WO2023061049A1 - Network-provisioning vehicle-mounted intelligent inspection robot system and methods based on same - Google Patents

Abstract

The present invention provides a network-provisioning vehicle-mounted intelligent inspection robot system and methods based on same. The system comprises a robot body provided on a vehicle and a detection apparatus carried on the robot body. The detection apparatus comprises at least an ultrasonic detection mechanism, an infrared detection mechanism and a visible light photographing mechanism which are fixed on a rotating gimbal, wherein the ultrasonic detection mechanism and the visible light photographing mechanism are coaxially and reversely arranged, or an included angle between the ultrasonic detection mechanism and the visible light photographing mechanism is in a preset value. The rotating gimbal, the ultrasonic detection mechanism and the visible light photographing mechanism are in communication with a control terminal of the robot body, and the control terminal of the robot body can perform inspection on a power distribution line and/or a power distribution tower and/or a power distribution device according to data transmitted by the detection apparatus. According to the present invention, non-stop inspection of the power distribution line can be implemented, comprehensive inspection of the power distribution line, the power distribution device and the power distribution tower is implemented, omission in the inspection of the power distribution line is avoided, and the accuracy of the inspection result of the power distribution line is greatly improved.

Description

A distribution network vehicle intelligent inspection robot system and method technical field

The invention relates to the technical field of power distribution inspection, in particular to a vehicle-mounted intelligent inspection robot system and method for a distribution network.

Background technique

The statements in this section merely provide background art related to the present invention and do not necessarily constitute prior art.

Power transmission and distribution lines are responsible for power transmission. The lines are long, have many equipments, and have a wide geographical range, which brings great difficulties to the operation and maintenance of the power grid and rapid fault repair.

At present, the inspection of power transmission and distribution lines mostly adopts the method of on-board inspection of distribution network. The automatic inspection of transmission and distribution lines is realized through various inspection equipment installed on inspection vehicles, which effectively avoids manual inspection. The inefficiency problem and the high cost problem of UAV inspection.

However, the inventors have found that the existing vehicle-mounted inspection methods for distribution networks still have the following problems:

(1) When the inspection vehicle is driving in a normal state, the ultrasonic equipment is directly used to inspect the discharge point, and it is impossible to visually display the source of the ultrasonic signal of the discharge point. If a visible light camera is simply added, due to mechanical interference, the axis where it is located It is inevitable that the picture at the source of the ultrasonic receiving signal cannot be collected under the condition that it is collinear with the axis of the ultrasonic equipment at the same time.

(2) Inspection vehicles are equipped with infrared detection equipment, and abnormal analysis is carried out through infrared images, but the overall distribution of infrared image gray levels is relatively low and concentrated, and due to the random interference brought by the surrounding environment to the infrared imaging process and the thermal imaging system itself The imperfection of the infrared image makes the signal-to-noise ratio and contrast of the infrared image relatively low; the traditional image segmentation algorithm based on threshold and edge is sensitive to noise, and it is easy to produce segmentation results with poor continuity, which reduces the accuracy of image segmentation and thus affects the fault. The result of recognition.

(3) Due to the complex distribution of power distribution lines, the distribution of poles and towers in each area is more staggered. When using inspection vehicles to conduct tower inspections, it is often necessary to stop and inspect multiple times, and the inspection efficiency is low; at the same time, it requires a lot of manpower to distinguish Distinguish information such as inspection areas, lines, pole towers, etc. Therefore, inspection personnel in the area are prone to problems such as missing poles, missing lines, and false inspections during the inspection process.

(4) The existing distribution line inspection strategy is that the daily inspection personnel need to plan which lines and towers to inspect this time before the inspection, and configure the lines and towers to be inspected into multiple inspection tasks, During the inspection, the inspection must start from the No. 1 tower of the specified task, and the task line cannot be changed in the middle. This inspection strategy needs to set up many inspection tasks in advance, and the workload of task configuration is heavy, and the inspection needs to follow the route planned by the task. It is less flexible to conduct inspections with pole towers.

(5) In the existing distribution network vehicle inspection strategy, the real-time position of the inspection vehicle and the tower is used to calculate the tracking angle. This calculation method is accurate when the vehicle is stationary, but it cannot be applied to the vehicle running state. detection; because in the driving state of the vehicle, according to the position of the vehicle and the tower at the current time node, it is calculated that the pan-tilt needs to track the rotation angle, and the rotation command is sent to the pan-tilt, and the pan-tilt rotates to the specified angle after receiving the command. In , the inspection vehicle is still in the driving state, and the position of the vehicle has changed. At this time, it is no longer accurate to use the previously calculated angle to track.

Contents of the invention

In order to solve the shortcomings of the existing technology, the present invention provides a vehicle-mounted intelligent inspection robot system and method for distribution network, which can realize non-stop inspection of power distribution lines, and realize the maintenance of power distribution lines, power distribution equipment and power distribution towers. The comprehensive and continuous automatic inspection avoids omissions in the inspection of distribution lines and improves the accuracy of inspection results of distribution lines.

In order to achieve the above object, the present invention adopts the following technical solutions:

The first aspect of the present invention provides a distribution network vehicle intelligent inspection robot system.

A vehicle-mounted intelligent inspection robot system for a distribution network, comprising: a robot body arranged on the vehicle and a detection device mounted on the robot body;

The detection device at least includes an ultrasonic detection mechanism, an infrared detection mechanism, and a visible light imaging mechanism fixed on the rotating platform, and the ultrasonic detection mechanism and the visible light imaging mechanism are arranged coaxially and oppositely, or the angle between the ultrasonic detection mechanism and the visible light imaging mechanism is default value;

The rotating pan/tilt, the ultrasonic detection mechanism and the visible light shooting mechanism all communicate with the control terminal of the robot body, and the control terminal of the robot body can conduct inspections of power distribution lines and/or power distribution towers and/or power distribution equipment according to the data transmitted by the detection device .

