CN114552601A - Binocular vision power transmission line oscillation monitoring and three-dimensional reconstruction method - Google Patents
Binocular vision power transmission line oscillation monitoring and three-dimensional reconstruction method Download PDFInfo
- Publication number
- CN114552601A CN114552601A CN202210182164.5A CN202210182164A CN114552601A CN 114552601 A CN114552601 A CN 114552601A CN 202210182164 A CN202210182164 A CN 202210182164A CN 114552601 A CN114552601 A CN 114552601A
- Authority
- CN
- China
- Prior art keywords
- power transmission
- image
- transmission line
- points
- point
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J3/00—Circuit arrangements for ac mains or ac distribution networks
- H02J3/24—Arrangements for preventing or reducing oscillations of power in networks
- H02J3/241—The oscillation concerning frequency
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/246—Analysis of motion using feature-based methods, e.g. the tracking of corners or segments
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/80—Analysis of captured images to determine intrinsic or extrinsic camera parameters, i.e. camera calibration
- G06T7/85—Stereo camera calibration
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30241—Trajectory
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T90/00—Enabling technologies or technologies with a potential or indirect contribution to GHG emissions mitigation
Abstract
The method introduces binocular stereo vision into monitoring of power transmission line oscillation, firstly calibrates a binocular stereo vision system to obtain internal and external parameters, prepares for three-dimensional point reconstruction, analyzes and splits Harris and SIFT corner point detection methods, fuses respective advantages, provides an optimized scale space SIFT link fusion corner point detection method, further provides a method for matching and combining the optimized scale space SIFT link fusion corner point detection method with MQS images based on analysis of MQS approximation degrees associated with pure scalar products, has obviously improved matching accuracy and speed compared with the traditional method, calculates a model of a three-dimensional space point, calculates three-dimensional coordinates of a spacer bar center point in a power transmission line scene, is free from temperature interference, simple and reliable in signal transmission and relatively low in cost, develops an automatic early-warning power grid power transmission line monitoring system, and has a good realization effect on the three-dimensional space point reconstruction system, has high engineering application value.
Description
Technical Field
The application relates to a power transmission line oscillation monitoring and three-dimensional reconstruction method, in particular to a binocular vision power transmission line oscillation monitoring and three-dimensional reconstruction method, and belongs to the technical field of power transmission line oscillation monitoring.
Background
The transmission line is an artery of a large-scale power grid, and the benefit and responsibility of the power department are directly related to the continuity, stability and safety of the transmission power of the power department. When the power transmission line is exposed to the natural environment for a long time, when the natural environment is severe, such as weather conditions of large day and night temperature difference, strong wind, acid rain, thunder, freezing and the like, part of the power transmission line can be greatly damaged, such as various physical and mechanical damages of inclination and collapse of a tower, short circuit of a power transmission wire, pollution flashover of an insulator and the like, so that the whole section or the whole power system can not normally operate or stop operating, because the power transmission line not only bears dead load, ice coating and the like, but also bears dynamic load generated by wind.
The power line can be in three modes in wind-induced vibration, namely breeze vibration, the frequency is 3-150Hz, the amplitude is 0.01-1d, d is the diameter of a power transmission wire and occurs at lower wind speed; secondly, oscillation occurs at a high wind speed with the frequency of 0.08-3Hz and the amplitude of 5-300 d; third is wake-induced oscillation, frequency 0.15-1OHz, amplitude centered on the wind speed that occurred. The electric transmission line oscillation caused by natural disasters has great harm to an electric power system, and in winter, when wind in the horizontal direction blows to a power transmission and transmission wire which is changed into a non-circular cross section due to ice coating, certain aerodynamic force is generated to induce the power transmission wire to generate a low-frequency and large-amplitude self-excited vibration phenomenon. In 1957 to 1994, 51 times of power transmission line oscillations with different strengths occur in China, 161 power transmission lines are affected in total, 69 power transmission lines are damaged, 6 power transmission lines are broken, and line short circuit trips for more than 119 times.
The damage of the power transmission line oscillation to the power transmission line is very large, and the monitoring and early warning of the power transmission line oscillation are particularly important. If three-dimensional information such as the position, the posture, the geometric shape and the like of the oscillation characteristic point of the power transmission line can be obtained in real time, the oscillation of the power transmission line can be analyzed and explained so as to better help technicians to detect and identify the motion trail of the characteristic point of the power transmission line and adopt corresponding measures to ensure the safety of the power transmission line. Therefore, there is an urgent need for an effective means for helping technicians identify transmission lines to reduce the oscillation loss of various transmission lines, reconstruct the positions, real-time changes and the like of characteristic points in the transmission lines, complete the analysis and understanding of the transmission lines, make a correct judgment, and finally determine whether to alarm and take corresponding measures to prevent dangers.
Computer vision is to use a single or multiple cameras and related devices such as computers to detect, identify, track and express objects in surrounding scenes by approximately replacing human eyes and brains, and then to process a single or multiple images, so that the images become images more convenient for human observation and expression. So that the computer can visually observe and analyze the external surrounding scene like a human.
The binocular stereo vision realizes the three-dimensional reconstruction of objects through two-dimensional images, which is a hotspot in the field of computer vision, two cameras are adopted to simulate two eyes to observe surrounding scene objects, specifically, two cameras positioned at different positions acquire target two-dimensional images at the same time or the same position, so as to acquire the three-dimensional coordinate information of the scene objects, and then reasonable analysis and judgment are made according to the three-dimensional coordinates or the change of the three-dimensional coordinates. The accurate recovery of the three-dimensional information of the objects in the surrounding scene is beneficial to detecting and identifying the required target information. Therefore, the computer binocular stereo vision technology is used for detecting and identifying the oscillation characteristics of the power transmission line, three-dimensional reconstruction is achieved, the motion trail of the oscillation characteristic spacer of the power transmission line can be analyzed, early warning is achieved, reference is provided for departments such as a power grid, and therefore losses of natural disasters to the power transmission line and the power industry are avoided or reduced.
In the prior art, some corner detection methods are provided, but after general classification, most methods are relatively long in time, the methods are relatively complex to reflect from the side, the number of the obtained corners is small under the condition of the same threshold value, the requirements of power transmission line oscillation monitoring are not met, and the real-time performance is relatively poor. Some methods not only consume long time, but also have harsh conditions, need to obtain angular points from different scale spaces, and obtain a large number of points under the same threshold value, wherein many points are useless points, and the large number of points can cause great difficulty for subsequent matching, and the prior art lacks an angular point detection method suitable for power transmission line oscillation monitoring.
