CN115240087A - Tree barrier positioning analysis method and system based on binocular stereo vision and laser point cloud - Google Patents
Tree barrier positioning analysis method and system based on binocular stereo vision and laser point cloud Download PDFInfo
- Publication number
- CN115240087A CN115240087A CN202210760831.3A CN202210760831A CN115240087A CN 115240087 A CN115240087 A CN 115240087A CN 202210760831 A CN202210760831 A CN 202210760831A CN 115240087 A CN115240087 A CN 115240087A
- Authority
- CN
- China
- Prior art keywords
- tree
- point cloud
- data
- barrier
- image
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/17—Terrestrial scenes taken from planes or by drones
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
- G01S13/86—Combinations of radar systems with non-radar systems, e.g. sonar, direction finder
- G01S13/867—Combination of radar systems with cameras
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/70—Determining position or orientation of objects or cameras
- G06T7/73—Determining position or orientation of objects or cameras using feature-based methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/255—Detecting or recognising potential candidate objects based on visual cues, e.g. shapes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/54—Extraction of image or video features relating to texture
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- 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
- G06T2207/10—Image acquisition modality
- G06T2207/10028—Range image; Depth image; 3D point clouds
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Remote Sensing (AREA)
- Multimedia (AREA)
- Radar, Positioning & Navigation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Software Systems (AREA)
- Medical Informatics (AREA)
- General Health & Medical Sciences (AREA)
- Databases & Information Systems (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Computer Networks & Wireless Communication (AREA)
- Image Processing (AREA)
Abstract
The invention provides a method and a system for positioning and analyzing a tree barrier based on binocular stereo vision and laser point cloud, belonging to the technical field of positioning and analyzing the tree barrier of an overhead line; the technical problem to be solved is as follows: the improvement of a tree barrier positioning analysis method based on binocular stereo vision and laser point cloud is provided; utilizing an unmanned aerial vehicle to acquire laser radar point cloud data and binocular camera images of a line channel in the range of the overhead transmission line protection area; classifying the point cloud data by using a neural network texture and shape algorithm, and preliminarily distinguishing tree barriers, wires, pole and tower point clouds and image information; then, carrying out image processing on the binocular stereoscopic vision photo by using a neural network texture and shape algorithm, and further judging the tree barrier type, the tree barrier three-dimensional coordinate position and the elevation; screening the judged tree obstacle information; drawing a three-dimensional space growth contour trend graph of the tree barrier, predicting the growth condition of the tree, and judging the risk level of the tree barrier; the positioning method is applied to positioning of the overhead line tree barrier.
Description
Technical Field
The invention provides a tree barrier positioning analysis method and system based on binocular stereo vision and laser point cloud, and belongs to the technical field of tree barrier positioning analysis methods and systems based on binocular stereo vision and laser point cloud.
Background
The safe operation of transmission overhead line often receives the serious threat of trees growth in its corridor, and especially some trees that grow soon, trunk height when safe distance is not enough, take place the tree easily and flash the trouble and trip out, cause the broken string even. Therefore, the ultra-high trees are one of the main factors influencing the safe operation of the line. In recent years, with the determination of the forest land operation right, local farmers plant a large number of fast-growing economic forests in the vacant space below the line, wherein the most planted poplar grows rapidly, the maximum annual growth amount of the poplar can reach more than 6m, the lumber height can reach 25m, the general line cannot meet the requirement of safe operation at all, and even if the line is designed to be high-span, the final height of the tree in the current design consideration is only 20m, and the requirement of safe operation cannot be met.
The long-term passageway tree barrier is apart from using subjective judgement distance to give first place to, and the error is very big, uses unmanned aerial vehicle three-dimensional modeling can improve the accuracy to centimetre level to accurate measure improves transmission line passageway fortune dimension efficiency.
The patent application with publication number CN110031818A has the following disadvantages of the method for extracting the contour line of the ground clearance from the power transmission line based on the point cloud data: 1. the method comprises the following steps of (1) carrying out point cloud data acquisition on a line by using a laser radar, and directly carrying out point cloud classification on the acquired point cloud data, wherein great misjudgment possibility exists; 2. the fitting manner of the wire is not preferable, leadLines are not just y = ax in mathematical expressions 2 The pattern of + bx + c (quadratic function), but
(cubic function), in the practical process, the calculation error of the maximum sag of the wire is extremely large and exceeds a meter level, and analysis cannot be performed at all; 3. the point cloud data can only analyze the object contour, and can not accurately judge whether the object contour is a tree or not and what kind of tree the object contour is; 4. the data collected by the laser radar are static data, the tree growth condition cannot be judged, and the early warning effect cannot be achieved. The paper 'power transmission line tree obstacle prediction model based on three-dimensional imaging laser radar technology' has the following defects: the method comprises the steps of determining the elevation division amount of tree barrier nodes by acquiring the point cloud data distribution condition of a tree barrier target of the power transmission line, and realizing extraction of the tree barrier target of the power transmission line, wherein the closer the position of the tree barrier nodes in a radar beam is to an imaging screen, the larger the node division proportion share related to the tree barrier nodes is, and the actual problems of huge analysis data amount, difficulty in predictive model analysis, incapability of effectively transmitting data and the like exist.