The second aspect of the present invention provides a method for locating power distribution equipment, including the following process:

The control terminal controls the rotation of the pan-tilt;

The rotating pan-tilt drives the ultrasonic detection mechanism to rotate, and the control terminal controls the ultrasonic detection mechanism to detect;

When the ultrasonic detection mechanism detects an abnormal signal, the rotating platform rotates 180° from the abnormal signal position or a preset angle so that the visible light shooting mechanism is facing the abnormal signal position;

The control terminal controls the visible light photographing mechanism to photograph the position of the abnormal signal.

The third aspect of the present invention provides a fault identification method for power distribution equipment, including the following process:

Obtain infrared images of power distribution equipment in the distribution network scenario;

According to the similarity between the equipment failure form of the infrared image of the power distribution equipment and the preset equipment failure form of other power scenarios, the equipment failure history images of other power scenarios are selected as the expanded samples;

Mark the defect position in the infrared image of the power distribution equipment and the extended sample, and cut it into a defect image block containing the defect position and a background image block not containing the defect position based on the defect position. For any background image block and any After splicing the defective image blocks, an image training set is obtained;

Based on the image training set, the pre-built segmentation module is trained by multiple upsampling fusion method, and the fault identification result is obtained by using the trained segmentation model for the infrared image of the power distribution equipment to be identified.

A fourth aspect of the present invention provides a method for pole tower inspection, including the following process:

According to the three-level galaxy topology structure of inspection line-inspection tower-tower inspection point, based on the principle of the same inspection person in charge and the principle of similar merger of inspection lines, the optimal inspection task strategy is determined;

Based on the optimal inspection task strategy and the three-dimensional space coordinate system of the vehicle pole, calculate the coordinate data of the preset compensation point position according to the quadrant where the inspection vehicle position is located and the driving distance of the vehicle within the preset interval;

Based on the coordinate data of the preset compensation point position, calculate the best tracking angle data of the vehicle-mounted pan-tilt tracking pole tower and convert it into a control command, so as to control the vehicle-mounted pan-tilt to always maintain the optimal positioning and tracking of the pole tower;

According to the comparison result between the operation data of the tower and the preset detection threshold obtained during the positioning and tracking of the tower, it is determined whether to issue an alarm prompt, so as to send the order to the inspection personnel to continue the inspection or to judge the order on the spot;

Based on all inspection data and on-site research and judgment results, the inspection results are obtained.

A fifth aspect of the present invention provides a tower inspection method, including the following process:

Obtain the geographical coordinate information of the distribution line tower in the set area;

According to the coordinate information of the distribution line towers, considering the constraints of the inspection path, every two levels of towers are interrelated as a minimum unit, and the same tower is included between two adjacent minimum units; traversing all the towers, all the smallest units form Rod-rod minimum unit inspection model;

Based on the inspection model, the next-level inspection tower associated with the current tower is automatically determined, and the task-free inspection of the power distribution line is performed.

The sixth aspect of the present invention provides a dynamic tracking method for towers, including the following process:

According to the set data acquisition frequency, obtain the current position information of the vehicle and the driving speed of the vehicle;

Calculate the driving distance of the vehicle in the interval between two data calculations, and calculate the coordinate data of the preset compensation point position based on the current position information of the vehicle;

Based on the coordinate data of the preset compensation point position, calculate the optimal tracking horizontal rotation angle and pitch angle of the cloud platform to the tower;

The control platform realizes the detection of the tower based on the optimal tracking horizontal angle and pitch angle.

Compared with prior art, the beneficial effect of the present invention is:

(1) The invention pioneered the development of a distribution network vehicle-mounted intelligent inspection robot inspection method, developed a distribution network vehicle-mounted intelligent inspection robot system, proposed a distribution network scene inspection tower dynamic tracking strategy, and constructed a lightweight The visible light appearance recognition model and the infrared tower detection model, designed the visual positioning technology of the discharge point of the distribution equipment, realized the non-stop inspection of the distribution line for the defect of the tower, and improved the real-time and accuracy of the defect detection of the distribution line.

(2) The present invention proposes a visual positioning method for the discharge point of power distribution equipment, constructs a pan-tilt coaxial flip model and a rotating scanning model, solves the problem of poor multi-axis detection accuracy, and realizes the location of the discharge point of power distribution equipment High-precision positioning improves the inspection efficiency and inspection quality of the distribution network vehicle inspection robot.

(3) The present invention proposes an infrared image fault identification method for power distribution equipment, and constructs a MobileNetv1-FCN infrared image equipment segmentation model. By splicing the background image of the distribution network scene with the defect image of other power scenes, the few-sample data is expanded; the pixel-level recognition of the power equipment in the infrared image is carried out through the semantic segmentation model based on the fully convolutional neural network ; Solve the problem of low defect recognition accuracy in the case of few samples, realize the effective segmentation of the foreground and background of power equipment in infrared images and the accurate extraction of equipment outline information, and improve the accuracy of equipment temperature information extraction.

(4) The present invention proposes an inspection task planning method for a distribution network vehicle-mounted intelligent inspection robot, constructs a galaxy topology model of distribution line towers, and designs an optimal inspection task strategy. According to the geographical coordinate information database of distribution line towers, the galaxy topology model of distribution line towers is established, and the patrol inspection tasks of distribution lines are planned through the established topology model, which solves the problem of complex distribution of distribution lines and towers, and various work areas. 1. Problems such as missing rods, missing lines, false inspections, and missed inspections caused by the interlacing of the lines belonging to the team; through the construction of a topology model, the distribution of each line can be clearly seen, so that the inspection personnel can More reasonable planning of inspection tasks improves inspection efficiency.

(5) The present invention innovatively proposes a task-free inspection method for a distribution network vehicle-mounted intelligent inspection robot, designs the minimum unit determination principle of a distribution line tower, and constructs a pole-pole minimum unit inspection model. By traversing all the towers, it is possible to establish the relationship between the coordinates of each tower in the area, and provide data support for the inspection path for the task-free inspection; it solves the problem of heavy workload and low flexibility in the inspection of power distribution lines, and realizes The decomposition of the network topology structure of distribution lines simplifies the model, removes the complex task configuration work, and greatly improves the flexibility of distribution line inspection.