In summary, the power transmission line oscillation monitoring and three-dimensional reconstruction methods in the prior art have many disadvantages, and the difficulties and problems to be solved in the present application mainly focus on the following aspects:
firstly, the damage of the power transmission line oscillation to the power transmission line is very large, and the monitoring and early warning of the power transmission line oscillation are particularly important, the prior art cannot accurately acquire the three-dimensional information of the position, the posture, the geometric shape and the like of the power transmission line oscillation characteristic point in real time, cannot analyze and explain the power transmission line oscillation in real time, and brings great difficulty for technicians to detect and identify the motion track of the power transmission line characteristic point and adopt corresponding measures to ensure the safety of the power transmission line, at present, an effective means capable of helping technicians to identify the power transmission line is urgently needed to reduce the oscillation loss of various power transmission lines, reconstruct the positions, real-time changes and the like of characteristic points in the power transmission line, complete the analysis and understanding of the power transmission line, make correct judgment, finally determine whether to alarm and take corresponding measures to prevent the occurrence of dangers, and the application needs to intensively solve the problems;
secondly, the oscillation of the power transmission line is a major hidden danger threatening the safety of the power grid, various oscillation monitoring methods in the prior art are mostly based on a GPS positioning method, an optical fiber sensor method and an acceleration sensor method, which have the defects of easy temperature interference, complex signal transmission, high cost and the like, most of the angular point detection methods in the prior art are relatively long in time and complex in method, the number of obtained angular points is very small, the method does not meet the requirement of monitoring the oscillation of the power transmission line, the real-time performance is poor, some methods not only consume long time, the angular points need to be obtained from different scale spaces, but also obtain a large number of points under the same threshold value, many of the points are useless points, great difficulty is caused to subsequent matching, the prior art lacks an angular point detection method suitable for monitoring the oscillation of the power transmission line, and based on the proposal of an intelligent power grid and the rapid development of a visual monitoring technology, the intelligent oscillation monitoring technology is applied to the operation of the power transmission line, and an automatic early warning power transmission line monitoring system of the power grid is developed, so that the intelligent oscillation monitoring system is a work with high application value;
thirdly, in the prior art, the SIFT corner detection method extracts a plurality of image extreme points, increases the amount of calculation in the image matching process, and prolongs the matching time, and is based on a detection method with unchanged scale, and the detected feature points are all extreme points in corresponding scales, so that the detected feature points can not be ensured to appear in different scales, that is, correct matching point pairs can not be obtained in the matching process; the prior art is not perfect when the approximation degree based on the correlation coefficient is used for monitoring the power transmission line oscillation, and has great defects in the aspects of accuracy, robustness and the like.
Disclosure of Invention
The method can obtain the three-dimensional information of the position, the posture, the geometric shape and the like of the oscillation characteristic point of the power transmission line in real time, can analyze and explain the oscillation of the power transmission line so as to better help technicians to detect and identify the motion trail of the characteristic point of the power transmission line, adopts corresponding measures to ensure the safety of the power transmission line, realizes the three-dimensional reconstruction of the spacer of the characteristic point of the power transmission line by using a series of methods such as binocular camera calibration, characteristic detection, characteristic matching, three-dimensional reconstruction and the like by using the principle of binocular stereo vision, thereby providing important reference for monitoring and early warning of the oscillation of the power transmission line, reducing the oscillation loss of various power transmission lines by using an effective means for helping technicians to identify the power transmission line, reconstructing the position, the real-time change and the like of the characteristic point in the power transmission line, and the analysis and understanding of the power transmission line are completed, the correct judgment is made, and finally whether the alarm is needed or not and the corresponding measures are taken to prevent the occurrence of danger are determined.
In order to achieve the effects of the above technology, the technical method adopted by the application is as follows:
a binocular vision power transmission line shock monitoring and three-dimensional reconstruction method is characterized in that binocular vision is introduced into power transmission line shock monitoring, firstly, a binocular vision system is calibrated, internal and external parameters are obtained to prepare for three-dimensional point reconstruction, a Harris and SIFT corner point detection method is analyzed and split, an optimized scale space SIFT link fusion corner point detection method is provided, a method for matching and combining the optimized scale space SIFT link fusion corner point detection method and an MQS image is further provided based on analysis of pure scalar product association MQS approximation, and finally, three-dimensional coordinates of a spacer bar center point in a power transmission line scene are calculated according to a calculation model of a three-dimensional space point provided in the binocular vision system; the specific method steps of the binocular vision power transmission line oscillation monitoring and three-dimensional reconstruction method comprise: firstly, calibrating a camera, secondly, detecting angular points of the spacers and matching images, and thirdly, three-dimensionally reconstructing the center points of the spacers;
the method comprises the steps of reconstructing three-dimensional points of a detected power transmission line, including camera calibration, obtaining parameters required by reconstruction, solving two-dimensional image coordinates of interest points in two-dimensional images obtained by a left camera and a right camera, wherein the points are characteristic points;
the binocular vision image matching is to find a corresponding point between images after angular point features of respective images are extracted by a left camera and a right camera, namely a certain feature point in one image is known, a matching point corresponding to the feature point is obtained from the other image, and projections and corresponding relations of the space point on different image planes are constructed to achieve spatial consistency.
A binocular vision power transmission line oscillation monitoring and three-dimensional reconstruction method is further a method for optimizing matching and combining a scale space SIFT link fusion corner point detection method and an MQS image: firstly, an SIFT corner point detection method is optimized to extract initial image extreme points, then secondary detection is carried out on the image extreme points extracted by the SIFT method by linking and fusing Harris corner point detection, the result after the secondary detection is used as the final image feature points, the number of the feature points is reduced, the image matching efficiency and the accuracy are improved, and the specific implementation process comprises the following steps:
the first step is as follows: constructing a scale space H (x, y, a) of the image through a DOG Gaussian difference operator, wherein the (x, y) is a space point image coordinate and represents the position of an image pixel, and a represents the mean square error of a Gaussian function;
the second step is that: detecting all local extreme points in an image scale space in a DOG Gaussian difference pyramid, and recording the detected image extreme points as initial characteristic points in the improved corner detection method;
the third step: for each initial feature point extracted by the SIFT corner point detection method, calculating a T value of a Harris corner point detection operator corresponding to the initial feature point:
Jxis a gradient in the x direction, JyIs the gradient in the y-direction;
the fourth step: after the T values of all initial feature points are calculated, the values of all local extreme points are arranged from large to small, the first M extreme points in the T value arrangement are selected as final feature points, and the value of M is 0.04 times of the number of the initial feature points.
