Power transmission line point cloud data updating method and system based on dynamic collection of unmanned aerial vehicle

Abstract

The application provides a power transmission line point cloud data updating method and system based on dynamic collection of an unmanned aerial vehicle, relates to the technical field of point cloud data processing, and comprises the following steps: acquiring current point cloud data, flight parameters and environmental parameters; dividing the current point cloud data into multiple sub-point cloud data; obtaining flight parameter features, environmental parameter features and quality evaluation features corresponding to the sub-point cloud data based on the flight parameters and the environmental parameters; generating a confidence degree for comprehensively evaluating the reliability of the sub-point cloud data; and finally updating the original point cloud data in combination with the confidence degree and the sub-point cloud data. Compared with a direct replacement method, the accuracy of the updated point cloud data is greatly improved, and the problem of excessive deviation between the updated point cloud data and the actual situation is avoided.

Description

Technical Field

[0001] This invention relates to the field of point cloud data processing technology, specifically to a method and system for updating point cloud data of power transmission lines based on dynamic collection by unmanned aerial vehicles. Background Technology

[0002] In recent years, with the development of data acquisition technologies such as drones and 3D laser scanning, the cost and time required to acquire 3D laser point cloud data have been effectively reduced. Therefore, 3D laser point cloud technology has been widely used in the monitoring of power transmission lines. For example, the authorization announcement number "CN115100364B" describes a method, device, equipment, and medium for visualizing laser point cloud data of power transmission lines (classification number G06F). This method acquires the point cloud data structure corresponding to the laser point cloud data of the target power transmission line to be displayed. For each point cloud data block in the point cloud data structure, the portion located outside the field of view is then visualized. Alternatively, all point cloud data blocks can be removed from the point cloud data structure. Then, point cloud resource files are loaded based on the target point cloud data in the removed point cloud data structure, and the target point cloud data is rendered to achieve the visualization of laser point cloud data. This technology removes point cloud data blocks located outside the field of view, avoids loading and rendering invisible point cloud data, reduces the amount of point cloud data loaded and rendered, thereby improving the efficiency of point cloud data visualization. It can also reduce the workload of computer graphics cards and processors, thereby reducing the requirements for computer hardware performance and facilitating the realization of point cloud visualization.

[0003] However, the existing technologies mentioned above, when applying 3D point cloud technology to power transmission line monitoring, neglect the technical challenge of updating point cloud data. Power transmission lines are often exposed to harsh external environments, and over time, they are prone to developing hidden dangers. Therefore, updating the point cloud data of power transmission lines is particularly important. However, current point cloud data updates often involve directly replacing the original point cloud with newly acquired point cloud data, neglecting the crucial question of whether the newly acquired point cloud data is reliable. If the newly acquired point cloud data is unreliable, blindly using it to replace the original point cloud will lead to a significant deviation between the updated point cloud data and the actual situation.

[0004] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0005] The purpose of this invention is to provide a method and system for updating power transmission line point cloud data based on dynamic collection by unmanned aerial vehicles (UAVs), so as to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] A method for updating power transmission line point cloud data based on dynamic collection by UAVs, comprising the following steps:

[0008] S1, based on a predetermined flight route, uses a drone to collect current point cloud data of the power transmission line and obtains the drone's flight parameters and environmental parameters in real time during the collection process. The flight parameters include flight coordinates, flight altitude and flight speed, and the environmental parameters include temperature, humidity, wind speed, light intensity and electromagnetic interference power density.

[0009] S2, the current point cloud data is segmented to obtain multiple sub-point cloud data, and the flight parameters and environmental parameters during the acquisition process of the sub-point cloud data are extracted to obtain flight parameter features and environmental parameter features. The flight parameter features include the mean values ​​of flight altitude and flight speed, flight coordinates, deviation values ​​of flight altitude and flight speed, and flight speed fluctuation values. The environmental parameter features include the mean values ​​of temperature, humidity, wind speed, light intensity, and electromagnetic interference power density.

[0010] S3. Construct and train a point cloud quality feature extraction model that takes flight parameter features and environmental parameter features as inputs and outputs quality assessment features as outputs. Input the flight parameter features and environmental parameter features of the sub-point cloud data into the model to obtain the quality assessment features. The quality assessment features include the point cloud density, data coverage, sampling accuracy error and data integrity of the sub-point cloud data.

[0011] S4 generates a confidence score based on flight parameter characteristics, environmental parameter characteristics, quality assessment characteristics, and the time interval between the acquisition of the current point cloud data and the original point cloud data. The confidence score is used to comprehensively evaluate the reliability of the sub-point cloud data.

[0012] S5 updates the original point cloud data based on confidence level and sub-point cloud data.

[0013] Furthermore, the formulas for calculating the average flight altitude and flight speed are as follows:

[0014]

[0015] In the formula, H(i) and V(i) represent the flight altitude and flight speed of the UAV at the i-th acquisition time during the acquisition of sub-point cloud data, respectively. Let i and m represent the average flight altitude and average flight speed of the UAV during the process of collecting sub-point cloud data, respectively. Let i be the index of the collection time of the UAV during the process of collecting sub-point cloud data, and i∈[1,m], where m is the total number of collection times of the UAV during the process of collecting sub-point cloud data.

[0016] The formula for calculating the flight coordinate deviation value is as follows:

[0017]

[0018] In the formula, lon(i) and lat(i) represent the longitude and latitude of the UAV at the i-th acquisition time during the acquisition of sub-point cloud data, respectively. The longitude lon(i) and latitude lat(i) together constitute the UAV's flight coordinates. b (i),lat b (i) represents the longitude and latitude of the position that the UAV should be at according to the predetermined flight path at the i-th collection time during the process of collecting sub-point cloud data, respectively. Δlon(i) and Δlat(i) represent the longitude difference and latitude difference between the position of the UAV at the i-th collection time and the position it should be at during the process of collecting sub-point cloud data, respectively.

[0019] In the formula, dlon(i) and dlat(i) represent the longitude deviation distance and latitude deviation distance between the position of the UAV at the i-th acquisition time and the position it should be during the acquisition of sub-point cloud data, respectively, and Cpc represents the flight coordinate deviation value of the UAV during the acquisition of sub-point cloud data.

[0020] The formula for calculating the flight altitude deviation value is as follows:

[0021]

[0022] In the formula, H(i) and H b (i) represent the flight altitude of the UAV at the i-th acquisition time during the acquisition of sub-point cloud data, and the flight altitude of the UAV at the position it should be according to the predetermined flight route, respectively. Hpc represents the flight altitude deviation value of the UAV during the acquisition of sub-point cloud data.

[0023] The formula for calculating the flight speed deviation is as follows:

[0024]

[0025] In the formula, V(i) and V b (i) represent the flight speed of the UAV at the i-th acquisition time during the acquisition of sub-point cloud data, and the flight speed at the position it should be at according to the predetermined flight route, respectively. Vpc represents the flight speed deviation value of the UAV during the acquisition of sub-point cloud data.

[0026] The formula for calculating the flight speed fluctuation value is as follows:

[0027]

[0028] In the formula, Vbd represents the flight speed fluctuation value of the human-machine interface during the process of collecting sub-point cloud data;

[0029] The formulas for calculating the average values ​​of temperature, humidity, wind speed, light intensity, and electromagnetic interference power density are as follows:

[0030]

[0031] In the formula, T(i), RH(i), V f I(i), E(i), and I(i) represent the temperature, humidity, wind speed, light intensity, and electromagnetic interference power density of the environment at the i-th acquisition time during the acquisition of sub-point cloud data by the UAV, respectively. These represent the average temperature, average humidity, average wind speed, average light intensity, and average electromagnetic interference power density of the environment in which the drone is located during the collection of sub-point cloud data.

