in the data. At this time, dynamic analysis and adjustment of the abnormal points are required to ensure a reasonable evaluation of the abnormal points. Therefore, in this solution, based on the actual primary abnormal evaluation, the proportion of abnormal nodes in the graph structure can be introduced for dynamic serialization prediction analysis. When the proportion is greater than the expected value, dynamically adjust the abnormal evaluation process, and combine forward prediction and reverse prediction. Here, the forward sequence data predicts the future data trend in chronological order, and the reverse is to reverse the data based on the time dimension to realize the prediction of past data. Combine the two data to form a complete data set, and use the Isolation Forest to judge whether the primary abnormal points are real abnormalities to improve the discrimination accuracy. If the proportion is within the expected value, perform abnormal evaluation based on methods such as the Z-Score standard score method to realize the dynamic regulation of the evaluation mode and ensure the accurate evaluation efficiency and analysis efficiency of the lidar-collected data.

[0128] The third aspect of the present invention also provides a computer-readable storage medium, which includes an air detection and analysis program based on lidar. When the air detection and analysis program based on lidar is executed by a processor, the steps of the air detection and analysis method based on lidar as described in any one of the above are realized.

[0129] The present invention discloses an air detection and analysis method and system based on lidar, which uses lidar technology for air detection and collects multi-dimensional air quality big data in a preset area within a predetermined time period. Set multiple nodes based on different air parameters and store the data in the form of a graph structure. Traverse the data points through the breadth-first search algorithm to mark the primary abnormal data points. Subsequently, extract and serialize the continuous data for each abnormal data point, use the LSTM prediction model for prediction analysis and secondary evaluation to generate an abnormal evaluation result. Based on this result, process the air quality big data to generate a secondary monitoring scheme based on lidar to achieve more accurate and reliable air quality monitoring and management.

[0130] In several embodiments provided by the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined, or can be integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the displayed or discussed components can be through some interfaces. The indirect coupling or communication connection of devices or units can be electrical, mechanical, or other forms.

[0131] The units described above as separate components may or may not be physically separated. The components displayed as units may or may not be physical units. They can be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0132] In addition, in each embodiment of the present invention, the functional units can all be integrated in a processing unit, or each unit can be separately used as a unit, or two or more units can be integrated in a unit. The above integrated units can be implemented in the form of hardware or in the form of a combination of hardware and software functional units.

[0133] Those of ordinary skill in the art can understand that all or part of the steps of implementing the above method embodiments can be completed by hardware related to program instructions. The foregoing program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps including the above method embodiments. The foregoing storage medium includes various media that can store program codes, such as removable storage devices, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), magnetic disks, or optical discs.

[0134] Alternatively, if the above integrated units of the present invention are implemented in the form of software functional modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the embodiments of the present invention, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the methods described in the various embodiments of the present invention. The foregoing storage medium includes: various media such as removable storage devices, ROM, RAM, magnetic disks, or optical disks that can store program codes.

[0135] As described above, the above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered by the protection scope of the present invention.

Claims

1. An air detection and analysis method based on laser radar, characterized in that: include: Air detection is carried out through LiDAR technology to collect multi-dimensional air quality big data in a preset area within a predetermined time period; Based on different air parameters, the air quality big data is set into multiple nodes, each node corresponds to an air parameter, and combined with the graph structure, the air quality big data is stored in the form of a graph structure to obtain graph structure big data; In the graph structure big data, each node data is used as a data point. By introducing the breadth-first search algorithm, the preset abnormal judgment conditions are set, each data point is traversed and the abnormal data point is evaluated once, and the abnormal data point is marked to obtain an abnormal data point; Based on an abnormal data point, continuous data within a preset time range is extracted for each data point, and the extracted data is serialized to form serialized data. The LSTM prediction model is used to perform predictive analysis on the serialized data and secondary evaluation of the abnormal data points to generate an abnormal evaluation result. Based on the abnormal assessment results, the air quality big data is processed and a secondary monitoring plan based on lidar is generated.

2. The air detection and analysis method based on laser radar according to claim 1 is characterized in that: The air detection is performed by using laser radar technology to collect multi-dimensional air quality big data in a preset area within a predetermined time period, specifically: Set a scheduled time period and preset area; Real-time air detection and data collection through LiDAR technology; Collect multi-dimensional air quality big data in a preset area within a predetermined time period, and perform data cleaning and standardization preprocessing on the air quality big data.

