CN110634123A - Track fastener loosening detection method adopting depth image - Google Patents
Track fastener loosening detection method adopting depth image Download PDFInfo
- Publication number
- CN110634123A CN110634123A CN201810583193.6A CN201810583193A CN110634123A CN 110634123 A CN110634123 A CN 110634123A CN 201810583193 A CN201810583193 A CN 201810583193A CN 110634123 A CN110634123 A CN 110634123A
- Authority
- CN
- China
- Prior art keywords
- fastener
- depth image
- movable part
- region
- height
- 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
- 238000001514 detection method Methods 0.000 title claims abstract description 36
- 238000000034 method Methods 0.000 claims abstract description 47
- 238000003384 imaging method Methods 0.000 claims abstract description 32
- 238000012876 topography Methods 0.000 claims abstract description 8
- 238000005070 sampling Methods 0.000 claims description 16
- 238000001914 filtration Methods 0.000 claims description 9
- 230000000877 morphologic effect Effects 0.000 claims description 9
- 230000011218 segmentation Effects 0.000 claims description 9
- 238000004364 calculation method Methods 0.000 claims description 6
- 230000008030 elimination Effects 0.000 claims description 3
- 238000003379 elimination reaction Methods 0.000 claims description 3
- 238000007689 inspection Methods 0.000 description 7
- 241001669679 Eleotris Species 0.000 description 4
- 238000001931 thermography Methods 0.000 description 3
- 229910000831 Steel Inorganic materials 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 239000010959 steel Substances 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000011514 reflex Effects 0.000 description 1
- 230000002787 reinforcement Effects 0.000 description 1
- 125000006850 spacer group Chemical group 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000004861 thermometry Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- 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
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
- G06T2207/30164—Workpiece; Machine component
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Quality & Reliability (AREA)
- Length Measuring Devices By Optical Means (AREA)
Abstract
The invention discloses a track fastener loosening detection method based on a depth image, and belongs to the field of railway infrastructure detection. The method mainly comprises the following steps: the method comprises the steps of acquiring three-dimensional topography data of the track fastener by using a three-dimensional imaging system, converting the three-dimensional topography data into a two-dimensional depth image by using a track plane as a depth image horizontal reference plane, extracting a fastener region in the two-dimensional depth image, calculating the floating height and the rotating angle of a movable part of the fastener in the fastener region, and judging whether the fastener is loosened or not by comprehensively utilizing the floating height and the rotating angle. The detection method provided by the invention is convenient to use, can not change the structure of the existing railway fastener, can effectively detect the looseness of bolts or nuts of various types of fasteners, and ensures the safe operation of the railway.
Description
Technical Field
The invention relates to the field of railway infrastructure detection, in particular to a track fastener loosening detection method adopting a depth image.
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. The bolt and the nut are the key for ensuring that the fastener is fixed firmly, and once the bolt or the nut floats, the fastener is inevitably loosened, so that serious potential safety hazards are caused.
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 image-based fastener detection defect detection methods have appeared at home and abroad, and fastener defects are mainly identified by shooting fastener images through a linear array camera and through an image processing algorithm. However, none of these prior art fastener detection systems identify and detect whether a fastener is loose.
Patent CN201580000881X discloses a railway fastener bolt is not hard up automatic display device, and the device is through including two-layer gasket from top to bottom, and when frictional force between upper gasket and the lower floor's gasket was less than the tension of spring, both upper gasket and lower floor's gasket overlapped the part and part under spring tension effect and separated, expose reflex reflector or illuminator to make check out test set or measurement personnel discover. Although this patent can detect whether the railway fastener bolt is not hard up, its main shortcoming lies in: these spacers need to be installed during rail construction to modify the existing clip structure. Moreover, the method needs manual work to participate in detection, and the loosening of the fastener cannot be automatically judged.
Patent CN2012101926412 introduces a railway fastener looseness high-speed detection system and method based on infrared thermal imaging: the method comprises the steps of directly obtaining an infrared thermal imaging image generated by contact stress of a fastener and a steel rail by using a built-in infrared camera, judging whether the fastener is loosened or missing or not according to an infrared image gray value, and carrying out automatic early warning. However, this method has the following disadvantages: according to the method, whether the fastener is loosened or not is judged by acquiring the temperature change of heat generated by the fastener under the fatigue stress through infrared thermal imaging, the method can be only used for tracks which are run by trains, and for tracks which are not run by trains, the method has no fatigue stress, does not have temperature change and cannot detect the track; this limits the range of applications of infrared thermometry to fastener loosening detection. That is, this method of inspection can only be used on trains or large rail inspection vehicles to impact and heat the fasteners.
The above-mentioned patent is difficult to measure the fastener not hard up degree, but can not be used for fastener early warning (early warning, can in time discover before the fastener is not totally released to take reinforcement measures). Therefore, a convenient and efficient rail fastener loosening detection method which can accurately measure the loosening degree of the fastener, reliably detect the loosening of the fastener, can be mounted on a train or a rail inspection vehicle and can be mounted on a daily inspection rail trolley is urgently needed.
Disclosure of Invention
The invention aims to provide a track fastener loosening detection method based on three-dimensional topography data, which is used for accurately detecting whether a track fastener is loosened or not so as to solve the problems of poor precision, low efficiency and limited application range of the conventional track fastener loosening detection method.
In order to solve the technical problems, the technical scheme of the invention is as follows:
the utility model provides an adopt track fastener looseness detecting method of degree of depth image which characterized in that: the method comprises the following steps:
step 1: acquiring three-dimensional topography data of the track fastener by adopting a three-dimensional imaging system, and converting the three-dimensional topography data into a two-dimensional depth image by taking a track plane as a horizontal reference plane of the depth image;
step 2: in the two-dimensional depth image, detecting a fastener area by adopting a threshold segmentation method according to the fastener height prior;
and step 3: extracting a fastener area according to the position of a fastener in the two-dimensional depth image and the fastener area;
and 4, step 4: calculating the height h of the movable part in the two-dimensional depth image and the fastener areacAnd angle betac；
And 5: calculating the floating height delta h of the movable part as hc-hb，hbIs the reference height of the movable part, and the rotation angle of the movable part is calculated as [ Delta beta ]c-βb|，βbIs the movable part reference angle;
step 6: setting a height determination threshold ThSum angle determination threshold TβWhen the floating height of the movable part is delta h > ThOr a rotation angle Delta beta > TβJudging that the fastener is loosened;
the fastener comprises a fixed part and a movable part, wherein the fixed part comprises a threaded hole seat or a threaded rod, and the movable part comprises a bolt or a nut.
The three-dimensional imaging system in the step 1 comprises a line structured light scanning three-dimensional imaging system, a surface structured light three-dimensional imaging system, a monocular laser speckle three-dimensional imaging system, a binocular stereoscopic vision three-dimensional imaging system, a TOF three-dimensional imaging system and a light field imaging three-dimensional imaging system; when a linear structured light scanning three-dimensional imaging system is adopted, the scanning direction calibration needs to be carried out on the obtained three-dimensional shape data so as to ensure that the physical sizes represented by the horizontal coordinates and the vertical coordinates of the pixels in the converted two-dimensional depth image are equal.
The movable part is a bolt, and the height h of the movable part is calculated in the step 4 by adopting the following stepscThe first method of (1) is:
step a-4-1: and (3) carrying out threshold segmentation on pixels in the fastener region to obtain a region R:
where f (x, y) represents the depth image pixel grayscale value at (x, y) in the fastener region, vmaxThe maximum value of the pixel gray scale of the depth image in the fastener area is obtained, a is a fixed constant, and the value range of a is 0-50;
step a-4-2: calculating a pixel histogram of the depth image in the region R, and taking a non-zero pixel value with the largest quantity in the pixel histogram as the height h of the movable partc。
The movable part is a bolt, and the height h of the movable part is calculated in step 4cThe second method of (2) is:
in the step a-4-2, the mean value or the median value of the depth image pixels in the region R is calculated, or regular polygon fitting is carried out on non-zero elements in the region R, and the mean value or the median value of the depth image pixels in the regular polygon or the circle inscribed region of the regular polygon is extracted and used as the height h of the movable partcThe regular polygon comprises a regular hexagon and a quadrangle.
The movable part is a nut, and the movable part is calculated in step 4Height h of the partcThe first method of (1) is:
step b-4-1: same as step a-4-1;
step b-4-2: performing morphological filtering on the region R, eliminating isolated noise, calculating the center c (x, y) and the radius R1 of a circumscribed circle of the region R when R1>R2+ e2, calculating the mean or median of the depth image pixels in the region R as the height h of the movable partcOtherwise, turning to the step b-4-3;
step b-4-3: c (x, y) is taken as a center, concentric rings with the radius r5 and r6 are set, r5 is r2+ e3, r4 is r3-e3, and the mean value or the median value of pixels in the concentric ring area is taken as the height h of the movable part on the depth imagec；
Wherein e2 and e3 are deviation values, and the value range is 0-50; r2 is the screw radius, and r3 is the screw cap circumcircle radius, which is calculated in advance.