The binocular vision power transmission line shock monitoring and three-dimensional reconstruction method is characterized in that a correlation coefficient is introduced based on analysis of MQS (magnitude of interference) approximation degrees of pure scalar product correlation, and changes of image brightness or grade are obeyed by adopting a pure scalar product correlation method;
assuming that the pixel sizes of the two images to be matched are both N M, in image J2In which image J is searched by constructing a rectangular search window of n x m1At a certain angle n1(v1,u1) In image J2Corresponding matching point n in2(v2,u2) And carrying out correlation calculation on the searched matching point pairs to obtain the pure scalar product correlation image matching defined formula (4-21):
wherein (i, J) is the two images to be matched in the respective image Jk(k is 1,2),
The MQS value is between positive and negative 1, namely-1 is larger than or equal to MQS and smaller than or equal to 1, the closer the MQS value is to 1, the higher the similarity of the two acquired images is, the MQS value is 1, the same the two acquired images is, the MQS value is-1, the two completely different images are shown, no public area is included, the MQS critical value is set autonomously, and the matching candidate feature point is considered to be higher than the designed value.
A binocular vision power transmission line oscillation monitoring and three-dimensional reconstruction method further comprises the steps of adopting an optimized scale space SIFT link fusion angular point detection method to greatly reduce the number of detected characteristic points, completing characteristic angular point matching between two images to be matched by combining the optimized scale space SIFT link fusion angular point detection method and MQS image matching, obtaining corresponding image matching point pairs, and using the corresponding image matching point pairs for subsequent reconstruction of three-dimensional space points monitored by power transmission lines. And calculating respective projection matrixes of the two corresponding cameras by combining the internal and external parameters calibrated by the cameras to obtain a three-dimensional reconstruction result of the power transmission line monitoring space point after the two images to be matched are matched in the binocular stereoscopic vision system.
The binocular vision power transmission line oscillation monitoring and three-dimensional reconstruction method comprises the following steps: the power transmission line oscillation monitoring comprises the steps that angular point detection is carried out on intersection points of a spacer and a power transmission conductor, image coordinates of the center point of the spacer are obtained through fitting, the coordinates of the center point of the spacer are finally obtained, the power transmission conductor can be indirectly tracked, the motion rule of the power transmission conductor is analyzed, oscillation of the power transmission conductor is monitored, and a series of steps of binocular vision power transmission line oscillation monitoring are buckled with one another and cannot be omitted;
the installation of two cameras of this application is not parallel structure, but there is the angle between the two in the mounting structure to on same high plane, place two cameras symmetry, and make the contained angle of optical axis and base line distance all equal to 45 degrees approximately.
A binocular vision power transmission line oscillation monitoring and three-dimensional reconstruction method is further provided, and the camera calibration step comprises the following steps: calibrating a camera to obtain internal and external parameters of the two cameras, wherein all the parameters are used in the final three-dimensional reconstruction of power transmission line oscillation monitoring;
firstly, data acquisition is carried out, a stable ground is selected, cameras are arranged, the postures of the two cameras are placed to be at about 45 degrees as much as possible, then the parameters of the cameras are adjusted to carry out synchronous shooting, 24 groups of image pairs are shot totally, graying, denoising and equal-proportion scaling are carried out on all the images to obtain processed images, the preprocessed images are input by adopting a calibration tool box, the internal and external parameters of the cameras are solved, and a calibration result is obtained.
A binocular vision power transmission line oscillation monitoring and three-dimensional reconstruction method comprises the following steps of: firstly, identifying and locking a spacer of a power transmission wire, keeping a camera still after the camera is calibrated, synchronously acquiring videos for 5-9 minutes, then converting the videos into image pairs for subsequent processing by 20 frames per second, adopting the method for matching and combining the SIFT link fusion angular point detection method and the MQS images in the optimized scale space, extracting the image pairs containing the same spacer to be detected in proportion for each processing, removing mismatched matched points by using the RANSAC method, obtaining at least three angular point image coordinates of a plurality of wire holes of the spacer and the power transmission line in the image plane of the double cameras respectively, and then carrying out circle fitting on the same spacer and a plurality of angular points of the power transmission line to obtain the image coordinate of the central point of the spacer.
A binocular vision power transmission line oscillation monitoring and three-dimensional reconstruction method is further characterized in that the specific steps of the method for optimizing the matching combination of a scale space SIFT link fusion corner point detection method and an MQS image are as follows:
step 1: determining a shooting visual angle, selecting a stable ground, arranging cameras, arranging the postures of the two cameras to enable the two cameras to be at an angle of about 45 degrees as much as possible, and preparing for power transmission line data acquisition;
step 2: calibrating a camera to obtain internal and external parameters of the two cameras, and synchronously shooting videos by the two cameras;
and 3, step 3: converting the collected video into a picture pair;
and 4, step 4: graying, denoising and scaling all the image pairs to obtain a processed image;
and 5, step 5: extracting the same part in the image pair according to the proportion;
and 6, step 6: extracting feature points by adopting an optimized scale space SIFT link fusion corner point detection method;
and 7, step 7: carrying out image matching by using an image matching method of pure scalar product correlation approximation;
and 8, step 8: RANSAC removes the mismatching;
step 9: performing circle fitting on the points matched with the left and right spacers respectively to obtain central point image coordinates;
step 10: and (5) carrying out oscillation monitoring and three-dimensional reconstruction on the transmission line, analyzing the monitoring data and obtaining a conclusion.
A binocular vision power transmission line oscillation monitoring and three-dimensional reconstruction method comprises the following steps: and finally obtaining the spatial three-dimensional coordinates of the center points of the spacers by adopting a binocular stereo imaging method after obtaining the fitted image coordinates of the left camera and the right camera at the center of the spacers, and respectively calculating the spatial three-dimensional coordinates of the center points of the spacers in each frame.