[0032] Furthermore, the process of obtaining quality assessment features for sub-point cloud data is as follows:

[0033] 1.1) Obtain sample point cloud data with multiple known quality assessment features, flight parameter features and environmental parameter features, and construct parameter feature vectors for each sample point cloud data based on the flight parameter features and environmental parameter features;

[0034] The method for obtaining the quality assessment features of the sample point cloud data is as follows:

[0035]

[0036] In the formula, D represents the point cloud density, N represents the number of data points in the point cloud data, A represents the area covered by the point cloud data, O represents the data coverage rate, Ltotal is the planned coverage length when collecting point cloud data, Lovelap is the length of the overlapping part in the point cloud data, Δxk, Δyk, and Δzk are the horizontal distance, vertical distance, and depth distance of the k-th reference point and the sampling point in the point cloud data corresponding to the reference point, respectively, k is the index of the reference point, and k∈[1,K], K is the number of reference points, CYwc is the sampling accuracy error, Nth is the expected number of data points to be collected for the point cloud data, and WZpg is the data integrity of the point cloud data;

[0037] 1.2) Construct a point cloud quality feature extraction model based on a deep neural network, with the input being a parameter feature vector and the output being a quality assessment feature. The parameter feature vector of the sample point cloud data is used as the input, and the corresponding quality assessment feature is used as the output label to train the point cloud quality feature extraction model.

[0038] 1.3) The flight parameter features and environmental parameter features of the sub-point cloud data are concatenated to form a parameter feature vector, and then the parameter feature vector is input into the point cloud quality feature extraction model after training to obtain the quality assessment features of the sub-point cloud data.

[0039] Furthermore, the confidence level calculation process includes the following steps:

[0040] 2.1) Based on flight parameter characteristics, a flight state evaluation coefficient is generated to assess the quality of the UAV's flight state when collecting sub-point cloud data. The calculation formula is as follows:

[0041]

[0042] In the formula, FXpg is the flight status evaluation coefficient. The preset flight speed deviation threshold, The preset flight coordinate deviation threshold is ω1, ω2, and ω3, which are the weights of flight altitude, flight speed, and flight coordinate factors in the calculation of flight status evaluation coefficient, respectively, and ω1+ω2+ω3=1, and ω2>ω1=ω3;

[0043] 2.2) Based on environmental parameter characteristics, environmental parameter evaluation coefficients are generated to assess the quality of environmental parameters when the UAV collects sub-point cloud data. The calculation formula is as follows:

[0044]

[0045] In the formula, HJpg is the environmental parameter evaluation coefficient, T*,RH*,Vf*,I*,E* are the suitable working temperature, suitable working humidity, standard wind resistance value, suitable light intensity and standard electromagnetic interference resistance value when the UAV collects point cloud data, respectively, and λ1,λ2,λ3,λ4,λ5 are the weights of temperature, humidity, wind speed, light intensity and electromagnetic interference factors in the calculation of environmental parameter evaluation coefficients, respectively, and the specific values ​​of λ1,λ2,λ3,λ4,λ5 are determined by the analytic hierarchy process.

[0046] 2.3) Based on the quality assessment features, a point cloud quality assessment coefficient is generated to evaluate the quality of the sub-point cloud data collected by the UAV. The calculation formula is as follows:

[0047]

[0048] In the formula, DYpg is the point cloud quality evaluation coefficient, D* and O* are the preset suitable values ​​for point cloud density and data coverage, respectively, and γ1, γ2, γ3, and γ4 are the weights of point cloud density, data coverage, sampling accuracy error, and data integrity in the calculation of the point cloud quality evaluation coefficient, respectively. The specific values ​​of γ1, γ2, γ3, and γ4 are also determined by the analytic hierarchy process.

[0049] 2.4) Based on the flight status evaluation coefficient, environmental parameter evaluation coefficient, quality evaluation coefficient, and the time interval between the acquisition of the current point cloud data and the original point cloud data, a confidence score is generated. The calculation formula is as follows:

[0050]

[0051] In the formula, ZXD is the confidence level of the sub-point cloud data, χ1, χ2, and χ3 are the weights of the flight status evaluation coefficient, environmental parameter evaluation coefficient, and quality evaluation coefficient in the confidence level evaluation calculation, respectively, and χ1+χ2+χ3=1, and χ3>χ1=χ2, t is the acquisition time interval between the current point cloud data and the original point cloud data, and t0 is the preset update time interval threshold.

[0052] Furthermore, the method for updating the original point cloud data is as follows:

[0053] 3.1) Acquire multiple test transmission line areas, and collect point cloud data of the test transmission line areas sequentially based on UAVs. Use the point cloud data obtained in the previous step as the original point cloud data for the test, and use the point cloud data obtained in the next step as the current point cloud data for the test. When collecting the current point cloud data for the test, collect the actual point cloud data of the test transmission line areas on the ground based on ground facilities. Register the original point cloud data, the current point cloud data, and the actual point cloud data for the test based on the iterative nearest point algorithm, and calculate the confidence level of the current point cloud data for the test using the same method.

[0054] 3.2) Construct a point cloud update model based on a convolutional neural network, using the registered original point cloud data for the experiment, the current point cloud data for the experiment and their corresponding confidence levels as inputs, and the registered actual point cloud data for the experiment as outputs, to train the point cloud update model;

[0055] 3.3) Extract the corresponding sub-point cloud data to be updated from the original point cloud data based on the sub-point cloud data, and register the sub-point cloud data and the sub-point cloud data to be updated. Input the registered sub-point cloud data to be updated, the sub-point cloud data and their corresponding confidence scores into the point cloud update model to obtain the updated sub-point cloud data. Input the updated sub-point cloud data into the original point cloud data to replace the sub-point cloud data to be updated.

[0056] A power transmission line point cloud data update system based on UAV dynamic collection, used in the aforementioned power transmission line point cloud data update method based on UAV dynamic collection, includes:

[0057] The data acquisition module uses a drone to collect current point cloud data of the power transmission line based on a predetermined flight route, and acquires the drone's flight parameters and environmental parameters in real time during the acquisition process. The flight parameters include flight coordinates, flight altitude and flight speed, and the environmental parameters include temperature, humidity, wind speed, light intensity and electromagnetic interference power density.

[0058] The feature extraction module is used to segment the current point cloud data to obtain multiple sub-point cloud data, and to extract features from the flight parameters and environmental parameters during the acquisition process of the sub-point cloud data to obtain flight parameter features and environmental parameter features. The flight parameter features include the mean values ​​of flight altitude and flight speed, flight coordinates, deviation values ​​of flight altitude and flight speed, and flight speed fluctuation values. The environmental parameter features include the mean values ​​of temperature, humidity, wind speed, light intensity, and electromagnetic interference power density.

[0059] The feature prediction module is used to build and train a point cloud quality feature extraction model that takes flight parameter features and environmental parameter features as inputs and outputs quality assessment features as outputs. The flight parameter features and environmental parameter features of the sub-point cloud data are input into the model to obtain the quality assessment features, which include the point cloud density, data coverage, sampling accuracy error and data integrity of the sub-point cloud data.

[0060] The confidence calculation module generates confidence scores based on flight parameter characteristics, environmental parameter characteristics, quality assessment characteristics, and the time interval between the acquisition of current point cloud data and original point cloud data. The confidence scores are used to comprehensively evaluate the reliability of sub-point cloud data.

[0061] The point cloud update module updates the original point cloud data based on confidence level and sub-point cloud data.

