Dangerous source space measuring system and measuring method thereof

Abstract

The invention discloses a dangerous source space measurement system and method, and relates to the technical field of power transmission channel safety monitoring, and the system comprises a point cloud collection module, a monocular image collection module, a point cloud processing module, a joint calibration module, a binocular measurement subsystem, a space measurement module, a data fusion module, and a control module. According to the dangerous source space measurement system and the measurement method thereof, a point cloud processing link is combined with scene feature optimization algorithms such as wire linearity and construction machinery volume, interference filtering is more accurate, and the system and the method are different from a general point cloud processing technology; the calibration module takes known parameters of the lead as a reference object, and calibration objects do not need to be additionally deployed, so that the field scene calibration problem is solved; dynamic distortion correction and multi-scale lead center line extraction are newly added in binocular measurement, the anti-interference capability in the complex environment is improved, and the limitation of existing fixed parameter correction is broken through, so that the measurement reliability and adaptability in the field complex environment are improved, and a special calibration object does not need to be additionally deployed.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power transmission channel safety monitoring, in particular to a dangerous source space measurement system and a measurement method thereof. BACKGROUND

[0002] The power transmission line is an important part of the power system, and the mechanical construction such as excavator and crane operation in the channel is one of the main reasons leading to line breakage accidents. Accurate measurement of the minimum space distance between the dangerous source such as construction machinery and the conductor is the core prerequisite for realizing the power transmission channel anti-breakage early warning.

[0003] The existing dangerous source space measurement scheme mainly includes point cloud and monocular image measurement, but it mainly uses general point cloud processing algorithms such as direct filtering and clustering, and does not optimize the specific geometric features of the conductor, tower and construction machinery in the power transmission channel, resulting in incomplete filtering of interference point clouds such as trees and birds, low target classification accuracy, and dependence on general calibration objects such as checkerboards in the calibration process. In the field power transmission channel scene, the calibration object is difficult to deploy and is easily disturbed by the environment, resulting in low accuracy of the point cloud-image mapping relationship, and ultimately affecting the measurement reliability. SUMMARY

[0004] In view of the deficiencies of the prior art, the present application provides a dangerous source space measurement system and a measurement method thereof, which solves the problems raised in the background art.

[0005] To achieve the above purpose, the present application is realized by the following technical scheme: a dangerous source space measurement system, comprising:

[0006] A point cloud acquisition module is used to acquire three-dimensional laser point cloud data of the power transmission channel, and the point cloud data contains three-dimensional space information of the conductor, tower and construction machinery type dangerous source;

[0007] A monocular image acquisition module is used to acquire real-time monocular images of the power transmission channel, and the acquisition area of the monocular image coincides with that of the point cloud acquisition module;

[0008] A point cloud processing module is used to filter and classify the three-dimensional laser point cloud data;

[0009] A joint calibration module is used to establish the mapping relationship between the point cloud coordinate system and the monocular image coordinate system;

[0010] A binocular measurement subsystem comprising two symmetrically deployed binocular cameras, a dynamic distortion correction module and a conductor center line extraction module;

[0011] A space measurement module is used to calculate the minimum space distance between the dangerous source and the conductor based on the point cloud-monocular data and the binocular data respectively;

[0012] The data fusion module is configured to perform confidence-weighted fusion on the point cloud-monocular measurement result and the binocular measurement result, and the confidence is dynamically adjusted according to the intensity of ambient light, wherein the confidence weight of the binocular measurement is 0.6-0.8 when the light intensity is greater than or equal to 500 lux, and the confidence weight of the point cloud-monocular measurement is 0.7-0.9 when the light intensity is less than 500 lux;

[0013] The control module is configured to control the cooperative work of the modules, receive the measurement result of the data fusion module, and trigger a warning signal if the measurement result is less than a preset warning threshold, wherein the warning threshold of the 110kV line is 5m, the warning threshold of the 220kV line is 8m, and the warning threshold of the 500kV line is 12m.

[0014] Further, the point cloud processing module filters out the point cloud disturbed by trees and birds by using an improved RANSAC algorithm based on the target geometric features of the power transmission channel, and classifies the effective point cloud into conductor point cloud, tower point cloud and hazard source point cloud, wherein the geometric features are linear features of the conductor and volume features of construction machinery.

[0015] The joint calibration module takes a conductor segment with known physical parameters in the power transmission channel as a calibration reference, combines the three-dimensional coordinates of the conductor point cloud with the pixel coordinates of the conductor in the monocular image, and solves the internal and external parameter matrices of the monocular camera and the point cloud-image projection transformation matrix.

[0016] Further, the dynamic distortion correction module adaptively adjusts the distortion correction coefficient according to the light intensity and contrast parameters of the images collected by the binocular camera in real time, and the conductor center line extraction module extracts the pixel coordinates of the conductor center line after removing the tower and tree disturbance edges by using multi-scale Canny edge detection combined with least squares curve fitting.

