A distribution network line inspection method based on target tracking and related devices

Abstract

The present application discloses a distribution network line inspection method and related devices based on target tracking, the method comprising: collecting samples of preset line inspection images through a preset rotation matrix to obtain multiple line sample images; using the line sample images to train an initial target detector to obtain a preset target detector, the initial target detector is constructed based on the KCF algorithm; performing target detection in the current inspection image based on the preset target detector to obtain a target detection frame, the target detection frame includes key point coordinates; determining the center of mass position of the target object in the current inspection image according to the preset center of mass tracking algorithm and the target detection frame to obtain the center of mass coordinates; adjusting the position and posture of the drone according to the center of mass coordinates to complete precise focus shooting. The present application can solve the technical problem of the prior art that it is difficult to achieve automatic focus and attitude adjustment after the drone changes its position and shooting angle.

Description

Technical Field

[0001] The present application relates to the field of drone inspection technology, and in particular to a distribution network line inspection method and related devices based on target tracking. Background Art

[0002] Drones play an important role in the automatic inspection of distribution network lines. For each target inspection point, multiple photos can be taken and sent back to the control and operation platform. However, for the same target inspection point, the drone needs to adjust different angles and even distances to obtain multi-directional images in order to accurately reflect the faults or other conditions on the target line.

[0003] In the current drone inspection process, in order to ensure that the drone can obtain accurate images of the cleaned distribution network lines every time, a fixed distance shooting method is generally adopted to reduce the change of the drone's posture, facilitate focusing, and obtain clear images; however, in the actual working process, if the drone needs to change its posture to a large extent and adjust the shooting angle, there is no better mechanism to meet the needs of automatic adjustment of the drone's posture and focus. Summary of the invention

[0004] The present application provides a distribution network line inspection method and related devices based on target tracking, which are used to solve the technical problem in the prior art that it is difficult to achieve automatic focus and attitude adjustment after the drone changes its position and shooting angle.

[0005] In view of this, the first aspect of the present application provides a distribution network line inspection method based on target tracking, comprising:

[0006] Samples of preset line inspection images are collected through a preset rotation matrix to obtain multiple line sample images;

[0007] Using the line sample image to train an initial target detector to obtain a preset target detector, wherein the initial target detector is constructed based on a KCF algorithm;

[0008] Performing target detection in the current inspection image based on the preset target detector to obtain a target detection frame, wherein the target detection frame includes key point coordinates;

[0009] Determine the centroid position of the target object in the current inspection image according to a preset centroid tracking algorithm and the target detection frame to obtain the centroid coordinates;

[0010] The position and posture of the drone is adjusted according to the center of mass coordinates to complete precise focus shooting.

[0011] Preferably, determining the centroid position of the target object in the current inspection image according to a preset centroid tracking algorithm and the target detection frame to obtain the centroid coordinates includes:

[0012] Calculate a first centroid position of the target object in the current inspection image according to the target detection frame;

[0013] Acquire a second centroid position of the target object in the current inspection image based on a preset centroid tracking algorithm;

[0014] Calculating the Euclidean distance between the first centroid position and the second centroid position;

[0015] The second centroid position corresponding to the smallest Euclidean distance is selected as the updated centroid position to obtain the centroid coordinates.

[0016] Preferably, the posture of the drone is adjusted according to the centroid coordinates to complete precise focus shooting, and then further includes:

[0017] A position association relationship is constructed based on the plurality of centroid coordinates to obtain a plurality of associated annotation data.

[0018] Preferably, the position association relationship is constructed based on the plurality of centroid coordinates to obtain a plurality of associated annotation data, and then further comprises:

[0019] The associated annotation data is structured and then stored.

[0020] The second aspect of the present application provides a distribution network line inspection device based on target tracking, comprising:

[0021] A sampling module is used to collect samples of preset line inspection images through a preset rotation matrix to obtain multiple line sample images;

[0022] A training module, used for training an initial target detector using the line sample image to obtain a preset target detector, wherein the initial target detector is constructed based on a KCF algorithm;

[0023] A detection module, configured to perform target detection in the current inspection image based on the preset target detector to obtain a target detection frame, wherein the target detection frame includes key point coordinates;

[0024] A positioning module, used to determine the centroid position of the target object in the current inspection image according to a preset centroid tracking algorithm and the target detection frame, and obtain the centroid coordinates;

[0025] The adjustment module is used to adjust the posture of the drone according to the center of mass coordinates to complete precise focus shooting.

