CN107564056B - Optimal data frame selection method for three-dimensional point cloud data of contact net supporting device - Google Patents
Optimal data frame selection method for three-dimensional point cloud data of contact net supporting device Download PDFInfo
- Publication number
- CN107564056B CN107564056B CN201710615521.1A CN201710615521A CN107564056B CN 107564056 B CN107564056 B CN 107564056B CN 201710615521 A CN201710615521 A CN 201710615521A CN 107564056 B CN107564056 B CN 107564056B
- Authority
- CN
- China
- Prior art keywords
- point cloud
- data frame
- data
- supporting device
- dimensional
- 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.)
- Expired - Fee Related
Links
Images
Abstract
The invention discloses a method for selecting an optimal data frame of three-dimensional point cloud data of a contact net supporting device, which comprises the steps of continuously shooting the contact net supporting device through a three-dimensional scanner arranged on a detection car roof through a mounting frame, and collecting a three-dimensional point cloud image of the contact net supporting device in real time; carrying out density statistics on point cloud data in the acquired point cloud image of the contact net supporting device, analyzing the change rule of the point cloud data, and extracting an effective data frame of the point cloud data; and carrying out three-dimensional centroid calculation on the extracted effective frame of the point cloud data, tracking and positioning the point cloud centroid, analyzing the change rule of the point cloud centroid, extracting the optimal data frame of the point cloud data, and counting the change rule by using the number of the point clouds of the contact net supporting device to improve the extraction timeliness of the optimal data frame. The invention effectively overcomes the problems of poor data validity and processing real-time performance of the system, improves the processing efficiency of the system, and better meets the requirements of high-speed contact network on real-time performance and validity of data selection in online detection.
Description
Technical Field
The invention relates to the technical field of online detection of a high-speed railway contact network, in particular to a method for selecting an optimal data frame of three-dimensional point cloud data of a contact network supporting device.
Background
With the development of high-speed electrified trains, the safety of railway transportation becomes more and more important. The contact network is one of the most important devices of a high-speed railway power supply system, and once a fault occurs, the safe operation of the railway is directly damaged. With the adoption of the image-based non-contact type overhead contact system safety protection detection system (6C system) which is the leading one of the original railway ministry, the guarantee is provided for the safe, stable and efficient operation of a railway power supply system. However, since the system mainly detects and evaluates the state of the contact network based on image data, the most direct problem is how to eliminate redundant data from massive image data and extract effective analysis data, thereby ensuring the real-time performance and accuracy of subsequent data processing and analysis and avoiding the problem of storage of a large amount of data. This is one of the key problems that currently restrict the system detection efficiency and the system integration level.
Disclosure of Invention
Aiming at the problems, the invention aims to provide a method for selecting the optimal data frame of the three-dimensional point cloud data of the high-speed railway contact net supporting device, which can remove a large amount of redundant data, reduce unnecessary system analysis and processing time, improve the timeliness of the system and realize the online detection of the system, and the technical scheme is as follows:
a method for selecting an optimal data frame of three-dimensional point cloud data of a contact net supporting device comprises the following steps:
step A: the mounting frame is arranged on a three-dimensional scanner for detecting the roof, the contact net supporting device is continuously shot, and a three-dimensional point cloud image of the contact net supporting device is collected in real time;
and B: carrying out density statistics on point cloud data in the acquired point cloud image of the contact net supporting device, analyzing the change rule of the point cloud data, and extracting an effective data frame of the point cloud data;
and C: and performing three-dimensional centroid calculation on the extracted effective frame of the point cloud data, tracking and positioning the point cloud centroid, analyzing the change rule of the point cloud centroid, and extracting the optimal data frame of the point cloud data.
Further, the specific process of step B is as follows:
establishing point cloud density function of point cloud image data frame
In the formula, PiFor the ith calculation point, Pi nearestIs a point PiNum is all the nearest points in the current point cloud image data frame
Summing the points;
setting a threshold value tau when
And judging that the current point cloud image data frame enters a detection area, and extracting the current point cloud image data frame as an effective data frame.