[0017] Further, the space measurement module projects the pixel coordinates of the hazard source in the monocular image to the point cloud coordinate system through the projection transformation matrix based on the point cloud-monocular data, combines the conductor space function fitted by the conductor point cloud, and calculates the minimum distance; based on the binocular data, the three-dimensional reconstruction is performed on the conductor center line corresponding point and the hazard source point, and the minimum distance is calculated after fitting the conductor space function.

[0018] Further, the filtering process of the point cloud processing module includes:

[0019] Step A1: voxel grid down-sampling is performed on the original point cloud to reduce the data volume;

[0020] Step A2: based on the linear features of the conductor, a straight line is fitted by using an improved RANSAC algorithm, conductor candidate point cloud is screened, and points with a distance threshold of less than or equal to 0.05m from the straight line are set as conductor candidate points.

[0021] The specific process of screening the conductor candidate point cloud is as follows:

[0022] A set of point clouds is defined as where represents the three-dimensional coordinates of the i-th point in the point cloud;

[0023] Randomly extract n points from the set of point clouds For line fitting, denoted as , a line is fitted through these n points The equation of the line is expressed in parametric form:

[0024]

[0025] where is a point on the line, such as , is the direction vector of the line, determined by and ;

[0026] For each point in the set of point clouds , calculate its distance to the line using the point-to-line distance formula:

[0027]

[0028] where , , , denotes the cross product of vectors, denotes the modulus of the vector, and a distance threshold is set. If , the point is considered an inlier, and the set of inliers is denoted as ;

[0029] Count the number of inliers in the set of inliers , and if is greater than the number of inliers of the current optimal model, record the current line and the set of inliers as the optimal model; Repeat the above process for iterations, where

[0030] is determined by:

[0031] ​​​​​

[0032] wherein is the probability of obtaining the correct model, is the proportion of inliers in the point cloud, and the inlier set obtained after the iteration process is the selected conductor candidate point cloud;

[0033] Step A3: screening the hazard source candidate point cloud based on the volume characteristics of the construction machinery;

[0034] Step A4: removing the interference point cloud outside the conductor candidate point cloud, the hazard source candidate point cloud and the tower point cloud.

[0035] Further, the calibration process of the joint calibration module includes:

[0036] Step B1: selecting a known span , of a conductor segment with a known diameter , as a calibration reference;

[0037] Step B2: obtaining the three-dimensional point cloud of the conductor segment through the point cloud acquisition module, and fitting the spatial straight line equation of the conductor:

[0038]

[0039] wherein represents the three-dimensional coordinates of a point in space, which is a variable and changes with the position of the point;

[0040] represents the three-dimensional coordinates of a known fixed reference point in space, which is a determined value and is used as the starting reference for calculation;

[0041] are three components of the direction vector, which together determine the direction of a straight line, wherein the direction vector is ;

[0042] Step B3: obtaining the monocular image containing the conductor segment through the monocular image acquisition module, and extracting the pixel straight line equation of the conductor segment:

[0043]

[0044] wherein, is the slope of the straight line, is the independent variable, is the dependent variable, and the intercept of the straight line on the axis;

[0045] Step B4: Establishing the three-dimensional coordinates of the wire point cloud , and substituting the corresponding relationship with the pixel coordinates into the camera imaging model, combining the physical constraints of the wire diameter , solving the camera intrinsic matrix and the extrinsic matrix , and then obtaining the point cloud-image projection transformation matrix .

[0046] Further, the extraction process of the wire center line extraction module includes:

[0047] Step C1: Gray-scale and Gaussian filter preprocessing are performed on the left and right views collected by the binocular camera;

[0048] Step C2: Multi-scale Canny edge detection is used to obtain the image edges;

[0049] Step C3: Based on the rectangular features of the tower and the irregular edge features of the trees, morphological operation is used to remove non-wire edges;

[0050] Step C4: Least squares curve fitting is performed on the remaining wire edge points to obtain the wire center line pixel coordinates.

[0051] A dangerous source space measurement method applied to the above-mentioned dangerous source space measurement system measurement system, comprising the following steps:

[0052] S1, system deployment:

[0053] Deploy the point cloud acquisition module, monocular image acquisition module and binocular measurement subsystem to ensure that the acquisition areas of the three coincide, complete the calibration through the joint calibration module, and set the early warning threshold of the transmission line with different voltage levels;

[0054] S2, data acquisition:

[0055] The point cloud acquisition module acquires real-time three-dimensional point cloud data of the transmission channel, the monocular image acquisition module acquires real-time monocular images, and the binocular measurement subsystem acquires real-time left and right views;

[0056] S3, point cloud processing and point cloud-monocular measurement:

[0057] The point cloud processing module filters and classifies the three-dimensional point cloud to obtain the wire point cloud and the dangerous source point cloud; the space measurement module projects the dangerous source pixel coordinates in the monocular image to the point cloud coordinate system based on the projection transformation matrix, and calculates the first minimum distance D1 between the dangerous source and the wire after fitting the wire space function.