(6) The present invention innovatively proposes a dynamic tracking method for the pole tower of the vehicle-mounted intelligent inspection robot for the distribution network, and designs a calculation algorithm for the position coordinates of the preset compensation points in the direction of vehicle travel, and constructs the spatial position relationship between the cloud platform and the pole tower spatial coordinate system. By inspecting the real-time vehicle speed of the vehicle, dynamically calculate the position of the preset compensation point, determine the optimal detection angle of the pan/tilt based on the coordinates of the preset compensation point, and calculate the vehicle-mounted pan/tilt detection angle through the calculation method of the spatial position relationship between the pan/tilt and the tower The tracking angle of the pan/tilt solves the problem of needing to stop and inspect the tower in the current vehicle inspection mode, and the rotation of the pan/tilt does not require manual operation, which reduces the burden on the inspection personnel and greatly improves the inspection efficiency.

Description of drawings

The accompanying drawings constituting a part of the present invention are used to provide a further understanding of the present invention, and the schematic embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations to the present invention.

FIG. 1 is a schematic structural diagram of a vehicle-mounted intelligent inspection robot system for a distribution network provided by Embodiment 1 of the present invention.

Fig. 2 is a schematic flowchart of a method for locating a discharge point of a power distribution device according to Embodiment 2 of the present invention.

FIG. 3 is a schematic flow chart of a method for identifying a fault in a power distribution device according to Embodiment 3 of the present invention.

FIG. 4 is a schematic flowchart of a tower inspection method provided in Embodiment 4 of the present invention.

FIG. 5 is a schematic diagram of calculating the pan-tilt tracking horizontal angle when the pole tower is located in the first quadrant and the vehicle travel direction is west-north provided by Embodiment 4 of the present invention.

FIG. 6 is a schematic diagram of a calculation method for a tower pitch tracking angle provided by Embodiment 4 of the present invention.

FIG. 7 is a schematic flowchart of a tower inspection method provided in Embodiment 5 of the present invention.

Fig. 8 is a schematic diagram of the pole-rod smallest unit inspection model provided by Embodiment 5 of the present invention.

FIG. 9 is a flow chart of a dynamic tracking method for poles and towers based on traveling vehicles provided in Embodiment 6 of the present invention.

FIG. 10 is a schematic diagram of calculating a preset compensation distance during vehicle running according to Embodiment 6 of the present invention.

Among them, 1-vehicle; 2-ultrasonic detection mechanism; 3-visible light detection mechanism; 4-infrared detection mechanism.

Detailed ways

The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

It should be noted that the following detailed description is exemplary and intended to provide further explanation of the present invention. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

It should be noted that the terminology used here is only for describing specific embodiments, and is not intended to limit exemplary embodiments according to the present invention. As used herein, unless the context clearly dictates otherwise, the singular is intended to include the plural, and it should also be understood that when the terms "comprising" and/or "comprising" are used in this specification, they mean There are features, steps, operations, means, components and/or combinations thereof.

In the case of no conflict, the embodiments and the features in the embodiments of the present invention can be combined with each other.

Example 1:

As shown in Figure 1, Embodiment 1 of the present invention provides a vehicle-mounted intelligent inspection robot system for a distribution network, including: a robot body installed on a vehicle 1 and a detection device mounted on the robot body;

The detection device at least includes an ultrasonic detection mechanism 2, an infrared detection mechanism 4, and a visible light imaging mechanism 3 fixed on the rotating platform, and the ultrasonic detection mechanism and the visible light imaging mechanism are arranged coaxially and oppositely, or the ultrasonic detection mechanism and the visible light imaging mechanism are arranged in opposite directions. The included angle is the default value.

It can be understood that in some other embodiments, the detection device may also include other sensors, such as temperature and humidity sensors, air pressure sensors, etc., which can be selected by those skilled in the art according to specific working conditions, and will not be repeated here.

The rotating pan/tilt, the ultrasonic detection mechanism and the visible light shooting mechanism all communicate with the control terminal of the robot body, and the control terminal of the robot body can conduct inspections of power distribution lines and/or power distribution towers and/or power distribution equipment according to the data transmitted by the detection device .

The horizontal rotation angle range of the swivel head is 0-360°, and the vertical rotation angle range of the swivel head is -90°-90°. In this embodiment, the preferred pitch angle is 0-90°. Those skilled in the art It can be selected according to specific working conditions, and will not be repeated here.

In this embodiment, the rotating pan-tilt is connected to the base of the pan-tilt for fixing, and the base for fixing is used for connecting with the vehicle.

Example 2:

Embodiment 2 of the present invention provides a method for locating discharge points of power distribution equipment. As shown in FIG. 2, using the vehicle-mounted intelligent inspection robot system for distribution network described in Embodiment 1 of the present invention includes the following process:

S201: the motor of the rotating pan/tilt receives a control command from the control terminal;

S202: The rotating pan/tilt drives the ultrasonic detection mechanism to rotate, and the ultrasonic detection mechanism performs detection according to the received control command from the control terminal;

S203: When the ultrasonic detection mechanism detects an abnormal signal, the rotating platform rotates 180° from the position of the abnormal signal so that the visible light imaging mechanism faces the position of the abnormal signal;

S204: The visible light photographing mechanism photographs the position of the abnormal signal according to the control instruction of the control terminal.

In S203, when the ultrasonic detection mechanism detects an abnormal signal, the rotating pan/tilt scans with the current orientation as the center, and every time it reaches a predetermined point, the current ultrasonic value is automatically recorded and drawn, and the corresponding point corresponding to the maximum value of the ultrasonic value is the point of partial discharge.