In conclusion, the distribution rule and the characteristic rule of the potential tree obstacle hazards of the power transmission line cannot be effectively mastered in the process of patrolling the power transmission overhead line, so that the method aims to complete the research on the growth speed rule of common trees in the governed line channel, perform spatial three-dimensional reconstruction on the trees in the range of the line protection area by matching with unmanned aerial vehicle laser radar point cloud and binocular stereoscopic vision, construct a line digital channel, calculate the distance of the tree line, and finally realize intelligent early warning through a potential tree obstacle early warning platform of the power transmission line channel.
Disclosure of Invention
In order to overcome the defects in the prior art, the invention aims to solve the technical problems that: the improvement of the tree barrier positioning analysis method based on binocular stereo vision and laser point cloud is provided.
In order to solve the technical problems, the invention adopts the technical scheme that: a tree barrier positioning analysis method based on binocular stereo vision and laser point cloud comprises the following steps:
s1: utilizing an unmanned aerial vehicle to acquire laser radar point cloud data and binocular camera images of a line channel in the range of the overhead transmission line protection area;
s2: the method comprises the steps of transmitting point cloud data of the laser radar and data of a binocular camera back to a data processing terminal in real time, classifying the point cloud data by utilizing a neural network texture and shape algorithm, and preliminarily distinguishing tree barriers, wires, pole tower point cloud and image information;
s3: analyzing the acquired binocular stereoscopic vision photo, performing image processing on the acquired image by utilizing a neural network texture and shape algorithm, and further judging the tree barrier type, the three-dimensional coordinate position and the elevation of the tree barrier;
s4: screening the judged tree barrier information, comparing the calculated position of the tree barrier with a plane where the circuit lead sag points are located and parallel to the ground, and analyzing and judging the distance of the tree barrier;
s5: according to a tree species growth cycle table corresponding to tree species and tree species distribution conditions in different areas, time, tree species, positions and boundaries are used as main elements, a tree barrier three-dimensional space growth contour trend graph is drawn, the tree growth condition is predicted, the tree barrier risk grade is judged according to the common tree growth condition and the tree line distance of a line channel, and early warning is achieved.
The step of classifying the point cloud data in step S2 is as follows:
the method comprises the steps of transmitting laser radar data and image data collected by a laser radar camera back to a data processing terminal, carrying out three-dimensional point cloud mathematical modeling, carrying out image processing on the collected image data, carrying out contour fitting on the processed image data, training by combining the size and shape of the existing trees, wires and towers with big data information deep learning, and analyzing by using neural network textures and shape algorithms to obtain tree barriers, wires and tower classifications.
The specific steps of the neural network texture and shape algorithm in the step S2 and the step S3 are as follows:
determining an identification area, randomly initializing an image template value, carrying out data coding on input data, using a network training error as a regulation result, obtaining image gray scale information, obtaining image texture and shape characteristic quantity, confirming data information, calculating a regulation value, judging whether the recalculated regulation value reaches an expected precision value, obtaining an optimal image template when the recalculated regulation value reaches the expected precision value, calculating an error, updating the image template, finishing training when a condition is met, calculating a texture energy value, outputting a result, outputting the result, if the condition is not met, recalculating the error and then updating the template until the condition is met and finishing training;
when the recalculated adjustment value does not reach the expected precision value, the subsequent steps need to be repeated after the image gray scale information is reacquired.
The image texture and shape feature quantities are calculated as follows:
for an image H (x, y), if H 1 ，H 2 ，……，H n Representing a set of templates, the texture energy component used to express texture properties in each pixel neighborhood in the image is convolved by G n ＝h*H n N =1,2, \8230, N is obtained, and if the size of the template is selected to be k × k, the texture image corresponding to the nth template is:
for each pixel position (x, y), there is a texture feature vector [ F ] 1 (x，y) F 2 (x，y) … F N (x，y)] T And after the image matrix is convolved with all templates respectively, texture energy information and shape feature identification information are obtained.