The movable part is a nut, and the height h of the movable part is calculated in step 4cThe second method of (2) is:
step c-4-1: same as step a-4-1;
step c-4-2: performing morphological filtering on the region R, eliminating isolated noise, calculating the central position of the region R, setting a circle by taking the central position as a circular point and giving a radius R1, sampling a depth image on the circle, and obtaining a sampling sequence S, wherein the value range of R1 is as follows: r2< r1< r3, wherein r2 is the radius of the screw, and r3 is the radius of the circumcircle of the screw cap, and is obtained by calculation in advance;
step c-4-3: carrying out thresholding treatment on the sampling sequence S to obtain a new sampling sequence S';
step c-4-4: calculating the mean or median of the new sampling sequence S' as the height h of the movable partc。
The movable part is a bolt, and the angle beta of the movable part is calculated in step 4cThe method comprises the following steps: when the top of the bolt is marked by numbers or characters: adopting a template matching method, and taking the image direction of the template under the optimal matching condition as the angle beta of the movable partc(ii) a When no marks such as numbers or characters exist on the top of the bolt: execution stepa-4-1, obtaining a region R, performing morphological filtering and isolated noise elimination on the region R, then performing regular polygon fitting on the region R, and taking a vertex connecting line with the smallest included angle with the track direction in symmetrical vertex connecting lines of the fitted regular polygon as an angle beta of a movable partcAnd the polygon comprises a hexagon and a quadrangle.
The movable part is a nut, and the angle beta of the movable part is calculated in step 4cThe method comprises the following steps:
step d-4-1: same as step a-4-1;
step d-4-2: same as step c-4-2;
step d-4-3: taking the mean value or the median value of the sampling sequence S as a threshold value, and carrying out threshold segmentation on depth image pixels in the fastener region to obtain a nut region G;
step d-4-4: fitting regular polygon to the nut region G, and taking the vertex connecting line with the minimum included angle with the track direction in the symmetrical vertex connecting lines of the fitted regular polygon as the angle beta of the movable partcAnd the polygon comprises a hexagon and a quadrangle.
The reference height h of the movable part in the step 4bAnd a reference angle betabIs the measured value, the measuring method and the calculation h of the current fastener under the condition of not looseningc、βcThe method is the same; during detection, the number K of the current fastener is obtained by counting the sleepers or the fasteners, and then the reference height h with the number K is extracted from the reference height data set and the reference angle data setbAnd a reference angle betabAnd (4) finishing.
The invention has the beneficial effects that: according to the method, the three-dimensional imaging system is adopted to obtain the depth image of the fastener, the floating height and the rotating angle of the bolt or the nut are measured in the depth image at the same time, and the height and angle change information is utilized to judge whether the bolt or the nut is loosened, so that the reliability of the detection of the looseness of the fastener can be obviously improved, and the requirement on the depth measurement precision of the three-dimensional imaging system is reduced. The concrete embodiment is as follows: in theory, it is possible to determine whether or not the fastener is loosened by simply using the floating height value of the bolt or the nut, but it is difficult to detect slight loosening when the depth resolution of the three-dimensional imaging system is insufficient. Taking a certain type of fastener as an example, the nut loosens for one circle and floats for 0.5mm, and when the nut loosens for 0.25 circle, the height change is 0.125mm, which draws high requirements on the precision of the three-dimensional imaging system. On the two-dimensional depth image, the bolt rotates 0.25 circles, and a large angle change is generated: 45 degrees, the bolt looseness is easily detected through angle change. However, it is difficult to distinguish the problem of angular coincidence of regular polygon structures using only angular variations. Taking a hexagon as an example, the patterns are overlapped when rotated by 60 degrees, 120 degrees, and 180 degrees. When the surface of the bolt or the nut does not have identification information such as numbers or letters, the rotation angle is difficult to accurately calculate by simply depending on a pattern matching method. In the depth image, when the polygonal patterns of the rotated bolt or nut are overlapped, the height values thereof are necessarily different, and thus the ambiguity of the angular rotation can be eliminated by using the height values. In summary, reliable detection of fastener loosening can be achieved as long as the three-dimensional imaging system can distinguish height variation values under the condition of 60-degree rotation (60-degree hexagonal rotation coincidence angle and 90-degree quadrilateral rotation coincidence angle).
Compared with the existing fastener loosening detection method based on images, the method provided by the invention has the advantages that the bolt or nut loosening detection is carried out by directly utilizing the height value and the angle value change of the bolt or nut, and the method is simple, visual, stable and reliable; compared with the CN201580000881X method, the method does not change the structure of the existing rail fastener, can be directly applied to the detection of the constructed railway fastener, and can realize the automatic detection of the loosening of the screw cap; compared with methods such as CN2012101926412 and CN2013101590001, the method provided by the invention does not need to adopt a train or a large-scale rail inspection vehicle to extrude and do work on the steel rail, can be used for installing a three-dimensional imaging system on a daily inspection trolley, a train or a large-scale rail inspection vehicle, and has the advantages of wider application range and more convenient and flexible use mode.
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 hexagonal bolt of a W-shaped fastener.
FIG. 4 is a schematic view of a quadrilateral bolt of a W-shaped fastener.
Fig. 5 is a schematic view of a hexagonal nut for a W-type fastener.
In the figure, 1 is a line structured light projector, 2 is an area array camera, 3 is a rail, 4 is a clip, 5 is a hexagonal bolt, 6 is a nut washer, 7 is a spring bar, 8 is a quadrangular bolt, 9 is a screw, 10 is a hexagonal nut, and 11 is a clip support hexagonal bolt.
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 loosening of a fastener of a track fastener by using a depth image, and the specific implementation manner is as follows:
example 1
The present embodiment will be described by taking a hexagonal bolt shown in fig. 3 as an example.
Step 1: a three-dimensional imaging system is adopted to obtain three-dimensional shape data of the track fastener, and the three-dimensional shape data is converted into a two-dimensional depth image by taking a track plane as a horizontal datum plane.
The linear structured light scanning three-dimensional imaging system shown in fig. 2 is adopted to obtain the three-dimensional topography data of the track fastener. As shown in fig. 2, the linear structure light scanning three-dimensional imaging system includes a linear structure light projector 1 and an area array camera 2, the position and the angle of the linear structure light projector 1 and the area array camera 2 are kept fixed, the linear structure light projector 1 generates a linear structure light sheet light, the perpendicular to rail 3 projects onto a fastener 4, a section profile is formed on the surface of the fastener 4, the section profile is shot by the area array camera 2, according to the geometric position relation between the linear structure light sheet light plane and the area array camera 2, the section profile coordinate can be calculated, the three-dimensional imaging system scans along the track direction, and the three-dimensional shape data of the fasteners 4 on the two sides of the track can be obtained. When the three-dimensional shape data is obtained, the scanning direction of the collected three-dimensional shape data needs to be calibrated so as to ensure that the physical sizes represented by the unit pixels of the abscissa and the ordinate in the converted two-dimensional depth image are equal. The bit width of the two-dimensional depth image is 16bits, and the height resolution is 0.5 mm.
Step 2: in the two-dimensional depth image, detecting a fastener area by adopting a threshold segmentation method according to the fastener height prior;
and step 3: extracting a fastener area according to the position of a fastener in the two-dimensional depth image and the fastener area;
and 4, step 4: calculating bolt height h 'in the two-dimensional depth image and fastener area'cAnd angle beta'c；
Step a-4-1: and (3) carrying out threshold segmentation on pixels in the fastener region to obtain a region R:
where f (x, y) represents the depth image pixel grayscale value at (x, y) in the fastener region, vmaxThe maximum value of the pixel gray scale of the depth image in the fastener area is obtained, a is a fixed constant, and the value range of a is 0-50;
step a-4-2: calculating a depth image pixel histogram in the region R, and taking the non-zero pixel value with the maximum number in the depth image pixel histogram as the bolt height h'c。
When the top of the bolt is marked by numbers or characters: adopting a template matching method, and taking the image direction of the template under the optimal matching condition as the bolt angle betac(ii) a When no marks such as numbers or characters exist on the top of the bolt: step a-4-1 is executed to obtain a region R, morphological filtering and isolated noise elimination are carried out on the region R, regular polygon fitting is carried out on the region R, and a vertex connecting line with the smallest included angle with the track direction in symmetrical vertex connecting lines of the fitted regular polygon is used as a bolt angle beta'cAnd the polygon is a hexagon.
And 5: calculating bolt floating height delta h ═ h'c-h′b，h′bIs a boltCalculating the bolt rotation angle delta beta ═ beta 'according to the height'c-β′b|，β′bIs a bolt reference angle;
the reference height h'bAnd a reference angle of β'bIs the measured value, the measuring method and the calculation h 'of the current bolt under the condition of not loosening'c、β′cThe method is the same; during detection, counting the sleeper or the fastener to obtain the number K of the current fastener, and extracting the reference height h 'with the number K from the reference height data set and the reference angle data set'bAnd a reference angle of β'bAnd (4) finishing.
Step 6: setting a height determination threshold ThSum angle determination threshold TβWhen the floating height of the bolt is delta h' > ThOr a rotation angle Δ β' > TβJudging that the bolt is loosened; wherein the height decision threshold value ThAngle decision threshold T of 10β＝10。
Example 2
The difference from embodiment 1 is that, in step a-4-2, the mean value or median value of the depth image pixels in the region R is calculated, or regular polygon fitting is performed on non-zero elements in the region R, and the mean value or median value of the depth image pixels in the regular polygon or regular polygon inscribed circle region is extracted as the bolt height h'cAnd the regular polygon is a regular hexagon.
Example 3
The difference from embodiment 1 is that the quadrangular bolt shown in fig. 4, which is a square, is processed.