[0026] Preferably, the positioning module is specifically used for:

[0027] Calculate a first centroid position of the target object in the current inspection image according to the target detection frame;

[0028] Acquire a second centroid position of the target object in the current inspection image based on a preset centroid tracking algorithm;

[0029] Calculating the Euclidean distance between the first centroid position and the second centroid position;

[0030] The second centroid position corresponding to the smallest Euclidean distance is selected as the updated centroid position to obtain the centroid coordinates.

[0031] Preferably, it also includes:

[0032] The association module is used to construct a position association relationship based on the multiple centroid coordinates to obtain multiple associated annotation data.

[0033] Preferably, it also includes:

[0034] The storage module is used to store the associated annotation data after structured processing.

[0035] A third aspect of the present application provides a distribution network line inspection device based on target tracking, the device comprising a processor and a memory;

[0036] The memory is used to store program code and transmit the program code to the processor;

[0037] The processor is used to execute the distribution network line inspection method based on target tracking described in the first aspect according to the instructions in the program code.

[0038] A fourth aspect of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium is used to store program code, and the program code is used to execute the distribution network line inspection method based on target tracking described in the first aspect.

[0039] It can be seen from the above technical solutions that the embodiments of the present application have the following advantages:

[0040] In the present application, a distribution network line inspection method based on target tracking is provided, including: collecting samples of preset line inspection images through a preset rotation matrix to obtain multiple line sample images; using the line sample images to train an initial target detector to obtain a preset target detector, and the initial target detector is constructed based on a KCF algorithm; based on the preset target detector, target detection is performed in the current inspection image to obtain a target detection frame, and the target detection frame includes key point coordinates; according to a preset center of mass tracking algorithm and the target detection frame, the center of mass position of the target object in the current inspection image is determined to obtain the center of mass coordinates; according to the center of mass coordinates, the posture of the drone is adjusted to complete precise focus shooting.

[0041] The present application provides a distribution network line inspection method based on target tracking. The target detector obtained through training performs target positioning on the current image acquired. The target detection frame can clearly identify the approximate position of the target object, and the preset center of mass tracking algorithm can determine the exact position of the target object. The position and posture of the drone are adjusted based on the change of the center of mass coordinates, and then the target object is focused to obtain a clear target image. Therefore, the present application can solve the technical problem of the prior art that it is difficult to achieve automatic focus and attitude adjustment after the drone changes its position and shooting angle. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 A flow chart of a distribution network line inspection method based on target tracking provided in an embodiment of the present application;

[0043] Figure 2 A schematic diagram of the structure of a distribution network line inspection device based on target tracking provided in an embodiment of the present application;

[0044] Figure 3 A schematic diagram of a target object being framed by a target detection frame provided in an embodiment of the present application;

[0045] Figure 4 A schematic diagram of the centroid position structure after establishing an association provided in an embodiment of the present application. DETAILED DESCRIPTION

[0046] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0047] For easier understanding, see Figure 1 , an embodiment of a distribution network line inspection method based on target tracking provided by the present application includes:

[0048] Step 101: Sample collection is performed on preset line inspection images through a preset rotation matrix to obtain a plurality of line sample images.

[0049] The cyclic rotation of the rotation matrix is ​​completed by cycling each layer in the rotation matrix separately. For example, when a certain layer is cyclically rotated once, each element in the layer will replace its adjacent element in the counterclockwise direction. The number of cycles k and the direction of the cycle can be configured according to actual needs and are not limited here. In this embodiment, the preset rotation matrix is ​​used to sample the line inspection image, which means that the image is moved up and down by different pixels to obtain a new sample image, that is, the line sample image, which is also the result obtained after the preset number of cycles. This sampling method can directly increase the number of samples to ensure the data requirements for research and analysis.

[0050] Step 102: Use the line sample image to train the initial target detector to obtain a preset target detector, where the initial target detector is constructed based on the KCF algorithm.

[0051] The KCF (Kernel Correlation Filter) algorithm is a tracking method with recognition capabilities. This method generally trains a target detector during the tracking process, uses the target detector to detect whether the predicted position of the next frame is the target, and then uses the new detection results to update the training set and then update the target detector. When training the target detector, the target area is generally selected as a positive sample, and the surrounding area of ​​the target is a negative sample. Of course, the closer the area to the target is, the greater the possibility of being a positive sample. Line sample images can be divided into positive samples and negative samples, and the target detector is trained at the same time to improve its recognition ability.

[0052] Step 103: Perform target detection in the current inspection image based on a preset target detector to obtain a target detection frame, which includes key point coordinates.