Example 3:

Embodiment 3 of the present invention provides a fault identification method for power distribution equipment. As shown in FIG. 3, using the vehicle-mounted intelligent inspection robot system for distribution network described in Embodiment 1 of the present invention includes the following process:

S301: Obtain the historical infrared images of power distribution equipment in the distribution network scene and other preset power scenes, and select the equipment in other power scenes according to the similarity between the equipment failure form of the power distribution equipment infrared image and the equipment failure form in other power scenes Fault history images are used as extended samples;

Preferably, the structural similarity algorithm is used to select samples with similar defect forms for expansion; for example, there are many bird nests in the power transmission scene, but there are very few bird nests in the power distribution scene, so a new image can be obtained by splicing the bird nests in the power transmission scene and the power distribution background pictures , to expand the number of samples and reduce the false detection rate.

Preferably, other power scenarios may include power transmission scenarios, power transformation scenarios, and the like.

S302: Mark the defect position in the infrared image of the power distribution equipment and the extended sample, and crop it into a defect image block containing the defect position and a background image block not containing the defect position based on the defect position, and for any background image block and After splicing any defective image block, a new image set is obtained;

Preferably, the few-sample defect data in the distribution network scene and the defect image data in other power scenes are cut out according to the defect coordinate positions and stored in a unified manner, and the background image blocks are stored separately.

Preferably, all defect images are cropped into image blocks by using functions in a proprietary image processing library according to defect coordinate positions.

In this embodiment, the defective image blocks are subjected to random image transformation operations of geometric transformation and pixel transformation, so as to further expand the defective image blocks and increase the number of defective image blocks;

Preferably, said geometric transformation includes flipping, translation, shearing, rotation and scaling.

Preferably, said pixel transformation includes brightness, contrast, saturation, channel transformation and adding noise.

In this embodiment, the defect image block and the background image block are synthesized through an image splicing method to construct an image training set;

The image stitching method includes: randomly selecting a square area in the background image block, the coordinate position of the square area is used as the insertion position of the defect image block in the background image block, and the transformed defect image block and the background image block are synthesized by image stitching , as new training data to train the segmentation model.

S303: Based on the image training set, the pre-built segmentation module is trained using a multi-upsampling fusion method, and the infrared image of the power distribution equipment to be identified is obtained using the trained segmentation model to obtain a fault identification result.

Preferably, the inspection target area in the image training set is calibrated, and the position and label of the power distribution equipment in the image are marked.

Preferably, in order to meet the input requirements of different network models, the image size can be adjusted.

In this embodiment, the construction process of the segmentation model is as follows: constructing the MobileNetv1-FCN semantic segmentation model based on the full convolutional neural network, using the lightweight convolutional neural network structure MobileNetv1 as the backbone network of the segmentation model, and training based on images set to train the segmentation model.

Preferably, the training process includes: after performing 5 convolution operations on the image training set, using multiple upsampling and fusion methods to obtain the feature map;

Specifically, since the size of the image becomes 1/32 of the original after 5 convolution operations, if the 32 times upsampling is used directly, the segmentation accuracy will be greatly reduced; therefore, in this embodiment, the final conv_dw5 layer is first obtained The feature map is 2 times upsampled, and the feature map corresponding to conv_dw4 is fused to obtain a 1/16 feature map; similarly, this feature map is 2 times upsampled, and corresponding to conv_dw3 The feature maps are fused to obtain 1/8 feature maps; finally, the obtained feature maps are 8 times upsampled and calculated; the infrared image is segmented at the pixel level, and the contour information of the inspection equipment is extracted to improve the accuracy of contour segmentation.

The core idea of MobileNetv1 is to use depthwise separable convolution (DepthWise separable Convolution) to build a lightweight deep convolutional neural network, and decompose the standard convolution operation into a Depthwise Convolution and a Pointwise Convolution convolution operation to ensure the accuracy of segmentation. At the same time, it reduces the computational complexity of the model and ensures the real-time performance of the algorithm, thereby realizing real-time infrared contour segmentation of power distribution equipment.

Preferably, the specific structure of the segmentation model includes 5 conv_dw (depth separable convolution) and 3 upsampling operations, including 2 double upsampling operations and one 8x upsampling operation.

Example 4:

Embodiment 4 of the present invention provides a pole tower inspection method. As shown in FIG. 4, using the distribution network vehicle-mounted intelligent inspection robot system described in Embodiment 1 of the present invention includes the following process:

S401: According to the three-level galaxy topology structure of inspection line-inspection tower-tower inspection point, based on the principle of the same inspection person in charge and the principle of similar merger of inspection lines, determine the optimal inspection task strategy.

In the specific implementation, the geographic information coordinates of the distribution line towers are obtained by using the geographic information positioning equipment to collect the longitude and latitude coordinate information of the distribution line towers.

Among them, the geographic information positioning equipment is centimeter-level precision. In this embodiment, geographic information coordinates of distribution line towers with centimeter-level accuracy are obtained.

The three-level galaxy topology structure of the inspection line-inspection tower-tower inspection point is constructed according to the geographical information coordinates of the power distribution line tower.

Among them, each inspection line includes multi-level towers of power distribution lines, and each level of towers includes multiple inspection points.

In this embodiment, the principle of the same person in charge of inspection is: the inspection lines belonging to the same person in charge of inspection are combined and inspected preferentially.

In the specific implementation, the principle of similar merging of inspection lines is as follows:

Select the two inspection lines with the shortest distance between inspection poles and towers as similar inspection lines, and give priority to combined inspections.

This embodiment builds a distribution line tower galaxy topology model, which solves the problem of missing poles, missing lines, false inspections, and missed inspections during the inspection process caused by the complex distribution of distribution lines and towers, and the interlacing of lines belonging to each work area and team. The problem is to obtain the geographical information coordinates of the distribution line towers, establish the geographical information coordinate library of the distribution line, establish the galaxy topology model structure of the distribution line according to the geographical information coordinate library, and plan the optimal inspection task according to the topology model Strategy, by building a topology model, you can clearly see the distribution of each line, so that inspectors can plan inspection tasks more reasonably according to the distribution of lines, and improve inspection efficiency.