Further, the specific process of step C is as follows:
establishing a position function of a three-dimensional centroid of a point cloud image data frame:
in the formula, miFor the mass of each point it is,
the coordinate of the three-dimensional centroid of the point cloud image data frame on the y axis is shown, and H is the height of the image; n is the total number of points;
judging whether the position of the three-dimensional centroid of the current point cloud image data frame is in the center of the image;
and if the current point cloud image data frame is in the center of the image, extracting the current point cloud image data frame as an optimal data frame.
Further, before the step C, the method further comprises:
establishing a gradient calculation formula of point cloud number change, and calculating the point cloud number change of a data frame n in a valid data frame of point cloud data
Chemical gradient PNMGn：
In the formula, PNnIs the point cloud number, PN, of the data frame nn-1Is the point cloud number, PN, of the data frame n-1n-2Is the point cloud number of the data frame n-2;
is the time interval between data frame n and data frame n-1,
is the time interval between data frame n-1 and data frame n-2;
the Peak moment when the point cloud number has a Peak value is obtained by calculating the product of the gradients of the adjacent data framesn-1：
In the formula, PNMGn-1A point cloud number change gradient for data frame n-1;
namely: when the gradient changes from positive to negative, the product is negative, and the data frame at the moment is judged as the data frame with the point cloud data peak value;
and then, carrying out three-dimensional centroid calculation on the data frame with the point cloud data peak value to extract the optimal data frame of the point cloud data.
The invention has the beneficial effects that:
1) according to the method, the data frame appearing on the contact net supporting device is found by utilizing the density change characteristic of the three-dimensional point cloud data of the contact net supporting device and according to the corresponding relation between the point cloud density and the appearance of the contact net supporting device, so that the effective data frame of the contact net point cloud is accurately extracted, the redundant data of the system is reduced, and the effectiveness of system data processing is ensured;
2) according to the method, the optimal frame of the point cloud data of the contact net supporting device is extracted by positioning a data frame when the three-dimensional centroid of the point cloud data of the contact net supporting device appears in the right center area of the image according to the position change rule of the three-dimensional centroid of the three-dimensional point cloud data image of the contact net supporting device; the optimal point cloud data frame of the contact net supporting device is effectively positioned, the system processing time is reduced, and the system processing efficiency is greatly improved;
3) according to the method, the optimal data frame is extracted by performing three-dimensional centroid calculation on the data frame near the moment by detecting the moment of change of the point cloud data peak of the contact net supporting device by utilizing the point cloud number change rule of the three-dimensional point cloud data of the contact net supporting device. The invention reduces the processing time of the system and improves the timeliness of the system on the basis of ensuring the selection of the optimal data frame.
Drawings
FIG. 1 is a simplified diagram of a data acquisition device of the present invention.
Fig. 2 is a three-dimensional point cloud data image of the complete catenary supporting device.
Fig. 3 is a point cloud density curve diagram of the contact net supporting device.
Fig. 4a to 4d are diagrams illustrating an effect of tracking and positioning a three-dimensional centroid of a point cloud of a supporting device of an overhead line system in an embodiment, where fig. 4a is a 22 th frame, fig. 4b is a 46 th frame, fig. 4c is an 82 th frame, and fig. 4d is a 94 th frame.
Fig. 5 is a point cloud number statistical result diagram of a point cloud image of a catenary supporting device.
Detailed Description
The invention is described in further detail below with reference to the figures and specific embodiments. Fig. 1 is a schematic view of the installation of the data acquisition device of the present invention. The working principle is that a three-dimensional scanner is erected above a detection vehicle, and a three-dimensional point cloud data image of a contact net supporting device is acquired in real time. And the optimal data frame selection of the contact net three-dimensional point cloud image is realized through a three-dimensional centroid algorithm and a mathematical statistic method. Fig. 2 is a collected and extracted three-dimensional point cloud data image of the catenary supporting device.
