A method for detecting pantograph dynamic envelope
Through the scanning point cloud data and feature convolution calculation of lidar, the dynamic envelope of the pantograph of the electric locomotive is quickly measured, which solves the problems of low efficiency and difficulty in the existing technology, and realizes efficient and accurate dynamic envelope detection .
Patent Information
- Application Number
- CN202111138952.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-09-27
- Publication Date
- 2024-12-06
- Estimated Expiration
- 2041-09-27
Smart Images

Figure CN113885009B_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of pantographs, and in particular to a method for detecting a dynamic envelope of a pantograph. Background Art
[0002] When the electric locomotive is running at high speed, the pantograph moves up, down, left and right continuously, and its maximum dynamic lift and maximum dynamic swing constitute the dynamic envelope of the pantograph. There must not be any faulty objects in the dynamic envelope, which is a key indicator for the safe operation of the train. During the acceptance and daily operation and maintenance of the contact network, the dynamic envelope of the pantograph needs to be inspected to ensure that the contact network suspension parts and foreign objects do not invade the dynamic envelope. At present, the railway-related departments mostly use static laser measurement to complete the inspection of the dynamic envelope of the pantograph, and manually calculate whether the locator device, suspension parts or foreign objects invade the dynamic envelope. This method is inefficient, labor-intensive, and difficult to operate at night. Therefore, there is an urgent need for an accurate and efficient measurement method. Summary of the invention
[0003] In view of the problems existing in the prior art, the present invention proposes a method for detecting the dynamic envelope of a pantograph. The method utilizes point cloud data scanned by a laser radar, firstly extracts positioning point information, and then analyzes and processes the dynamic envelope cross-section data with the positioning point as a reference point and performs iterative calculations, so as to quickly measure the key indicators of the dynamic envelope and ensure the measurement accuracy.
[0004] To achieve the above objectives, the following solutions are provided:
[0005] A method for detecting a pantograph dynamic envelope comprises the following steps:
[0006] S1: Use the dynamic measuring instrument of laser radar to continuously scan the contact network dynamically, use the dynamic measurement method of the contact network locator positioning point to detect the positioning point data, and quickly locate the position of the contact network locator device component;
[0007] S2: A two-stage method based on overall features quickly separates the overall contour detection data of the locator device;
[0008] S3: Perform convolution operation on the features of each component of the locator device, calculate the location of the key parts of the target through feature convolution and region growth calculation, and extract the key measurement points;
[0009] S4: Calculate the measurement parameters of the key points of the dynamic envelope, construct the limits of the dynamic envelope diagram, and use this as a basis, combined with the driving speed and the curve, to determine whether the positioner device, suspension parts or foreign objects invade the dynamic envelope.
[0010] Further, the steps for quickly locating the contact network locator device assembly are as follows:
[0011] A1: Install a high-precision laser radar on the detection device, and install a mileage sensor on the detection device. When measuring the dynamic envelope of the pantograph, the detection device moves along the track and automatically and continuously scans and measures the contact network device;
[0012] A2: The measurement point data continuously scanned by the laser radar and the position data returned by the mileage encoder synchronously form the original three-dimensional point cloud data of the overhead line device and send it to the data processing center of the detection device;
[0013] A3: De-noise the original 3D point cloud data of the contact network device and construct a 3D point cloud model of the contact network;
[0014] A4: Use the dynamic measurement method of the contact network locator to quickly detect the locating point and calculate the guide height and pull-out value data of the locating point.
[0015] Further, the steps of separating the overall contour detection data of the locator device in step S2 are as follows:
[0016] B1: Filter the point cloud data scanned by the radar, remove noise points, take the positioning point as the reference point, and locate the three-dimensional point cloud position of the contact network locator device according to the spatial position relationship;
[0017] B2: Segment the locator device data offline, import the locator device models for positive positioning and reverse positioning respectively, and extract the overall features of the locator device;
[0018] B3: Utilize the overall characteristics of the positive or negative positioning locator device to match the locator device online and quickly separate the overall contour detection data of the locator device.