[0058] S4, binocular measurement:

[0059] The dynamic distortion correction module performs adaptive distortion correction on the left and right binocular views; the wire center line extraction module extracts wire center line pixel coordinates; the spatial measurement module performs three-dimensional reconstruction on the wire center line corresponding points and the dangerous source points, and after fitting the wire spatial function, the second minimum distance D2 between the dangerous source and the wire is calculated;

[0060] S5, data fusion:

[0061] The data fusion module determines the confidence weight of D1 and D2 according to the real-time environmental light intensity, and calculates the fused minimum distance D = w1 x D1 + w2 x D2, wherein (w1 + w2 = 1), w1 is the confidence weight of D1, and w2 is the confidence weight of D2;

[0062] S6, early warning judgment: the control module compares D with the preset early warning threshold, and if D < early warning threshold, the early warning operation of triggering the scene shouting, the warning light flashing and sending information to the operation and maintenance personnel is triggered.

[0063] Further, in the step S3, the fitting process of the wire spatial function includes: using a cubic polynomial to fit the wire point cloud to obtain a wire spatial curve equation: x = f(z), y = g(z), wherein z is a vertical direction coordinate; and the calculation process of the first minimum distance D1 includes: traversing each point (xq, yq, zq) in the dangerous source point cloud, calculating the distance from the point to the wire spatial curve, and taking the minimum value as D1.

[0064] Further, in the step S4, the three-dimensional reconstruction process includes: based on the internal and external parameter matrices of the binocular camera, performing stereo matching on the wire center line corresponding points and the dangerous source points in the left and right views, and calculating the three-dimensional coordinates of the corresponding points by using the triangulation method; the fitting process of the wire spatial function is consistent with that in the step S3, and the second minimum distance D2 is the minimum distance from the dangerous source three-dimensional coordinates to the wire spatial curve.

[0065] In the step S5, the determination rule of the confidence weight is: when the environmental light intensity I is greater than or equal to 500 lux, w1 = 0.2-0.4 and w2 = 0.6-0.8; when 300 lux < I < 500 lux, w1 = 0.5 and w2 = 0.5; and when I < 300 lux, w1 = 0.7-0.9 and w2 = 0.1-0.3.

[0066] The present application provides a dangerous source spatial measurement system and a measurement method thereof, which have the following beneficial effects:

[0067] 1. The hazard source space measurement system and its measurement method, the point cloud processing link combines the scene characteristics such as the linear of the conductor and the volume of the construction machinery to optimize the algorithm, filters out the interference more accurately, and is different from the general point cloud processing technology; the calibration module takes the known parameters of the conductor itself as the reference, does not need to deploy additional calibration objects, and solves the problem of field scene calibration; the binocular measurement adds dynamic distortion correction and multi-scale conductor center line extraction, improves the anti-interference ability in complex environment, and breaks through the limitation of the existing fixed parameter correction; a data fusion module is further added, the measurement weight is adjusted dynamically combined with light, so that the measurement reliability and adaptability in the complex environment are improved, the minimum space distance between the construction machinery type hazard source and the conductor can be accurately measured; without additional deployment of special calibration objects, the equipment can run independently without the background, and can trigger multi-mode early warning in time for different voltage grade transmission lines, and meets the monitoring needs of the mechanical external damage of the transmission channel. BRIEF DESCRIPTION OF DRAWINGS

[0068] Figure 1 It is a step flow diagram of the hazard source space measurement method of the application. DETAILED DESCRIPTION

[0069] The embodiments of the application will be further described in detail below in combination with the drawings and examples. The following examples are used to illustrate the application, but cannot be used to limit the scope of the application.

[0070] As shown in the drawings, Figure 1 The application provides a technical scheme: a hazard source space measurement system, comprising:

[0071] A point cloud acquisition module is used to acquire three-dimensional laser point cloud data of the transmission channel, and the point cloud data contains three-dimensional space information of the conductor, the tower and the construction machinery type hazard source;

[0072] A monocular image acquisition module is used to acquire monocular images of the transmission channel in real time, and the acquisition area of the monocular image coincides with that of the point cloud acquisition module;

[0073] A point cloud processing module is used to filter and classify the three-dimensional laser point cloud data, and the module adopts an improved RANSAC algorithm to filter out the interference point cloud such as trees and birds based on the target geometric characteristics of the transmission channel, and classifies the effective point cloud into conductor point cloud, tower point cloud and hazard source point cloud, wherein the geometric characteristics are the linear characteristics of the conductor and the volume characteristics of the construction machinery;

[0074] The filtering process of the point cloud processing module includes:

[0075] Step A1: voxel grid downsampling is performed on the original point cloud to reduce the data amount;

[0076] Step A2: Based on the linear characteristics of the traverse, the improved RANSAC algorithm is used to fit the straight line, and candidate point clouds of the traverse are screened. Points with a distance of ≤0.05m from the straight line are set as candidate points of the traverse.

[0077] The specific process for selecting candidate point clouds for traverse lines is as follows:

[0078] Define the point cloud set as ,in Represents the first point in the point cloud The three-dimensional coordinates of each point;

[0079] From point cloud collection Randomly selected from For a point, and for line fitting, , recorded as Through this A straight line is fitted from the points. ,straight line The equation can be expressed in parametric form:

[0080]

[0081] in For example, a point on a straight line , Let be the direction vector of the line, by and Sure;

[0082] For point cloud collections Each point in Calculate its distance to the line distance Using the formula for the distance from a point to a line:

[0083]

[0084] in , , , Represents the cross product of vectors. The magnitude of the vector is defined by a distance threshold. ,if Then the point Consider them as interior points, and denote the set of interior points as... ;

[0085] Statistical set of interior points Number of interior points ,like If the number of interior points is greater than the current optimal model, then the current line... and interior point set Recorded as the optimal model;

[0086] Repeat the above process to perform The next iteration, in which Determined by the following formula:

[0087]

[0088] in It represents the probability of obtaining the correct model. It represents the proportion of interior points in the point cloud. After the above iterative process, the final set of interior points is obtained. This refers to the selected candidate point cloud for the conductor.

[0089] Step A3: Based on the volumetric characteristics of construction machinery, screen candidate point clouds of hazard sources;

[0090] Step A4: Remove interfering point clouds other than the candidate point clouds of conductors, candidate point clouds of hazards, and point clouds of towers;

[0091] The joint calibration module is used to establish the mapping relationship between the point cloud coordinate system and the monocular image coordinate system. The module uses the conductor segment with known physical parameters in the power transmission channel as the calibration reference. Combining the three-dimensional coordinates of the conductor point cloud and the pixel coordinates of the conductor in the monocular image, the module solves the intrinsic and extrinsic parameter matrix of the monocular camera and the point cloud-image projective transformation matrix.

[0092] The calibration process of the joint calibration module includes:

[0093] Step B1: Select a known span within the transmission channel. , Given diameter , The conductor segment is used as a calibration reference.

[0094] Step B2: Obtain the 3D point cloud of the traverse segment using the point cloud acquisition module, and fit the spatial straight line equation of the traverse. :

[0095]

[0096] in The three-dimensional coordinates of a point in space are variables that change as the point's position changes;

[0097] The three-dimensional coordinates of a known fixed reference point in space are definite values ​​used as the starting reference for calculations.

[0098] Let be the three components of the direction vector, which together determine the direction of a straight line, where the direction vector is . ;

[0099] Step B3: Obtain a monocular image containing the conductor segment through a monocular image acquisition module, and extract a pixel straight line equation of the conductor segment :

[0100]

[0101] wherein, is a slope of the straight line, is an independent variable, is a dependent variable, an intercept of the straight line on an axis;

[0102] Step B4: Establish a three-dimensional coordinate of the conductor point cloud , and a corresponding relationship of the pixel coordinate is substituted into a camera imaging model, combined with a physical constraint of a conductor diameter , an internal parameter matrix and an external parameter matrix of the camera are solved, and a point cloud-image projection transformation matrix is obtained;

[0103] A binocular measurement system includes two symmetrically deployed binocular cameras, a dynamic distortion correction module and a conductor center line extraction module, the dynamic distortion correction module adaptively adjusts a distortion correction coefficient according to light intensity and contrast parameters of images collected by the binocular cameras in real time, and the conductor center line extraction module extracts pixel coordinates of a conductor center line by using multi-scale Canny edge detection combined with least square curve fitting after removing interference edges of a tower, a tree and the like.

[0104] The extraction process of the conductor center line extraction module includes:

[0105] Step C1: Perform grayscale and Gaussian filter pretreatment on left and right views collected by the binocular cameras;

[0106] Step C2: Obtain image edges by using multi-scale Canny edge detection;

[0107] Step C3: Remove non-conductor edges by using morphological operation based on rectangular features of a tower and irregular edge features of a tree;

[0108] Step C4: Obtain pixel coordinates of a conductor center line by using least square curve fitting on remaining conductor edge points;

[0109] ​a spatial measurement module, configured to calculate the minimum spatial distance between the dangerous source and the conductor based on the point cloud-monocular data and the binocular data respectively, when the module is based on the point cloud-monocular data, the pixel coordinates of the dangerous source in the monocular image are back-projected to the point cloud coordinate system through the projection transformation matrix, and the minimum distance is calculated in combination with the conductor spatial function fitted based on the conductor point cloud; when the module is based on the binocular data, the minimum distance is calculated after the three-dimensional reconstruction of the dangerous source point and the conductor center line corresponding point and the fitting of the conductor spatial function;

[0110] a data fusion module, configured to perform confidence weighted fusion on the point cloud-monocular measurement result and the binocular measurement result, wherein the confidence is dynamically adjusted according to the environmental light intensity, and the binocular measurement confidence weight is 0.6-0.8 when the light intensity is greater than or equal to 500 lux, and the point cloud-monocular measurement confidence weight is 0.7-0.9 when the light intensity is less than 500 lux;

[0111] a control module, configured to control the collaborative work of the modules, receive the measurement result of the data fusion module, and trigger a warning signal if the measurement result is less than a preset warning threshold, wherein the warning threshold of the 110 kV line is 5 m, the warning threshold of the 220 kV line is 8 m, and the warning threshold of the 500 kV line is 12 m;

[0112] a measurement method, which is measured by the dangerous source spatial measurement system described above, and includes the following steps:

[0113] S1, system deployment:

[0114] deploy the point cloud acquisition module, the monocular image acquisition module and the binocular measurement subsystem to ensure that the acquisition areas of the three modules coincide, complete the calibration through the joint calibration module, and set the warning thresholds of different voltage grade transmission lines;

[0115] S2, data acquisition:

[0116] the point cloud acquisition module acquires real-time three-dimensional point cloud data of the transmission channel, the monocular image acquisition module acquires real-time monocular images, and the binocular measurement subsystem acquires real-time left and right views;

[0117] S3, point cloud processing and point cloud-monocular measurement:

[0118] the point cloud processing module filters and classifies the three-dimensional point cloud to obtain the conductor point cloud and the dangerous source point cloud; the spatial measurement module back-projects the pixel coordinates of the dangerous source in the monocular image to the point cloud coordinate system based on the projection transformation matrix, and calculates the first minimum distance D1 between the dangerous source and the conductor after fitting the conductor spatial function;

[0119] The fitting process of the wire space function includes: using a cubic polynomial to fit the wire point cloud to obtain a wire space curve equation: x=f(z), y=g(z), wherein z is a vertical direction coordinate; the first minimum distance D1 is calculated by traversing each point (xq, yq, zq) in the dangerous source point cloud, calculating the distance from the point to the wire space curve, and taking the minimum value as D1;

[0120] S4, binocular measurement:

[0121] The dynamic distortion correction module performs adaptive distortion correction on the left and right views of the binocular; the wire center line extraction module extracts the wire center line pixel coordinates; the space measurement module performs three-dimensional reconstruction on the wire center line corresponding points and the dangerous source points, fits the wire space function, and calculates the second minimum distance D2 between the dangerous source and the wire;

[0122] The three-dimensional reconstruction process includes: based on the internal and external parameter matrices of the binocular camera, performing stereo matching on the wire center line corresponding points and the dangerous source points in the left and right views, and calculating the three-dimensional coordinates of the corresponding points by using the triangulation method; the fitting process of the wire space function is consistent with step S3, and the second minimum distance D2 is the minimum distance from the three-dimensional coordinates of the dangerous source to the wire space curve;

[0123] S5, data fusion:

[0124] The data fusion module determines the confidence weight of D1 and D2 according to the real-time environmental light intensity, and calculates the fused minimum distance D=w1xD1+w2xD2, wherein (w1+w2=1), w1 is the confidence weight of D1, and w2 is the confidence weight of D2;

[0125] The determination rule of the confidence weight is: when the environmental light intensity I is greater than or equal to 500 lux, w1=0.2-0.4, and w2=0.6-0.8; when 300 lux≤I<500 lux, w1=0.5, and w2=0.5; when I<300 lux, w1=0.7-0.9, and w2=0.1-0.3;

[0126] S6, warning judgment: the control module compares D with the preset warning threshold, and if D<warning threshold, the warning operation of triggering the site shouting, the warning light flashing and sending information to the operation and maintenance personnel is performed.