Compared with the prior art, the innovation points and advantages of the application are as follows:
firstly, the damage of the power transmission line oscillation to the power transmission line is extremely large, and the monitoring and early warning of the power transmission line oscillation are particularly important, the method provided by the application can obtain the three-dimensional information such as the position, the posture, the geometric shape and the like of the power transmission line oscillation characteristic point in real time, can analyze and explain the power transmission line oscillation so as to better help technicians to detect and identify the motion trail of the power transmission line characteristic point, adopt corresponding measures to ensure the safety of the power transmission line, and realize the three-dimensional reconstruction of the power transmission line characteristic point spacer by using a series of methods such as binocular camera calibration, characteristic detection, characteristic matching, three-dimensional reconstruction and the like by utilizing the principle of binocular stereo vision, thereby providing important reference for the power transmission line oscillation monitoring and early warning, reducing various power transmission line oscillation losses by using effective means for helping technicians to identify the power transmission line, reconstructing the position, real-time change and the like of the characteristic point in the power transmission line, the analysis and understanding of the power transmission line are completed, correct judgment is made, and finally whether alarming is needed or not and corresponding measures are taken to prevent danger;
secondly, the method introduces binocular stereo vision into the monitoring of the power transmission line oscillation, firstly calibrates a binocular stereo vision system to obtain internal and external parameters, prepares for three-dimensional point reconstruction, analyzes and splits Harris and SIFT corner point detection methods, fuses respective advantages, provides an optimized scale space SIFT link fusion corner point detection method, further provides a method for matching and combining the optimized scale space SIFT link fusion corner point detection method with MQS images based on the analysis of pure scalar product association MQS approximation degree, proves that the matching accuracy and speed of the method are remarkably improved compared with the traditional method through collected images, calculates the three-dimensional coordinates of the center point of the spacer in the power transmission line scene through a calculation model of the provided three-dimensional space point, is free from temperature interference, simple and reliable in signal transmission and relatively low in cost, and adopts an intelligent oscillation monitoring technology to the operation of the power transmission line, the power grid power transmission line monitoring system with automatic early warning is developed, the three-dimensional space point reconstruction system is better in realization effect, and the engineering application value is very high;
thirdly, the method extracts the characteristic corners in the images to be matched by comparing the advantages and the disadvantages of various classical corner detection methods and simultaneously provides an optimized scale space SIFT link fusion corner detection method and constructs a correspondingly improved image matching method in order to ensure that the extracted characteristic corners are more stable and effective, adopts an image matching method with higher accuracy and stronger robustness and a pure scalar product correlation approximation degree by combining the approximation degree of image matching, uses the obtained matched image pixel points in the reconstruction of three-dimensional space points, firstly extracts initial image extreme points by optimizing the SIFT corner detection method, then carries out secondary detection on the image extreme points extracted by the SIFT method by linking and fusing Harris corner detection, takes the result after the secondary detection as final image characteristic points, reduces the number of the characteristic points and improves the efficiency and the accuracy of image matching, the method of pure scalar product correlation is adopted to obtain a better effect, the change of image brightness or grade is overcome, the used time is short, and the method is an image matching method which can be used for monitoring power transmission line oscillation.
Drawings
Fig. 1 is a graph of experimental results of SIFT corner point detection based matching with SIFT images and RANSAC processing.
FIG. 2 is a graph of the combination of the optimized scale space SIFT chaining fusion corner point detection method and MQS and results of RANSAC processing.
FIG. 3 is a schematic diagram showing data comparison of image matching results between different methods.
Fig. 4 is a schematic diagram of a placement model between two cameras in field power line oscillation monitoring.
Fig. 5 is a schematic diagram of the graying, denoising, and scaling processes.
FIG. 6 shows instantaneous displacement DeltaX of spacer in Y-directioniAnd (5) a queue spectrum analysis result graph.
FIG. 7 is a diagram of the steps of a method for matching and combining an optimized scale space SIFT link fusion corner detection method with an MQS image.
FIG. 8 is an image coordinate v of the spacer center point in the left camera1A trend graph of the change.
FIG. 9 is an image coordinate v of the spacer center point in the right camera2A trend graph of the change.
Fig. 10 is a diagram of the two cycles of the motion trajectories of the image coordinates of the center point of the spacer in the left camera.
Fig. 11 is a diagram of the two-cycle motion trajectories of the image coordinates of the spacer center point in the right camera.
Fig. 12 is a spatial three-dimensional coordinate motion trail diagram of the center point of the power transmission line spacer.
Detailed description of the invention
The technical scheme of the binocular vision power transmission line shock monitoring and three-dimensional reconstruction method provided by the application is further described below with reference to the accompanying drawings, so that a person skilled in the art can better understand the application and can implement the application.
With the gradual expansion of the range and the scale of various power grids, the number of power grid disaster accidents is increased year by year, and the oscillation of a power transmission line becomes a great hidden danger threatening the safety of the power grid. Various oscillation monitoring methods in the prior art are mostly based on a GPS (global positioning system) positioning method, an optical fiber sensor method and an acceleration sensor method, and the methods have the defects of high possibility of being interfered by temperature, complex signal transmission, high cost and the like. Based on the intelligent power grid and the rapid development of the visual monitoring technology, the intelligent oscillation monitoring technology is applied to the operation of the power transmission line, and the power grid power transmission line monitoring system capable of automatically early warning is developed, so that the power grid power transmission line monitoring system is a work with high application value.
The binocular stereo vision observes the same scene through two visual angles, and obtains the two-dimensional information of two images, thereby recovering the three-dimensional information of an object. The application introduces binocular stereo vision into monitoring of power transmission line oscillation, firstly calibrates a binocular stereo vision system to obtain internal and external parameters for preparing for three-dimensional point reconstruction, analyzes and splits Harris and SIFT corner point detection methods, fuses respective advantages, provides an optimized scale space SIFT link fusion corner point detection method, further provides a method for matching and combining the optimized scale space SIFT link fusion corner point detection method and an MQS image based on analysis of pure scalar product association MQS approximation, proves that the matching accuracy and speed of the method are remarkably improved compared with the traditional method through collected images, and finally calculates the three-dimensional coordinates of the center point of a spacer in a power transmission line scene according to a calculation model of the three-dimensional space point provided in the binocular stereo vision system, and provides specific method steps of the binocular vision power transmission line oscillation monitoring and three-dimensional reconstruction method, the result shows that the method provided by the application has a good effect of realizing the three-dimensional space point reconstruction system and has a high engineering application value.