[0062] Compared with the prior art, the beneficial effects of the present invention are:

[0063] The present invention relates to a method and system for updating power transmission line point cloud data based on dynamic collection by unmanned aerial vehicles (UAVs). After collecting the current point cloud data, the current point cloud data is first divided into multiple sub-point cloud data. Then, the quality of the sub-point cloud data, the flight status of the UAV when collecting the point cloud data, and the environment are comprehensively evaluated to obtain a confidence level reflecting the reliability of the sub-point cloud data. Finally, the original point cloud data is updated based on the confidence level and the sub-point cloud data. Compared with the direct replacement method, this method greatly improves the accuracy of the updated point cloud data and avoids the problem of excessive deviation between the updated point cloud data and the actual situation. Attached Figure Description

[0064] Figure 1 This is a schematic diagram of the overall method flow of the present invention;

[0065] Figure 2 This is a schematic diagram of the overall system modules of the present invention. Detailed Implementation

[0066] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0067] It should be noted that, unless otherwise defined, the technical or scientific terms used in this invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0068] Example:

[0069] Please see Figure 1 This invention provides a method for updating point cloud data of power transmission lines based on dynamic collection by unmanned aerial vehicles (UAVs), the specific steps of which include:

[0070] S1, based on a predetermined flight route, uses a drone to collect current point cloud data of the power transmission line and obtains the drone's flight parameters and environmental parameters in real time during the collection process. The flight parameters include flight coordinates, flight altitude and flight speed, and the environmental parameters include temperature, humidity, wind speed, light intensity and electromagnetic interference power density.

[0071] It should be noted that the specific planning method for the flight route is as follows: based on the layout of the power transmission lines to be measured, the flight route is designed using UAV flight route planning software (such as Mission Planner, Pix4D, etc.), and the designed flight route is ensured to cover all areas of all power transmission lines. This is existing technology and will not be elaborated here.

[0072] Among them, the drone collects the current point cloud data of the power transmission line through its own LiDAR, and the flight coordinates, flight altitude and flight speed can be obtained by the drone's built-in GPS module and inertial measurement unit (IMU).

[0073] Among them, the environmental parameters are the environmental parameters of the environment in which the UAV is located during flight. Temperature and humidity are obtained by digital temperature and humidity sensors (such as DHT22, SHT31, etc.) installed on the UAV. Wind speed can be obtained by insertion anemometers (such as thermal anemometers, ultrasonic anemometers, etc.) installed on the UAV. Light intensity can be obtained by light sensors (such as BH1750, TSL2561, etc.) installed on the UAV. Electromagnetic interference power density can be obtained by electromagnetic field detectors installed on the UAV. The units of temperature, humidity, wind speed, light intensity and electromagnetic interference power density are degrees Celsius, %, meters per second, Lux, and milliwatts per square meter, respectively.

[0074] It should be noted that electromagnetic interference power density is generally obtained indirectly through an electromagnetic field detector. Specifically, this involves first measuring the magnetic field strength of the environment in which the UAV is located using the electromagnetic field detector, and then multiplying the square of the magnetic field strength by the characteristic resistance of free space to obtain the electromagnetic interference power density. The characteristic resistance of free space is generally taken as 377 ohms, which is existing technology and will not be elaborated here.

[0075] In step S1, flight parameters and environmental parameters are collected at the same collection frequency to facilitate subsequent data analysis. The interval between adjacent collection times can be 3 seconds, 5 seconds, 10 seconds, etc., and there is no restriction here. Of course, after collecting flight parameters and environmental parameters at different collection frequencies, the timestamp can also be unified by data processing methods such as interpolation. The specific settings are determined by the staff according to the actual situation and there is no restriction here.

[0076] S2, the current point cloud data is segmented to obtain multiple sub-point cloud data, and the flight parameters and environmental parameters during the acquisition process of the sub-point cloud data are extracted to obtain flight parameter features and environmental parameter features. The flight parameter features include the mean values ​​of flight altitude and flight speed, flight coordinates, deviation values ​​of flight altitude and flight speed, and flight speed fluctuation values. The environmental parameter features include the mean values ​​of temperature, humidity, wind speed, light intensity, and electromagnetic interference power density.

[0077] It should be noted that the data can be segmented according to the collection time or the collection length. The specific choice should be made by the staff based on the actual situation. There is no restriction here. The segmentation method facilitates subsequent point cloud data update operations. Compared with the method of updating the point cloud data as a whole, segmenting the point cloud data first and then updating each sub-point cloud data with the corresponding original point cloud data reduces the workload of a single update and can achieve the effect of quickly updating point cloud data.

[0078] The method for segmenting based on the acquisition time is as follows: A reasonable time interval is set in advance (such as every 5 minutes, every 10 minutes, etc.). After the point cloud data acquisition is completed, all point cloud data are traversed, and the timestamp of each data point is recorded. The minimum timestamp is determined as the first timestamp. The timestamp that is one time interval away from the minimum timestamp is taken as the second timestamp. The timestamp that is two time interval away from the minimum timestamp is taken as the third timestamp, and so on. Data points whose timestamps are between the first and second timestamps are extracted to form the first sub-point cloud data. Data points whose timestamps are between the second and third timestamps are extracted to form the second sub-point cloud data, and so on to complete the segmentation.

[0079] The method for segmenting based on the collection length is as follows: A reasonable length interval is set in advance (such as every 100 meters, every 200 meters, etc.). After the point cloud data collection is completed, one end is selected as the starting point and the other end as the ending point. The extension direction is from the starting point to the ending point. The point that is one length interval away from the starting point along the extension direction is taken as the first segment point. The point that is one length interval away from the first segment point along the extension direction is taken as the second segment point. And so on. Data points located between the starting point and the first segment point are extracted to form the first sub-point cloud data. Data points located between the first segment point and the second segment point are extracted to form the second sub-point cloud data. The segmentation is completed in this way.

[0080] The formulas for calculating the average flight altitude and flight speed are as follows:

[0081]

[0082] In the formula, H(i) and V(i) represent the flight altitude and flight speed of the UAV at the i-th acquisition time during the acquisition of sub-point cloud data, respectively. Let i and m represent the average flight altitude and average flight speed of the UAV during the process of collecting sub-point cloud data, respectively. Let i be the index of the collection time of the UAV during the process of collecting sub-point cloud data, and i∈[1,m], where m is the total number of collection times of the UAV during the process of collecting sub-point cloud data.

[0083] The formula for calculating the flight coordinate deviation is as follows:

[0084]

[0085] In the formula, lon(i) and lat(i) represent the longitude and latitude of the UAV at the i-th acquisition time during the acquisition of sub-point cloud data, respectively. The longitude lon(i) and latitude lat(i) together constitute the UAV's flight coordinates. b (i),lat b(i) represents the longitude and latitude of the position that the UAV should be at according to the predetermined flight path at the i-th collection time during the process of collecting sub-point cloud data, respectively. Δlon(i) and Δlat(i) represent the longitude difference and latitude difference between the position of the UAV at the i-th collection time and the position it should be at during the process of collecting sub-point cloud data, respectively.

[0086] In the formula, dlon(i) and dlat(i) represent the longitude deviation distance and latitude deviation distance (in meters) between the location of the UAV at the i-th acquisition time and the expected location during the acquisition of sub-point cloud data. The above formula for converting longitude difference and latitude difference into longitude deviation distance and latitude deviation distance is existing technology and will not be explained here.

[0087] In the formula, CPC represents the flight coordinate deviation value of the UAV during the process of collecting sub-point cloud data. It is used to characterize the severity of the UAV's flight coordinate deviation from the predetermined flight path during the process of collecting sub-point cloud data. The larger the longitude deviation distance and latitude deviation distance, the higher the degree of deviation of the UAV from the predetermined flight path during the process of collecting sub-point cloud data, the larger the flight coordinate deviation value, and the worse the reliability of the sub-point cloud data collected by the UAV.

[0088] The formula for calculating the flight altitude deviation is as follows:

[0089]

[0090] In the formula, H(i) and H b (i) represent the flight altitude of the UAV at the i-th acquisition time during the acquisition of sub-point cloud data, and the flight altitude of the UAV according to the predetermined flight path, respectively. Hpc represents the flight altitude deviation value of the UAV during the acquisition of sub-point cloud data, which is used to characterize the severity of the UAV's flight altitude deviating from the predetermined flight path during the acquisition of sub-point cloud data. H(i) and H b The larger the absolute difference (i), the greater the degree to which the UAV deviates from the predetermined flight path during the collection of sub-point cloud data, the greater the flight altitude deviation value, and the worse the reliability of the sub-point cloud data collected by the UAV.