Example 4
The difference from example 1 is that the hexagonal nut shown in fig. 5 was processed, and the nut height h ″' was calculated in step 4 by the following procedurec：
Step b-4-1: same as step a-4-1;
step b-4-2: performing morphological filtering on the region R, eliminating isolated noise, calculating the center c (x, y) and the radius R1 of a circumscribed circle of the region R when R1>R2+ e2, calculating the mean or median of the depth image pixels in the region R as the nut height hcOtherwise, turning to the step b-4-3;
step b-4-3: c (x, y) is taken as a center, concentric rings with the radiuses of r5 and r6 are set, r5 is r2+ e3, r4 is r3-e3, and the mean value or the median value of pixels in the concentric ring area is taken as the height h of the nut on the depth imagec″；
Wherein e 2-5 and e 3-5; r2 is the screw radius, and r3 is the screw cap circumcircle radius, which is calculated in advance.
The method comprises calculating the angle beta of the nutc″：
Step d-4-1: same as step a-4-1;
step d-4-2: same as step c-4-2;
step d-4-3: taking the mean value or the median value M of the sampling sequence S as a reference, and carrying out threshold segmentation on depth image pixels in the fastener area to obtain a nut area G;
wherein f (x, y) represents the depth image pixel grayscale value at (x, y) in the fastener region, b-5;
step d-4-4: performing regular hexagon fitting on the nut area G, and taking the vertex connecting line with the smallest included angle with the track direction in the symmetrical vertex connecting lines of the fitted regular hexagon as the nut angle betac″。
Example 5
The difference from embodiment 4 is that the nut height h is calculated in step 4 by the following methodc″：
Step c-4-1: same as step a-4-1;
step c-4-2: performing morphological filtering on the region R to eliminate isolated noise, calculating the central position of the region R, setting a circle by taking the central position as a circular point and giving a radius R1, sampling the depth image on the circle, and obtaining a sampling sequence S ═ S { S }1,s2,...,snThe value range of r1 is as follows: r2<r1<r3, wherein r2 is the radius of the screw, and r3 is the radius of the circumcircle of the screw cap, which is obtained by calculation in advance;
step c-4-3: setting a threshold valueFor sampling sequence S ═ S1,s2,...,snCarry out thresholding
Obtaining a new sampling sequence S' ═ S1',s'2,…,s'n}；
Step c-4-4: calculating a new sample sequence S' ═ S1',s'2,…,s'nThe mean or median value of }, as the nut height h ″c。
While the invention has been described with reference to a preferred embodiment, it will be understood by those skilled in the art that various changes may be made and equivalents may be substituted for elements thereof without departing from the scope of the invention.
Claims (9)
1. The utility model provides an adopt track fastener looseness detecting method of degree of depth image which characterized in that: the method comprises the following steps:
step 1: acquiring three-dimensional topography data of the track fastener by adopting a three-dimensional imaging system, and converting the three-dimensional topography data into a two-dimensional depth image by taking a track plane as a horizontal reference plane of the depth image;
step 2: in the two-dimensional depth image, detecting a fastener area by adopting a threshold segmentation method according to the fastener height prior;
and step 3: extracting a fastener area according to the position of a fastener in the two-dimensional depth image and the fastener area;
and 4, step 4: calculating the height h of the movable part in the two-dimensional depth image and the fastener areacAnd angle betac；
And 5: calculating the floating height delta h of the movable part as hc-hb，hbIs the reference height of the movable part, and the rotation angle of the movable part is calculated as [ Delta beta ]c-βb|，βbIs the movable part reference angle;
step 6: setting a height determination threshold ThSum angle determination threshold TβWhen the floating height of the movable part is delta h > ThOr a rotation angle Delta beta > TβJudging that the fastener is loosened;
the fastener comprises a fixed part and a movable part, wherein the fixed part comprises a threaded hole seat or a threaded rod, and the movable part comprises a bolt or a nut.
2. The track fastener loosening detection method using depth image according to claim 1, characterized in that: the three-dimensional imaging system in the step 1 comprises a line structured light scanning three-dimensional imaging system, a surface structured light three-dimensional imaging system, a monocular laser speckle three-dimensional imaging system, a binocular stereoscopic vision three-dimensional imaging system, a TOF three-dimensional imaging system and a light field imaging three-dimensional imaging system; when a linear structured light scanning three-dimensional imaging system is adopted, the scanning direction calibration needs to be carried out on the obtained three-dimensional shape data so as to ensure that the physical sizes represented by the horizontal coordinates and the vertical coordinates of the pixels in the converted two-dimensional depth image are equal.
3. The track fastener loosening detection method using the depth image according to claim 1 or 2, characterized in that: when the movable part is a bolt, the height h of the movable part is calculated in step 4 by adopting the following stepsc：
Step a-4-1: carrying out threshold segmentation on the depth image pixels in the fastener region to obtain a region R:
where f (x, y) represents the depth image pixel grayscale value at (x, y) in the fastener region, vmaxThe maximum value of the pixel gray scale of the depth image in the fastener area is obtained, a is a fixed constant, and the value range of a is 0-50;
step a-4-2: computing a histogram of depth image pixels located in region RTaking the non-zero pixel value with the largest number in the pixel histogram as the height h of the movable partc。
4. The track fastener loosening detection method using depth image according to claim 3, characterized in that: when the movable part is a screw cap, the height h of the movable part is calculated in the step 4 by adopting the following stepsc：
Step b-4-1: same as step a-4-1;
step b-4-2: performing morphological filtering on the region R, eliminating isolated noise, calculating the center c (x, y) and the radius R1 of a circumscribed circle of the region R when R1>R2+ e2, calculating the mean or median of the depth image pixels in the region R as the nut height hcOtherwise, turning to the step b-4-3;
step b-4-3: c (x, y) is taken as a center, concentric rings with the radiuses of r5 and r6 are set, r5 is r2+ e3, r4 is r3-e3, and the average value or the median value of pixels in the concentric ring area is taken as the height h of the movable part on the depth imagec；
Wherein e2 and e3 are deviation values, and the value range is 0-50; r2 is the screw radius, and r3 is the screw cap circumcircle radius, which is calculated in advance.
5. The track fastener loosening detection method using depth image according to claim 3, characterized in that: when the movable part is a bolt, in the step a-4-2, calculating the mean value or the median value of the depth image pixels in the region R, or performing regular polygon fitting on non-zero elements in the region R, and extracting the mean value or the median value of the depth image pixels in a circle inscribed region of a regular polygon or the regular polygon as the height h of the movable partcThe regular polygon comprises a regular hexagon and a quadrangle.
6. The track fastener loosening detection method using depth image according to claim 3, characterized in that: the movable part is a nut, and the height h of the movable part is calculated by adopting the following stepsc：
Step c-4-1: same as step a-4-1;
step c-4-2: performing morphological filtering on the region R, eliminating isolated noise, calculating the central position of the region R, setting a circle by taking the central position as a circular point and giving a radius R1, sampling a depth image on the circle, and obtaining a sampling sequence S, wherein the value range of R1 is as follows: r2< r1< r3, wherein r2 is the radius of the screw, and r3 is the radius of the circumcircle of the screw cap, and is obtained by calculation in advance;
step c-4-3: carrying out thresholding treatment on the sampling sequence S to obtain a new sampling sequence S';
step c-4-4: the mean or median of the new sample sequence S' is calculated as the active part height.
7. The track fastener loosening detection method using depth image according to claim 3, characterized in that: the movable part is a bolt, and the angle beta of the movable part is calculatedcThe method comprises the following steps: when the top of the bolt is marked by numbers or characters: adopting a template matching method, and taking the image direction of the template under the optimal matching condition as the angle beta of the movable partc(ii) a When no marks such as numbers or characters exist on the top of the bolt: step a-4-1 is executed to obtain a region R, morphological filtering and isolated noise elimination are carried out on the region R, regular polygon fitting is carried out on the region R, and the vertex connecting line with the smallest included angle with the track direction in the symmetrical vertex connecting lines of the fitted regular polygon is used as the angle beta of the movable partcAnd the polygon comprises a hexagon and a quadrangle.
8. The track fastener loosening detection method adopting the depth image according to any one of claims 3 to 7, characterized in that: the movable part is a nut, and the angle beta of the movable part is calculatedcThe method comprises the following steps:
step d-4-1: same as step a-4-1;
step d-4-2: same as step c-4-2;
step d-4-3: taking the mean value or the median value of the sampling sequence S as a threshold value, and carrying out threshold segmentation on depth image pixels in the fastener region to obtain a nut region G;
step d-4-4: fitting regular polygon to the nut region G, and taking the vertex connecting line with the minimum included angle with the track direction in the symmetrical vertex connecting lines of the fitted regular polygon as the angle beta of the movable partcAnd the polygon comprises a hexagon and a quadrangle.