The method comprises the steps of reconstructing three-dimensional points of a detected power transmission line, including camera calibration, obtaining parameters required by reconstruction, solving two-dimensional image coordinates of interest points in two-dimensional images obtained by a left camera and a right camera, wherein the points are characteristic points, firstly detecting the characteristic points, and detecting by regarding intersection points of spacers and power transmission lines as angular points.
The binocular vision image matching is to find a corresponding point between images after angular point features of respective images are extracted by a left camera and a right camera, namely a certain feature point in one image is known, a matching point corresponding to the feature point is obtained from the other image, and projections and corresponding relations of the space point on different image planes are constructed to achieve spatial consistency.
Method for matching and combining optimized scale space SIFT link fusion corner point detection method and MQS image
By comparing advantages and disadvantages of various classical corner detection methods and simultaneously aiming at enabling extracted characteristic corners to be more stable and effective, an optimized scale space SIFT link fusion corner detection method is provided for extracting characteristic corners in an image to be matched and constructing a corresponding improved image matching method, an image matching method with higher accuracy and stronger robustness and pure scalar product correlation approximation is adopted in combination with the approximation of image matching, and the obtained matched image pixel point pairs are used for reconstruction of three-dimensional space points.
The SIFT corner detection method is based on a detection method with unchanged scale, the detected feature points are all extreme points in corresponding scales, the detected feature points cannot be ensured to appear in different scales, and correct matching point pairs cannot be obtained in the matching process.
Therefore, the method comprises the steps of firstly optimizing an SIFT corner detection method to extract initial image extreme points, then linking and fusing Harris corner detection to carry out secondary detection on the image extreme points extracted by the SIFT method, taking the result after the secondary detection as a final image feature point, reducing the number of the feature points, improving the image matching efficiency and accuracy, and specifically realizing the following processes:
the first step is as follows: constructing a scale space H (x, y, a) of an image by using a DOG Gaussian difference operator, wherein the (x, y) is a space point image coordinate and represents the position of an image pixel, and a represents the mean square error of a Gaussian function, and the application takes 1 to 10;
the second step: detecting all local extreme points in an image scale space in a DOG Gaussian difference pyramid, and recording the detected image extreme points as initial characteristic points in the improved corner detection method;
the third step: for each initial feature point extracted by the SIFT corner point detection method, calculating a T value of a Harris corner point detection operator corresponding to the initial feature point:
Jxis a gradient in the x direction, JyIs the gradient in the y-direction;
the fourth step: after the T values of all initial feature points are calculated, the values of all local extreme points are arranged from large to small, the first M extreme points in the T value arrangement are selected as final feature points, and the value of M is 0.04 times of the number of the initial feature points.
The correlation coefficient is one of the most widely used approximations in image matching, reflecting the degree of correlation between two images. Although the approximation degree based on the correlation coefficient is a method which is widely and popular at present, the method is not perfect and has certain influence on the aspects of accuracy, robustness and the like. Therefore, in order to improve the accuracy and enhance the robustness, a pure scalar product correlation method is adopted to obtain a better effect, and the change of the brightness or the grade of the image is overcome.
Assuming that the pixel sizes of the two images to be matched are both N M, in image J2In which image J is searched by constructing a rectangular search window of n x m1At a certain angle n1(v1,u1) In image J2Corresponding matching point n in2(v2,u2) And performing correlation calculation on the searched matching point pairs to obtain pure scalar product correlation image matching, wherein the definition of the pure scalar product correlation image matching is shown as a formula 2:
wherein (i, J) is the two images to be matched in respective image Jk(k is 1,2),
The MQS value is between positive and negative 1, namely-1 is larger than or equal to MQS and smaller than or equal to 1, the closer the MQS value is to 1, the higher the similarity of the two acquired images is, the MQS value is 1, the same the two acquired images is, the MQS value is-1, the two completely different images are shown, no public area is included, the MQS critical value is set autonomously, and the matching candidate feature point is considered to be higher than the designed value.
The MQS needs to gradually calculate the similarity of all the feature points, so that a large amount of time is consumed to obtain good precision, time delay occurs, and in order to solve the serious problems, an optimized scale space SIFT link fusion corner point detection method is adopted, so that the number of detected feature points is greatly reduced, and a large amount of time is saved.
After the feature corner matching between two images to be matched is completed by a method of combining an optimized scale space SIFT link fusion corner detection method and MQS image matching, a corresponding image matching point pair is obtained and is used for reconstructing a three-dimensional space point for subsequent power transmission line monitoring. And calculating respective projection matrixes of the two corresponding cameras by combining the internal and external parameters calibrated by the cameras to obtain a three-dimensional reconstruction result of the power transmission line monitoring space point after the two images to be matched are matched in the binocular stereoscopic vision system.
Fig. 1 is a graph of experimental results of matching with SIFT images based on SIFT corner detection and processed by RANSAC, fig. 2 is a graph of results of combining an optimized scale space SIFT link fusion corner detection method with MQS and processed by RANSAC, and fig. 3 is a graph of comparison of image matching results between different methods. As can be seen from fig. 1,2 and 3, the method for optimizing the combination of the scale space SIFT link fusion corner detection method and the MQS image matching is more accurate, more robust and shorter in time compared with the conventional SIFT corner detection and SIFT image matching method, and is an image matching method for power transmission line oscillation monitoring.
Second, binocular vision power transmission line oscillation monitoring and three-dimensional reconstruction step
The main target of power transmission line oscillation monitoring is to detect the corner point of the intersection point of a spacer and a power transmission conductor, and to fit the intersection point to obtain the image coordinate of the central point of the spacer, and finally to obtain the coordinate of the central point of the spacer, the power transmission conductor can be indirectly tracked, the motion rule of the power transmission conductor is analyzed, the oscillation of the power transmission conductor is monitored, a series of steps of binocular vision power transmission line oscillation monitoring are linked and linked, and the steps of each step are as follows.
The installation of two cameras of this application is not parallel structure, but there is the angle between the two in the mounting structure to not on same high plane, place two cameras symmetry, and make the contained angle of optical axis and base line distance all approximately equal to 45 degrees, figure 4 is the model of settling between two cameras when the field.