S402: Based on the optimal inspection task strategy and the three-dimensional space coordinate system of the pole, calculate the coordinate data of the preset compensation point position according to the quadrant where the inspection vehicle is located and the vehicle's driving distance within a preset interval.

As a preferred implementation manner, the accuracy of the longitude and latitude coordinate information of the vehicle and the power distribution line tower is centimeter level. Centimeter-level positioning can improve the accuracy of angle calculation and tower tracking.

Use geographic information positioning equipment with centimeter-level accuracy to collect the longitude and latitude coordinates of distribution line towers, and store the collected longitude and latitude coordinates of towers, the line they belong to, and tower numbers and other information into the system database to build a coordinate library for distribution line towers. At the same time, during the inspection process, the inspection control system (for example, the positioning equipment mounted on the roof) collects the real-time latitude and longitude coordinate data of the inspection vehicle at a set frequency (for example: 5 times/second).

It should be noted here that those skilled in the art may also select other collection frequencies according to actual conditions, but the collection frequency is not lower than 2 times/second.

In the specific implementation, the coordinates of the distribution line tower are taken as the origin, the longitude, latitude and height are used as the x-axis, y-axis and z-axis, and the position points of the inspection vehicles are used as points in the coordinate system to construct a three-dimensional space coordinate system for the vehicle pole.

S403: Based on the coordinate data of the preset compensation point position, calculate the optimal tracking angle data of the vehicle-mounted pan-tilt to track the tower and convert it into a control command, so as to control the vehicle-mounted pan-tilt to always maintain the optimal positioning and tracking of the pole.

Wherein, the optimal tracking angle data of the vehicle-mounted pan/tilt tracking tower includes the optimal tracking horizontal rotation angle and the optimal tracking pitch angle.

Specifically, the calculation process of the optimal tracking horizontal rotation angle of the tower includes:

The preset compensation point position is used as the real-time position of the vehicle, and based on the preset compensation point position and pole tower position, a three-dimensional space coordinate system of the vehicle pole is constructed;

Calculate the distance between the vehicle and the tower and the vertical distance between the real-time position of the vehicle and the x-axis, and calculate the angle between the line between the vehicle and the tower and the y-axis based on the distance;

Calculate the angle between the vehicle's driving direction and the y-axis;

According to the quadrant where the vehicle is located, calculate the horizontal rotation angle of the gimbal tracking.

According to the latitude and longitude coordinates (Lng ₁ , Lat ₁ ) of the tower and the latitude and longitude coordinates (Lng ₂ , Lat ₂ ) of the vehicle, calculate the straight-line distance L between the vehicle and the pole. The calculation formula is as follows:

(1)radLat ₁ ＝Lat ₁ *Math.PI/180.0

(2) radLat ₂ =Lat ₂ *Math.PI/180.0

(3) Lng＝(Lng ₁ -Lng ₂ )*Math.PI/180.0

(4) Lat＝(Lat ₁ -Lat ₂ )*Math.PI/180.0

(5) L=(2*Math.Asin(Math.Sqrt(Math.Pow(Math.Sin(Lng/2),2)+Math.Cos(radLat1)*Math.Cos(radLat2)*Math.Pow( Math.Sin(Lng/2),2))))*Earth Radius

Among them, Math.PI refers to the mathematical function, pi; Math.Asin mathematical trigonometric function, arcsine function; Math.Sqrt mathematical function, square root; Math.Pow mathematical function, power function; Math.Sin mathematical function, sine Function; Math.Lng is the calculation result of formula (3); Math.Cos mathematical function, cosine function; Math.radLat ₂ is the calculation result of formula (2).

Calculate the vertical distance Y_D between the vehicle position and the x-axis according to the latitude and longitude coordinates of the tower and the vehicle;

According to L and Y_D, calculate the angle ∠β between the connecting line of the car and the pole and the north direction of the y-axis, and the calculation formula is as follows:

∠β＝Math.ASin(Y_D/L)*(180/Math.PI)

Calculate the angle ∠α between the driving direction of the vehicle and the true north direction of the y-axis according to the data of the vehicle-mounted dual positioning equipment:

According to the quadrant where the vehicle is located, different formulas are used to calculate the rotation angle of the gimbal tracking. For example, the detailed calculation rules are shown in Figure 5, and X_D represents the vertical distance between the vehicle position and the tower.

Specifically, the calculation process of the optimal tracking pitch angle includes:

The preset compensation point position is used as the real-time position of the vehicle, and based on the preset compensation point position and pole tower position, a three-dimensional space coordinate system of the vehicle pole is constructed;

Calculate the distance between the vehicle and the tower, and the difference between the height of the tower head and the height of the on-board inspection equipment components;

Based on the distance and the height difference, the pitch angle of the gimbal is calculated in the three-dimensional space coordinate system of the vehicle pole.

Calculate the pitch angle data of tower tracking, the calculation method is shown in Figure 6:

Calculate the linear distance L between the vehicle and the pole according to the latitude and longitude coordinates of the pole tower and the vehicle;

According to the height H ₁ of the tower head and the height H ₂ of the on-board inspection equipment components, the height difference h between the two can be obtained, and the calculation formula is as follows:

h＝H ₁ -H ₂

Calculate the pitch angle ∠γ of the gimbal according to the height difference h and the straight-line distance L between the vehicle and the pole. The calculation formula is as follows:

∠γ=Math.Atan(h/L)*(180/Math.PI).

Step S404: According to the comparison result between the operation data of the tower acquired during the positioning and tracking of the tower and the preset detection threshold, determine whether to issue an alarm prompt, so as to send an order to continue the inspection or an order for on-site research and judgment to the inspectors.

In the specific implementation, according to the PELCO-D protocol, the tracking motion angle data of the vehicle-mounted pan-tilt is converted into control commands to guide the vehicle-mounted pan-tilt to automatically perform horizontal and pitch movements.

The converted angle data is converted into the corresponding hexadecimal control command according to the protocol format, and the command is sent to the detection pan/tilt, and the pan/tilt is controlled to move according to the specified angle. The format of the control command is:

Byte 1 is the start byte, byte 2 is the PTZ address, byte 3 is command word 1, byte 4 is command word 2, byte 5 is data 1, byte 6 is data 2, byte 7 is the end byte.