Fig. 3 is a point cloud density curve of the contact net supporting device. Because the three-dimensional scanner is installed on a traveling detection roof, the contact net supporting device gradually appears in a point cloud data image from local to global in the shooting process of the three-dimensional scanner from the absence to the existence. The density change of the point cloud data just reflects the whole shooting process. From fig. 3, it can be seen that the point cloud density is small before the detection vehicle enters the detection area, the density gradually increases with the proximity of the detection area, and the density is large and the change tends to be stable after the detection vehicle enters the detection area. The point cloud density gradually decreases as it gradually leaves the detection area. Therefore, a threshold τ is set, and when the density is greater than τ, the detection area is considered to be entered, and a valid data frame is extracted. Point cloud density function
The definition is as follows.
Description of the above formula, Pi nearestIs a point PiOf (2) aThe proximity point is determined by the Kd-tree method. num is the sum of all points in the current point cloud data frame.
Fig. 4 a-4 d are diagrams of the contact net supporting device point cloud three-dimensional centroid tracking and positioning effect. Due to the continuity of the three-dimensional scanner acquisition system and the characteristics of traveling shooting of the three-dimensional scanner acquisition system, the contact net supporting device gradually appears and then gradually disappears in the whole imaging acquisition system. Then, the three-dimensional centroid of the contact network point cloud data image can regularly move in the image along with the change. As can be seen from fig. 4a, 4b, 4c, and 4d, the three-dimensional centroid of the catenary point cloud data image moves up and down regularly as the catenary supporting device continuously shoots in the image. When the centroid appears at the very center of the image, the data structure acquired is the most complete, as shown in fig. 4 c. Therefore, a position function of the three-dimensional centroid of the point cloud image data frame can be established, the centroid position of the continuous frame data is positioned through calculation, whether the position of the centroid position is in the middle of the image or not is judged, and the optimal data frame of the point cloud data is extracted.
Description of the above formula, miThe quality of each point is represented by,
is the three-dimensional centroid of the point cloud image data frame,
is the coordinate of the centroid on the y-axis, and H is the height of the image.
Because the data volume of the single-frame three-dimensional point cloud is large, the time consumption is high when the three-dimensional centroid of the image is processed and calculated, and the real-time detection of the system is not facilitated. In the statistical process of the number of the contact network point cloud data in fig. 5, it can be found that the contact network supporting device point cloud data gradually decreases after reaching the peak value, and the corresponding relationship exists between the peak value and the position change rule of the three-dimensional centroid. Therefore, the variation gradient of the point cloud is obtained by establishing a gradient calculation formula of the point cloud number variation;
in the formula, PNnIs the point cloud number, PN, of the data frame nn-1Is the point cloud number, PN, of the data frame n-1n-2Is the point cloud number of the data frame n-2;
is the time interval between data frame n and data frame n-1,
is the time interval between data frame n-1 and data frame n-2;
the Peak moment when the point cloud number has a Peak value is obtained by calculating the product of adjacent local gradientsn-1：
When the gradient changes from positive to negative, the product is negative, the data frame with the point cloud data peak can be accurately positioned, then the three-dimensional centroid calculation of the point cloud data frame is carried out, the optimal data frame of the point cloud data is extracted, and the timeliness of the process of extracting the optimal data frame can be improved.
Claims (2)
1. A method for selecting an optimal data frame of three-dimensional point cloud data of a contact net supporting device is characterized by comprising the following steps:
step A: the mounting frame is arranged on a three-dimensional scanner for detecting the roof, the contact net supporting device is continuously shot, and a three-dimensional point cloud image of the contact net supporting device is collected in real time;
and B: carrying out density statistics on point cloud data in the acquired point cloud image of the contact net supporting device, analyzing the change rule of the point cloud data, and extracting an effective data frame of the point cloud data;
and C: performing three-dimensional centroid calculation on the extracted effective frame of the point cloud data, tracking and positioning the point cloud centroid, analyzing the change rule of the point cloud centroid, and extracting an optimal data frame of the point cloud data;
the specific process of the step B is as follows:
establishing point cloud density function of point cloud image data frame
In the formula, PiFor the ith calculation point, Pi nearestIs a point PiNum is the sum of all points in the current point cloud image data frame;
setting a threshold value tau when
Judging to enter a detection area, and extracting a current point cloud image data frame as an effective data frame; the specific process of the step C is as follows:
establishing a position function of a three-dimensional centroid of a point cloud image data frame:
in the formula, miFor the mass of each point it is,
the coordinate of the three-dimensional centroid of the point cloud image data frame on the y axis is shown, and H is the height of the image; n is the total number of points;
judging whether the position of the three-dimensional centroid of the current point cloud image data frame is in the center of the image;
and if the current point cloud image data frame is in the center of the image, extracting the current point cloud image data frame as an optimal data frame.
2. The method for selecting the optimal data frame of the three-dimensional point cloud data of the catenary supporting device according to claim 1, wherein the step C comprises the following steps:
establishing a gradient calculation formula of point cloud number change, and calculating a point cloud number change gradient PNMG of a data frame n in a point cloud data effective data framen：
In the formula, PNnIs the point cloud number, PN, of the data frame nn-1Is the point cloud number, PN, of the data frame n-1n-2Is the point cloud number of the data frame n-2;
is the time interval between data frame n and data frame n-1,
is the time interval between data frame n-1 and data frame n-2;
the Peak moment when the point cloud number has a Peak value is obtained by calculating the product of the gradients of the adjacent data framesn-1：
In the formula, PNMGn-1A point cloud number change gradient for data frame n-1;
namely: when the gradient changes from positive to negative, the product is negative, and the data frame at the moment is judged as the data frame with the point cloud data peak value;
and then, carrying out three-dimensional centroid calculation on the data frame with the point cloud data peak value to extract the optimal data frame of the point cloud data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710615521.1A CN107564056B (en) | 2017-07-26 | 2017-07-26 | Optimal data frame selection method for three-dimensional point cloud data of contact net supporting device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710615521.1A CN107564056B (en) | 2017-07-26 | 2017-07-26 | Optimal data frame selection method for three-dimensional point cloud data of contact net supporting device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107564056A CN107564056A (en) | 2018-01-09 |
| CN107564056B true CN107564056B (en) | 2020-12-18 |
Family
ID=60973868
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710615521.1A Expired - Fee Related CN107564056B (en) | 2017-07-26 | 2017-07-26 | Optimal data frame selection method for three-dimensional point cloud data of contact net supporting device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107564056B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108564575B (en) * | 2018-04-11 | 2022-03-15 | 西南交通大学 | Non-contact type contact network parameter detection method based on three-dimensional point cloud data |
| CN108765393B (en) * | 2018-05-21 | 2022-03-15 | 西南交通大学 | Method for detecting vibration behavior of high-speed railway contact net |
| CN109636904B (en) * | 2018-08-13 | 2023-03-10 | 长沙理工大学 | Noise processing technology based on UAV aerial survey terrain data |
| CN113295143B (en) * | 2020-03-12 | 2023-04-25 | 广东中科如铁技术有限公司 | Dynamic measurement method for gradient of overhead line system locator |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103020654A (en) * | 2012-12-12 | 2013-04-03 | 北京航空航天大学 | Synthetic aperture radar (SAR) image bionic recognition method based on sample generation and nuclear local feature fusion |