[0091] The formula for calculating the flight speed deviation is as follows:

[0092]

[0093] In the formula, V(i) and V b(i) represents the flight speed of the UAV at the i-th acquisition time during the acquisition of sub-point cloud data, and the flight speed at the position it should be at according to the predetermined flight path, respectively. Vpc represents the flight speed deviation value of the UAV during the acquisition of sub-point cloud data, which is used to characterize the severity of the UAV's flight speed deviating from the predetermined flight path during the acquisition of sub-point cloud data. V(i) and V b The larger the absolute difference (i), the greater the degree to which the UAV deviates from the predetermined flight path during the collection of sub-point cloud data, the greater the flight speed deviation, and the worse the reliability of the sub-point cloud data collected by the UAV.

[0094] The formula for calculating the flight speed fluctuation value is as follows:

[0095]

[0096] In the formula, Vbd represents the flight speed fluctuation value of the drone during the process of collecting sub-point cloud data. It is used to characterize the fluctuation of the drone's flight speed during the process of collecting sub-point cloud data. The larger the flight speed fluctuation value, the more unstable the drone's flight speed is during the process of collecting sub-point cloud data, and the greater the adverse impact on the drone's process of collecting sub-point cloud data, which makes the reliability of the sub-point cloud data collected by the drone worse.

[0097] The formulas for calculating the average values ​​of temperature, humidity, wind speed, light intensity, and electromagnetic interference power density are as follows:

[0098]

[0099] In the formula, T(i), RH(i), V f I(i), E(i), and I(i) represent the temperature, humidity, wind speed, light intensity, and electromagnetic interference power density of the environment at the i-th acquisition time during the acquisition of sub-point cloud data by the UAV, respectively. These represent the average temperature, average humidity, average wind speed, average light intensity, and average electromagnetic interference power density of the environment in which the drone is located during the collection of sub-point cloud data.

[0100] S3. Construct and train a point cloud quality feature extraction model that takes flight parameter features and environmental parameter features as inputs and outputs quality assessment features as outputs. Input the flight parameter features and environmental parameter features of the sub-point cloud data into the model to obtain the quality assessment features. The quality assessment features include the point cloud density, data coverage, sampling accuracy error and data integrity of the sub-point cloud data.

[0101] The process of constructing and training the point cloud quality feature extraction model is as follows:

[0102] 1.1) Obtain sample point cloud data with multiple known quality assessment features, flight parameter features, and environmental parameter features. Construct parameter feature vectors for each sample point cloud data based on the flight parameter features and environmental parameter features. The parameter feature vectors are represented as follows:

[0103]

[0104] In the formula, CSxl represents the parameter feature vector, which represents the state of the point cloud data collected by the UAV from two dimensions: flight parameters and environmental parameters. This facilitates the subsequent prediction and analysis of the quality assessment features of the sample point cloud data through these parameter features.

[0105] It should be noted that the sample point cloud data are point cloud data from other transmission lines. Because they are all transmission lines, they share similarities, making the constructed point cloud quality feature extraction model more targeted. This improves the accuracy of the subsequent point cloud quality feature extraction model in predicting the quality assessment features of the sub-point cloud data. The methods for obtaining the flight parameter features and environmental parameter features corresponding to the sample point cloud data are the same as described above and will not be repeated here. The method for obtaining the quality assessment features is as follows:

[0106]

[0107] In the formula, D represents the point cloud density, N represents the number of data points in the point cloud data, which can be directly recorded by the sensor during the point cloud data acquisition process, and A represents the area covered by the point cloud data, which can be obtained by point cloud data processing software such as pix4D and QGIS. The higher the point cloud density, the more data points there are per unit area.

[0108] In the formula, O represents the data coverage, Ltotal is the planned coverage length (i.e., the total length of the flight path) when collecting point cloud data, and Lovelap is the length of the overlapping part in the point cloud data, which can be obtained by point cloud data processing software such as Pix4D and Agisoft Metashap. The larger the data coverage, the more duplicate data points there are in the point cloud data.

[0109] In the formula, Δxk, Δyk, and Δzk represent the horizontal, vertical, and depth distances of the k-th reference point and the corresponding sampling point in the point cloud data, respectively. k is the index of the reference point, and k∈[1,K], where K is the number of reference points. The selection method for Δxk, Δyk, and Δzk is as follows: K reference points are randomly selected in the real area corresponding to the point cloud data, and their position coordinates are obtained using high-precision ground measurement data. Then, the position coordinate information of the reference point in the point cloud data is determined based on point cloud data processing software such as Pix4D and Agisoft Metashap. The two types of position coordinates are then transformed into the real three-dimensional coordinate system to obtain Δxk, Δyk, and Δzk. The setting is used to convert the absolute error into a relative error relative to the entire point cloud data coverage length, so as to more intuitively reflect the sampling accuracy error. CYwc is the sampling accuracy error.

[0110] In the formula, Nth represents the expected number of data points to be collected corresponding to the point cloud data. It can be planned and determined before the point cloud data is collected, or the product of the average collection frequency and the collection time during the point cloud data collection process of the UAV can be used as the expected number of data points to be collected. WZpg represents the data integrity of the point cloud data. The larger the value, the more the point cloud data collection meets the expectations and the more complete the collected point cloud data is.

[0111] 1.2) A point cloud quality feature extraction model is constructed based on a deep neural network, taking the parameter feature vector as input and the quality assessment feature as output. The parameter feature vector of the sample point cloud data is used as input, and the corresponding quality assessment feature is used as the output label to train the point cloud quality feature extraction model. The specific training process is as follows:

[0112] Each quality assessment feature is labeled onto its corresponding parameter feature vector. The labeled parameter feature vectors are then combined into a dataset, which is randomly divided into a training set and a test set in an 80:20 ratio. The point cloud quality feature extraction model is trained using the training set, with mean squared error as the loss function. The model parameters are iteratively optimized by minimizing the difference between the model's predicted and actual values. Cross-validation is used to select the optimal model hyperparameters to avoid overfitting. After a predetermined number of iterations, the processing time prediction model is tested using the test set. If the test requirements are met, the training is considered complete; otherwise, training is repeated until the test requirements are met. Specific test requirements can be set to an accuracy of the model's predicted values ​​of no less than 90%, 92%, 95%, etc., without any restrictions.

[0113] 1.3) The flight parameter features and environmental parameter features of the sub-point cloud data are concatenated to form a parameter feature vector, and then the parameter feature vector is input into the point cloud quality feature extraction model after training to obtain the quality assessment features of the sub-point cloud data.

[0114] It should be noted that, in order to improve the convergence speed and performance of the point cloud quality feature extraction model, the data belonging to the same class in the dataset and the sub-point cloud data (such as the point cloud density of the sub-point cloud data and the point cloud density of the sample point cloud data in the dataset belonging to the same class) can be subjected to max-min normalization to unify the dimensions of various types of data, thereby improving the convergence speed and performance of the subsequent point cloud quality feature extraction model.

[0115] S4, based on flight parameter characteristics, environmental parameter characteristics, quality assessment characteristics, and the acquisition time interval between the current point cloud data and the original point cloud data, generates a confidence score. The confidence score is used to comprehensively evaluate the reliability of the sub-point cloud data, including the following steps:

[0116] 2.1) Based on flight parameter characteristics, a flight state evaluation coefficient is generated to assess the quality of the UAV's flight state when collecting sub-point cloud data. The calculation formula is as follows:

[0117]

[0118] In the formula, FXpg is the flight status evaluation coefficient, which combines the mean of flight altitude and flight speed, flight coordinates, deviation of flight altitude and flight speed, and flight speed fluctuation value to comprehensively evaluate the flight status of the UAV when collecting sub-point cloud data from three levels: flight altitude factor, flight coordinate factor, and flight speed factor. The larger the flight status evaluation index, the worse the flight status of the UAV when collecting sub-point cloud data, which means that the reliability of the sub-point cloud data collected by the UAV is lower.