3. The air detection and analysis method based on laser radar according to claim 1 is characterized in that: Based on different air parameters, the air quality big data is set to multiple nodes, each node corresponds to an air parameter, and the air quality big data is stored in the form of a graph structure in combination with the graph structure to obtain the graph structure big data, specifically: Based on air quality big data, obtain different types of air parameters; According to the graph structure storage mode, multiple nodes are set for different types of air parameters, and the relationship between each node is represented by an edge. The air quality big data is stored in the form of a graph structure to obtain graph structure big data.

4. The air detection and analysis method based on laser radar according to claim 3 is characterized in that: In the graph structure big data, each node data is used as a data point, and a breadth-first search algorithm is introduced to set a preset abnormal judgment condition, traverse each data point and evaluate the abnormal data point once, mark the abnormal data point, and obtain an abnormal data point once, specifically: In the graph structure big data, each node data is used as a data point, the distance between the starting node and the preset node is set, and the breadth-first search algorithm is introduced to traverse each node; During the traversal process, based on the current node, multiple associated nodes are obtained, and the distance between the associated nodes and the current node is within the preset node distance; The data points corresponding to the current node and all associated nodes are extracted to form an intermediate data set. In the intermediate data set, the Z-Score standard score method is used to determine whether the current node is an abnormal data point. After an evaluation, the abnormal data point is marked to obtain an abnormal data point.

5. The air detection and analysis method based on laser radar according to claim 4 is characterized in that: Based on the abnormal data point once, continuous data within a preset time range is extracted for each data point, and the extracted data is serialized to form serialized data. The serialized data is predicted and analyzed and the abnormal data point is secondary evaluated through the LSTM prediction model to generate an abnormal evaluation result, which is specifically: Based on an abnormal data point, continuous data within a preset range is extracted for each data point in the air quality big data to obtain multiple data segments; Based on the time dimension, multiple data segments are serialized to form multiple serialized data; Import multiple serialized data into the LSTM prediction model for data prediction, and generate prediction sequence data corresponding to each abnormal data point; The predicted sequence data is subjected to data analysis and secondary evaluation of abnormal data points to obtain abnormal evaluation results.

6. The air detection and analysis method based on laser radar according to claim 5 is characterized in that: Based on the abnormal assessment results, the air quality big data is processed and a secondary monitoring plan based on laser radar is generated, specifically: According to the abnormal evaluation results, the primary abnormal data points are updated, the abnormal data points that do not meet the requirements are eliminated, and the secondary abnormal data points are retained; According to the secondary abnormal data points, data retrieval is performed through graph structure big data, the corresponding data collection process is optimized and evaluated, and a secondary monitoring plan based on lidar is generated.

7. An air detection and analysis system based on laser radar, characterized in that: The system includes: a memory and a processor, wherein the memory includes an air detection and analysis program based on a laser radar, and when the air detection and analysis program based on a laser radar is executed by the processor, the following steps are implemented: Air detection is carried out through LiDAR technology to collect multi-dimensional air quality big data in a preset area within a predetermined time period; Based on different air parameters, the air quality big data is set into multiple nodes, each node corresponds to an air parameter, and combined with the graph structure, the air quality big data is stored in the form of a graph structure to obtain graph structure big data; In the graph structure big data, each node data is used as a data point. By introducing the breadth-first search algorithm, the preset abnormal judgment conditions are set, each data point is traversed and the abnormal data point is evaluated once, and the abnormal data point is marked to obtain an abnormal data point; Based on an abnormal data point, continuous data within a preset time range is extracted for each data point, and the extracted data is serialized to form serialized data. The LSTM prediction model is used to perform predictive analysis on the serialized data and secondary evaluation of the abnormal data points to generate an abnormal evaluation result. Based on the abnormal assessment results, the air quality big data is processed and a secondary monitoring plan based on lidar is generated.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a laser radar-based air detection and analysis program. When the laser radar-based air detection and analysis program is executed by a processor, the steps of the laser radar-based air detection and analysis method as described in any one of claims 1 to 6 are implemented.

Images