Specifically, the control command format is shown in Table 1:

Table 1: Control command format

S405: Obtain an inspection report based on all inspection data and on-site research and judgment results.

Among them, the on-site research and judgment results are obtained by the inspection personnel getting off the vehicle and holding the detection equipment to conduct a second inspection on the pole tower, so as to confirm whether the abnormal alarm of the system is true.

Specifically, if the tower operation data acquired during the positioning and tracking of the tower exceeds the preset detection threshold, an alarm is issued; otherwise, no alarm is issued.

When an alarm prompt is issued, an on-site research and judgment order is sent to the inspector; when an alarm prompt is not issued, an inspection order is sent to the inspector.

Example 5:

Embodiment 5 of the present invention provides a pole tower inspection method. As shown in FIG. 7, using the distribution network vehicle-mounted intelligent inspection robot system described in Embodiment 1 of the present invention includes the following process:

S501: Obtain geographic coordinate information of distribution line towers within the set area;

In this embodiment, use the geographic information positioning equipment (such as: GPS, Beidou, Galileo, etc.) that the accuracy is centimeter level, collect the latitude coordinate information of distribution line tower longitude, build the tower information database of distribution line; In this database It also stores information such as the tower number and the height of the tower.

S502: According to the coordinate information of distribution line towers, considering the constraints of the inspection path, every two levels of towers are related to each other as a minimum unit, and the same tower is included between two adjacent minimum units, and then two adjacent minimum units are established. The connection between units; traverse all towers, and all the smallest units form a pole-tower minimum unit inspection model;

Referring to Figure 8, the coordinate data of the lines and towers of the distribution lines are sorted out. Each two-level tower is used as a minimum unit, and the distribution line is divided into many minimum units. These minimum units are combined to form a pole-pole minimum unit. Inspection model.

In this embodiment, considering the limiting factors of the inspection path, the next-level tower is selected based on the principle of determining the minimum unit to form a minimum unit for mutual correlation;

Take the determined next-level tower as the current tower, and continue to select the next-level tower of the current tower to form another minimum unit;

By analogy, all poles and towers are traversed to obtain the pole-tower minimum unit inspection model.

Among them, the minimum unit determination principle mainly includes the following aspects:

1) The principle of priority on the same line: the towers in the same line are given priority to keep the smallest unit in the line;

2) The principle of adjacency and close distance: the towers of the first and last cross-line towers are given priority to the towers of adjacent lines or other lines with closer distances;

3) The principle of the same person in charge: priority is given to the selection of towers belonging to the area of the same person in charge to form the smallest unit.

The information of each level of tower includes tower number, coordinates and height information; when the current tower is inspected, the position of the tower for the next inspection is determined according to the next tower information associated with it in the smallest unit.

In this embodiment, the form of the pole-tower minimum unit inspection model in the database is shown in Table 2. When the current tower is detected, the data of the next tower can be extracted from the database.

Table 2: The form of the pole-rod minimum unit inspection model in the database

S503: Based on the inspection model, automatically determine a tower for next-level inspection, and perform task-free inspection of distribution lines.

In this embodiment, by comparing the real-time position of the vehicle with the position of the tower of the power distribution line, find the tower closest to the vehicle as the current inspection tower; Through this method of using the current tower to drive the next level of tower inspection, the task-free inspection mode of the power distribution line is realized.

Specifically, the set tower is used as the initial tower, and the tower that forms the smallest tower unit with the initial tower is used as the tower for the next inspection;

The set tower is used as the starting tower, and the tower that forms the smallest unit with the starting tower is used as the tower for the next inspection;

After reaching the next level of towers, in addition to the upper level of towers,

If there is only one smallest tower unit in the next-level tower, the tower that forms the smallest tower unit with it will be used as the tower for inspection at the next level;

If there are two or more minimum tower units in the next-level tower, select the tower that is closest to the inspection equipment and has not been inspected as the tower for the next-level inspection;

If there is no minimum tower unit in the next-level tower, the inspection ends.

The pole-rod minimum unit inspection model is stored in the inspection vehicle, and the determined coordinate information of the next-level inspection pole tower is used as the target position of the inspection vehicle to determine the running path of the inspection vehicle. In addition, the inspection vehicle records the positions of the towers that have been inspected at the end of the last inspection. During this inspection, the towers that have not been inspected are given priority as the towers for the next inspection.

In this embodiment, by constructing the inspection model of the pole-pole minimum unit, no one performs the configuration of the inspection task, so that the automatic inspection without tasks can be realized, the configuration task is simplified, and the inspection efficiency is improved.

Embodiment 6:

Embodiment 6 of the present invention provides a dynamic tracking method for pole towers. As shown in FIG. 9, using the distribution network vehicle-mounted intelligent inspection robot system described in Embodiment 1 of the present invention includes the following process:

S601: Obtain the current position information of the vehicle and the driving speed of the vehicle according to the set data acquisition frequency;

Specifically, using centimeter-level precision geographic information positioning equipment (such as: GPS, Beidou, Galileo, etc.), according to the set data collection frequency, real-time acquisition of vehicle geographic information coordinate data;

Calculate the real-time driving speed of the vehicle according to the change distance of the vehicle coordinate positions obtained twice adjacently.

For example: the first data collection is at location point A, and the second data collection arrives at location point B. Calculate the driving speed of the vehicle through the distance L between A-B and the time interval t between two data receptions (t is 200ms). v=L/t.

S602: Calculate the driving distance of the vehicle in the interval between two data calculations, and calculate the coordinate data of the preset compensation point position based on the current position information of the vehicle;

Specifically, in combination with Figure 10, according to the obtained real-time vehicle driving speed, combined with the frequency of data acquisition, calculate the driving distance of the vehicle in the interval between two adjacent data acquisitions, use the current vehicle position coordinates, and then superimpose the driving distance in the interval , to obtain the coordinate data of the preset compensation point position.