9. The track fastener loosening detection method using the depth image according to any one of claims 1 to 8, characterized in that: the reference height h of the movable part in the step 4bAnd a reference angle betabIs the measured value, the measuring method and the calculation h of the current fastener under the condition of not looseningc、βcThe method is the same; during detection, the number K of the current fastener is obtained by counting the sleepers or the fasteners, and then the reference height h with the number K is extracted from the reference height data set and the reference angle data setbAnd a reference angle betabAnd (4) finishing.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810583193.6A CN110634123A (en) | 2018-06-05 | 2018-06-05 | Track fastener loosening detection method adopting depth image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810583193.6A CN110634123A (en) | 2018-06-05 | 2018-06-05 | Track fastener loosening detection method adopting depth image |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110634123A true CN110634123A (en) | 2019-12-31 |
Family
ID=68967531
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810583193.6A Pending CN110634123A (en) | 2018-06-05 | 2018-06-05 | Track fastener loosening detection method adopting depth image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110634123A (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111311560A (en) * | 2020-02-10 | 2020-06-19 | 中国铁道科学研究院集团有限公司基础设施检测研究所 | Method and device for detecting state of steel rail fastener |
| CN111666943A (en) * | 2020-04-20 | 2020-09-15 | 国网浙江省电力有限公司 | Matching detection method and system for bolt and nut of power transmission line based on image recognition |
| CN112149519A (en) * | 2020-09-03 | 2020-12-29 | 武汉纵横天地空间信息技术有限公司 | Sleeper identification method and system based on image identification and identification code identification |
| CN112365461A (en) * | 2020-11-06 | 2021-02-12 | 北京格灵深瞳信息技术有限公司 | Fastener loosening identification method, system, terminal and storage medium |
| CN112419297A (en) * | 2020-12-04 | 2021-02-26 | 中冶建筑研究总院（深圳）有限公司 | Bolt looseness detection method, device, equipment and storage medium |
| CN112950562A (en) * | 2021-02-22 | 2021-06-11 | 杭州申昊科技股份有限公司 | Fastener detection algorithm based on line structured light |
| CN114387268A (en) * | 2022-03-22 | 2022-04-22 | 中国长江三峡集团有限公司 | Bolt looseness detection method and device |
| CN114742820A (en) * | 2022-05-11 | 2022-07-12 | 西南交通大学 | Bolt looseness detection method and system based on deep learning and storage medium |
| CN116993676A (en) * | 2023-07-03 | 2023-11-03 | 中铁九局集团电务工程有限公司 | Subway rail fastener counting and positioning method based on deep learning |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US4918627A (en) * | 1986-08-04 | 1990-04-17 | Fmc Corporation | Computer integrated gaging system |
| JPH0815062A (en) * | 1994-06-30 | 1996-01-19 | East Japan Railway Co | Relaxation detection device of rail fastener |
| DE102013002554A1 (en) * | 2013-02-15 | 2014-08-21 | Jungheinrich Aktiengesellschaft | Method for detecting objects in a warehouse and / or for spatial orientation in a warehouse |
| CN205382373U (en) * | 2015-12-23 | 2016-07-13 | 李社军 | Static railway rails subassembly dimensions and disease optical detection system |
| CN106600581A (en) * | 2016-12-02 | 2017-04-26 | 北京航空航天大学 | Train operation fault automatic detection system and method based on binocular stereoscopic vision |
| CN107688024A (en) * | 2017-10-13 | 2018-02-13 | 成都精工华耀机械制造有限公司 | A kind of railway rail clip abnormality detection system based on monocular vision and laser speckle |
| US20180061034A1 (en) * | 2016-08-30 | 2018-03-01 | Microsoft Technology Licensing, Llc | Deformation Detection and Automatic Calibration for a Depth Imaging System |
| CN207248768U (en) * | 2017-10-13 | 2018-04-17 | 成都精工华耀机械制造有限公司 | A kind of railway rail clip abnormal detector based on linear array thermal camera |
-
2018
- 2018-06-05 CN CN201810583193.6A patent/CN110634123A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US4918627A (en) * | 1986-08-04 | 1990-04-17 | Fmc Corporation | Computer integrated gaging system |
| JPH0815062A (en) * | 1994-06-30 | 1996-01-19 | East Japan Railway Co | Relaxation detection device of rail fastener |
| DE102013002554A1 (en) * | 2013-02-15 | 2014-08-21 | Jungheinrich Aktiengesellschaft | Method for detecting objects in a warehouse and / or for spatial orientation in a warehouse |
| CN205382373U (en) * | 2015-12-23 | 2016-07-13 | 李社军 | Static railway rails subassembly dimensions and disease optical detection system |
| US20180061034A1 (en) * | 2016-08-30 | 2018-03-01 | Microsoft Technology Licensing, Llc | Deformation Detection and Automatic Calibration for a Depth Imaging System |
| CN106600581A (en) * | 2016-12-02 | 2017-04-26 | 北京航空航天大学 | Train operation fault automatic detection system and method based on binocular stereoscopic vision |
| CN107688024A (en) * | 2017-10-13 | 2018-02-13 | 成都精工华耀机械制造有限公司 | A kind of railway rail clip abnormality detection system based on monocular vision and laser speckle |
| CN207248768U (en) * | 2017-10-13 | 2018-04-17 | 成都精工华耀机械制造有限公司 | A kind of railway rail clip abnormal detector based on linear array thermal camera |
Non-Patent Citations (1)
| Title |
|---|
| 黄志文;张学习;: "输电线杆塔横担处的螺栓缺陷检测技术研究" * |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111311560A (en) * | 2020-02-10 | 2020-06-19 | 中国铁道科学研究院集团有限公司基础设施检测研究所 | Method and device for detecting state of steel rail fastener |
| CN111311560B (en) * | 2020-02-10 | 2023-09-12 | 中国铁道科学研究院集团有限公司基础设施检测研究所 | Method and device for detecting state of steel rail fastener |
| CN111666943A (en) * | 2020-04-20 | 2020-09-15 | 国网浙江省电力有限公司 | Matching detection method and system for bolt and nut of power transmission line based on image recognition |
| CN112149519B (en) * | 2020-09-03 | 2021-06-22 | 武汉纵横天地空间信息技术有限公司 | Sleeper identification method and system based on image identification and identification code identification |
| CN112149519A (en) * | 2020-09-03 | 2020-12-29 | 武汉纵横天地空间信息技术有限公司 | Sleeper identification method and system based on image identification and identification code identification |
| CN112365461A (en) * | 2020-11-06 | 2021-02-12 | 北京格灵深瞳信息技术有限公司 | Fastener loosening identification method, system, terminal and storage medium |
| CN112419297A (en) * | 2020-12-04 | 2021-02-26 | 中冶建筑研究总院（深圳）有限公司 | Bolt looseness detection method, device, equipment and storage medium |
| CN112419297B (en) * | 2020-12-04 | 2024-06-04 | 中冶建筑研究总院（深圳）有限公司 | Bolt loosening detection method, device, equipment and storage medium |
| CN112950562A (en) * | 2021-02-22 | 2021-06-11 | 杭州申昊科技股份有限公司 | Fastener detection algorithm based on line structured light |
| CN114387268A (en) * | 2022-03-22 | 2022-04-22 | 中国长江三峡集团有限公司 | Bolt looseness detection method and device |
| CN114742820A (en) * | 2022-05-11 | 2022-07-12 | 西南交通大学 | Bolt looseness detection method and system based on deep learning and storage medium |
| CN116993676A (en) * | 2023-07-03 | 2023-11-03 | 中铁九局集团电务工程有限公司 | Subway rail fastener counting and positioning method based on deep learning |
| CN116993676B (en) * | 2023-07-03 | 2024-05-07 | 中铁九局集团电务工程有限公司 | Subway rail fastener counting and positioning method based on deep learning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110634123A (en) | Track fastener loosening detection method adopting depth image | |
| CN110567680B (en) | Track fastener looseness detection method based on angle comparison | |
| CN110634121A (en) | Track fastener loosening detection method based on texture and depth images | |
| CN109029283B (en) | Track fastener bolt floating detection method based on height comparison | |
| AU2016385541B2 (en) | Object surface deformation feature extraction method based on line scanning three-dimensional point Cloud | |
| CN108797241B (en) | Track fastener nut looseness detection method based on height comparison | |
| CN106091951B (en) | A kind of municipal rail train wheel rim parameter on-line detecting system and method | |
| CN104792790A (en) | Tunnel state detection device and detection method | |
| CN109059775B (en) | Steel rail abrasion detection method with image edge extraction step | |
| CN103149222A (en) | Flaw detection device in real-time imaging of ray | |
| CN101126638A (en) | Pick-up measuring method for checking road surface planeness | |
| CN108177660B (en) | Steel rail abrasion detection method with laser image processing step | |
| AU2022349296A1 (en) | Road surface technical condition detection method and device based on three-dimensional contour | |
| CN109613546A (en) | Converter furnace chamber method for three-dimensional measurement and measuring device based on three-dimensional laser radar auxiliary positioning | |
| JP2003207319A (en) | Sectional shape of rail measuring device | |
| CN108819980B (en) | Device and method for online dynamic measurement of geometric parameters of train wheels | |
| CN204495733U (en) | Tunnel state pick-up unit | |
| CN108839674B (en) | Online dynamic measurement device and measurement method for geometric parameters of train wheels | |
| CN106813569A (en) | A kind of automobile tire 3-D positioning method based on line-structured light | |
| CN107117188B (en) | The vertically arranged tramcar wheel footpath online test method of linear sensor | |
| CN108844465B (en) | Online dynamic measurement device and measurement method for geometric parameters of train wheels | |
| CN113947583B (en) | Weld joint nondestructive testing method based on deep learning two-dimensional time sequence image | |
| CN110567397B (en) | Fastener spring tongue separation detection method | |
| CN109115172B (en) | Machine vision-based method for detecting pose of body of coal mining machine | |
| CN108830891B (en) | Method for detecting looseness of steel rail fishplate fastener |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20191231 |
|
| WD01 | Invention patent application deemed withdrawn after publication |