Camera calibration procedure
And calibrating the cameras to obtain internal and external parameters of the two cameras, wherein all the parameters are used in the final three-dimensional reconstruction of power transmission line oscillation monitoring.
Firstly, data acquisition is carried out, a stable ground is selected, cameras are arranged, the postures of the two cameras are placed to be at about 45 degrees as much as possible, then the parameters of the cameras are adjusted to carry out synchronous shooting, 24 groups of image pairs are shot in total, all the images are subjected to graying, denoising and equal-proportion scaling to obtain processed images as shown in fig. 5, the preprocessed images are input by a calibration tool box, the internal and external parameters of the cameras are solved, the specific flow is shown in fig. 6, and a calibration experiment result is obtained.
(II) spacer angular point detection and image matching step
The Harris corner detection method and the SIFT corner detection method have respective advantages and disadvantages, and the optimized scale space SIFT link fusion corner detection method is provided by the application, so that the effect is better, and matching is performed on the basis of corner detection.
In order to realize monitoring of power transmission line oscillation, firstly, a spacer of a power transmission line is identified and locked, after a camera is calibrated, the camera is kept still, videos are synchronously acquired for 5-9 minutes, then the videos are converted into image pairs at 20 frames per second for subsequent processing, the method for matching and combining SIFT link fusion corner point detection method and MQS images is adopted, and the image pairs containing the same spacer to be detected are extracted in proportion in each processing. The matching point after mismatching is removed by using a RANSAC method, because the peripheral wire holes of 8 split spacers are the same, a certain wire hole is difficult to be respectively formed in the motion process of the power transmission line, the three-dimensional reconstruction of the certain wire hole and the corner point of the power transmission line is very difficult, the wire hole track of the spacers is inaccurate, and the motion rule of the power transmission line is difficult to master, and simultaneously, because the 8 wire holes of the spacers are symmetrical, a plurality of wire holes of the spacers in a double-camera image plane and at least three corner point image coordinates of the power transmission line are respectively obtained, and then the image coordinates of the center point of the spacers are obtained by performing circle fitting on the same spacer and a plurality of corner points of the power transmission line. The specific steps are shown in FIG. 7.
The method for optimizing the matching combination of the SIFT link fusion corner point detection method and the MQS image comprises the following specific steps:
step 1: determining a shooting visual angle, selecting a stable ground, arranging cameras, arranging the postures of the two cameras to enable the two cameras to be at an angle of about 45 degrees as much as possible, and preparing for power transmission line data acquisition;
step 2: calibrating a camera to obtain internal and external parameters of the two cameras, and synchronously shooting videos by the two cameras;
and 3, step 3: converting the collected video into a picture pair;
and 4, step 4: graying, denoising and scaling all the image pairs to obtain a processed image;
and 5, step 5: extracting the same part in the image pair according to the proportion;
and 6, step 6: extracting feature points by adopting an optimized scale space SIFT link fusion corner point detection method;
and 7, step 7: carrying out image matching by using an image matching method of pure scalar product correlation approximation;
and 8, step 8: RANSAC removing mismatching;
step 9: performing circle fitting on the points matched with the left and right spacers respectively to obtain central point image coordinates;
step 10: and (5) carrying out oscillation monitoring and three-dimensional reconstruction on the transmission line, analyzing the monitoring data and obtaining a conclusion.
(III) three-dimensional reconstruction step of central point of spacer
And finally obtaining the spatial three-dimensional coordinates of the center points of the spacers by adopting a binocular stereo imaging method after obtaining the fitted image coordinates of the left camera and the right camera at the center of the spacers, and respectively calculating the spatial three-dimensional coordinates of the center points of the spacers in each frame.
Thirdly, analyzing the oscillation monitoring and three-dimensional reconstruction simulation result of the power transmission line
From the obtained image coordinates and three-dimensional coordinates of the center point of the spacer rod in each of 140 frames in the left and right cameras, the image coordinates v of the center point of the spacer rod in the left and right cameras are respectively reflected as shown in fig. 8 and 91、v2The change trend of (2) is similar to V-shaped. Since the video is converted into an image at 20 frames per second, the motion period of the spacer can be obtained from the imageIs about 60 frames, i.e. 3s, and the line spacer is stable in motion without large fluctuations in the case of a breezy weather of 3.3m/s to 5.6 m/s. Knowing the motion period of the power transmission line spacer, the motion frequency of the power transmission line spacer can be calculated to be 0.33HZ, and meanwhile, because the influence of ice coating does not exist, the vibration of the power transmission line can be judged to be the sub-span vibration.
Only the type of the vibration of the power transmission conductor can be judged according to the vibration frequency of the power transmission line from the one-dimensional image coordinate, the motion trail of two periods of the image coordinate of the central point of the spacer in the left camera is reflected in the graph 10, and the motion trail of two periods of the image coordinate of the central point of the spacer in the right camera is reflected in the graph 11. The amplitudes of the left and right image coordinates obtained from the image are 5.60 pixels and 5.62 pixels, so that the motion of the transmission line can be evaluated according to the amplitude, and the method has a very important role.
The binocular stereoscopic vision solves the problem that the two-dimensional image coordinates are not vivid, and the final purpose is to obtain the three-dimensional coordinates of the center of the spacer of the power transmission line, replace the motion condition of the whole spacer with the three-dimensional point coordinate motion of the center of the spacer and indirectly reflect the influence of the external natural environment on the power transmission line. From the change in the spacer center point world coordinate X, Y, Z, the amplitude of the X coordinate value was approximately 20cm, the amplitude of the Y coordinate value was approximately 10cm, the amplitude of the Z coordinate value was approximately 36cm, and the variation pattern of the X coordinate value was observed. From these figures and data, the actual variation of the spacer center point can be obtained. Two periodic motion tracks of the power transmission line spacer can be visually seen from fig. 12, the power transmission line in the graph is not influenced by strong wind, the motion is very smooth, the amplitude is about 0.437m, the motion state of the power transmission line is tracked in real time by an image, and technicians can predict the motion of the following power transmission line in advance according to the previous motion trend.
Claims (9)
1. The binocular vision power transmission line shock monitoring and three-dimensional reconstruction method is characterized in that binocular stereo vision is introduced into power transmission line shock monitoring, a binocular stereo vision system is calibrated firstly, internal and external parameters are obtained to prepare for three-dimensional point reconstruction, Harris and SIFT corner point detection methods are analyzed and split, an optimized scale space SIFT link fusion corner point detection method is provided, a method for matching and combining the optimized scale space SIFT link fusion corner point detection method with MQS images is further provided based on analysis of pure scalar product associated MQS approximation, and finally, three-dimensional coordinates of a spacer bar center point in a power transmission line scene are calculated according to a calculation model of a three-dimensional space point provided in the binocular vision system; the specific method steps of the binocular vision power transmission line oscillation monitoring and three-dimensional reconstruction method comprise: firstly, calibrating a camera, secondly, detecting angular points of the spacers and matching images, and thirdly, three-dimensionally reconstructing the center points of the spacers;
the method comprises the steps of reconstructing three-dimensional points of a detected power transmission line, including camera calibration, obtaining parameters required by reconstruction, solving two-dimensional image coordinates of interest points in two-dimensional images obtained by a left camera and a right camera, wherein the points are characteristic points;
the binocular vision image matching is to find a corresponding point between images after angular point features of respective images are extracted by a left camera and a right camera, namely a certain feature point in one image is known, a matching point corresponding to the feature point is obtained from the other image, and projections and corresponding relations of the space point on different image planes are constructed to achieve spatial consistency.
2. The binocular vision power transmission line shock monitoring and three-dimensional reconstruction method of claim 1, wherein a method for optimizing a scale space SIFT link fusion corner point detection method and matching and combining MQS images is as follows: firstly, an SIFT corner point detection method is optimized to extract initial image extreme points, then secondary detection is carried out on the image extreme points extracted by the SIFT method by linking and fusing Harris corner point detection, the result after the secondary detection is used as the final image feature points, the number of the feature points is reduced, the image matching efficiency and the accuracy are improved, and the specific implementation process comprises the following steps:
the first step is as follows: constructing a scale space H (x, y, a) of the image through a DOG Gaussian difference operator, wherein the (x, y) is a space point image coordinate and represents the position of an image pixel, and a represents the mean square error of a Gaussian function;
the second step is that: detecting all local extreme points in an image scale space in a DOG Gaussian difference pyramid, and recording the detected image extreme points as initial characteristic points in the improved corner detection method;
the third step: for each initial feature point extracted by the SIFT corner point detection method, calculating a T value of a Harris corner point detection operator corresponding to the initial feature point:
Jxis a gradient in the x direction, JyIs the gradient in the y-direction;
the fourth step: after the T values of all initial feature points are calculated, the values of all local extreme points are arranged from large to small, the first M extreme points in the T value arrangement are selected as final feature points, and the value of M is 0.04 times of the number of the initial feature points.
3. The binocular vision power transmission line shock monitoring and three-dimensional reconstruction method of claim 2, wherein correlation coefficients are introduced based on analysis of pure scalar product correlation MQS approximation, and a pure scalar product correlation method is adopted to accommodate changes in image brightness or level;
assuming that the pixel sizes of the two images to be matched are both N M, in image J2In which image J is searched by constructing a rectangular search window of n x m1At a certain angle n1(v1,u1) In image J2Corresponding matching point n in2(v2,u2) And carrying out correlation calculation on the searched matching point pairs to obtain the pure scalar product correlation image matching defined formula (4-21):
wherein (i, J) is the two images to be matched in the respective image Jk(k is 1,2),
the MQS value is between positive and negative 1, namely-1 is larger than or equal to MQS and smaller than or equal to 1, the closer the MQS value is to 1, the higher the similarity of the two acquired images is, the MQS value is 1, the same the two acquired images is, the MQS value is-1, the two completely different images are shown, no public area is included, the MQS critical value is set autonomously, and the matching candidate feature point is considered to be higher than the designed value.
4. The binocular vision power transmission line shock monitoring and three-dimensional reconstruction method according to claim 1, characterized in that an optimized scale space SIFT link fusion corner point detection method is adopted, the number of detected feature points is greatly reduced, after the feature corner point matching between two images to be matched is completed through the optimized scale space SIFT link fusion corner point detection method and MQS image matching combination method, corresponding image matching point pairs are obtained and are used for subsequent reconstruction of three-dimensional space points monitored by the power transmission line, and after internal and external parameters calibrated by cameras are combined, respective projection matrixes of the two corresponding cameras are calculated, and then three-dimensional reconstruction results of the power transmission line monitoring space points of the two images to be matched in the binocular stereoscopic vision system can be obtained.
5. The binocular vision power transmission line shock monitoring and three-dimensional reconstruction method according to claim 1, wherein the binocular vision power transmission line shock monitoring and three-dimensional reconstruction step comprises: the power transmission line oscillation monitoring comprises the steps that angular point detection is carried out on intersection points of a spacer and a power transmission conductor, image coordinates of the center point of the spacer are obtained through fitting, the coordinates of the center point of the spacer are finally obtained, the power transmission conductor can be indirectly tracked, the motion rule of the power transmission conductor is analyzed, oscillation of the power transmission conductor is monitored, and a series of steps of binocular vision power transmission line oscillation monitoring are buckled with one another and cannot be omitted;
the installation of two cameras of this application is not parallel structure, but there is the angle between the two in the mounting structure to on same high plane, place two cameras symmetry, and make the contained angle of optical axis and base line distance all equal to 45 degrees approximately.
6. The binocular vision power transmission line oscillation monitoring and three-dimensional reconstruction method according to claim 1, wherein the camera calibration step comprises: calibrating a camera to obtain internal and external parameters of the two cameras, wherein all the parameters are used in the final three-dimensional reconstruction of power transmission line oscillation monitoring;
firstly, data acquisition is carried out, a stable ground is selected, cameras are arranged, the postures of the two cameras are placed to be at about 45 degrees as much as possible, then the parameters of the cameras are adjusted to carry out synchronous shooting, 24 groups of image pairs are shot totally, graying, denoising and equal-proportion scaling are carried out on all the images to obtain processed images, the preprocessed images are input by adopting a calibration tool box, the internal and external parameters of the cameras are solved, and a calibration result is obtained.