[0053] See also Figure 3 The target detection frame can frame the target object in the current inspection image to track the position of the target object. The framed target object is marked with coordinates to obtain the specific outline of the target object. The key point coordinates of the target detection frame generally refer to the coordinates of the four corners.

[0054] Step 104: determine the centroid position of the target object in the current inspection image according to a preset centroid tracking algorithm and a target detection frame to obtain the centroid coordinates.

[0055] Furthermore, step 104 includes:

[0056] Calculate the first centroid position of the target object in the current inspection image according to the target detection frame;

[0057] Acquire the second centroid position of the target object in the current inspection image based on a preset centroid tracking algorithm;

[0058] Calculating the Euclidean distance between the first centroid position and the second centroid position;

[0059] The second centroid position corresponding to the smallest Euclidean distance is selected as the updated centroid position to obtain the centroid coordinates.

[0060] The preset center of mass tracking algorithm can process the bounding box and key point coordinates of each target object in each frame image, thereby calculating the center of mass position of each target object, and then performing distance analysis with the center of mass position determined by the target detection box. The executions with the closest distance are regarded as the same target object, and the second execution position is used as the updated center of mass position, which is more accurate and reliable.

[0061] Step 105: Adjust the position of the drone according to the center of mass coordinates to complete precise focus shooting.

[0062] After accurate object positioning, the drone’s position and posture can be adjusted according to the positioning data. Even if the angle or position of the drone changes significantly, causing the target object to be lost, the target object on the distribution line can be accurately photographed through precise target tracking and drone posture adjustment.

[0063] Furthermore, step 105 further includes:

[0064] A position association relationship is constructed based on multiple centroid coordinates to obtain multiple associated annotation data.

[0065] See also Figure 4 ,Multiple centroid coordinates can determine the position transformation of a specific target object, and ,establishing an association relationship can ensure that the UAV can accurately focus and shoot the same target object.

[0066] Furthermore, a position association relationship is constructed based on the multiple centroid coordinates to obtain multiple associated annotation data, and then it also includes:

[0067] The associated annotation data is stored after being structured. The storage of associated annotation data is structured storage, which contains a variety of data information and can reflect the relevance of the photographed objects. The stored data can reduce the pressure of later image screening and image annotation work and improve work efficiency.

[0068] The embodiment of the present application provides a distribution network line inspection method based on target tracking. The target detector obtained through training performs target positioning on the current image acquired. The target detection frame can clearly identify the approximate position of the target object, and the preset center of mass tracking algorithm can determine the exact position of the target object. The position and posture of the drone are adjusted based on the change of the center of mass coordinates, and then the target object is focused to obtain a clear target image. Therefore, the embodiment of the present application can solve the technical problem of the prior art that it is difficult to achieve automatic focus and attitude adjustment after the drone changes its position and shooting angle.

[0069] For easier understanding, see Figure 2 The present application provides an embodiment of a distribution network line inspection device based on target tracking, including:

[0070] The sampling module 201 is used to collect samples of the preset line inspection images through a preset rotation matrix to obtain multiple line sample images;

[0071] A training module 202 is used to train an initial target detector using a line sample image to obtain a preset target detector, where the initial target detector is constructed based on a KCF algorithm;

[0072] A detection module 203 is used to perform target detection in the current inspection image based on a preset target detector to obtain a target detection frame, where the target detection frame includes key point coordinates;

[0073] A positioning module 204 is used to determine the centroid position of the target object in the current inspection image according to a preset centroid tracking algorithm and a target detection frame to obtain the centroid coordinates;

[0074] The adjustment module 205 is used to adjust the posture of the drone according to the center of mass coordinates to achieve precise focus shooting.

[0075] Furthermore, the positioning module 204 is specifically configured to:

[0076] Calculate the first centroid position of the target object in the current inspection image according to the target detection frame;

[0077] Acquire the second centroid position of the target object in the current inspection image based on a preset centroid tracking algorithm;

[0078] Calculating the Euclidean distance between the first centroid position and the second centroid position;

[0079] The second centroid position corresponding to the smallest Euclidean distance is selected as the updated centroid position to obtain the centroid coordinates.

[0080] Furthermore, it also includes:

[0081] The association module 206 is used to construct a position association relationship based on multiple centroid coordinates to obtain multiple associated annotation data.

[0082] Furthermore, it also includes:

[0083] The storage module 207 is used to store the associated annotation data after structured processing.