S603: Based on the coordinate data of the preset compensation point position, calculate the optimal tracking horizontal angle and pitch angle of the pan-tilt to the tower. For the specific calculation method, refer to Embodiment 4, which will not be repeated here.

S604: Realize the detection of the tower based on the optimal tracking horizontal angle and pitch angle.

Convert the optimal tracking horizontal rotation angle and pitch angle data of the vehicle-mounted pan/tilt into control commands, so as to control the pan/tilt to perform automatic movement according to the control commands, and realize accurate detection of the tower without stopping.

The above descriptions are only preferred embodiments of the present invention, and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and changes. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.

Claims (23)

1. A vehicle-mounted intelligent inspection robot system for a distribution network, characterized in that,

   Including: the robot body installed on the vehicle and the detection device mounted on the robot body;

   The detection device at least includes an ultrasonic detection mechanism, an infrared detection mechanism, and a visible light imaging mechanism fixed on the rotating platform, and the ultrasonic detection mechanism and the visible light imaging mechanism are arranged coaxially and oppositely, or the angle between the ultrasonic detection mechanism and the visible light imaging mechanism is default value;

   The rotating pan/tilt, the ultrasonic detection mechanism and the visible light camera mechanism all communicate with the control terminal of the robot body, and the control terminal of the robot body conducts inspections of power distribution lines and/or power distribution towers and/or power distribution equipment according to the data transmitted by the detection device.
2. A method for locating power distribution equipment based on the vehicle-mounted intelligent inspection robot system for distribution network according to claim 1, characterized in that,

   The method includes the following processes:

   The control terminal controls the rotation of the pan-tilt;

   The rotating pan-tilt drives the ultrasonic detection mechanism to rotate, and the control terminal controls the ultrasonic detection mechanism to detect;

   When the ultrasonic detection mechanism detects an abnormal signal, the rotating platform rotates 180° from the abnormal signal position or a preset angle so that the visible light shooting mechanism is facing the abnormal signal position;

   The control terminal controls the visible light photographing mechanism to photograph the position of the abnormal signal.
3. The positioning method according to claim 2, characterized in that,

   When the ultrasonic detection mechanism detects an abnormal signal, the rotating pan/tilt rotates around the current azimuth, and every time it reaches a predetermined point, the current ultrasonic value is automatically recorded and drawn, and the corresponding point corresponding to the maximum value of the ultrasonic value is the partial discharge point.
4. The positioning method according to claim 2, characterized in that,

   Perform image enhancement processing on captured images, including the following processes:

   Extract the color information and grayscale information from the obtained distribution line inspection image, and use the trained image quality evaluation network to obtain the image quality score based on the color information and grayscale information;

   According to the image quality score, the distribution line inspection image is screened, and the color space conversion is performed on the distribution line inspection image obtained after screening, and the brightness component is extracted;

   The luminance component is enhanced based on the pre-built enhancement model based on the attention residual module, and the enhanced luminance component, original hue component and original saturation component are converted into color space to obtain an enhanced power distribution line inspection image.
5. A fault identification method for power distribution equipment based on the vehicle-mounted intelligent inspection robot system for distribution network according to claim 1, characterized in that,

   The method includes the following processes:

   Obtain infrared images of power distribution equipment in the distribution network scenario;

   According to the similarity between the equipment failure form of the infrared image of the power distribution equipment and the preset equipment failure form of other power scenarios, select the equipment failure history images of other power scenarios as the expanded sample;

   Mark the defect position in the infrared image of the power distribution equipment and the extended sample, and cut it into a defect image block containing the defect position and a background image block not containing the defect position based on the defect position. For any background image block and any After splicing the defective image blocks, an image training set is obtained;

   Based on the image training set, the pre-built segmentation module is trained by multiple upsampling fusion method, and the fault identification result is obtained by using the trained segmentation model for the infrared image of the power distribution equipment to be identified.
6. The fault identification method according to claim 5, characterized in that,

   Based on the image training set, the pre-built segmentation module is trained using multiple upsampling fusion methods, including:

   Using the lightweight convolutional neural network structure MobileNetv1 as the backbone network of the segmentation model, the MobileNetv1-FCN semantic segmentation model based on the full convolutional neural network is constructed. After multiple convolution operations on the image training set, multiple upsampling fusion is adopted. method for training.
7. A pole tower inspection method based on the vehicle-mounted intelligent inspection robot system for distribution network according to claim 1, characterized in that,

   The method includes the following processes:

   According to the three-level galaxy topology structure of inspection line-inspection tower-tower inspection point, based on the principle of the same inspection person in charge and the principle of similar merger of inspection lines, the optimal inspection task strategy is determined;

   Based on the optimal inspection task strategy and the three-dimensional space coordinate system of the vehicle pole, calculate the coordinate data of the preset compensation point position according to the quadrant where the inspection vehicle position is located and the driving distance of the vehicle within the preset interval;

   Based on the coordinate data of the preset compensation point position, calculate the best tracking angle data of the vehicle-mounted pan-tilt tracking pole tower and convert it into a control command, so as to control the vehicle-mounted pan-tilt to always maintain the optimal positioning and tracking of the pole tower;

   According to the comparison result between the operation data of the tower and the preset detection threshold obtained during the positioning and tracking of the tower, it is determined whether to issue an alarm prompt, so as to send the order to the inspection personnel to continue the inspection or to judge the order on the spot;

   Based on all inspection data and on-site research and judgment results, the inspection results are obtained.
8. The tower inspection method according to claim 7, wherein,

   The three-level galaxy topology structure of the inspection line-inspection tower-tower inspection point is constructed according to the geographical information coordinates of the power distribution line tower.
9. The tower inspection method according to claim 7, wherein,

   The principle of the person in charge of the same inspection is as follows:

   Prioritize combined inspection for the inspection lines belonging to the same inspection person in charge.
10. The tower inspection method according to claim 7, wherein,

    The principle of similar merging of inspection lines is as follows:

