CN112329781A - Method for detecting loss of pull rivet pin collar of truck brake beam strut based on image restoration - Google Patents
Method for detecting loss of pull rivet pin collar of truck brake beam strut based on image restoration Download PDFInfo
- Publication number
- CN112329781A CN112329781A CN202011217314.9A CN202011217314A CN112329781A CN 112329781 A CN112329781 A CN 112329781A CN 202011217314 A CN202011217314 A CN 202011217314A CN 112329781 A CN112329781 A CN 112329781A
- Authority
- CN
- China
- Prior art keywords
- image
- brake beam
- beam strut
- area
- restoration
- 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.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
- G06V10/267—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion by performing operations on regions, e.g. growing, shrinking or watersheds
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60T—VEHICLE BRAKE CONTROL SYSTEMS OR PARTS THEREOF; BRAKE CONTROL SYSTEMS OR PARTS THEREOF, IN GENERAL; ARRANGEMENT OF BRAKING ELEMENTS ON VEHICLES IN GENERAL; PORTABLE DEVICES FOR PREVENTING UNWANTED MOVEMENT OF VEHICLES; VEHICLE MODIFICATIONS TO FACILITATE COOLING OF BRAKES
- B60T17/00—Component parts, details, or accessories of power brake systems not covered by groups B60T8/00, B60T13/00 or B60T15/00, or presenting other characteristic features
- B60T17/18—Safety devices; Monitoring
- B60T17/22—Devices for monitoring or checking brake systems; Signal devices
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Multimedia (AREA)
- Geometry (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Transportation (AREA)
- Mechanical Engineering (AREA)
- Image Analysis (AREA)
Abstract
A method for detecting the loss of a riveting pin sleeve ring of a brake beam strut of a truck based on image restoration relates to the technical field of image recognition, aims at the problems of low detection efficiency and poor accuracy when the failure of the riveting pin sleeve ring of the brake beam strut is manually detected in the prior art, and utilizes an automatic image recognition mode to replace manual detection, so that the detection efficiency and the accuracy are improved. Different from the conventional fault detection methods such as template matching, feature extraction and the like, the fault detection is carried out by adopting the idea of image restoration. The characteristics of bilateral symmetry of the brake beam support and uniform gray distribution on the rivet pin lantern ring are fully considered when the image is repaired, and meanwhile, neighborhood information and similar structural information of the image are utilized, and the repairing effect is improved. And horizontally overturning and splicing the left and right pillar images on the same brake beam, performing GMS matching by using the extracted vanishing point estimation, and calculating affine correction parameters to serve as constraint conditions for image restoration, thereby greatly improving the speed of image restoration.
Description
Technical Field
The invention relates to the technical field of image recognition, in particular to a method for detecting the loss of a riveting pin sleeve of a truck brake beam strut based on image restoration.
Background
A brake beam pillar rivet pin lantern ring loss fault is a fault which endangers driving safety, and fault detection is carried out on a traditional brake beam station in a mode of automatically acquiring images and manually checking the images. With the continuous increase of the freight volume of the railway freight car, a large amount of car inspection personnel are needed to ensure the train inspection time, and the labor intensity is extremely high. And the car inspection personnel need to face massive images every day, so that fatigue, missed inspection and false inspection are easily caused, and the driving safety is endangered.
Different from the situation that expensive GPU equipment investment is required for deep learning, automatic identification of loss faults of brake beam strut rivet pin sleeves can be achieved on inherent equipment by adopting an image processing method, manual work only needs to confirm alarm results, machine inspection operation difficulty and personnel investment are further reduced, vehicle inspection cost is reduced on the basis of ensuring safety, dynamic vehicle inspection quality and working efficiency are improved, and vehicle operation safety is ensured.
Disclosure of Invention
The purpose of the invention is: aiming at the problems of low detection efficiency and poor accuracy when the defect of loss of a brake beam strut rivet pin sleeve ring is manually detected in the prior art, the method for detecting the loss of the brake beam strut rivet pin sleeve ring based on image restoration is provided.
The technical scheme adopted by the invention to solve the technical problems is as follows:
the method for detecting the loss of the pull rivet pin collar of the brake beam strut of the truck based on image restoration comprises the following steps:
the method comprises the following steps: acquiring a linear array image of the railway wagon;
step two: carrying out coarse positioning on a brake beam strut part aiming at the obtained linear array image of the railway wagon;
step three: and (3) carrying out binarization processing on the roughly positioned image, and positioning the position and the direction of the brake beam strut according to the area of a black area in the binarized image, wherein the specific steps are as follows:
step three, firstly: sequentially traversing the black areas, and selecting two areas with the largest areas as the edges of the brake beam strut;
step three: projecting the selected edge to obtain a convex image, wherein the position of the convex image is the position of the brake beam strut angle iron, and the opposite direction of the convex image is the direction of the rivet pin lantern ring;
step four: splicing the brake beam pillar image on one side as the upper part and the brake beam pillar image on the other side as the lower part;
step five: positioning the rivet pin lantern ring in the spliced image according to the direction of the rivet pin lantern ring and by combining the edge of the brake beam strut;
step six: generating a shielding image for shielding the rivet pin sleeve ring in the rivet pin sleeve ring image;
step seven: carrying out image restoration on the shielded rivet pin collar image;
step eight: calculating the absolute value of the image gray difference before and after image restoration to obtain a difference image, and performing binarization processing on the difference image;
step nine: and judging whether the blind rivet pin collar is in failure or not according to the area of the white area in the image after the binarization processing of the difference image.
Further, the brake beam support column position rough positioning in the step two is carried out through axle distance information and the prior position information of the brake beam support column.
Further, the specific steps of the seventh step are as follows:
step seven one: calculating affine correction parameters between two brake beam supports;
step seven and two: obtaining candidate positions of the shielding images according to the positions of the shielding masks in the shielding images and the affine correction parameters, and then carrying out geometric transformation according to the candidate positions of the shielding images;
step seven and three: and expanding the candidate position of the shielding image after the geometric conversion to the upper side, the lower side, the left side and the right side to obtain an expanded image area, wherein the expanded image area is larger than the shielding image, and finally obtaining a repairing image according to the expanded image area.
Further, the candidate positions of the occlusion image after the geometric transformation in the seventh step and the sixth step are expanded to the upper, lower, left and right directions as follows: and expanding the candidate positions of the shielding images after the geometric transformation to the neighborhood of 3 pixels respectively from top to bottom, left to right.
Furthermore, the affine correction parameters are obtained by performing feature matching on the features between the line segments by using the GMS,
extracting characteristics among the line segments according to the line segments to obtain the characteristics;
the line segments are obtained by estimating vanishing points of the shielded images.
Further, vanishing point estimation is performed by LSD line segment detection and J-Linkage ensemble clustering.
Further, the specific steps of obtaining the restored image according to the expanded image area in the seventh step and the third step are as follows:
step seven, three and one: establishing a multi-scale feature map according to the expanded image area;
step seven, three, two: calculating a matching error between the source Patch and the target Patch in a candidate region of the feature map of each scale;
step seven, step three: searching a plurality of scale spaces, repeatedly executing the step seven, the step three and the step two, when the matching error value is lower than a target value, selecting a compensation position from the rivet pin collar image according to the matching error value at the moment, and then solving a fusion boundary by utilizing a graph-cut algorithm;
step seven, three and four: and obtaining a restored image after the fusion boundary is processed by standard Poisson fusion.
Further, the target value was 0.2.
Further, a multi-scale feature map is established through mask information of the expanded image and the shielding image.
Further, the specific step of judging whether the blind rivet pin collar is in fault or not according to the area of the white area in the image after the difference image binarization processing in the ninth step is as follows: and if the area of the white area is larger than 100 pixel points, determining that the white area is a fault, otherwise, determining that the white area is a non-fault.
The invention has the beneficial effects that:
1. and the automatic image identification mode is used for replacing manual detection, so that the detection efficiency and accuracy are improved.
2. Different from the conventional fault detection methods such as template matching, feature extraction and the like, the fault detection is carried out by adopting the idea of image restoration.
3. The characteristics of bilateral symmetry of the brake beam support and uniform gray distribution on the rivet pin lantern ring are fully considered when the image is repaired, and meanwhile, neighborhood information and similar structural information of the image are utilized, and the repairing effect is improved.
4. And horizontally overturning and splicing the left and right pillar images on the same brake beam, performing GMS matching by using the extracted vanishing point estimation, and calculating affine correction parameters to serve as constraint conditions for image restoration, thereby greatly improving the speed of image restoration.
Drawings
FIG. 1 is a flow chart of fault identification;
FIG. 2 is an original image;
FIG. 3 is a binarized image;
FIG. 4 is a normal image stitching image A;
FIG. 5 is a rivet pin collar loss image stitching image B;
FIG. 6 is an occlusion map A;
FIG. 7 is an occlusion map B;
FIG. 8 shows vanishing point detection result A;
fig. 9 is a vanishing point detecting result B;
FIG. 10 is image A after restoration;
FIG. 11 is image B after restoration;
FIG. 12 is a binarization result A of the difference value of the image before and after restoration;
fig. 13 is a binarization result B of the image difference value before and after restoration.