[0084] For ease of understanding, the present application also provides a distribution network line inspection device based on target tracking, the device including a processor and a memory;

[0085] The memory is used to store the program code and transmit the program code to the processor;

[0086] The processor is used to execute the distribution network line inspection method based on target tracking in the above method embodiment according to the instructions in the program code.

[0087] For ease of understanding, the present application also provides a computer-readable storage medium, which is used to store program code, and the program code is used to execute the distribution network line inspection method based on target tracking in the above method embodiment.

[0088] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0089] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0090] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0091] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for executing all or part of the steps of the method described in each embodiment of the present application through a computer device (which can be a personal computer, a server, or a network device, etc.). The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (full name in English: Read-Only Memory, English abbreviation: ROM), random access memory (full name in English: Random Access Memory, English abbreviation: RAM), disk or optical disk and other media that can store program codes.

[0092] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A distribution network line inspection method based on target tracking, characterized in that: include: Samples of preset line inspection images are collected through a preset rotation matrix to obtain multiple line sample images; Using the line sample image to train an initial target detector to obtain a preset target detector, wherein the initial target detector is constructed based on a KCF algorithm; Performing target detection in the current inspection image based on the preset target detector to obtain a target detection frame, wherein the target detection frame includes key point coordinates; Determine the centroid position of the target object in the current inspection image according to a preset centroid tracking algorithm and the target detection frame to obtain the centroid coordinates; The position and posture of the drone is adjusted according to the center of mass coordinates to complete precise focus shooting.

2. The distribution network line inspection method based on target tracking according to claim 1 is characterized in that: Determining the centroid position of the target object in the current inspection image according to the preset centroid tracking algorithm and the target detection frame to obtain the centroid coordinates includes: Calculate a first centroid position of the target object in the current inspection image according to the target detection frame; Acquire a second centroid position of the target object in the current inspection image based on a preset centroid tracking algorithm; Calculating the Euclidean distance between the first centroid position and the second centroid position; The second centroid position corresponding to the smallest Euclidean distance is selected as the updated centroid position to obtain the centroid coordinates.

3. The distribution network line inspection method based on target tracking according to claim 1 is characterized in that: The method further includes adjusting the posture of the drone according to the centroid coordinates to complete precise focus shooting, and then: A position association relationship is constructed based on the plurality of centroid coordinates to obtain a plurality of associated annotation data.

4. The distribution network line inspection method based on target tracking according to claim 3 is characterized in that: The position association relationship is constructed based on the plurality of centroid coordinates to obtain a plurality of associated annotation data, and then the method further includes: The associated annotation data is structured and then stored.

5. A distribution network line inspection device based on target tracking, characterized in that: include: A sampling module is used to collect samples of preset line inspection images through a preset rotation matrix to obtain multiple line sample images; A training module, used for training an initial target detector using the line sample image to obtain a preset target detector, wherein the initial target detector is constructed based on a KCF algorithm; A detection module, configured to perform target detection in the current inspection image based on the preset target detector to obtain a target detection frame, wherein the target detection frame includes key point coordinates; A positioning module, used to determine the centroid position of the target object in the current inspection image according to a preset centroid tracking algorithm and the target detection frame, and obtain the centroid coordinates; The adjustment module is used to adjust the posture of the drone according to the center of mass coordinates to complete precise focus shooting.

6. The distribution network line inspection device based on target tracking according to claim 5 is characterized in that: The positioning module is specifically used for: Calculate a first centroid position of the target object in the current inspection image according to the target detection frame; Acquire a second centroid position of the target object in the current inspection image based on a preset centroid tracking algorithm; Calculating the Euclidean distance between the first centroid position and the second centroid position; The second centroid position corresponding to the smallest Euclidean distance is selected as the updated centroid position to obtain the centroid coordinates.

7. The distribution network line inspection device based on target tracking according to claim 5 is characterized in that: Also includes: The association module is used to construct a position association relationship based on the multiple centroid coordinates to obtain multiple associated annotation data.

8. The distribution network line inspection device based on target tracking according to claim 7 is characterized in that: Also includes: The storage module is used to store the associated annotation data after structured processing.

9. A distribution network line inspection device based on target tracking, characterized in that: The device comprises a processor and a memory; The memory is used to store program code and transmit the program code to the processor; The processor is used to execute the distribution network line inspection method based on target tracking according to any one of claims 1-4 according to the instructions in the program code.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium is used to store program code, and the program code is used to execute the distribution network line inspection method based on target tracking as described in any one of claims 1-4.

Images