7. The binocular vision power transmission line shock monitoring and three-dimensional reconstruction method according to claim 1, wherein the spacer corner point detection and image matching step comprises: firstly, identifying and locking a spacer of a power transmission wire, keeping a camera still after the camera is calibrated, synchronously acquiring videos for 5-9 minutes, then converting the videos into image pairs for subsequent processing by 20 frames per second, adopting the method for matching and combining the SIFT link fusion angular point detection method and the MQS images in the optimized scale space, extracting the image pairs containing the same spacer to be detected in proportion for each processing, removing mismatched matched points by using the RANSAC method, obtaining at least three angular point image coordinates of a plurality of wire holes of the spacer and the power transmission line in the image plane of the double cameras respectively, and then carrying out circle fitting on the same spacer and a plurality of angular points of the power transmission line to obtain the image coordinate of the central point of the spacer.
8. The binocular vision power transmission line shock monitoring and three-dimensional reconstruction method according to claim 1, wherein the specific steps of the method for optimizing the combination of a scale space SIFT link fusion corner point detection method and MQS image matching are as follows:
step 1: determining a shooting visual angle, selecting a stable ground, arranging cameras, arranging the postures of the two cameras to enable the two cameras to be at an angle of about 45 degrees as much as possible, and preparing for power transmission line data acquisition;
step 2: calibrating the cameras to obtain internal and external parameters of the two cameras, and synchronously shooting videos by the two cameras;
and 3, step 3: converting the collected video into a picture pair;
and 4, step 4: graying, denoising and scaling all the image pairs to obtain a processed image;
and 5, step 5: extracting the same part in the image pair according to the proportion;
and 6, a step of: extracting feature points by adopting an optimized scale space SIFT link fusion corner point detection method;
and 7, step 7: carrying out image matching by using an image matching method of pure scalar product correlation approximation;
and 8, step 8: RANSAC removes the mismatching;
step 9: performing circle fitting on the points matched with the left and right spacers respectively to obtain central point image coordinates;
step 10: and (5) carrying out oscillation monitoring and three-dimensional reconstruction on the transmission line, analyzing the monitoring data and obtaining a conclusion.
9. The binocular vision power transmission line shock monitoring and three-dimensional reconstruction method according to claim 1, wherein the spacer center point three-dimensional reconstruction step: and finally obtaining the spatial three-dimensional coordinates of the center points of the spacers by adopting a binocular stereo imaging method after obtaining the fitted image coordinates of the left camera and the right camera at the center of the spacers, and respectively calculating the spatial three-dimensional coordinates of the center points of the spacers in each frame.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210182164.5A CN114552601A (en) | 2022-02-25 | 2022-02-25 | Binocular vision power transmission line oscillation monitoring and three-dimensional reconstruction method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210182164.5A CN114552601A (en) | 2022-02-25 | 2022-02-25 | Binocular vision power transmission line oscillation monitoring and three-dimensional reconstruction method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114552601A true CN114552601A (en) | 2022-05-27 |
Family
ID=81679056
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210182164.5A Pending CN114552601A (en) | 2022-02-25 | 2022-02-25 | Binocular vision power transmission line oscillation monitoring and three-dimensional reconstruction method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114552601A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115731303A (en) * | 2022-11-23 | 2023-03-03 | 江苏濠汉信息技术有限公司 | Large-span transmission conductor sag three-dimensional reconstruction method based on bidirectional binocular vision |
-
2022
- 2022-02-25 CN CN202210182164.5A patent/CN114552601A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115731303A (en) * | 2022-11-23 | 2023-03-03 | 江苏濠汉信息技术有限公司 | Large-span transmission conductor sag three-dimensional reconstruction method based on bidirectional binocular vision |
| CN115731303B (en) * | 2022-11-23 | 2023-10-27 | 江苏濠汉信息技术有限公司 | Large-span transmission conductor sag three-dimensional reconstruction method based on bidirectional binocular vision |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106356757B (en) | A kind of power circuit unmanned plane method for inspecting based on human-eye visual characteristic | |
| CN110415342B (en) | Three-dimensional point cloud reconstruction device and method based on multi-fusion sensor | |
| Sohn et al. | Automatic powerline scene classification and reconstruction using airborne lidar data | |
| Meng et al. | Morphology-based building detection from airborne LIDAR data | |
| KR20200007084A (en) | Ship detection method and system based on multi-dimensional features of scene | |
| CN109870108B (en) | Method and device for detecting icing of power transmission line | |
| CN110675387B (en) | Method and device for detecting structural defects of transmission line iron tower | |
| CN108389256B (en) | Two-three-dimensional interactive unmanned aerial vehicle electric power tower inspection auxiliary method | |
| CN112001958B (en) | Virtual point cloud three-dimensional target detection method based on supervised monocular depth estimation | |
| Yang et al. | Deep learning‐based bolt loosening detection for wind turbine towers | |
| CN112686938B (en) | Power transmission line clear distance calculation and safety alarm method based on binocular image ranging | |
| CN110084785B (en) | Power transmission line vertical arc measuring method and system based on aerial images | |
| CN104966063A (en) | Mine multi-camera video fusion method based on GPU and CPU cooperative computing | |
| CN115809986A (en) | Multi-sensor fusion type intelligent external damage detection method for power transmission corridor | |
| CN114552601A (en) | Binocular vision power transmission line oscillation monitoring and three-dimensional reconstruction method | |
| CN114998245A (en) | Method for detecting galloping of power transmission line based on binocular distance measurement and image segmentation | |
| CN114758215A (en) | Distribution network live working robot grounding ring and bolt identification and positioning method thereof | |
| WO2024067435A1 (en) | Video-based multi-object displacement tracking monitoring method and apparatus | |
| CN112762841A (en) | Bridge dynamic displacement monitoring system and method based on multi-resolution depth features | |
| Wu et al. | HeightFormer: Explicit Height Modeling without Extra Data for Camera-only 3D Object Detection in Bird's Eye View | |
| CN114396921B (en) | Method for measuring tidal height and propagation speed of Yangtze river on basis of unmanned aerial vehicle | |
| CN111354028B (en) | Binocular vision-based power transmission channel hidden danger identification and tracking method | |
| CN115393387A (en) | Building displacement monitoring method and device | |
| Wang et al. | An effective method for sensing power safety distance based on monocular vision depth estimation | |
| CN112766313A (en) | Crystal segmentation and positioning method, device, equipment and medium based on U-net structure |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |