阅读说明：本技术 一种基于双目视觉的输电杆塔倾斜的测量方法 (Binocular vision-based transmission tower inclination measurement method ) 是由 乔铁柱 焦晨浩 阎高伟 吕玉祥 杨毅 于 2019-08-16 设计创作，主要内容包括：本发明一种基于双目视觉的输电杆塔倾斜的测量方法,属于机器视觉、电力系统在线监测技术领域；由于输电线路覆盖地貌复杂多变,对输电杆塔的安全性要求越来越高,人工定时巡检费时费力,利用倾斜传感器测量杆塔倾斜数据,不能够完全反应倾斜杆塔实地情况；本发明提供一种基于双目视觉的输电杆塔倾斜的测量方法,双目相机标定后,根据对旋转、尺度缩放、亮度变化保持不变性的杆塔特征点计算其三维世界坐标,由三维坐标计算出杆塔倾斜角和倾斜方位角,若杆塔处于异常状态,将倾斜的杆塔信息传输至后台监控室；远程监控室实时全面了解杆塔信息,有利于实现自动化安全监控,具有重大的社会意义。(The invention relates to a binocular vision-based transmission tower inclination measuring method, belonging to the technical field of machine vision and electric power system online monitoring; the power transmission line coverage landform is complex and changeable, the requirement on the safety of a power transmission tower is higher and higher, manual timing inspection wastes time and labor, and the inclination sensor is used for measuring the inclination data of the tower, so that the field situation of the inclined tower cannot be completely reflected; the invention provides a binocular vision-based method for measuring the inclination of a transmission tower, which comprises the steps of calculating three-dimensional world coordinates of the transmission tower according to tower characteristic points with rotation, scale scaling and unchanged brightness change after a binocular camera is calibrated, calculating the inclination angle and the inclination azimuth angle of the tower according to the three-dimensional coordinates, and transmitting the information of the inclined tower to a background monitoring room if the tower is in an abnormal state; the remote monitoring room can comprehensively know the tower information in real time, is favorable for realizing automatic safety monitoring and has great social significance.)

1. A binocular vision-based transmission tower inclination measurement method is characterized by comprising the following steps: the method comprises the following steps:

s10: calibrating a binocular CCD camera consisting of two cameras, and acquiring a tower image through the calibrated binocular CCD camera;

s20: extracting feature points for maintaining invariance to rotation, scale scaling and brightness change from the tower left camera image and the tower right camera image acquired in S10; completing feature point matching capable of reflecting the same point in a sceneWorking, selecting a pair of matched characteristic points belonging to the tower body of the tower as target observation points, and calculating the three-dimensional coordinates (X) of the target observation points in a world coordinate system through the pixel coordinate points of the selected target observation points in the left camera image and the right camera image_W，Y_W，Z_W)；

S30: by the three-dimensional coordinates (X) of the feature points in the world coordinate system of S20_W，Y_W，Z_W) Calculating an inclination angle R and an inclination azimuth angle B of the transmission tower, and specifically implementing the following steps:

s31: the tower inclination angle R is found by using the formula (8) based on the three-dimensional coordinates of the target observation point obtained in the process of S20,

s32: in the monitoring process, if the inclination angle R of the tower is equal to the inclination angle R of the tower in the normal state₀The difference is more than 15 degrees, the inclination azimuth angle B is calculated according to the three-dimensional coordinates of the target observation point, and if the inclination angle R of the tower and the inclination angle R of the tower in the normal state are larger than the same, the inclination angle R of the tower is smaller than the inclination angle R of the tower in the normal state₀The difference is less than 15 °, skipping this step, the tilt azimuth B is calculated as (9):

s40: and (4) judging the tower state according to the inclined azimuth angle B: if the difference value of the inclination azimuth angle B is more than 15 degrees, the tower is in a dangerous state, early warning is sent to a monitoring background, and the inclination angle R, the inclination azimuth angle B and the captured real-time image are transmitted to the monitoring background; the inclination azimuth angle B is less than 15 degrees, the tower is in a normal state, and the monitoring is continued.

2. The binocular vision based transmission tower inclination measurement method according to claim 1, wherein: the specific operation of calibrating the binocular CCD camera in S10 is as follows:

s11: placing a chessboard pattern in a binocular CCD visual field;

s12: respectively observing the chessboard patterns placed in the visual field from five different angles, and acquiring an image of one chessboard pattern at each angle;

s13: and detecting the corner points on the chessboard image to finish the camera calibration.

3. The binocular vision based transmission tower inclination measurement method according to claim 1, wherein: in S20, three-dimensional coordinates (X) of the target viewpoint in the world coordinate system are acquired_W，Y_W，Z_W) The method comprises the following specific operations:

s21, in the same world coordinate system, establishing a left camera coordinate system C by respectively taking two cameras as reference origins_LAnd the right camera coordinate system C_RAnd using the world coordinate system C_WExpressed, as in equation (2):

R_L、t_Lis an external parameter of the left camera, R_R、t_RIs an extrinsic parameter of the right camera, the two camera coordinate systems are represented by formula (3) and formula (4):

respectively representing the rotation and translation relations of the left camera relative to the right camera;

s22, adding an auxiliary camera to fill in the depth of field information Zc,

wherein (x)_L，y_L) Is the coordinates of the projection of the target feature point on the left camera imaging plane, (x)_R，y_R) Is the coordinates of the projection of the target feature point on the right camera imaging plane, f_xL、f_yLIs the focal length of the left camera in pixels, f_xR、f_yRIs the focal length of the right camera in pixels;

s23, calculating to obtain the three-dimensional coordinates of the target characteristic points by using a formula (6),

obtaining three-dimensional coordinates (X) of the target characteristic point in a world coordinate system according to a formula (7)_W，Y_W，Z_W)，

(X_C，Y_C，Z_C) Is the coordinate of the target feature point in the left camera coordinate system, R_L、t_LIs an external parameter of the left camera.

Technical Field

The invention belongs to the technical field of machine vision and on-line monitoring of an electric power system, and relates to a binocular vision-based method for measuring the inclination of a transmission tower.

Background

With the rapid development of economy, high-voltage transmission lines distributed nationwide are more and more. Because the ground condition environment covered by the transmission line tends to be complex, the requirement on the safety of the transmission tower is higher and higher, and therefore, the research on the state detection system of the tower is necessary. In order to monitor the inclination angles of the towers in adverse geological areas such as debris flow, landslide regions, coal mine mining subsidence areas and the like in real time, relevant responsible personnel can regularly patrol the inclination conditions of the towers in the responsible areas every day, and the towers in abnormal states can call maintainers to repair the towers in time, so that the method is time-consuming and labor-consuming; the monitoring host collects tower inclination data measured by the inclination sensor, the monitoring host is connected with the sensor through RS485 under the specific condition that the power supply mode is that the solar cell supplies power, the data is transmitted to the monitoring host from the sensor through RS485, and finally the data is transmitted to a back-end system through 3G/4G by the monitoring host. Therefore, how to simply, reliably, quickly and automatically detect the inclination condition of the tower is a key technical problem for preventing the overhead line accident caused by the inclination of the tower.

Disclosure of Invention

The invention aims to provide a binocular vision-based method for measuring the inclination of a transmission tower, and solves the problems that the existing method is large in workload, complex in device and poor in monitoring effect when used for monitoring the tower angle in real time.

The technical scheme adopted by the invention is that the binocular vision-based method for measuring the inclination of the transmission tower comprises the following steps:

s10: calibrating a binocular CCD camera consisting of two cameras, and acquiring a tower image through the calibrated binocular CCD camera.

S20: extracting feature points for maintaining invariance to rotation, scale scaling and brightness change from the tower left camera image and the tower right camera image acquired in S10; completing the matching work of the feature points which can reflect the same point in the scene, selecting a pair of matched feature points which belong to the tower body of the tower as target observation points, and using the selected target observation points to obtain a left camera image and a right camera imageThe three-dimensional coordinates (X) of the target observation point in the world coordinate system are obtained from the pixel coordinate points in (1)_W，Y_W，Z_W)。

S30: by the three-dimensional coordinates (X) of the feature points in the world coordinate system of S20_W，Y_W，Z_W) Calculating an inclination angle R and an inclination azimuth angle B of the transmission tower, and specifically implementing the following steps:

s31: the tower inclination angle R is found by using the formula (8) based on the three-dimensional coordinates of the target observation point obtained in the process of S20,

s32: in the monitoring process, if the inclination angle R of the tower is equal to the inclination angle R of the tower in the normal state₀The difference is more than 15 degrees, the inclination azimuth angle B is calculated according to the three-dimensional coordinates of the target observation point, and if the inclination angle R of the tower and the inclination angle R of the tower in the normal state are larger than the same, the inclination angle R of the tower is smaller than the inclination angle R of the tower in the normal state₀The difference is less than 15 °, skipping this step, the tilt azimuth B is calculated as (9):

s40: and (4) judging the tower state according to the inclined azimuth angle B: if the difference value of the inclination azimuth angle B is more than 15 degrees, the tower is in a dangerous state, early warning is sent to a monitoring background, and the inclination angle R, the inclination azimuth angle B and the captured real-time image are transmitted to the monitoring background; the inclination azimuth angle B is less than 15 degrees, the tower is in a normal state, and the monitoring is continued.