[0119] It should be noted that a larger flight altitude deviation value indicates a greater deviation between the drone's flight altitude and the planned flight altitude when collecting sub-point cloud data, thus indicating a greater degree of deviation from the flight path during drone flight. The ratio... The setting is used to evaluate the relative deviation of flight altitude using the average flight altitude as the evaluation benchmark, and the ratio. The larger the ratio, the greater the relative deviation in flight altitude of the drone when collecting sub-point cloud data, and the worse the reliability of the collected sub-point cloud data. Therefore, the ratio is... As a component in calculating the flight status evaluation coefficient;

[0120] It should be noted that excessively low flight speeds during UAV data collection result in a large number of unnecessary duplicate data points, increasing the difficulty of subsequent data processing. Conversely, excessively high flight speeds lead to too few data points, hindering subsequent analysis of power line conditions based on point cloud data. Both excessively high and low flight speeds increase flight speed deviation, thereby reducing the reliability of the UAV-collected sub-point cloud data. Furthermore, the reliability of the sub-point cloud data decreases rapidly as the flight speed deviation increases. Therefore, flight speed deviation is considered a crucial factor in UAV flight status assessment, as shown in the formula. A preset flight speed deviation threshold is set by staff based on actual conditions. For example, it can be set to 10% to 20% of the average flight speed along a predetermined flight path. When the flight speed deviation exceeds this threshold, it indicates that the flight speed is extremely unreasonable. If the deviation continues to increase, the reliability of the sub-point cloud data collected by the drone drops rapidly. Therefore, a [further action is required]. The nonlinear impact of flight speed deviation on flight status is characterized by a certain form. As the flight speed deviation increases, the unreliability of the sub-point cloud data collected by the UAV gradually increases. Furthermore, once the flight speed deviation exceeds the flight speed deviation threshold, further increases in flight speed deviation will lead to further unreliability. The rapid increase reflects the phenomenon that the reliability of sub-point cloud data drops rapidly due to excessive deviation in flight speed.

[0121] Furthermore, when analyzing flight speed factors to assess the flight status of a drone, flight speed stability is also an important consideration. Stable flight speed helps the drone collect high-quality sub-point cloud data, while unstable flight speed leads to a decrease in the reliability of the collected sub-point cloud data. The ratio is used to reflect the relative volatility of sub-point cloud data collected by drones. The larger the value, the worse the stability of the drone when collecting sub-point cloud data, and the lower the reliability of the collected sub-point cloud data. Therefore, we introduce... To Corrections are made to comprehensively evaluate flight speed factors when assessing the flight status of drones;

[0122] It should be noted that a larger flight coordinate deviation value indicates a greater deviation between the UAV's flight coordinates and the planned flight coordinates of the predetermined flight path when collecting sub-point cloud data. This means the UAV deviates significantly from its flight path during flight, resulting in lower reliability of the collected sub-point cloud data. This is a preset flight coordinate deviation threshold, the value of which is set by the staff according to the actual situation. It can be set to a fixed value of 1 meter, or it can be set to between one-thousandth and one-hundredth of the length of the area where the sub-point cloud data is located. Its function is to provide a reference for the flight coordinate deviation value, so that the staff can quantify the degree of deviation represented by the flight coordinate deviation value, and the ratio. The settings are used to evaluate the relative deviation of flight coordinates based on a flight coordinate deviation threshold, and the ratio. The larger the value, the greater the relative deviation of the drone's flight coordinates when collecting sub-point cloud data. Therefore, the ratio is... As a component in calculating the flight status evaluation coefficient;

[0123] In the formula, ω1, ω2, and ω3 represent the weights of flight altitude, flight speed, and flight coordinates in the calculation of the flight status evaluation coefficient, respectively. Since the quality of sub-point cloud data acquisition largely depends on flight speed, flight speed is the most sensitive influencing factor. Flight speed also largely determines flight altitude and flight coordinate deviations. Therefore, flight speed is given the largest weight. Flight altitude and flight coordinate deviations reflect the distance the UAV deviates from the predetermined flight path. Therefore, flight altitude and flight coordinates are given the same weight. Based on ω1 + ω2 + ω3 = 1, let ω2 > ω1 = ω3. The specific values ​​of the three factors can be adjusted and used by the staff according to the actual situation. No further restrictions are imposed here.

[0124] 2.2) Based on environmental parameter characteristics, environmental parameter evaluation coefficients are generated to assess the quality of environmental parameters when the UAV collects sub-point cloud data. The calculation formula is as follows:

[0125]

[0126] In the formula, HJpg is the environmental parameter evaluation coefficient, which comprehensively evaluates the environment in which the UAV collects sub-point cloud data from five aspects: temperature, humidity, wind speed, light intensity and electromagnetic interference power density. The larger the environmental parameter evaluation coefficient, the worse the environment in which the UAV collects sub-point cloud data, which means that the reliability of the sub-point cloud data collected by the UAV is lower.

[0127] In the formula, T*, RH*, Vf*, I*, and E* represent the suitable operating temperature, suitable operating humidity, standard wind resistance value, suitable light intensity, and standard electromagnetic interference resistance value when the UAV collects point cloud data, respectively.

[0128] It should be noted that suitable operating temperature, suitable operating humidity, and suitable light intensity are all environmental parameters suitable for drones to collect point cloud data. Under these conditions, drones can collect point cloud data most accurately. The suitable operating temperature is between 20 and 25 degrees Celsius, the suitable operating humidity is between 30% and 50%, and the suitable light intensity (q) is between 4000 and 7000 lux. The specific values ​​should be set by the staff according to the actual situation.

[0129] It's important to note that the standard wind resistance value indicates that the drone can maintain good stability and ensure good data collection accuracy when the wind speed is below this value. However, when the wind speed exceeds this value, the drone's stability will decrease significantly due to the increased wind speed, leading to a sharp decline in the reliability of the collected data. The specific value of the standard wind resistance value is determined according to the type of drone. For example, for small consumer drones, the standard wind resistance value is 5 meters per second, meaning that at wind speeds below 5 meters per second, the small consumer drone can still maintain good stability and ensure good data collection accuracy. However, when the wind speed exceeds 5 meters per second, the stability of the small consumer drone will decrease significantly due to the increased wind speed. Increased wind speed and a significant decrease in wind speed can cause a rapid decline in the reliability of the collected data. Similarly, for medium-sized commercial drones, the standard wind resistance value is 9 meters per second, and for large industrial drones, the standard wind resistance value is 12 meters per second. The standard electromagnetic interference resistance value indicates that when the electromagnetic interference power density is below this value, the drone can still maintain good anti-interference capabilities to ensure good data collection accuracy. However, when the electromagnetic interference power density exceeds this value, the stability of the drone will decrease significantly due to the increased wind speed, causing a rapid decline in the reliability of the collected data. The standard electromagnetic interference resistance value can be between 0.5 and 1.5 milliwatts per square meter.

[0130] It should be noted that, This value is used to characterize the deviation of the average temperature from the suitable operating temperature. The larger the value, the worse the temperature factor, which is less conducive to UAVs collecting sub-point cloud data and results in lower reliability of the collected sub-point cloud data. This value is used to characterize the deviation of the mean humidity from the suitable working humidity. The larger the value, the worse the humidity factor, which is less conducive to UAVs collecting sub-point cloud data, and the lower the reliability of the collected sub-point cloud data. It is used to characterize the deviation of the average light intensity from the suitable light intensity. The larger the value, the worse the light intensity factor is, the less conducive it is to the UAV to collect sub-point cloud data, and the worse the reliability of the collected sub-point cloud data.

[0131] It should be noted that, This is used to characterize the quality of wind speed in the environment where the drone collects sub-point cloud data. The higher the average wind speed, the worse the wind speed in the environment where the drone collects sub-point cloud data, and the less reliable the collected sub-point cloud data is. The formation of this metric is used to characterize the nonlinear impact of wind speed on the accuracy of UAV data acquisition. It indicates that as wind speed increases, the adverse effect of wind speed on the accuracy of UAV data acquisition initially increases slowly until the average wind speed exceeds the standard wind resistance value. After that, as the average wind speed increases further, the accuracy of UAV data acquisition rapidly declines, leading to a rapid decrease in the reliability of the acquired sub-point cloud data. Similarly... This is used to characterize the quality of electromagnetic interference (EMI) power density in the environment where the UAV is collecting sub-point cloud data. The higher the average EMI power density, the worse the EMI power density in the environment where the UAV is collecting sub-point cloud data, and the less reliable the collected sub-point cloud data is. The formation of this metric is used to characterize the nonlinear effect of electromagnetic interference power density on the acquisition accuracy of UAVs. It indicates that as the electromagnetic interference power density increases, the adverse effect of electromagnetic interference on the acquisition accuracy of UAVs first increases slowly, until the average electromagnetic interference power density is greater than the standard electromagnetic interference resistance value. Then, as the average electromagnetic interference power density increases, the acquisition accuracy of UAVs drops rapidly, and the reliability of the acquired sub-point cloud data drops rapidly.

[0132] In the formula, λ1, λ2, λ3, λ4, and λ5 represent the weights of temperature, humidity, wind speed, light intensity, and electromagnetic interference in the calculation of environmental parameter evaluation coefficients, respectively. The specific values ​​of λ1, λ2, λ3, λ4, and λ5 are determined by the analytic hierarchy process (AHP), with the following logic:

[0133] Five factors—temperature, humidity, wind speed, light intensity, and electromagnetic interference—are labeled. The relative importance of each pair of factors is determined using the nine-scale method, and a judgment matrix is ​​constructed. The index of temperature is labeled 1, humidity 2, wind speed 3, light intensity 4, and electromagnetic interference 5. The constructed judgment matrix is ​​as follows:

[0134]

[0135] Where f and v both represent the indices of the evaluation values, and f∈[1,5], v∈[1,5], qfv represents the importance of the evaluation value with index f to the environmental parameter evaluation coefficient relative to the evaluation value with index v. The specific value of qfv is determined by relevant experts using a 1-9 scoring method. qfv=9 means that the evaluation value with index f is extremely important to the environmental parameter evaluation coefficient relative to the evaluation value with index v, and qfv=1 means that the evaluation value with index f is extremely unimportant to the environmental parameter evaluation coefficient relative to the evaluation value with index v.

[0136] Divide each element value in the judgment matrix by the sum of its columns to obtain a normalized judgment matrix. Calculate the mean of the element values ​​in each row of the normalized judgment matrix. Use the mean of the first row as the scaling factor for the temperature factor, the mean of the second row as the scaling factor for the humidity factor, the wind speed factor as the scaling factor for the corrected nose difference assessment value, the mean of the fourth row as the scaling factor for the light intensity factor, and the mean of the fifth row as the scaling factor for the electromagnetic interference factor. With the constraint that the sum of the scaled values ​​equals 1, scale the five scaling factors proportionally and use the scaled values ​​as the weights of the corresponding factors.

[0137] 2.3) Based on the quality assessment features, a point cloud quality assessment coefficient is generated to evaluate the quality of the sub-point cloud data collected by the UAV. The calculation formula is as follows:

[0138]

[0139] In the formula, DYpg is the point cloud quality evaluation coefficient, which evaluates the quality of the collected sub-point cloud data from four aspects: point cloud density, data coverage, sampling accuracy error and data integrity. The larger the point cloud quality evaluation coefficient, the worse the quality of the collected sub-point cloud data, that is, the worse the reliability of the sub-point cloud data.

[0140] In the formula, D* and O* are the preset suitable values ​​for point cloud density and data coverage, respectively. The suitable value for point cloud density is the point cloud density that is suitable for subsequent point cloud data processing. Its value is determined according to the exploration accuracy of the transmission line. For example, the suitable value for point cloud density can be 10 points per square meter for low-precision exploration, 20-50 points per square meter for medium-precision exploration, and 100 points per square meter for low-precision exploration. The specific value is set according to the actual situation. The suitable value for data coverage is generally between 20% and 30% to balance the amount of data points collected and the connectivity of data points.

[0141] It should be noted that, This value is used to characterize the deviation of point cloud density from its optimal value. A larger value indicates poorer quality of the collected sub-point cloud data. Similarly... This is used to characterize the degree of deviation between the data coverage rate and the appropriate value of the data coverage rate. The larger the value, the worse the quality of the collected sub-point cloud data. Similarly, the worse the sampling accuracy error, the worse the quality of the sub-point cloud data. On the other hand, the higher the data integrity, the more complete the sub-point cloud data and the better the quality of the sub-point cloud data. Therefore, a negative sign is used to characterize the negative correlation between data integrity and point cloud quality evaluation coefficient.

[0142] In the formula, γ1, γ2, γ3, and γ4 are the weights of point cloud density, data coverage, sampling accuracy error, and data integrity in the calculation of point cloud quality evaluation coefficient, respectively. The specific values ​​of γ1, γ2, γ3, and γ4 are also determined by the analytic hierarchy process, which will not be elaborated here.

[0143] 2.4) Based on the flight status evaluation coefficient, environmental parameter evaluation coefficient, quality evaluation coefficient, and the time interval between the acquisition of the current point cloud data and the original point cloud data, a confidence score is generated. The calculation formula is as follows:

[0144]

[0145] In the formula, ZXD is the confidence level of the sub-point cloud data. It combines the flight status evaluation coefficient, environmental parameter evaluation coefficient, quality evaluation coefficient, and the time interval between the current point cloud data and the original point cloud data to comprehensively evaluate the credibility of the sub-point cloud data. The higher the confidence level, the worse the credibility of the sub-point cloud data.

[0146] It should be noted that the flight status evaluation coefficient, environmental parameter evaluation coefficient, and quality evaluation coefficient assess the reliability of sub-point cloud data from different levels. Therefore, the confidence score is generated by weighted summation of the three coefficients and taking the reciprocal. The larger the three evaluation coefficients are, the worse the reliability of the sub-point cloud data is, and the lower the confidence score is. χ1, χ2, and χ3 are the weights of the flight status evaluation coefficient, environmental parameter evaluation coefficient, and quality evaluation coefficient in the confidence score calculation, respectively. Since the quality evaluation coefficient directly reflects the quality of the sub-point cloud data, it is given the largest weight. The flight status evaluation coefficient and environmental parameter evaluation coefficient are both indirectly evaluated from the external environment to predict the quality of the sub-point cloud data, so they are given the same weight. Therefore, based on χ1 + χ2 + χ3 = 1, we let χ3 > χ1 = χ2. The specific values ​​of the three coefficients are adjusted and used by the staff according to the actual situation, and no further restrictions are imposed here.

[0147] It should be noted that the original point cloud data is the point cloud data currently in use (i.e., the point cloud data to be updated). The transmission line itself will change continuously with the external environment, so it is necessary to update the original point cloud data in a timely manner. As the time without updating increases, the necessity and urgency of updating the original point cloud data also increase. Therefore, the collection time interval between the current point cloud data and the original point cloud data is introduced to correct the main part of the confidence score. In the formula, t is the collection time interval between the current point cloud data and the original point cloud data, and t0 is the preset update time interval threshold. The larger the ratio of t to t0, the higher the urgency of updating the original point cloud data. Therefore, the ratio of t to t0 is introduced to correct the main part of the confidence score, so that the final confidence score takes into account the urgency of updating. The specific value of the update time interval threshold is set by the staff according to the actual situation, such as 1 day, 3 days, etc., and is not restricted here.