[0019] Further, the steps of extracting key measurement points in step S3 are as follows:
[0020] C1: Extract the features of the locator, oblique wrist, positioning tube, insulator, and wrist base offline, and quickly separate the locator, oblique wrist, positioning tube, insulator, and wrist base by combining the feature convolution method;
[0021] C2: Dynamic adaptive region growing method is used to iteratively calculate the threshold of key position parameters;
[0022] C3: Using the key point threshold, gradually approach and find the measurement position of the key parts;
[0023] C4: Extract all the measurement points at the key positions, filter them, remove the edge points, and calculate the average value as the measurement points at the key positions.
[0024] Further, the calculation and judgment process of step S4 is as follows:
[0025] D1: Convert the rail plane to the calculation coordinate plane, and calculate the guide height value and pull-out value of the key points;
[0026] D2: Select key points and draw the actual dynamic envelope diagram based on driving speed and curves;
[0027] D3: Further detection and judgment of points around the dynamic envelope;
[0028] D4: Output dynamic envelope measurement results.
[0029] The working principle and advantages of the present invention are:
[0030] (1) The present invention adopts a non-contact detection method to continuously scan and measure the contact network locator device, etc., firstly using the positioning point to quickly locate the position of the locator device, and then based on the two-stage method of overall features, the overall contour detection data of the locator device is quickly separated, and the features of each component of the locator device are convoluted; through the calculation of feature convolution and regional growth, the key measurement points are accurately found;
[0031] (2) Compared with the traditional measurement method, the method of the present invention can quickly measure all the values of the dynamic envelope limit at one time, greatly improving the measurement efficiency. It is suitable for dynamic envelope measurement in various scenarios such as single-fixed single column, single-fixed double columns, double-fixed double columns, branch lines, anchor sections, stations, tunnels, etc.
[0032] The present invention uses a laser radar to continuously and dynamically scan the railway contact network, extracts the cross-section point cloud data of the contact network positioning device, removes noise points, and after threshold filtering and fitting, measures the key point data and gives the measurement results of the dynamic envelope safety. The method has the characteristics of convenient measurement, many measurement data points, high measurement accuracy, and fast measurement speed. BRIEF DESCRIPTION OF THE DRAWINGS
[0033] Figure 1 This is a flow chart of the dynamic envelope limit measurement of the present invention;
[0034] Figure 2 It is a point cloud diagram of the radar scanning locator device of the present invention;
[0035] Figure 3 This is a cloud diagram of the radar scanning locator device separated by the present invention. DETAILED DESCRIPTION
[0036] The following is a further detailed description through specific implementation methods:
[0037] like Figure 1 As shown, a method for detecting a pantograph dynamic envelope includes the following steps:
[0038] S1: Use the dynamic measuring instrument of laser radar to continuously scan the contact network dynamically, establish the three-dimensional point cloud data model of the locator device, and use the dynamic measurement method of the contact network locator positioning point to quickly detect the positioning point data;
[0039] To ensure measurement accuracy and efficiency, the detection device used in this method is further equipped with a high-precision radar with an angular resolution of 0.09 and a measurement range of 0.5 meters to 100 meters, a measurement frequency of 60Hz / second, and an mileage sensor with an accuracy of 1 / 20000. When measuring the dynamic envelope of the pantograph, the detection device moves along the rail at about 2-4 kilometers / hour. The number of data points of the locator device, tunnel davit or pillar, and tunnel measured by the laser radar per second is:
[0040] 1000×60=60000
[0041] According to the running speed of 2-4 km / h of the dynamic measuring instrument, it takes 1-3 seconds to measure a dynamic envelope, during which the number of measuring points is 60,000-180,000.
[0042] The measurement point data continuously scanned by the laser radar and the position data returned by the mileage encoder synchronously form the original three-dimensional point cloud data of the overhead line device and send it to the data processing center of the detection device;
[0043] The original 3D point cloud data of the contact network device is subjected to mean filtering and Gaussian filtering to remove unnecessary noise points, feature matching, and separate the point cloud data of the contact network device (such as Figure 2 );
[0044] Use the dynamic measurement method of the contact network locator to quickly detect the locating point (such as Figure 3 Point A), calculate the guide height and pull-out value data of the positioning point.
[0045] S2: The two-stage method based on the overall features quickly separates the overall contour detection data of the locator device:
[0046] Perform statistical filtering and Gaussian filtering on the point cloud scanned by the radar, and then use the positioning point as a reference point to locate the three-dimensional point cloud position of the contact network locator device according to the spatial position relationship, and segment the locator device data offline;
[0047] Import the locator device models of positive positioning and reverse positioning respectively. Considering that this method is a dynamic measurement, the point cloud density will change with the measurement speed, and there will be occlusion during the measurement process, and the characteristics of the locator device will change significantly. This method extracts the pillar or suspension column, positioning tube, locator, and oblique wrist arm as the overall feature target of the locator device model, and establishes the overall feature model of the locator device model of positive positioning and reverse positioning respectively;
[0048] Using the overall features of the positive or negative positioning locator device, online matching of the locator device, and further iterating and modifying the overall features of the locator device, the overall contour detection data of the locator device is quickly and integrally separated;
[0049] Scan m (m>20) positive locator devices and reverse locator devices respectively, scan each locator device at least n (n>20) times at different speeds, repeat verification and matching, and repeat step 2 until all sample data are successfully processed.
[0050] S3: Perform convolution operation on the features of each component of the locator device, calculate the location of the key parts of the target through feature convolution and region growth calculation, and extract the key measurement points:
[0051] Extract the features of the locator, oblique arm, positioning tube, insulator, and arm base offline as the initial feature value reference input;
[0052] The FCN method is used to further perform convolution calculation on the features of the locator, oblique wrist, positioning tube, insulator, and wrist base. The locator, oblique wrist, positioning tube, insulator, and wrist base are quickly separated by combining offline features and FCN convolution features.
[0053] According to the fast segmentation result of FCN, the smallest area in the middle of the target object is taken as the initial seed point S for region growth. 0 ;
[0054] According to the 8-neighborhood method, the seed point and the point M scanned by its eight neighbors are i To compare similarities,
[0055] The comparison distance value is calculated as follows:
[0056]
[0057] The similarity thresholds are as follows:
[0058] d≤θ×M d
[0059] Among them, M d is the maximum spacing value, θ is the adjustment factor, the value range is 0.5≤θ≤1, and the θ value is adjusted in the iterative process;
[0060] Repeat the above steps of region growing method to gradually approach and find the key parts (such as Figure 3 ) Measurement positions of B, C, D, and E;
[0061] All the measurement points at the key positions are extracted, filtered, edge points are removed, and the average value is calculated as the measurement points at the key positions.
[0062] S4: Calculate the measurement parameters of the key points of the dynamic envelope, construct the limits of the dynamic envelope diagram, and determine whether the positioner device, suspension parts or foreign objects invade the dynamic envelope in combination with the driving speed and the curve;
[0063] Using the conversion rail plane as the reference coordinate system, calculate the guide height (y) and pull-out value (x) of points A, B, C, D, and E:
[0064]
[0065] y i =y 0 +d i ×sinθ i
[0066] trackwidth is the track gauge, d i is the distance from the radar test to the contact line point at the hanging string point, θ i is the radar test angle, x 0 y 0 are the initial coordinates of the radar installation, and i is point A, B, C, D, and E.
[0067] Construct the dynamic envelope limit diagram with points A, B, C, D, and E;
[0068] According to the driving speed, the curve reference Figure 3 , draw the dynamic envelope of the train's pantograph. If the dynamic envelope limit constructed by A, B, C, D, and E is completely outside the dynamic envelope of the train's pantograph, then the dynamic envelope limit of the overhead contact network is safe. Otherwise, the positioner device or the davit needs to be adjusted.
[0069] In particular, the points on the insulators on the CD line are judged to see whether they will intrude into the dynamic envelope.
[0070] Finally, the measurement results of dynamic envelope limit safety are output.
[0071] The present invention adopts a non-contact detection method to quickly and continuously scan the locator device through a high-precision laser radar, and quickly and accurately measures the dynamic envelope line. The present invention adopts a non-contact detection method to continuously scan and measure the contact network locator device, etc., firstly uses the positioning point to quickly locate the position of the locator device, and then based on the two-stage method of overall features, the overall contour detection data of the locator device is quickly separated, and the features of each component of the locator device are convoluted; through the calculation of feature convolution and regional growth, the key measurement points are accurately found; compared with the traditional measurement method, the method of the present invention can quickly measure all values of the dynamic envelope line limit at one time, greatly improving the measurement efficiency, and is suitable for dynamic envelope line measurement in various scenarios such as single-fixed single column, single-fixed double columns, double-fixed double columns, branch lines, anchor sections, stations, tunnels, etc.
[0072] The above is only an embodiment of the present invention. The common sense such as the known specific structure and characteristics in the scheme is not described in detail here. The ordinary technicians in the relevant field know all the common technical knowledge in the technical field of the invention before the application date or priority date, can obtain all the existing technologies in the field, and have the ability to apply the conventional experimental means before that date. The ordinary technicians in the relevant field can improve and implement this scheme in combination with their own abilities under the enlightenment given by this application. Some typical known structures or known methods should not become obstacles for ordinary technicians in the relevant field to implement this application. It should be pointed out that for those skilled in the art, without departing from the structure of the present invention, several deformations and improvements can be made, which should also be regarded as the scope of protection of the present invention, which will not affect the effect of the implementation of the present invention and the applicability of the patent. The scope of protection required by this application shall be based on the content of its claims, and the specific implementation methods and other records in the specification can be used to interpret the content of the claims.
Claims
1. A method for detecting a pantograph dynamic envelope, characterized in that: The following steps are involved: S1: Use the dynamic measuring instrument of laser radar to continuously scan the contact network dynamically, use the dynamic measurement method of the contact network locator positioning point to detect the positioning point data, and quickly locate the position of the contact network locator device component; S2: A two-stage method based on overall features is used to quickly separate the overall contour detection data of the locator device; the steps for separating the overall contour detection data of the locator device are as follows: B1: Filter the point cloud data scanned by the radar, remove noise points, take the positioning point as the reference point, and locate the three-dimensional point cloud position of the contact network locator device according to the spatial position relationship; B2: Segment the locator device data offline, import the locator device models for positive positioning and reverse positioning respectively, and extract the overall features of the locator device; B3: Using the overall characteristics of the positive or negative positioning locator device, the locator device is matched online to quickly separate the overall contour detection data of the locator device; S3: Perform convolution operation on the features of each component of the locator device, calculate the location of the key parts of the target through feature convolution and region growth calculation, and extract the key measurement points; S4: Calculate the measurement parameters of the key points of the dynamic envelope, construct the limits of the dynamic envelope diagram, and use this as a basis, combined with the driving speed and the curve, to determine whether the positioner device, suspension parts or foreign objects invade the dynamic envelope.
2. The method for detecting the pantograph dynamic envelope according to claim 1, characterized in that: The steps for quickly locating the contact network locator device components are as follows: A1: Install a high-precision laser radar on the detection device, and install a mileage sensor on the detection device. When measuring the dynamic envelope of the pantograph, the detection device moves along the track and automatically and continuously scans and measures the contact network device; A2: The measurement point data continuously scanned by the laser radar and the position data returned by the mileage encoder synchronously form the original three-dimensional point cloud data of the overhead line device and send it to the data processing center of the detection device; A3: De-noise the original 3D point cloud data of the contact network device and construct a 3D point cloud model of the contact network; A4: Use the dynamic measurement method of the contact network locator to quickly detect the locating point and calculate the guide height and pull-out value data of the locating point.
3. The method for detecting the pantograph dynamic envelope according to claim 1, characterized in that: Step S3: The steps of extracting key measurement points are as follows: C1: Extract the features of the locator, oblique wrist, positioning tube, insulator, and wrist base offline, and quickly separate the locator, oblique wrist, positioning tube, insulator, and wrist base by combining the feature convolution method; C2: Dynamic adaptive region growing method is used to iteratively calculate the threshold of key position parameters; C3: Using the key point threshold, gradually approach and find the measurement position of the key parts; C4: Extract all the measurement points at the key positions, filter them, remove the edge points, and calculate the average value as the measurement points at the key positions.
4. The method for detecting the pantograph dynamic envelope according to claim 1, characterized in that: The calculation and judgment process of step S4 is as follows: D1: Convert the rail plane to the calculation coordinate plane, and calculate the guide height value and pull-out value of the key points; D2: Select key points and draw the actual dynamic envelope diagram based on driving speed and curves; D3: Further detection and judgment of points around the dynamic envelope; D4: Output dynamic envelope measurement results.