Further, the specific operation of calibrating the binocular CCD camera in S10 is as follows:

s11: a checkerboard picture was placed in the binocular CCD field of view.

S12: the checkerboard images placed in the field of view are viewed from five different angles, respectively, and an image of one checkerboard image is taken at each angle.

S13: and detecting the corner points on the chessboard image to finish the camera calibration.

Further, the target is acquired in S20Three-dimensional coordinates (X) of a viewpoint in a world coordinate system_W，Y_W，Z_W) The method comprises the following specific operations:

s21, in the same world coordinate system, establishing a left camera coordinate system C by respectively taking two cameras as reference origins_LAnd the right camera coordinate system C_RAnd using the world coordinate system C_WExpressed, as in equation (2):

R_L、t_Lis an external parameter of the left camera, R_R、t_RIs an extrinsic parameter of the right camera, the two camera coordinate systems are represented by formula (3) and formula (4):

representing the rotational and translational relationships of the left camera relative to the right camera, respectively.

S22, adding an auxiliary camera to fill in the depth of field information Zc,

wherein (x)_L，y_L) Is the coordinates of the projection of the target feature point on the left camera imaging plane, (x)_R，y_R) Is the coordinates of the projection of the target feature point on the right camera imaging plane, f_xL、f_yLIs the focal length of the left camera in pixels, f_xR、f_yRIs the focal length of the right camera in pixels.

S23, calculating to obtain the three-dimensional coordinates of the target characteristic points by using a formula (6),

obtaining three-dimensional coordinates (X) of the target characteristic point in a world coordinate system according to a formula (7)_W，Y_W，Z_W)，

(X_C，Y_C，Z_C) Is the coordinate of the target feature point in the left camera coordinate system, R_L、t_LIs an external parameter of the left camera.

Compared with the prior system, the invention has the advantages of less equipment, simple structure and low cost. The method can fully utilize the image processing technology to carry out remote global monitoring on the inclination condition of the power transmission line tower, and overcomes a plurality of defects of the existing system. The method has the greatest characteristic that a manager can know the inclination angle of the tower and the inclination direction of the tower, and finally, the manager can evaluate the safety of the tower according to the real-time tower image. Therefore, the method is beneficial to realizing automatic safety monitoring, thereby having great social significance.

Drawings

FIG. 1 is a checkerboard diagram for binocular CCD camera calibration;

FIG. 2 is a mathematical model for calculating the tower inclination angle and inclination azimuth angle; establishing a world three-dimensional coordinate system by taking a part where the tower is connected with the ground as a coordinate origin;

fig. 3 is a schematic diagram of a binocular vision-based transmission tower inclination measurement method.

In the figure, OT is the mathematical model of the inclined tower, T is the characteristic point of the tower body and the target observation point, R is the inclination angle, B is the inclination azimuth angle, and in the figure 3, three coordinate systems are provided, with the optical center O of the left camera_LCamera coordinate system with origin of coordinates, right camera optical center O_RCamera coordinate system as coordinate origin and world seat with inclined tower and ground connection point O as coordinate originA label system; the projection points of the characteristic point T of the tower body on the left image plane and the right image plane are respectively E_LAnd E_RPoints, the Z-axis of the left camera coordinate system and the Z-axis of the right camera coordinate system passing through the respective image planes at M_LAnd M_RAnd (4) point.

Detailed Description

The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. The invention relates to a binocular vision-based transmission tower inclination measuring method, which is implemented according to the following steps:

s10: and calibrating a binocular CCD consisting of two cameras, and acquiring a tower image through the calibrated binocular CCD camera.

The camera calibration is specifically implemented according to the following steps:

s11: placing the chessboard diagram shown in fig. 1 in a plane of which the world coordinate system Z is 0 as shown in fig. 2, and observing the chessboard diagram in the shooting field by using a binocular CCD;

s12: the angular points on the chessboard are observed from five different directions respectively, and at least three pictures are selected theoretically to solve internal parameters and external parameters, wherein the internal parameters comprise focal length and main image point coordinates, and the external parameters comprise translation variables and rotation variables. However, the error of the internal and external parameters obtained by selecting three pictures is relatively large, and in addition, it needs to be particularly noted that in at least three pictures, the optical axes of the cameras corresponding to any two pictures cannot be parallel, and five pictures are generally selected to solve the internal and external parameters.

S13: calculating internal parameters and external parameters of a camera, wherein the coordinates of corner points in the chessboard and actual three-dimensional coordinates are required, the corner points in each chessboard accord with the characteristic of coplanarity but non-collinearity, and on the basis of the characteristic, the plane where the chessboard is located is assumed to be a horizontal plane, namely Z is 0; the solved internal parameter matrix does not depend on the view of the scene, and the calculated internal parameters can be reused as long as the focal length of the camera is fixed. When solving the external parameters, the rotation angles of a plurality of points are taken and averaged to obtain the rotation angle with the precision meeting the measurement and calculation requirements, the external parameters are solved, the coordinates of the corner points in the chessboard pattern in the world coordinate system can be obtained, and the formula for calculating the internal and external parameters of the binocular CCD camera is as follows:

sm '═ A [ R | t ] M' or

Wherein (X, Y, Z) and M' both represent world coordinates of any corner point in the chessboard diagram, wherein Z is 0; both (u, upsilon) and m' represent coordinates of angular points projected on an image plane, and pixel points are taken as units; s is a calibration coefficient, A is called a camera matrix, or an intrinsic parameter matrix, [ R | t]Is an extrinsic parameter matrix composed of extrinsic parameters. (c)_x,c_y) Is a reference point, usually in the center of the image, (f)_x，f_y) Is the focal length in pixels.

S20: extracting feature points from the left camera image and the right camera image of the tower acquired through S10, wherein the feature points meet the requirements of rotation, scale scaling and invariant brightness change, completing matching operation on the feature points with the same attribute on the two images, and selecting the selected feature points to be the tower body of the tower, wherein the feature points have multiple groups, only the tower exists in an imaging view field, the virtual background does not have the feature points, selecting any pair of matched feature points as a target observation point, and calculating the three-dimensional coordinates of the target observation point in a world coordinate system through pixel coordinate points of the target observation point on the corresponding image.

And observing the condition of the target tower through the calibrated binocular CCD camera to obtain the three-dimensional coordinates of the tower characteristic points. However, one camera can only acquire planar two-dimensional information, and an auxiliary camera needs to be added to fill depth of field information in order to acquire three-dimensional coordinates of tower feature points. Both cameras are in the same world coordinate system, and the formula (2) is used for passing through the world coordinate system C_WRepresenting the left camera coordinate system C_LAnd the right camera coordinate system C_R。

The coordinate systems of the two cameras are mutually expressed by using a formula (3) and a formula (4), the depth of field information Zc is calculated by using a formula (5), then the three-dimensional coordinates of the target feature point of the tower can be calculated according to a formula (6), and the world coordinate system of the feature point is solved according to a formula (7)Three-dimensional coordinates (X) of (2)_W，Y_W，Z_W)。

S30: on the basis of the characteristic points obtained in step S20, an inclination angle R and an inclination azimuth B of the transmission tower are calculated, and the calculation is specifically performed according to the following steps:

s31: according to the three-dimensional coordinates of the target observation point obtained in the S20 process, the tower inclination angle R is calculated by using a formula (8), and a model is measured according to the inclination azimuth angle B of the tower shown in the figure 2;

s32: in the monitoring process, the inclination angle of the tower in a normal state is assumed to be R₀If the tower has the angle of inclination R and R₀If the difference is larger than 15 degrees, namely the tower is in an abnormal state, the inclination azimuth angle B is calculated by using a formula (9) according to the three-dimensional coordinate of the tower.

S40: and (4) judging the tower state according to the inclined azimuth angle B: if the difference value of the inclination azimuth angle B is more than 15 degrees, the tower is in a dangerous state, early warning is sent to a monitoring background, and the inclination angle R, the inclination azimuth angle B and the captured real-time image are transmitted to the monitoring background; the inclination azimuth angle B is less than 15 degrees, the tower is in a normal state, and the monitoring is continued.

The above description is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments, and all technical solutions belonging to the idea of the present invention belong to the protection scope of the present invention. It should be noted that modifications and embellishments within the scope of the invention may occur to those skilled in the art without departing from the principle of the invention, and are considered to be within the scope of the invention.