| CN103557788A (en) * | 2013-10-15 | 2014-02-05 | 西南交通大学 | High-speed rail catenary geometric parameter detection non-contact compensation and Kalman filtering correction method |
| CN104050712A (en) * | 2013-03-15 | 2014-09-17 | 索尼公司 | Method and apparatus for establishing three-dimensional model |
| CN105023014A (en) * | 2015-08-21 | 2015-11-04 | 马鞍山市安工大工业技术研究院有限公司 | Method for extracting tower target in unmanned aerial vehicle routing inspection power transmission line image |
| CN106951860A (en) * | 2017-03-20 | 2017-07-14 | 河南腾龙信息工程有限公司 | A kind of three-dimensional data intelligent identification Method based on a cloud |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103942751B (en) * | 2014-04-28 | 2017-06-06 | 中央民族大学 | A kind of video key frame extracting method |
-
2017
- 2017-07-26 CN CN201710615521.1A patent/CN107564056B/en not_active Expired - Fee Related
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103020654A (en) * | 2012-12-12 | 2013-04-03 | 北京航空航天大学 | Synthetic aperture radar (SAR) image bionic recognition method based on sample generation and nuclear local feature fusion |
| CN104050712A (en) * | 2013-03-15 | 2014-09-17 | 索尼公司 | Method and apparatus for establishing three-dimensional model |
| CN103557788A (en) * | 2013-10-15 | 2014-02-05 | 西南交通大学 | High-speed rail catenary geometric parameter detection non-contact compensation and Kalman filtering correction method |
| CN105023014A (en) * | 2015-08-21 | 2015-11-04 | 马鞍山市安工大工业技术研究院有限公司 | Method for extracting tower target in unmanned aerial vehicle routing inspection power transmission line image |
| CN106951860A (en) * | 2017-03-20 | 2017-07-14 | 河南腾龙信息工程有限公司 | A kind of three-dimensional data intelligent identification Method based on a cloud |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107564056A (en) | 2018-01-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107564056B (en) | Optimal data frame selection method for three-dimensional point cloud data of contact net supporting device | |
| CN107895376A (en) | Based on the solar panel recognition methods for improving Canny operators and contour area threshold value | |
| CN111754465B (en) | Insulator positioning and string dropping detection method | |
| CN103500324B (en) | Violent behavior recognition methods based on video monitoring | |
| CN105741291A (en) | Method for detecting faults of equipotential lines of high-speed railway overhead line system suspension devices | |
| CN103500329B (en) | Street lamp automatic extraction method based on vehicle-mounted mobile laser scanning point cloud | |
| CN107784652B (en) | Tower rapid detection method based on unmanned aerial vehicle image | |
| CN107481237B (en) | A kind of infrared array image hot spot detection method based on multiframe temperature characterisitic | |
| CN112883878B (en) | Automatic point cloud classification method under transformer substation scene based on three-dimensional grid | |
| Naccarato et al. | Improvements in the detection efficiency model for the Brazilian lightning detection network (BrasilDAT) | |
| CN111260668A (en) | Power line extraction method, system and terminal | |
| CN102194129B (en) | Vehicle-type-clustering-based video detection method for traffic flow parameters | |
| CN106886988B (en) | Linear target detection method and system based on unmanned aerial vehicle remote sensing | |
| CN112508386A (en) | Noctilucent and POI data fused urban built-up area extraction method | |
| CN112131924A (en) | Transformer substation equipment image identification method based on density cluster analysis | |
| CN109299700A (en) | Subway group abnormality behavioral value method based on crowd density analysis | |
| CN113516105B (en) | Lane detection method and device and computer readable storage medium | |
| CN115690062A (en) | Rail surface damage state detection method and device and electronic equipment | |
| Tran et al. | A robust traffic-aware city-scale multi-camera vehicle tracking of vehicles | |
| CN103336965B (en) | Based on profile difference and the histogrammic prospect of block principal direction and feature extracting method | |
| CN107369162A (en) | A kind of generation method and system of insulator candidate target region | |
| CN111260631B (en) | Efficient rigid contact line structure light bar extraction method | |
| Skadins et al. | Edge pre-processing of traffic surveillance video for bandwidth and privacy optimization in smart cities | |
| CN107240111B (en) | Edge communication segmentation passenger flow statistical method | |
| Hsieh et al. | Grid-based template matching for people counting |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20201218 Termination date: 20210726 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |