阅读说明：本技术 一种扣件弹舌离缝检测方法 (Fastener spring tongue separation detection method ) 是由 左丽玛 于 2018-06-05 设计创作，主要内容包括：本发明公开一种扣件弹舌离缝检测方法,属于铁路基础设施检测领域。本发明的主要步骤为：通过三维数据采集系统获取轨道两侧的三维形貌数据,并将其转换为二维深度图像,提取扣件所在区域的局部二维深度图像,并对其进行阈值化处理,提取扣件螺栓轮廓,计算螺栓中心位置,并以螺栓中心为起点,向钢轨一侧进行行采样,然后对采样序列进行阈值化处理,计算扣件弹舌高度和弹舌离缝高度,并根据阈值判断是否存在弹舌离缝危险。本发明所提出的检测方法可自动有效检测扣件弹舌离缝,及时消除列车运行安全隐患。(the invention discloses a fastener spring tongue separation detection method, and belongs to the field of railway infrastructure detection. The method mainly comprises the following steps: the method comprises the steps of obtaining three-dimensional appearance data of two sides of a track through a three-dimensional data acquisition system, converting the three-dimensional appearance data into two-dimensional depth images, extracting local two-dimensional depth images of a region where a fastener is located, thresholding the local two-dimensional depth images, extracting the outline of a fastener bolt, calculating the center position of the bolt, sampling the line of one side of a steel rail by taking the center of the bolt as a starting point, thresholding a sampling sequence, calculating the height of an elastic tongue of the fastener and the height of an open seam of the elastic tongue, and judging whether the danger of the open seam of the elastic tongue exists or not according. The detection method provided by the invention can automatically and effectively detect the elastic tongue separation of the fastener, and eliminate the potential safety hazard of train operation in time.)

1. a fastener spring tongue seam separation detection method is characterized in that: the method comprises the following steps:

Step 1: acquiring three-dimensional profile data of a track by adopting a three-dimensional imaging system, and converting the three-dimensional profile data into a two-dimensional depth image by taking a track plane as a horizontal plane reference of the depth image;

Step 2: determining the position of the fastener in the two-dimensional depth image, and extracting a local two-dimensional depth image of the fastener;

And step 3: setting a threshold T₁Thresholding is carried out on the local two-dimensional depth image of the fastener, and the bolt outline, T, of the fastener is extracted₁the value range of (A) is 0 to 1000;

and 4, step 4: calculating the center position C (x) of the fastener bolt_c,y_c)；

And 5: taking the central position of the bolt area as a starting point, sampling r pixel gray values by taking the central position of the bolt area as a starting point, wherein the central position faces one side of the steel rail and is vertical to the longitudinal direction of the steel rail, and obtaining a sampling sequence S-S₁,s₂,...,s_rThe value range of r is 5-100;

Step 6: setting a threshold T₃The sampling sequence is thresholded to obtain a new sampling sequence S '═ S'₁,s'₂,...,s'_r}，T₃The value range of (A) is 0 to 1000;

And 7: based on the new sampling sequence S '═ { S'₁,s'₂,...,s'_rCalculating the height h of the elastic tongue of the fastener;

And 8: based on the reference height h of the latch_sCalculating the spring tongue seam-separating height delta h ═h-h_s；

And step 9: setting a threshold T₄According to the height delta h of the open seam of the spring tongue and the threshold value T₄Judging whether the spring tongue is in danger of separating from the seam or not, T₄The value range of (A) is 0 to 1000.

2. The method for detecting the seam separation of the spring tongue of the fastener according to claim 1, wherein: the three-dimensional imaging system in the step 1 is a linear structured light three-dimensional scanning imaging system, and when the three-dimensional shape data of the track fastener is acquired, the scanning direction calibration needs to be carried out on the acquired three-dimensional shape data so as to ensure that the physical sizes represented by the horizontal coordinate and the vertical coordinate unit pixels in the converted two-dimensional depth image are equal.

3. The method for detecting the seam separation of the spring tongue of the fastener according to claim 1, wherein: the position of the fastener in the two-dimensional depth image can be determined in the step 2 by manual selection, or by automatic detection methods such as SVM (support vector machine) or deep learning or neural network classification.

4. The method for detecting the seam separation of the spring tongue of the fastener according to claim 1, wherein: the threshold value T in step 3₁＝v_maxa, wherein v_maxRepresenting the maximum gray value in the local two-dimensional depth image, wherein a is a fixed constant and takes the value as a positive integer; the thresholding processing method comprises the following steps:

Where f (x, y) and f' (x, y) represent the pixel grayscale values at pixel coordinates (x, y) in the local two-dimensional depth images before and after thresholding, respectively.

5. the method for detecting the open seam of the base fastener latch tongue as claimed in claim 1, wherein: the bolt center position C (x) in step 4_c,y_c) Adopting formula (2) or formula (3):

Wherein (x)_c,y_c) Indicates the center position of the bolt region, x_min、x_maxMinimum and maximum abscissa values, y, respectively representing the area of the bolt containing the pixel points_min、y_maxMinimum and maximum ordinate values, x, respectively representing the bolt region containing pixel points_i,y_iIs the pixel coordinate, S represents the bolt region, and N is the number of pixels in the bolt region.

6. The method for detecting the seam separation of the spring tongue of the fastener according to claim 1, wherein: the threshold valueThe thresholding processing method comprises the following steps:

7. The method for detecting the seam separation of the spring tongue of the fastener according to claim 1, wherein: the calculation formula of the spring tongue height h in the step 7 is as follows:

Wherein r represents a sampling sequence S '═ S'₁,s'₂,...,s'_rthe number of elements in the sequence, m, represents a sampling sequence S '═ S'₁,s'₂,...,s'_rThe number of non-zero elements in the test.

8. The method for detecting the open seam of the spring tongue of a fastener according to claim 1The method is characterized in that: the standard height value h of the spring tongue in the step 8_sThe height value of the current detection spring bolt under the condition that no seam separation occurs can be obtained from normal historical data; when in detection, the current spring tongue reference height value h_bAnd counting by a sleeper or a fastener to obtain the current spring tongue number K, and extracting the reference height value of the spring tongue with the number of K from the spring tongue reference height value data set.

9. The method for detecting the seam separation of the spring tongue of the fastener according to claim 1, wherein: in step 9, when the height delta h of the separation seam of the spring tongue is larger than the threshold value T₄And if not, the danger of the elastic tongue separating from the seam does not exist.

10. the method for detecting the seam separation of the spring tongue of the fastener according to claim 1, wherein: the fasteners are all types of fasteners with spring tongues, and the pixel bit width of the two-dimensional depth image is larger than 8bits so as to avoid reducing the three-dimensional imaging height resolution.

Technical Field

The invention relates to the field of railway infrastructure detection, in particular to a fastener spring tongue gap detection method.

Background

The fastener is an important part for connecting the rail and the sleeper, and has the functions of fixing the rail on the sleeper, keeping the track gauge and preventing the rail from moving longitudinally and transversely relative to the sleeper, so that the fastener plays an important role in ensuring the stability and reliability of the rail. Whether the elastic tongue of the fastener is separated from the seam is an important embodiment of whether the fastener is fixed firmly, and once the elastic tongue is separated from the seam, serious potential safety hazards can be caused. Therefore, the method for detecting whether the spring tongue of the fastener is separated from the seam in time is very important in the field of railway infrastructure detection.

on the aspect of railway infrastructure detection, China mainly uses manual and static detection for a long time, has high maintenance cost, high strength and poor safety, and puts higher requirements on automation and real-time performance of railway detection along with rapid development of high-speed railways. At present, some fastener detection technologies appear at home and abroad, mainly including: continuous scanning devices based on line lasers, such as the company mock, germany; computer vision inspection devices based on area array image sensors, such as the VIS system of the american ENSCO, the photoelectric rail inspection system developed by atlas electronic, germany, and the high-speed vehicular rail image recognition system of the beijing forsda company. However, the existing fastener detection systems cannot detect the open seam of the spring tongue of the fastener, and therefore, an effective and reliable method for detecting the open seam of the spring tongue of the fastener is needed and the potential safety hazard of the open seam of the spring tongue of the fastener is eliminated.

Disclosure of Invention

The invention aims to provide a fastener spring tongue gap detection method, which automatically detects whether a fastener has spring tongue gaps through an image processing-based method so as to solve the problem that the conventional fastener detection system cannot detect the fastener spring tongue gaps.

In order to solve the technical problems, the technical scheme of the invention is as follows: a method for detecting the seam separation of the spring tongue of a fastener. The method is characterized by comprising the following steps:

Step 1: acquiring three-dimensional profile data of a track by adopting a three-dimensional imaging system, and converting the three-dimensional profile data into a two-dimensional depth image by taking a track plane as a horizontal plane reference of the depth image;

The three-dimensional imaging system is a linear structured light three-dimensional scanning imaging system, and when three-dimensional shape data of the track fastener is acquired, the scanning direction of the acquired three-dimensional shape data needs to be calibrated so as to ensure that the physical sizes represented by the horizontal coordinate and the vertical coordinate unit pixel in the converted two-dimensional depth image are equal.

step 2: determining the position of the fastener in the two-dimensional depth image, and extracting a local two-dimensional depth image of the fastener;

The position of the fastener in the two-dimensional depth image can be manually selected, and automatic detection methods such as SVM (support vector machine) or deep learning or neural network classification can also be adopted.

And step 3: setting a threshold T₁Thresholding is carried out on the local two-dimensional depth image of the fastener to extract the fastener screwBolt profile, T₁The value range of (A) is 0 to 1000;

The threshold value T₁＝v_maxA, wherein v_maxRepresenting the maximum gray value in the local two-dimensional depth image, wherein a is a fixed constant and takes the value as a positive integer; the thresholding processing method comprises the following steps:

Where f (x, y) and f' (x, y) represent the pixel grayscale values at pixel coordinates (x, y) in the local two-dimensional depth images before and after thresholding, respectively.

And 4, step 4: calculating the center position C (x) of the fastener bolt_c,y_c)；

The bolt center position C (x)_c,y_c) Adopting formula (2) or formula (3):

Wherein (x)_c,y_c) Indicates the center position of the bolt region, x_min、x_maxMinimum and maximum abscissa values, y, respectively representing the area of the bolt containing the pixel points_min、y_maxMinimum and maximum ordinate values, x, respectively representing the bolt region containing pixel points_i,y_iIs the pixel coordinate, S represents the bolt region, and N is the number of pixels in the bolt region.

And 5: taking the central position of the bolt area as a starting point, sampling r pixel gray values by taking the central position of the bolt area as a starting point, wherein the central position faces one side of the steel rail and is vertical to the longitudinal direction of the steel rail, and obtaining a sampling sequence S-S₁,s₂,...,s_rThe value range of r is 5-100;

Step 6: setting a threshold T₃the sampling sequence is thresholded to obtain a new sampling sequence S '═ S'₁,s'₂,...,s'_r}，T₃The value range of (A) is 0 to 1000;

the threshold valueThe thresholding processing method comprises the following steps:

And 7: based on the new sampling sequence S '═ { S'₁,s'₂,...,s'_rcalculating the height h of the elastic tongue of the fastener;

The calculation formula of the spring tongue height h in the step 7 is as follows:

Wherein r represents a sampling sequence S '═ S'₁,s'₂,...,s'_rthe number of elements in the sequence, m, represents a sampling sequence S '═ S'₁,s'₂,...,s'_rThe number of non-zero elements in the test.

And 8: based on the reference height h of the latch_sCalculating the spring tongue seam-separating height delta h as h-h_s；

The standard height value h of the spring tongue_sThe height value of the current detection spring bolt under the condition that no seam separation occurs can be obtained from normal historical data; when in detection, the current spring tongue reference height value h_bAnd counting by a sleeper or a fastener to obtain the current spring tongue number K, and extracting the reference height value of the spring tongue with the number of K from the spring tongue reference height value data set.

And step 9: setting a threshold T₄According to the height delta h of the open seam of the spring tongue and the threshold value T₄Judging whether the spring tongue is in danger of separating from the seam or not, T₄The value range of (1) is 0-1000, when the height delta h of the spring tongue is greater than the threshold value T₄And if not, the danger of the elastic tongue separating from the seam does not exist.

the fastener refers to all types of fasteners having a spring tongue.

Preferably, the pixel bit width of the two-dimensional depth image is greater than 8bits to avoid degrading the three-dimensional imaging height resolution.

the invention has the beneficial effects that: the detection method provided by the invention can automatically detect whether the spring tongue of the fastener is separated from the seam by acquiring the three-dimensional topography data of the rail fastener by using a three-dimensional imaging system, converting the three-dimensional topography data into a two-dimensional depth image and adopting thresholding treatment, sampling and other methods on the two-dimensional depth image. Compared with the existing fastener detection system, the invention adopts three-dimensional shape data and utilizes a historical data comparison method to judge whether the height of the spring tongue changes or not so as to judge whether the spring tongue is separated from the seam or not, thereby effectively overcoming the problem that the existing fastener detection system based on images is difficult to detect the separation of the spring tongue and eliminating the potential safety hazard of train operation in time.

Drawings

FIG. 1 is a flow chart of the method of the present invention.

Fig. 2 is a schematic diagram of a three-dimensional imaging system architecture.

Fig. 3 is a schematic view of a W-shaped fastener structure.

Fig. 4 is a two-dimensional depth image.

Fig. 5 is a fastener region of interest.

Fig. 6 is a schematic drawing of the extracted bolt area.

Fig. 7 is a schematic view of calculating the center position of the bolt region.

Fig. 8 is a schematic diagram of the sampling process.

Wherein, 1 is line structure light projector, 2 is area array camera, 3 is the rail, 4 is the fastener, 5 is the bolt, 6 is the bullet tongue, 7 is the nut gasket, 8 is the bullet strip.

Detailed Description

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

As shown in fig. 1, the embodiment provides a method for detecting a seam separation of a fastener tongue, which is specifically implemented as follows:

step 1: adopt two three-dimensional imaging system to obtain the three-dimensional topography data of track both sides simultaneously, three-dimensional data acquisition system is as shown in fig. 2, constitute by line laser 1 and high-speed area array camera 2 machine, line laser 1 keeps fixed with high-speed area array camera 2's position and angle, line laser 1 perpendicular to track sends the scanning laser face of a branch of certain wavelength, form a section profile on 3 surfaces of track, shoot through high-speed area array camera and acquire the track three-dimensional topography data that contain W type fastener 4, and convert it into two-dimensional depth image, as shown in fig. 4. Wherein, the schematic view of the W-shaped fastener is shown in FIG. 3.

Step 2: the position of the W-shaped fastener in the two-dimensional depth image is manually selected, and a local two-dimensional depth image of the area where the W-shaped fastener is located is extracted, as shown in fig. 5.

And step 3: setting a threshold T₁＝v_maxa, wherein v_maxthe maximum gray value in the local two-dimensional depth image is represented, a is a fixed constant and is a positive integer, and in a specific embodiment, a is 10.

the local two-dimensional depth image is subjected to thresholding as follows,

Where f (x, y) and f' (x, y) represent the pixel grayscale values at pixel coordinates (x, y) in the local two-dimensional depth images before and after thresholding, respectively.

Through the thresholding process in formula (1), the bolt profile of the W-type fastener can be accurately extracted, as shown in fig. 6.

And 4, step 4: as shown in FIG. 7, the center position C (x) of the bolt is calculated_c,y_c) The concrete formula is as follows:

Wherein，x_min、x_maxMinimum and maximum abscissa values, y, respectively, of the bolt profile containing pixel points_min、y_maxthe minimum and maximum ordinate values of the bolt profile containing the pixel points are respectively.

And 5: for the line image at the center of the bolt, as shown in fig. 8, r pixel gray values are sampled with the rightmost side as the starting point, and the sampling sequence S ═ S is obtained₁,s₂,...,s_rIn the specific implementation process, r is 30.

Step 6: setting a threshold valuefor sampling sequence S ═ S₁,s₂,...,s_rCarry on the following thresholding:

The thresholding process can obtain a new sampling sequence S '═ { S'₁,s'₂,...,s'_r}。

And 7: based on the new sampling sequence S '═ { S'₁,s'₂,...,s'_rCalculating the height h of the elastic tongue of the W-shaped fastener, wherein the specific formula is as follows:

Wherein r represents a sampling sequence S '═ S'₁,s'₂,...,s'_rThe number of elements in the sequence, m, represents a sampling sequence S '═ S'₁,s'₂,...,s'_rthe number of non-zero elements in the test.

and 8: standard height value h of spring tongue based on historical data_scalculating the spring tongue seam-separating height delta h as h-h_s。

And step 9: setting a threshold T₄20mm, according to the height delta h of the open seam of the spring tongue and the threshold value T₄And judging whether the elastic tongue is in danger of seam separation. When the height delta h of the spring tongue is larger than the threshold value T₄When it is, there is a spring-tongue separationThe seam is dangerous, otherwise, the elastic tongue is not dangerous to leave the seam.

In the step 2, the position of the fastener in the two-dimensional depth image can also be determined by adopting an automatic detection method such as SVM (support vector machine) or deep learning or neural network classification.

While the invention has been described in connection with what is presently considered to be the most practical and preferred embodiment, it is to be understood that the invention is not to be limited to the disclosed embodiment, but on the contrary, is intended to cover various modifications and equivalent arrangements included within the spirit and scope of the appended claims.