The step of screening the tree obstacle information in the step S4 is as follows:
combining a conductor catenary equation and tower object ID basic data, converting two-dimensional position data acquired by a binocular camera into three-dimensional space data corresponding to laser radar point cloud space data through an algorithm, comparing the calculated position of a tree obstacle with a plane parallel to the ground where a line conductor sag point is located, and utilizing a distance formula between a midpoint of a space coordinate system and the platform
Obtaining a distance set L = [ d ] from each barrier point to the plane where the arc point of the wire is located 1 ,d 2 ,d 3 ……]Center point P of 0 (x 0 ,y 0 ,z 0 ) The plane equation of the sag point of the lead is Ax + By + Cz + D =0 at any point in space.
The formula for judging the tree obstacle risk level in the step S5 is as follows:
in the above formula: k 1 、K 2 Are all empirical values, where K 1 The value range is more than or equal to 0 and less than or equal to K for reliable coefficient 1 ≤1；K 2 Is a seasonal coefficient, and the value range is not less than 0.75 and not more than K 2 ≤1.25。
The protection area of the overhead transmission line is within the space ranges of 10 meters on two sides of a 110 kilovolt line, 15 meters on two sides of a 220 kilovolt line and 20 meters on two sides of a 500 kilovolt line.
Further comprising S6: preparing a corresponding digital channel picture and a field picture for each hidden danger of the tree barrier; after the hidden danger of the tree obstacle is cleared up, the hidden danger information of the tree obstacle is updated in the early warning platform of the hidden danger of the tree obstacle of the power transmission line channel, an unmanned aerial vehicle is reused for secondary inspection, latest tree line distance data are acquired, and closed-loop management and control of the hidden danger of the tree obstacle are achieved.
The tree barrier positioning analysis system based on binocular stereoscopic vision and laser point cloud comprises an unmanned aerial vehicle, a local server, a data processing terminal and a tree barrier potential warning platform of a power transmission line channel, wherein a laser radar and a binocular camera are arranged on the unmanned aerial vehicle, an intelligent image data information processor and a memory are arranged inside the unmanned aerial vehicle, the intelligent image data information processor and the memory are connected with a transceiver and a user interface respectively through connecting bus interfaces, the intelligent image data information processor realizes data information acquisition, preprocessing, packaging and transmission of the laser radar and the binocular camera, the local server stores data, a computer program of a tree barrier positioning analysis method based on the binocular stereoscopic vision and the laser point cloud is arranged inside the data processing terminal, tree barriers, leads and towers are classified, tree barrier information is screened and determined, risk level is judged, and the tree barrier potential warning platform of the power transmission line channel tree barrier displays a tree barrier judgment result and a tree barrier three-dimensional space growth profile trend graph.
The early warning platform for the potential tree obstacle of the power transmission line channel records the potential tree obstacle in real time and takes the potential tree obstacle at any time; preparing a corresponding digital channel picture and a field picture for each hidden danger of the tree barrier; after the hidden danger of the tree obstacle is cleared, the hidden danger information of the tree obstacle is updated in the early warning platform of the hidden danger of the tree obstacle of the power transmission line channel, and then the unmanned aerial vehicle is used for carrying out secondary inspection to obtain the latest tree line distance data.
Compared with the prior art, the invention has the following beneficial effects:
(one) use unmanned aerial vehicle laser radar to carry out three-dimensional point cloud and model building to utilize the parallax principle to improve trees position rate of accuracy and rate of identification, improve the accuracy to centimetre level, greatly promoted tree kind of recognition effect, thereby accurate arrangement improves transmission line passageway fortune dimension efficiency.
Secondly, carrying out image processing on the acquired binocular camera image by using a texture analysis and shape feature recognition algorithm, and accurately judging the tree barrier type and the three-dimensional coordinate position;
and thirdly, comparing the calculated position of the tree barrier with a plane where the circuit wire sag points are located and parallel to the ground, analyzing and judging the distance and the risk level of the tree barrier, generating a three-dimensional space growth profile trend chart of the tree barrier, and realizing the function of early warning by researching the common tree growth conditions of a circuit channel and combining the distance of a tree line.
The early warning platform can record the hidden danger of the tree obstacle in real time and take the hidden danger at any time; preparing a corresponding digital channel picture and a field picture for each hidden danger of the tree barrier; after the hidden danger of the tree obstacle is cleared up, the hidden danger information of the tree obstacle is updated in the system, the unmanned aerial vehicle is reused for secondary inspection, the latest tree line distance data is obtained, and closed-loop control of the hidden danger of the tree obstacle is achieved.
Drawings
The invention is further described with reference to the accompanying drawings:
fig. 1 is a schematic diagram of an internal circuit structure of an unmanned aerial vehicle adopted by the present invention;
FIG. 2 is a schematic diagram of an intelligent image data processor according to the present invention;
fig. 3 is a schematic structural view of an unmanned aerial vehicle employed in the present invention;
FIG. 4 is a schematic diagram of a data transmission system according to the present invention;
FIG. 5 is a flow chart of the classification of point cloud data according to the present invention;
FIG. 6 is a world coordinate system X of the present invention W Y W Z W Camera coordinate system X T Y T Z T A schematic diagram of transformation relation between image coordinate systems xy;
FIG. 7 is a schematic view of an imaging model of the binocular camera of the present invention;
FIG. 8 is a flow chart of a neural network texture and shape algorithm of the present invention;
FIG. 9 is a schematic diagram of the distance used to calculate wire sag and tree barrier according to the present invention;
FIG. 10 is a schematic representation of the present invention predicting tree growth;
in the figure: 1 is binocular camera, 2 is lidar, 3 is unmanned aerial vehicle.
Detailed Description
As shown in fig. 1-10, the tree obstacle positioning analysis system based on binocular stereo vision and laser point cloud provided by the invention adopts an unmanned aerial vehicle 3 to mount a laser radar 2 and a binocular camera 1, and an intelligent image data information processor comprises: the device comprises a data acquisition module, a data determination module, a data calculation module and a data processing module; the method comprises the following steps of utilizing an unmanned aerial vehicle to conduct laser radar point cloud data acquisition and binocular camera image acquisition on a line channel in a protection area range of a power transmission line, wherein the power transmission line protection area specifically refers to the space ranges of 10 meters on two sides of a 110 kilovolt line, 15 meters on two sides of a 220 kilovolt line and 20 meters on two sides of a 500 kilovolt line; the collected data comprises information of trees, wires and towers.
The system of the hardware part of the invention realizes the functions of data information acquisition, preprocessing, packaging, transmission and the like by using an intelligent image data information processor embedded in a front-end system of the unmanned aerial vehicle. The intelligent image data information processor comprises: the device comprises a data acquisition module, a data determination module, a data calculation module and a data processing module. Unmanned aerial vehicle carry device fixes in the unmanned aerial vehicle frame, and its effect can carry the front end equipment, is lidar and binocular camera respectively. As shown in fig. 1-3.
The software part of the invention mainly comprises: the data of the laser radar point cloud and the binocular camera are transmitted back to a data processing terminal in real time, the laser radar point cloud data are firstly classified, and tree barriers, wires, pole tower point clouds and image information are preliminarily distinguished by utilizing a texture analysis and shape feature recognition algorithm; analyzing and judging binocular stereoscopic vision positioning pictures acquired by the binocular cameras, and performing image processing on the acquired images by using texture analysis and shape feature recognition algorithms to obtain accurate tree barrier zone positions, lead positions, tower positions, tree barrier types and elevations; then screening tree obstacle information, judging the distance between a tree and a wire and a tower by using laser radar point cloud, identifying whether the tree exists by using a binocular camera, and improving the accuracy and identification rate of the tree position by using a parallax principle.
According to the tree species growth cycle table corresponding to the tree species and the tree species distribution condition in the Jincheng area, the tree barrier three-dimensional space growth contour trend graph is drawn by taking time, the tree species, the position, the boundary and the like as main elements, and the tree growth condition is predicted. The contents such as a piecewise function, a reliability coefficient, a seasonal coefficient and the like are introduced, the tree obstacle condition is judged, early warning information is obtained, a standing book is formed, the operation and maintenance work of a line is guided, the hidden danger of the tree obstacle is found in the process of unmanned aerial vehicle routing inspection twice in one year, and the hidden danger is timely processed. The data storage module, the data transmission module and the data processing module can perform the above function calculation on the data, but are not limited to the structure content of the design drawing.
Point cloud data classification
The method comprises the steps of firstly collecting point cloud data by using a laser radar, obtaining omnidirectional data and image information, returning a data processing terminal, namely a power transmission line channel tree barrier hidden danger early warning platform, preprocessing the point cloud data and images, coding the point cloud data, classifying the collected point cloud data through texture analysis and a shape feature recognition algorithm, distinguishing the point cloud information of the tree barrier, a lead and a tower, obtaining spatial position data, and marking the spatial position data as green (the tree barrier), silver gray (the lead) and dark gray (the tower). The data transmission process is shown in fig. 4, and the point cloud data classification is shown in fig. 5.
Binocular camera imaging model
The binocular camera calibration can acquire the depth information of a calibration object in a three-dimensional space, and the geometric shape of the surface of the object is reconstructed through the parallax principle. In the imaging model, each coordinate is a world coordinate system X W 、Y W 、Z W Camera coordinate system X T 、Y T 、Z T An image coordinate system xy. The transformation relationship between the respective coordinate systems is shown in fig. 6, in which the world coordinate system is converted into the camera coordinate system by the calibration transformation, and the camera coordinate system is converted into the image coordinates by the projection transformation.
Let the distance between the Z-axes of the left and right cameras be the baseline distance and be denoted as B, the focal length as f, and the binocular camera imaging model is shown in fig. 7. The coordinate of the point P in the world coordinate system is P (X) W ，Y W ，Z W ) The coordinate of which is P (X) in the camera coordinate system T ，Y T ，Z T ) The transformation relationship between the two coordinates is as follows:
in the formula: r is a 3 × 3 rotation matrix and is orthogonal; t is a translation vector of 3 × 1; the relationship between the camera coordinate system and the world coordinate system can be represented by R and T, and R and T vectors can be determined by calibration with RTK space position data of the unmanned aerial vehicle.
Object P shot at the same time in camera coordinate system X T Y T Z T Lower index P (X) T ，Y T ，Z T ) For points on the left and right cameras, p Left side of (x Left side of ,y Left side of )、p Right side (x Right side ,y Right side ). The left and right cameras are horizontally facing, y Left side of ＝y Right side = y, left-right image parallax D = x Left side of -x Right side . The coordinates of the feature points under the camera three-dimensional coordinate system are obtained as follows:
(III) neural network texture and shape algorithm
The method comprises the steps of calculating the distribution situation of image texture energy in an area by utilizing a template (namely a training set), obtaining gray change information, obtaining texture and shape characteristic quantities, and comparing the texture and shape characteristic quantities with known tree information to obtain tree information. For an image H (x, y), if H 1 ，H 2 ，……，H n Representing a set of templates, the texture energy component used to express texture properties in each pixel neighborhood in the image can be represented by convolution G n ＝h*H n N =1,2, \8230 \ 8230, and N is obtained. If the size of the template is k × k, the texture image corresponding to the nth template is:
thus for each pixel position (x, y) there is a texture feature vector [ F ] 1 (x，y) F 2 (x，y) … F N (x，y)] T And after the image matrix and all templates are respectively convolved, texture energy information and shape feature identification information can be obtained.
(IV) distance from each point of tree barrier to plane of conductor sag point
Combining the catenary equation of the conductor and the basic data of the tower object ID (the formula is
Wherein k and C 1 And C 2 The catenary model coefficient can be obtained by using basic data of tower real object ID; the basic data of the tower real object ID comes from the electronic data of the PMS2.0 system), and binocular shooting is carried outConverting two-dimensional position data acquired by the head into three-dimensional space data corresponding to the point cloud space data of the laser radar through an algorithm, comparing the calculated position of the tree obstacle with a plane parallel to the ground where the arc point of the line conductor is located, and utilizing a distance formula between the point in a space coordinate system and the platform
Obtaining a distance set L = [ d ] from each point of the tree obstacle to the plane where the arc point of the wire is located 1 ,d 2 ,d 3 ……]Center point P of 0 (x 0 ,y 0 ,z 0 ) At any point in space, the plane is Ax + By + Cz + D =0.
(V) drawing a tree barrier three-dimensional space growth contour trend chart
According to a tree species growth cycle table corresponding to tree species and the tree species distribution situation of the Jincheng area, time, tree species, positions, boundaries and the like are used as main elements, three months are used as one growth cycle, three growth cycles generate a tree barrier outline trend graph once, and as shown in fig. 10, if the innermost circle outline is a tree barrier position after three months, the middle outline is a tree barrier position after six months, and the outermost circle outline is a tree barrier position after nine months, a tree barrier three-dimensional space growth outline trend graph is drawn, and the tree growth situation is predicted. The following table 1 shows the growth cycle of common tree species, and the table 2 shows the tree species distribution in the Shanxi Jincheng region;
table 1 growth cycle of common tree species;
table 2 table of distribution of tree species in shanxi jin city.
(VI) generating tree growth early warning information (piecewise function)
In the formula, K 1 、K 2 Are all empirical values, where K 1 The value range is more than or equal to 0 and less than or equal to K for reliable coefficient 1 ≤1；K 2 Is a seasonal coefficient, and the value range is not less than 0.75 and not more than K 2 ≤1.25。
In the daily operation and maintenance work of the power transmission line, in order to realize the fine management of the hidden danger of the tree obstacle, the invention utilizes the point cloud data acquisition of the laser radar and the image acquisition of the binocular camera to establish a platform for early warning the hidden danger of the tree obstacle of the power transmission line channel, establish a tree obstacle analysis model, obtain the tree species through texture analysis and shape feature recognition algorithm, draw the three-dimensional outline of the tree growth according to the growth speed of each tree species, judge the risk level of the tree obstacle, provide control measures and provide effective suggestions for the information of a tree obstacle database.
It should be noted that, regarding the specific structure of the present invention, the connection relationship between the modules adopted in the present invention is determined and can be realized, except for the specific description in the embodiment, the specific connection relationship can bring the corresponding technical effect, and the technical problem proposed by the present invention is solved on the premise of not depending on the execution of the corresponding software program.
Finally, it should be noted that: the above embodiments are only used to illustrate the technical solution of the present invention, and not to limit the same; while the invention has been described in detail and with reference to the foregoing embodiments, it will be understood by those skilled in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some or all of the technical features may be equivalently replaced; and the modifications or the substitutions do not make the essence of the corresponding technical solutions depart from the scope of the technical solutions of the embodiments of the present invention.
Claims (10)
1. A tree barrier positioning analysis method based on binocular stereo vision and laser point cloud is characterized by comprising the following steps: the method comprises the following steps:
s1: utilizing an unmanned aerial vehicle to acquire laser radar point cloud data and binocular camera images of a line channel in the range of the overhead transmission line protection area;
s2: the method comprises the steps of transmitting point cloud data of the laser radar and data of a binocular camera back to a data processing terminal in real time, classifying the point cloud data by using neural network texture and shape algorithms, and preliminarily distinguishing tree barriers, wires, point clouds of poles and towers and image information;
s3: analyzing the acquired binocular stereoscopic vision photo, performing image processing on the acquired image by utilizing a neural network texture and shape algorithm, and further judging the tree barrier type, the three-dimensional coordinate position and the elevation of the tree barrier;
s4: screening the judged tree barrier information, comparing the calculated position of the tree barrier with a plane where the circuit lead sag points are located and parallel to the ground, and analyzing and judging the distance of the tree barrier;
s5: according to a tree species growth cycle table corresponding to tree species and tree species distribution conditions in different areas, time, tree species, positions and boundaries are taken as main elements, a three-dimensional space growth contour trend graph of the tree barrier is drawn, the tree growth condition is predicted, the tree barrier risk grade is judged according to the common tree growth condition and the tree line distance of a line channel, and early warning is achieved.
2. The tree barrier positioning analysis method based on binocular stereo vision and laser point cloud according to claim 1, wherein the tree barrier positioning analysis method comprises the following steps: the step of classifying the point cloud data in step S2 is as follows:
the method comprises the steps of transmitting laser radar data and image data collected by a laser radar camera back to a data processing terminal, carrying out three-dimensional point cloud mathematical modeling, carrying out image processing on the collected image data, carrying out contour fitting on the processed image data, training by combining the size and shape of the existing trees, wires and towers with big data information deep learning, and analyzing by using neural network textures and shape algorithms to obtain tree barriers, wires and tower classifications.
3. The tree barrier positioning analysis method based on binocular stereo vision and laser point cloud according to claim 1 or 2, characterized in that: the neural network texture and shape algorithm in the step S2 and the step S3 comprises the following specific steps:
determining an identification area, randomly initializing an image template value, carrying out data coding on input data, using a network training error as a regulation target, obtaining image gray information, obtaining image texture and shape characteristic quantity, confirming data information, calculating a regulation value, judging whether the recalculated regulation value reaches an expected precision value, obtaining an optimal image template when the calculated regulation value reaches the expected precision value, calculating an error, updating the image template, finishing training when a condition is met, calculating a texture energy value, outputting a result, outputting the result, if the condition is not met, recalculating the error, then updating the template until the condition is met, and finishing training;
when the recalculated adjustment value does not reach the expected precision value, the subsequent steps need to be repeated after the image gray scale information is reacquired.
4. The tree barrier positioning analysis method based on binocular stereo vision and laser point cloud according to claim 3, wherein the tree barrier positioning analysis method comprises the following steps: the image texture and shape feature quantities are calculated as follows:
for an image H (x, y), if H 1 ，H 2 ，……，H n Representing a set of templates, the texture energy component used to express texture properties in each pixel neighborhood in the image is convolved by G n ＝h*H n N =1,2, \8230, N, where the size of the template is k × k, the texture image corresponding to the nth template is:
for each pixel position (x, y), there is a texture feature vector [ F ] 1 (x、y) F 2 (x，y)…F N (x，y)] T And after the image matrix is convolved with all templates respectively, texture energy information and shape feature identification information are obtained.
5. The tree barrier positioning analysis method based on binocular stereo vision and laser point cloud according to claim 1, wherein the tree barrier positioning analysis method comprises the following steps: the step of screening the tree obstacle information in the step S4 is as follows:
combining a conductor catenary equation and tower object ID basic data, converting two-dimensional position data acquired by a binocular camera into three-dimensional space data corresponding to laser radar point cloud space data through an algorithm, comparing the calculated position of a tree obstacle with a plane parallel to the ground where a line conductor sag point is located, and utilizing a distance formula between a midpoint of a space coordinate system and the platform
Obtaining a distance set L = [ d ] from each barrier point to the plane where the arc point of the wire is located 1 ,d 2 ,d 3 ……]Center point P of 0 (x 0 ,y 0 ,z 0 ) The plane equation of the sag point of the lead is Ax + By + Cz + D =0 at any point in space.
6. The tree barrier positioning analysis method based on binocular stereo vision and laser point cloud according to claim 5, wherein the tree barrier positioning analysis method comprises the following steps: the formula for judging the tree obstacle risk level in the step S5 is as follows:
in the above formula: k is 1 、K 2 Are all empirical values, where K 1 The value range is more than or equal to 0 and less than or equal to K for reliable coefficient 1 ≤1；K 2 Taking values for seasonal factorsK is within the range of 0.75-K 2 ≤1.25。
7. The tree barrier positioning analysis method based on binocular stereo vision and laser point cloud according to claim 1, wherein the tree barrier positioning analysis method comprises the following steps: the protection area of the overhead transmission line is within the space ranges of 10 meters on two sides of a 110 kilovolt line, 15 meters on two sides of a 220 kilovolt line and 20 meters on two sides of a 500 kilovolt line.
8. The tree barrier positioning analysis method based on binocular stereo vision and laser point cloud according to claim 1, wherein the tree barrier positioning analysis method comprises the following steps: further comprising S6: preparing a corresponding digital channel picture and a field picture for each hidden danger of the tree barrier; after the hidden danger of the tree obstacle is cleared up, the hidden danger information of the tree obstacle is updated in the early warning platform of the hidden danger of the tree obstacle of the power transmission line channel, an unmanned aerial vehicle is reused for secondary inspection, latest tree line distance data are acquired, and closed-loop management and control of the hidden danger of the tree obstacle are achieved.
9. Tree barrier positioning analysis system based on binocular stereo vision and laser point cloud, its characterized in that: the tree obstacle positioning analysis method based on the binocular stereoscopic vision and the laser point cloud comprises an unmanned aerial vehicle, a local server, a data processing terminal and a power transmission line channel tree obstacle early warning platform, wherein a laser radar and a binocular camera are arranged on the unmanned aerial vehicle, an intelligent image data information processor and a memory are arranged inside the unmanned aerial vehicle, the intelligent image data information processor and the memory are connected with a transceiver and a user interface respectively through a connecting bus interface, the intelligent image data information processor is used for achieving data information acquisition, preprocessing, packaging and transmission of the laser radar and the binocular camera, the local server is used for storing data, a computer program of the tree obstacle positioning analysis method based on the binocular stereoscopic vision and the laser point cloud is arranged inside the data processing terminal, the tree obstacle, a lead and a tower are classified, tree obstacle information is screened and determined, risk level is judged, and the power transmission line channel tree obstacle early warning platform displays a tree obstacle judgment result and a tree obstacle three-dimensional space growth profile trend graph.
10. The tree barrier positioning analysis system based on binocular stereo vision and laser point cloud of claim 9, wherein: the early warning platform for the tree obstacle hidden danger of the power transmission line channel records the tree obstacle hidden danger in real time and takes the tree obstacle hidden danger at any time; preparing a corresponding digital channel picture and a corresponding field picture for each potential tree obstacle; after the hidden danger of the tree obstacle is cleared, the hidden danger information of the tree obstacle is updated in the early warning platform of the hidden danger of the tree obstacle of the power transmission line channel, and then the unmanned aerial vehicle is used for carrying out secondary inspection to obtain the latest tree line distance data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210760831.3A CN115240087A (en) | 2022-06-29 | 2022-06-29 | Tree barrier positioning analysis method and system based on binocular stereo vision and laser point cloud |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210760831.3A CN115240087A (en) | 2022-06-29 | 2022-06-29 | Tree barrier positioning analysis method and system based on binocular stereo vision and laser point cloud |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115240087A true CN115240087A (en) | 2022-10-25 |
Family
ID=83671202
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210760831.3A Pending CN115240087A (en) | 2022-06-29 | 2022-06-29 | Tree barrier positioning analysis method and system based on binocular stereo vision and laser point cloud |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115240087A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116592851A (en) * | 2023-07-13 | 2023-08-15 | 中国电力科学研究院有限公司 | Three-dimensional monitoring method and system for power transmission channel tree obstacle satellite fused with line body |
| CN116918593A (en) * | 2023-09-14 | 2023-10-24 | 众芯汉创(江苏)科技有限公司 | Binocular vision unmanned image-based power transmission line channel tree obstacle monitoring system |
| CN117554942A (en) * | 2024-01-11 | 2024-02-13 | 江苏翰林正川工程技术有限公司 | Method and device for monitoring tree line distance of transmission line based on millimeter wave radar |
| CN118003339A (en) * | 2024-04-08 | 2024-05-10 | 广州三拓智能装备有限公司 | Robot sorting control algorithm based on artificial intelligence |
-
2022
- 2022-06-29 CN CN202210760831.3A patent/CN115240087A/en active Pending
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116592851A (en) * | 2023-07-13 | 2023-08-15 | 中国电力科学研究院有限公司 | Three-dimensional monitoring method and system for power transmission channel tree obstacle satellite fused with line body |
| CN116592851B (en) * | 2023-07-13 | 2023-09-19 | 中国电力科学研究院有限公司 | Three-dimensional monitoring method and system for power transmission channel tree obstacle satellite fused with line body |
| CN116918593A (en) * | 2023-09-14 | 2023-10-24 | 众芯汉创(江苏)科技有限公司 | Binocular vision unmanned image-based power transmission line channel tree obstacle monitoring system |
| CN116918593B (en) * | 2023-09-14 | 2023-12-01 | 众芯汉创(江苏)科技有限公司 | Binocular vision unmanned image-based power transmission line channel tree obstacle monitoring system |
| CN117554942A (en) * | 2024-01-11 | 2024-02-13 | 江苏翰林正川工程技术有限公司 | Method and device for monitoring tree line distance of transmission line based on millimeter wave radar |
| CN117554942B (en) * | 2024-01-11 | 2024-03-22 | 江苏翰林正川工程技术有限公司 | Method and device for monitoring tree line distance of transmission line based on millimeter wave radar |
| CN118003339A (en) * | 2024-04-08 | 2024-05-10 | 广州三拓智能装备有限公司 | Robot sorting control algorithm based on artificial intelligence |
| CN118003339B (en) * | 2024-04-08 | 2024-06-07 | 广州三拓智能装备有限公司 | Robot sorting control algorithm based on artificial intelligence |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115240087A (en) | Tree barrier positioning analysis method and system based on binocular stereo vision and laser point cloud | |
| CN113902897B (en) | Training of target detection model, target detection method, device, equipment and medium | |
| WO2022188379A1 (en) | Artificial intelligence system and method serving electric power robot | |
| CN111340797A (en) | Laser radar and binocular camera data fusion detection method and system | |
| CN109141248A (en) | Pig weight measuring method and system based on image | |
| CN110349260B (en) | Automatic pavement marking extraction method and device | |
| Lou et al. | Accurate multi-view stereo 3D reconstruction for cost-effective plant phenotyping | |
| CN115187803B (en) | Positioning method for picking process of famous tea tender shoots | |
| CN115541030B (en) | Method and device for identifying blast furnace top charge level temperature distribution and storage medium | |
| CN115100741B (en) | Point cloud pedestrian distance risk detection method, system, equipment and medium | |
| CN115825984A (en) | Power transmission line inspection system and inspection method based on laser point cloud | |
| CN112906719A (en) | Standing tree factor measuring method based on consumption-level depth camera | |
| CN116739558A (en) | Tree obstacle hidden danger early warning method based on laser point cloud modeling | |
| CN113345094A (en) | Electric power corridor safety distance analysis method and system based on three-dimensional point cloud | |
| CN117409339A (en) | Unmanned aerial vehicle crop state visual identification method for air-ground coordination | |
| CN117952960B (en) | Electric power iron tower part defect detection method based on artificial intelligence | |
| CN115457001A (en) | Photovoltaic panel foreign matter detection method, system, device and medium based on VGG network | |
| CN114119606A (en) | Intelligent tree obstacle hidden danger analysis method based on visible light photo power line coloring | |
| CN118115564A (en) | Fruit tree canopy structure information measurement method and device | |
| CN109657540A (en) | Withered tree localization method and system | |
| CN117291443A (en) | Intelligent paying-off system based on multidimensional sensing technology | |
| CN116843738A (en) | Tree dumping risk assessment system and method based on TOF depth camera | |
| CN116740665A (en) | Point cloud target detection method and device based on three-dimensional cross-correlation ratio | |
| CN116721345A (en) | Morphology index nondestructive measurement method for pinus massoniana seedlings | |
| CN115761265A (en) | Method and device for extracting substation equipment in laser radar point cloud |
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 |