[0127] Embodiment:

[0128] Taking the 220kV power transmission line crossing the construction area as an example, the specific implementation process of the dangerous source space measurement system and method is as follows:

[0129] System Deployment: Point cloud acquisition modules (such as Riegl VZ-400i 3D laser scanners), monocular image acquisition modules (Basler acA2500-20gm industrial cameras with 12mm lenses), and binocular measurement subsystems (two Basler acA1600-20gm industrial cameras, with a baseline distance of 50cm) are deployed on both sides and at key locations along the 220kV transmission line. This ensures that the acquisition areas of all three modules completely overlap, and a specific range is selected. ,diameter Using the conductor segment as a calibration reference, the camera intrinsic parameter matrix is ​​obtained by solving the problem through the joint calibration module. extrinsic parameter matrix and point cloud-image projective transformation matrix At the same time, the warning threshold for this 220kV line is set as follows: ;

[0130] Data Acquisition: The point cloud acquisition module acquires 3D point cloud data of the power transmission channel in real time at a frequency of 10 points / second; the monocular image acquisition module acquires monocular images at a frame rate of 30 frames / second; and the binocular measurement subsystem simultaneously acquires left and right views. A unit with a volume of approximately [missing information] is located in the construction area. The aspect ratio is The cranes, when in operation, may pose a threat to power transmission lines;

[0131] Point Cloud Processing and Point Cloud-Monocular Measurement: The adaptive point cloud processing module performs voxel lattice downsampling on the original point cloud, reducing the data volume by 60%. Based on the linear features of the traverse lines, an improved RANSAC algorithm is adopted (setting the expected correct model probability). Initial interior point ratio Fit a straight line and filter out lines that are at a distance from the line threshold. Based on the volumetric characteristics of construction machinery, the candidate point cloud of the traverse was selected, and the crane was chosen as the candidate point cloud for hazard source. Other interfering point clouds were removed. The spatial measurement module back-projected the pixel coordinates of the crane in the monocular image to the point cloud coordinate system through a projective transformation matrix. The spatial curve equation of the traverse was obtained by fitting the traverse point cloud with a cubic polynomial. , Traverse each point in the crane point cloud, calculate its distance to the traverse space curve, and obtain the first minimum distance. ;

[0132] Binocular measurement: The dynamic distortion correction module adjusts the current light intensity based on the current light intensity. , the distortion correction coefficient is adaptively adjusted, the left and right views of binoculars are preprocessed, the wire center line extraction module adopts multi-scale Canny edge detection, combines morphological operation and least square curve fitting, and extracts wire center line pixel coordinates, based on the internal and external parameter matrices of binocular cameras, the corresponding points of the wire center line and the crane points in the left and right views are stereomatched, three-dimensional reconstruction is completed by using the triangulation method, after the wire space function is fitted, the second minimum distance D is calculated ;

[0133] Data fusion: the data fusion module determines w1=0.3 and w2=0.7 according to the light intensity , the minimum distance D after fusion is calculated by the formula D=w1xD1+w2xD2, and the minimum distance D after fusion is calculated by the formula D=w1xD1+w2xD2 ;

[0134] Early warning judgment: the control module compares D=7.29 with the preset warning threshold, since , the system immediately triggers the on-site voice warning, warning light high-frequency flicker, and sends the alarm information containing the specific position and dangerous distance to the operation and maintenance personnel's mobile phone.

[0135] Based on the above description, the point cloud processing link of the application optimizes the algorithm in combination with the scene characteristics such as wire linearity and construction machinery volume, filters out interference more accurately, and is different from general point cloud processing technology; the calibration module takes the known parameters of the wire itself as the reference, without the need for additional deployment of calibration objects, solving the problem of field scene calibration; the binocular measurement adds dynamic distortion correction and multi-scale wire center line extraction, improving the anti-interference ability in complex environments and breaking through the limitations of existing fixed parameter correction; a data fusion module is also added, which adjusts the measurement weight dynamically combined with light, thereby improving the measurement reliability and adaptability in complex outdoor environments, and accurately measuring the minimum spatial distance between the construction machinery type danger source and the wire; without the need for additional deployment of special calibration objects, the device can operate independently from the background, and can trigger multi-mode early warning in time for different voltage grade transmission lines, meeting the needs of transmission channel anti-mechanical damage monitoring.

[0136] Embodiments of the application are given for the purpose of illustration and description, and are not intended to be exhaustive or to limit the application to the forms disclosed. Many modifications and variations will be apparent to those of ordinary skill in the art. Embodiments are chosen and described in order to best explain the principles of the application and its practical application, and to enable others skilled in the art to understand the application for various embodiments with various modifications as are suited to the particular use contemplated.

Claims

1. A hazardous source spatial measurement system, characterized in that: include: The point cloud acquisition module is used to acquire three-dimensional laser point cloud data of the power transmission channel. The three-dimensional laser point cloud data includes three-dimensional spatial information of hazardous sources such as conductors, towers and construction machinery. The monocular image acquisition module is used to acquire monocular images of the power transmission channel in real time. The acquisition area of ​​the monocular image overlaps with that of the point cloud acquisition module. A binocular measurement subsystem is used to acquire binocular left and right views in real time. The joint calibration module is used to establish the mapping relationship between the point cloud coordinate system and the monocular image coordinate system, and to solve the point cloud-image projective transformation matrix H. The point cloud processing module is used to filter and classify 3D laser point cloud data to obtain conductor point clouds and hazard source point clouds. The spatial measurement module is used to back-project the pixel coordinates of the hazard source in the monocular image to the point cloud coordinate system based on the point cloud-image projective transformation matrix H, and after fitting the conductor space function, calculate the first minimum distance D1 between the hazard source and the conductor. The dynamic distortion correction module is used to perform adaptive distortion correction on the left and right views of the binoculars; The conductor centerline extraction module is used to extract the pixel coordinates of the conductor centerline and obtain the corresponding point of the conductor centerline; The spatial measurement module is also used to perform three-dimensional reconstruction of the points corresponding to the centerline of the conductor and the points of the hazard source. After fitting the spatial function of the conductor, it calculates the second minimum distance D2 between the hazard source and the conductor. The data fusion module is used to determine the confidence weights of D1 and D2 based on the real-time ambient light intensity, and calculate the minimum distance after fusion, D=w1×D1+w2×D2, where (w1+w2=1), w1 is the confidence weight of D1, and w2 is the confidence weight of D2. The control module is used to control the coordinated work of each module. It compares D with the preset warning threshold. If D < the warning threshold, it triggers on-site announcements, flashing warning lights, and sending information to maintenance personnel as warning operations.

2. The hazardous source space measurement system according to claim 1, characterized in that: The point cloud processing module uses an improved RANSAC algorithm to filter out tree and bird interference point clouds based on the geometric features of the power transmission channel target, and classifies the effective point clouds into conductor point clouds, tower point clouds and hazard source point clouds. The geometric features are conductor linear features and construction machinery volume features. The joint calibration module uses the conductor segment with known physical parameters in the transmission channel as the calibration reference. It combines the three-dimensional coordinates of the conductor point cloud with the pixel coordinates of the conductor in the monocular image to solve for the intrinsic and extrinsic parameter matrix of the monocular camera and the point cloud-image projective transformation matrix H.

3. The hazardous source space measurement system according to claim 1, characterized in that: The dynamic distortion correction module adaptively adjusts the distortion correction coefficient based on the illumination intensity and contrast parameters of the images acquired in real time by the binocular camera; the conductor centerline extraction module uses multi-scale Canny edge detection combined with least squares curve fitting to extract the pixel coordinates of the conductor centerline after removing interference edges from towers and trees.

4. A hazardous source space measurement system according to claim 3, characterized in that: When the spatial measurement module is based on point cloud-monocular data, it back-projects the coordinates of the hazard source pixels in the monocular image to the point cloud coordinate system through the projective transformation matrix, and calculates the minimum distance by combining the traverse spatial function fitted to the traverse point cloud; when based on binocular data, it performs three-dimensional reconstruction of the corresponding points of the traverse centerline and the hazard source points, and calculates the minimum distance after fitting the traverse spatial function.

5. A hazardous source space measurement system according to claim 2, characterized in that: The filtering process of the point cloud processing module includes: Step A1: Perform voxel lattice downsampling on the original point cloud to reduce the amount of data; Step A2: Based on the linear characteristics of the traverse, the improved RANSAC algorithm is used to fit the straight line, and candidate point clouds of the traverse are screened. Points with a distance of ≤0.05m from the straight line are set as candidate points of the traverse. The specific process for selecting candidate point clouds for traverse lines is as follows: Define the point cloud set as ,in Represents the first point in the point cloud The three-dimensional coordinates of each point; From point cloud collection Randomly selected For a point, and for line fitting, , recorded as Through this A straight line is fitted from the points. ,straight line The equation can be expressed in parametric form: ； in For example, a point on a straight line , Let be the direction vector of the line, by and Sure; For point cloud collections Each point in Calculate its distance to the line distance Using the formula for the distance from a point to a line: ； in , , , Represents the cross product of vectors. The magnitude of the vector is defined by a distance threshold. ,if Then the point Consider them as interior points, and denote the set of interior points as... ; Statistical set of interior points Number of interior points ,like If the number of interior points is greater than the current optimal model, then the current line... and interior point set Recorded as the optimal model; Repeat the above process to perform The next iteration, in which Determined by the following formula: ； in It represents the probability of obtaining the correct model. It represents the proportion of interior points in the point cloud. After the above iterative process, the final set of interior points is obtained. This refers to the selected candidate point cloud for the conductor. Step A3: Based on the volumetric characteristics of construction machinery, screen candidate point clouds of hazard sources; Step A4: Remove interfering point clouds other than the candidate point clouds of conductors, candidate point clouds of hazards, and point clouds of towers.