    Select the two inspection lines with the shortest distance between inspection poles and towers as similar inspection lines, and give priority to combined inspections.
11. The tower inspection method according to claim 7, wherein,

    Each inspection line includes multi-level towers of power distribution lines, and each level of towers includes multiple inspection points.
12. The tower inspection method according to claim 7, wherein,

    The three-dimensional space coordinate system of the vehicle pole is constructed with the coordinates of the distribution line pole and tower as the origin, the longitude, latitude and height as the x-axis, y-axis and z-axis, and the inspection vehicle position points as points in the coordinate system.
13. The tower inspection method according to claim 7, wherein,

    The optimal tracking angle data of the vehicle-mounted pan/tilt tracking tower includes an optimal tracking horizontal rotation angle and an optimal tracking pitch angle.
14. The tower inspection method according to claim 13, characterized in that,

    The calculation process of the optimal tracking horizontal rotation angle of the tower includes:

    The preset compensation point position is used as the real-time position of the vehicle, and based on the preset compensation point position and pole tower position, a three-dimensional space coordinate system of the vehicle pole is constructed;

    Calculate the distance between the vehicle and the tower and the vertical distance between the real-time position of the vehicle and the x-axis, and calculate the angle between the line between the vehicle and the tower and the y-axis based on the distance;

    Calculate the angle between the vehicle's driving direction and the y-axis;

    According to the quadrant where the vehicle is located, calculate the horizontal rotation angle of the gimbal tracking.
15. The tower inspection method according to claim 14, wherein the calculation process of the optimal tracking pitch angle comprises:

    The preset compensation point position is used as the real-time position of the vehicle, and based on the preset compensation point position and pole tower position, a three-dimensional space coordinate system of the vehicle pole is constructed;

    Calculate the distance between the vehicle and the tower, and the difference between the height of the tower head and the height of the on-board inspection equipment components;

    Based on the distance and the height difference, the pitch angle of the gimbal is calculated in the three-dimensional space coordinate system of the vehicle pole.
16. A pole tower inspection method based on the vehicle-mounted intelligent inspection robot system for distribution network according to claim 1, wherein the method includes the following processes:

    Obtain the geographical coordinate information of the distribution line tower in the set area;

    According to the coordinate information of the distribution line towers, considering the constraints of the inspection path, every two levels of towers are interrelated as a minimum unit, and the same tower is included between two adjacent minimum units; traversing all the towers, all the smallest units form Rod-rod smallest unit inspection model;

    Based on the inspection model, the next-level inspection tower associated with the current tower is automatically determined, and the task-free inspection of the power distribution line is performed.
17. The tower inspection method according to claim 16, characterized in that,

    All the smallest elements form the rod-rod smallest element inspection model, including:

    Considering the limiting factors of the inspection path, select the next-level tower based on the minimum unit determination principle to form a minimum unit for mutual correlation;

    Take the determined next-level tower as the current tower, and continue to select the next-level tower of the current tower to form another minimum unit;

    By analogy, all poles and towers are traversed to obtain the pole-tower minimum unit inspection model.
18. The tower inspection method according to claim 16, characterized in that,

    The principles for determining the smallest unit include:

    The principle of priority in the same line, the towers in the same line are given priority to keep the smallest unit in the line;

    Adjacent and close to the principle of distance, the first and last cross-line towers give priority to the towers of adjacent lines or other lines with closer distances;

    Based on the principle of the same person in charge, priority is given to selecting towers within the area of the same person in charge to form the smallest unit.
19. The tower inspection method according to claim 16, characterized in that,

    Based on the inspection model, the task-free inspection of distribution lines is carried out, and the specific process includes:

    The set tower is used as the starting tower, and the tower that forms the smallest unit with the starting tower is used as the tower for the next inspection;

    After reaching the next level of towers, in addition to the upper level of towers,

    If there is only one smallest tower unit in the next-level tower, the tower that forms the smallest tower unit with it will be used as the tower for inspection at the next level;

    If there are two or more minimum tower units in the next-level tower, select the tower that is closest to the inspection equipment and has not been inspected as the tower for the next-level inspection;

    If there is no minimum tower unit in the next-level tower, the inspection ends.
20. The tower inspection method according to claim 16, characterized in that,

    The pole-rod minimum unit inspection model is stored in the inspection vehicle, and the determined coordinate information of the next-level inspection pole tower is used as the target position of the inspection vehicle to plan the vehicle running path;

    The inspection vehicles record the positions of the towers that have been inspected at the end of the last inspection. In this inspection, the towers that have not been inspected are given priority to be inspected as the towers of the next inspection.
21. A method for dynamically tracking poles and towers based on the vehicle-mounted intelligent inspection robot system for distribution network according to claim 1, comprising:

    According to the set data acquisition frequency, obtain the current position information of the vehicle and the driving speed of the vehicle;

    Calculate the driving distance of the vehicle in the interval between two data calculations, and calculate the coordinate data of the preset compensation point position based on the current position information of the vehicle;

    Based on the coordinate data of the preset compensation point position, calculate the optimal tracking horizontal rotation angle and pitch angle of the cloud platform to the tower;

    The control platform realizes the detection of the tower based on the optimal tracking horizontal angle and pitch angle.
22. The tower dynamic tracking method as claimed in claim 21, comprising:

    The described acquisition of the traveling speed of the vehicle specifically includes:

    Obtain the current geographic information coordinate data of the vehicle based on the geographic information positioning device;

    Calculate the real-time driving speed of the vehicle according to the change distance of the vehicle coordinate positions obtained twice adjacently.
23. The tower dynamic tracking method as claimed in claim 21, comprising:

    The calculation of the coordinate data of the preset compensation point position based on the current position information of the vehicle specifically includes:

    According to the obtained vehicle driving speed and data calculation frequency, calculate the driving distance of the vehicle in the time interval between two adjacent data acquisitions, use the current vehicle position coordinates, calculate the driving distance within the superimposed interval time, and obtain the preset compensation point The coordinate data of the location.

Images