[0148] S5 updates the original point cloud data based on confidence level and sub-point cloud data. The specific update method is as follows:

[0149] 3.1) Acquire multiple test transmission line areas, and collect point cloud data of the test transmission line areas sequentially based on UAVs. Use the point cloud data obtained in the previous step as the original point cloud data for the test, and use the point cloud data obtained in the next step as the current point cloud data for the test. When collecting the current point cloud data for the test, collect the actual point cloud data of the test transmission line areas on the ground based on ground facilities. Register the original point cloud data, the current point cloud data, and the actual point cloud data for the test based on the iterative nearest point algorithm, and calculate the confidence level of the current point cloud data for the test using the same method.

[0150] It should be noted that a high-precision ground laser scanner (TLS) was selected for the ground facilities to collect actual point cloud data of the test transmission line area. Compared with UAVs flying in the air, the ground laser scanner has a more stable acquisition environment, which ensures the authenticity and accuracy of the point cloud data. It is also equipped with a high-precision GPS receiver and IMU (inertial measurement unit) to improve the spatial positioning accuracy of the point cloud data. In the acquisition process, a step-by-step scanning and multi-angle acquisition method is adopted to ensure that the actual point cloud data of the test is accurate data that closely matches the real situation.

[0151] 3.2) A point cloud update model is constructed based on a convolutional neural network. The registered original point cloud data for the experiment, the current point cloud data for the experiment, and their corresponding confidence scores are used as inputs, and the registered actual point cloud data for the experiment is used as outputs to train the point cloud update model. The specific training process is as follows:

[0152] The registered original point cloud data for the experiment, the current point cloud data for the experiment, and their corresponding confidence scores are used to form the input features. The registered actual point cloud data for the experiment are then labeled onto the corresponding input features. The labeled input features are combined into a dataset, which is then randomly divided into a training set and a test set in an 80:20 ratio. The point cloud update model is trained using the training set, with mean squared error as the loss function. The model parameters are iteratively optimized by minimizing the difference between the model's predicted values ​​and the actual values. Cross-validation is used to select the optimal model hyperparameters to avoid overfitting. After reaching a predetermined number of iterations, the processing time prediction model is tested using the test set. If the test requirements are met, the training is considered complete; otherwise, the training is repeated until the test requirements are met. Specific test requirements can be set to the accuracy of the model's predicted values ​​being no less than 80%, 85%, 90%, etc., without any restrictions.

[0153] 3.3) Extract the corresponding sub-point cloud data to be updated from the original point cloud data based on the sub-point cloud data, and register the sub-point cloud data and the sub-point cloud data to be updated. Input the registered sub-point cloud data to be updated, the sub-point cloud data and their corresponding confidence scores into the point cloud update model to obtain the updated sub-point cloud data. Input the updated sub-point cloud data into the original point cloud data to replace the sub-point cloud data to be updated.

[0154] Example 2:

[0155] Please see Figure 2 This invention provides a transmission line point cloud data update system based on UAV dynamic collection, used in the transmission line point cloud data update method based on UAV dynamic collection in Embodiment 1 above, comprising:

[0156] The data acquisition module uses a drone to collect current point cloud data of the power transmission line based on a predetermined flight route, and acquires the drone's flight parameters and environmental parameters in real time during the acquisition process. The flight parameters include flight coordinates, flight altitude and flight speed, and the environmental parameters include temperature, humidity, wind speed, light intensity and electromagnetic interference power density.

[0157] The feature extraction module is used to segment the current point cloud data to obtain multiple sub-point cloud data, and to extract features from the flight parameters and environmental parameters during the acquisition process of the sub-point cloud data to obtain flight parameter features and environmental parameter features. The flight parameter features include the mean values ​​of flight altitude and flight speed, flight coordinates, deviation values ​​of flight altitude and flight speed, and flight speed fluctuation values. The environmental parameter features include the mean values ​​of temperature, humidity, wind speed, light intensity, and electromagnetic interference power density.

[0158] The feature prediction module is used to build and train a point cloud quality feature extraction model that takes flight parameter features and environmental parameter features as inputs and outputs quality assessment features as outputs. The flight parameter features and environmental parameter features of the sub-point cloud data are input into the model to obtain the quality assessment features, which include the point cloud density, data coverage, sampling accuracy error and data integrity of the sub-point cloud data.

[0159] The confidence calculation module generates confidence scores based on flight parameter characteristics, environmental parameter characteristics, quality assessment characteristics, and the time interval between the acquisition of current point cloud data and original point cloud data. The confidence scores are used to comprehensively evaluate the reliability of sub-point cloud data.

[0160] The point cloud update module updates the original point cloud data based on confidence level and sub-point cloud data.

[0161] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0162] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented in software, the above embodiments can be implemented, in whole or in part, as a computer program product. Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution.

[0163] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0164] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A method for updating point cloud data of transmission lines based on dynamic collection by unmanned aerial vehicles, characterized in that, The specific steps include: S1, based on a predetermined flight route, uses a drone to collect current point cloud data of the power transmission line and obtains the drone's flight parameters and environmental parameters in real time during the collection process. The flight parameters include flight coordinates, flight altitude and flight speed, and the environmental parameters include temperature, humidity, wind speed, light intensity and electromagnetic interference power density. S2, the current point cloud data is segmented to obtain multiple sub-point cloud data, and the flight parameters and environmental parameters during the acquisition process of the sub-point cloud data are extracted to obtain flight parameter features and environmental parameter features. The flight parameter features include the average flight altitude, average flight speed, flight coordinate deviation, flight altitude deviation, flight speed deviation, and flight speed fluctuation. The environmental parameter features include the average temperature, average humidity, average wind speed, average light intensity, and average electromagnetic interference power density. S3. Construct and train a point cloud quality feature extraction model that takes flight parameter features and environmental parameter features as inputs and outputs quality assessment features as outputs. Input the flight parameter features and environmental parameter features of the sub-point cloud data into the model to obtain the quality assessment features. The quality assessment features include the point cloud density, data coverage, sampling accuracy error and data integrity of the sub-point cloud data. S4 generates a confidence score based on flight parameter characteristics, environmental parameter characteristics, quality assessment characteristics, and the time interval between the acquisition of the current point cloud data and the original point cloud data. The confidence score is used to comprehensively evaluate the reliability of the sub-point cloud data. S5 updates the original point cloud data based on confidence level and sub-point cloud data.

2. The method for updating transmission line point cloud data based on UAV dynamic collection according to claim 1, characterized in that: The formulas for calculating the average flight altitude and average flight speed are as follows: In the formula, , Let $\alpha$ and $\alpha$ represent the flight altitude and speed of the UAV at the $i$-th data acquisition moment during the acquisition of sub-point cloud data, respectively. , Let represent the average flight altitude and average flight speed of the UAV during the sub-point cloud data acquisition process, respectively, and let i be the index of the acquisition time during the sub-point cloud data acquisition process. m is the total number of data collection moments during the process of collecting sub-point cloud data by the UAV; The formula for calculating the flight coordinate deviation value is as follows: In the formula, , These represent the longitude and latitude of the UAV's location at the i-th data collection moment during the process of collecting sub-point cloud data. and latitude Together they form the flight coordinates of the drone. , These represent the longitude and latitude of the location the UAV should be at according to the predetermined flight path at the i-th acquisition time during the acquisition of sub-point cloud data. , These represent the difference in longitude and latitude between the location of the UAV at the i-th acquisition moment and its expected location during the acquisition of sub-point cloud data, respectively. In the formula, , These represent the longitude and latitude deviations between the current position of the UAV at the i-th acquisition moment and its expected position during the acquisition of sub-point cloud data, respectively. This represents the flight coordinate deviation value of the UAV during the process of collecting sub-point cloud data; The formula for calculating the flight altitude deviation value is as follows: In the formula, , Let $\alpha$ and $\alpha$ represent the flight altitude of the UAV at the i-th data acquisition moment during the acquisition of sub-point cloud data, respectively, and the flight altitude of the UAV according to the predetermined flight path. This indicates the flight altitude deviation value of the drone during the process of collecting sub-point cloud data; The formula for calculating the flight speed deviation is as follows: In the formula, , Let $\frac{i}{i}$ represent the flight speed of the UAV at the i-th data acquisition moment during the acquisition of sub-point cloud data, and $\frac{i}{i}$ represent the flight speed of the UAV at the position it should be at according to the predetermined flight path. This indicates the deviation in flight speed of the drone during the process of collecting sub-point cloud data; The formula for calculating the flight speed fluctuation value is as follows: In the formula, This represents the fluctuation value of the flight speed of the human-machine interface during the process of collecting sub-point cloud data; The formulas for calculating the average temperature, average humidity, average wind speed, average light intensity, and average electromagnetic interference power density are as follows: In the formula, , , , , These represent the temperature, humidity, wind speed, light intensity, and electromagnetic interference power density of the environment in which the UAV is located at the i-th acquisition moment during the acquisition of sub-point cloud data, respectively. , , , , These represent the average temperature, average humidity, average wind speed, average light intensity, and average electromagnetic interference power density of the environment in which the drone is located during the collection of sub-point cloud data.

3. The method for updating transmission line point cloud data based on UAV dynamic collection according to claim 2, characterized in that: The process of obtaining quality assessment features for sub-point cloud data is as follows: 1.1) Obtain sample point cloud data with multiple known quality assessment features, flight parameter features and environmental parameter features, and construct parameter feature vectors for each sample point cloud data based on the flight parameter features and environmental parameter features; The method for obtaining the quality assessment features of the sample point cloud data is as follows: In the formula, D represents the point cloud density, N represents the number of data points in the point cloud data, A represents the area covered by the point cloud data, and O represents the data coverage rate. The planned coverage length for point cloud data acquisition. The length of the overlapping portion in the point cloud data. , , Let be the horizontal distance, vertical distance, and depth distance of the k-th reference point and the corresponding sampling point in the point cloud data, respectively, where k is the index of the reference point. K is the number of reference points. To account for sampling accuracy error, This represents the estimated number of data points to be collected corresponding to the point cloud data. To ensure the data integrity of point cloud data; 1.2) Construct a point cloud quality feature extraction model based on a deep neural network, with the input being a parameter feature vector and the output being a quality assessment feature. The parameter feature vector of the sample point cloud data is used as the input, and the corresponding quality assessment feature is used as the output label to train the point cloud quality feature extraction model. 1.3) The flight parameter features and environmental parameter features of the sub-point cloud data are concatenated to form a parameter feature vector, and then the parameter feature vector is input into the point cloud quality feature extraction model after training to obtain the quality assessment features of the sub-point cloud data.

4. The method for updating transmission line point cloud data based on UAV dynamic collection according to claim 3, characterized in that: The confidence level calculation process includes the following steps: 2.1) Based on flight parameter characteristics, a flight state evaluation coefficient is generated to assess the quality of the UAV's flight state when collecting sub-point cloud data. The calculation formula is as follows: In the formula, For flight status assessment coefficient, The preset flight speed deviation threshold, The preset flight coordinate deviation threshold, , , These represent the weights of flight altitude, flight speed, and flight coordinates in the calculation of the flight status evaluation coefficient, respectively. + + =1, and ; 2.2) Based on environmental parameter characteristics, environmental parameter evaluation coefficients are generated to assess the quality of environmental parameters when the UAV collects sub-point cloud data. The calculation formula is as follows: In the formula, For environmental parameter evaluation coefficients, These are the suitable operating temperature, suitable operating humidity, standard wind resistance value, suitable light intensity, and standard electromagnetic interference resistance value for drones collecting point cloud data. These represent the weights of temperature, humidity, wind speed, light intensity, and electromagnetic interference in the calculation of environmental parameter evaluation coefficients, respectively. The specific value is determined by the analytic hierarchy process (AHP). 2.3) Based on the quality assessment features, a point cloud quality assessment coefficient is generated to evaluate the quality of the sub-point cloud data collected by the UAV. The calculation formula is as follows: In the formula, This is the point cloud quality evaluation coefficient. These are the preset suitable values ​​for point cloud density and data coverage, respectively. These represent the weights of point cloud density, data coverage, sampling accuracy error, and data integrity in the calculation of the point cloud quality evaluation coefficient, respectively. The specific value is also determined by the analytic hierarchy process (AHP). 2.4) Based on the flight status evaluation coefficient, environmental parameter evaluation coefficient, quality evaluation coefficient, and the time interval between the acquisition of the current point cloud data and the original point cloud data, a confidence score is generated. The calculation formula is as follows: In the formula, ZXD represents the confidence level of the sub-point cloud data. These represent the weights of the flight status assessment coefficient, environmental parameter assessment coefficient, and quality assessment coefficient in the confidence assessment calculation, respectively. ,and t is the time interval between the current point cloud data and the original point cloud data, and t0 is the preset update time interval threshold.

5. The method for updating transmission line point cloud data based on UAV dynamic collection according to claim 1, characterized in that: The method for updating the original point cloud data is as follows: 3.1) Acquire multiple test transmission line areas, and collect point cloud data of the test transmission line areas sequentially based on UAVs. Use the point cloud data obtained in the previous step as the original point cloud data for the test, and use the point cloud data obtained in the next step as the current point cloud data for the test. When collecting the current point cloud data for the test, collect the actual point cloud data of the test transmission line areas on the ground based on ground facilities. Use the iterative nearest point algorithm to register the original point cloud data, the current point cloud data, and the actual point cloud data for the test. Calculate the confidence level of the current point cloud data for the test using the same method. 3.2) Construct a point cloud update model based on a convolutional neural network, using the registered original point cloud data for the experiment, the current point cloud data for the experiment and their corresponding confidence levels as inputs, and the registered actual point cloud data for the experiment as outputs, to train the point cloud update model; 3.3) Extract the corresponding sub-point cloud data to be updated from the original point cloud data based on the sub-point cloud data, and register the sub-point cloud data and the sub-point cloud data to be updated. Input the registered sub-point cloud data to be updated, the sub-point cloud data and their corresponding confidence scores into the point cloud update model to obtain the updated sub-point cloud data. Input the updated sub-point cloud data into the original point cloud data to replace the sub-point cloud data to be updated.

6. A transmission line point cloud data update system based on UAV dynamic collection, used in the transmission line point cloud data update method based on UAV dynamic collection as described in any one of claims 1-5, characterized in that, include: The data acquisition module uses a drone to collect current point cloud data of the power transmission line based on a predetermined flight route, and acquires the drone's flight parameters and environmental parameters in real time during the acquisition process. The flight parameters include flight coordinates, flight altitude and flight speed, and the environmental parameters include temperature, humidity, wind speed, light intensity and electromagnetic interference power density. The feature extraction module is used to segment the current point cloud data to obtain multiple sub-point cloud data, and to extract features from the flight parameters and environmental parameters during the acquisition process of the sub-point cloud data to obtain flight parameter features and environmental parameter features. The flight parameter features include the average flight altitude, average flight speed, flight coordinate deviation, flight altitude deviation, flight speed deviation, and flight speed fluctuation. The environmental parameter features include the average temperature, average humidity, average wind speed, average light intensity, and average electromagnetic interference power density. The feature prediction module is used to build and train a point cloud quality feature extraction model that takes flight parameter features and environmental parameter features as inputs and outputs quality assessment features as outputs. The flight parameter features and environmental parameter features of the sub-point cloud data are input into the model to obtain the quality assessment features, which include the point cloud density, data coverage, sampling accuracy error and data integrity of the sub-point cloud data. The confidence calculation module generates confidence scores based on flight parameter characteristics, environmental parameter characteristics, quality assessment characteristics, and the time interval between the acquisition of current point cloud data and original point cloud data. The confidence scores are used to comprehensively evaluate the reliability of sub-point cloud data. The point cloud update module updates the original point cloud data based on confidence level and sub-point cloud data.

Images