6. A hazardous source space measurement system according to claim 2, characterized in that: The calibration process of the joint calibration module includes: Step B1: Select a known span within the transmission channel. , Given diameter , The conductor segment is used as a calibration reference. Step B2: Obtain the 3D point cloud of the traverse segment using the point cloud acquisition module, and fit the spatial straight line equation of the traverse. : ； in The three-dimensional coordinates of a point in space are variables that change as the point's position changes; The three-dimensional coordinates of a known fixed reference point in space are definite values ​​used as the starting reference for calculations. Let be the three components of the direction vector, which together determine the direction of a straight line, where the direction vector is . ; Step B3: Acquire a monocular image containing the conductor segment using the monocular image acquisition module, and extract the pixel line equation of the conductor segment. : ； in, Let be the slope of the line. As the independent variable, As the dependent variable, The straight line is Intercept on the axis; Step B4: Establish the three-dimensional coordinates of the guide point cloud , and pixel coordinates The corresponding relationship is substituted into the camera imaging model, combined with the wire diameter. Physical constraints, solving for the camera intrinsic parameter matrix and extrinsic parameter matrix This leads to the point cloud-image projective transformation matrix. .

7. A hazardous source space measurement system according to claim 3, characterized in that: The extraction process of the conductor centerline extraction module includes: Step C1: Perform grayscale conversion and Gaussian filtering preprocessing on the left and right views acquired by the binocular camera; Step C2: Use multi-scale Canny edge detection to obtain image edges; Step C3: Based on the rectangular features of the towers and the irregular edge features of the trees, morphological operations are used to remove non-conductor edges; Step C4: Use least squares curve fitting on the remaining edge points of the conductor to obtain the pixel coordinates of the conductor centerline.

8. A method for measuring the space of a hazardous source, applied to a measurement system of a hazardous source space measurement system as described in any one of claims 1-7, characterized in that: Includes the following steps: S1. System Deployment: Deploy a point cloud acquisition module, a monocular image acquisition module, and a binocular measurement subsystem to ensure that the acquisition areas of the three overlap. Complete the calibration through a joint calibration module and set early warning thresholds for transmission lines of different voltage levels. S2, Data Acquisition: The point cloud acquisition module acquires 3D point cloud data of the power transmission channel in real time, the monocular image acquisition module acquires monocular images in real time, and the binocular measurement subsystem acquires left and right views in real time. S3. Point Cloud Processing and Point Cloud Monocular Measurement: The point cloud processing module filters and classifies the 3D point cloud to obtain the conductor point cloud and the hazard source point cloud; the spatial measurement module back-projects the hazard source pixel coordinates in the monocular image to the point cloud coordinate system based on the projective transformation matrix, fits the conductor space function, and calculates the first minimum distance D1 between the hazard source and the conductor. S4. Binocular Measurement: The dynamic distortion correction module performs adaptive distortion correction on the binocular left and right views; the conductor centerline extraction module extracts the pixel coordinates of the conductor centerline; the spatial measurement module performs three-dimensional reconstruction of the corresponding points of the conductor centerline and the hazard source points, and after fitting the conductor spatial function, calculates the second minimum distance D2 between the hazard source and the conductor. S5, Data Fusion: The data fusion module determines the confidence weights of D1 and D2 based on the real-time ambient light intensity, and calculates the minimum distance after fusion, D = w1 × D1 + w2 × D2, where (w1 + w2 = 1), w1 is the confidence weight of D1, and w2 is the confidence weight of D2. S6. Warning Judgment: The control module compares D with the preset warning threshold. If D < warning threshold, it triggers on-site announcements, flashing warning lights, and sending information to maintenance personnel as warning operations.

9. The measurement method according to claim 8, characterized in that: In step S3, the fitting process of the traverse space function includes: fitting the traverse point cloud with a cubic polynomial to obtain the traverse space curve equation: x=f(z), y=g(z), where z is the vertical coordinate; the calculation process of the first minimum distance D1 is: traversing each point (xq, yq, zq) in the hazard source point cloud, calculating the distance from the point to the traverse space curve, and taking the minimum value as D1.

10. A measurement method according to claim 8, characterized in that: In step S4, the three-dimensional reconstruction process includes: based on the intrinsic and extrinsic parameter matrix of the binocular camera, performing stereo matching on the corresponding points of the conductor centerline and the hazard source points in the left and right views, and calculating the three-dimensional coordinates of the corresponding points using triangulation; the fitting process of the conductor space function is the same as in step S3, and the second minimum distance D2 is the minimum distance from the three-dimensional coordinates of the hazard source to the conductor space curve. The confidence weights in step S5 are determined as follows: when the ambient light intensity I ≥ 500 lux, w1 = 0.2-0.4, w2 = 0.6-0.8; when 300 lux ≤ I < 500 lux, w1 = 0.5, w2 = 0.5; when I < 300 lux, w1 = 0.7-0.9, w2 = 0.1-0.3.

Images