Detailed Description
It should be noted that, in the present invention, the embodiments disclosed in the present application may be combined with each other without conflict.
The first embodiment is as follows: specifically describing the embodiment with reference to fig. 1, the method for detecting the loss of the rivet pin collar of the truck brake beam strut based on image restoration comprises the following steps:
the method comprises the following steps: acquiring a linear array image of the railway wagon;
step two: carrying out coarse positioning on a brake beam strut part aiming at the obtained linear array image of the railway wagon;
step three: and (3) carrying out binarization processing on the roughly positioned image, and positioning the position and the direction of the brake beam strut according to the area of a black area in the binarized image, wherein the specific steps are as follows:
step three, firstly: sequentially traversing the black areas, and selecting two areas with the largest areas as the edges of the brake beam strut;
step three: projecting the selected edge to obtain a convex image, wherein the position of the convex image is the position of the brake beam strut angle iron, and the opposite direction of the convex image is the direction of the rivet pin lantern ring;
step four: splicing the brake beam pillar image on one side as the upper part and the brake beam pillar image on the other side as the lower part;
step five: positioning the rivet pin lantern ring in the spliced image according to the direction of the rivet pin lantern ring and by combining the edge of the brake beam strut;
step six: generating a shielding image for shielding the rivet pin sleeve ring in the rivet pin sleeve ring image;
step seven: carrying out image restoration on the shielded rivet pin collar image;
step eight: calculating the absolute value of the image gray difference before and after image restoration to obtain a difference image, and performing binarization processing on the difference image;
step nine: and judging whether the blind rivet pin collar is in failure or not according to the area of the white area in the image after the binarization processing of the difference image.
The brake beam can generate certain change along with brake, and the images of different brake beams at different stations are slightly different. In the traditional image processing methods, such as template matching, feature extraction and other fault detection methods, it is difficult to select proper templates and features to be suitable for all images; deep learning methods require a large amount of data to be collected for training. The invention fully utilizes the inherent characteristics of the component and adopts an image restoration mode to detect the fault.
Image restoration is initially to remove occlusions or to restore a broken image. The existing information in the image is analyzed to restore the missing part in the image, so that the quality of the image is improved. Such as improving the sharpness of the image, removing image noise, or replacing erroneous data in the image with replacement data. Image restoration requires a plurality of premise assumptions, and the optimization criterion is established under certain constraint conditions, so that the image restoration is realized.
The detection problem of the loss of the pull rivet pin collar of the brake beam strut of the railway wagon is the loss of parts in a detection image, and is equivalent to the inverse problem of image restoration. Namely, we assume that the component is lost, namely, the component area is shielded, the image is repaired, and whether the lost fault occurs is judged by comparing the difference before and after the repair.
Secondly, the difficulty of image restoration is how to complement this occlusion part with other information. Therefore, the reliability of the information source is critical to the repair effect. The left part and the right part of the brake beam are symmetrical, and the gray level distribution on the rivet pin lantern ring is uniform, so that the image can be repaired by using peripheral pixels and structural information at the same time. The unique component information of the rivet pin lantern ring can enable image restoration to be more reliable, and therefore restoration effect is improved. And therefore when the loss fault does not occur, the information is fully utilized, so that the image change before and after the repair is smaller; on the contrary, when a failure occurs, the image change before and after the repair is large. Therefore, whether the brake beam strut rivet pin lantern ring is lost or not can be judged according to the difference.
Image stitching
Because the left and right pillar images of the same brake beam (as shown in fig. 2) have symmetry and high similarity, the right image is horizontally flipped and spliced below the left image (as shown in fig. 4 and 5). Therefore, in the same image, the structure and texture information which are relatively similar exist, and compared with the condition that the matching is carried out by using a uniform template, the anti-interference performance is improved. In addition, similar structure information in the spliced images can provide important basis for image restoration.
Pull rivet pin lantern ring loss fault judgment
And calculating the image difference value before and after restoration, carrying out binarization on the image difference value, and judging whether a fault occurs according to the size of a white area in the binarized image. As shown in fig. 12 and 13.
The second embodiment is as follows: the present embodiment is a further description of the first embodiment, and the difference between the present embodiment and the first embodiment is that the coarse positioning of the brake beam strut position in the second step is performed by using the wheel base information and the prior position information of the brake beam strut.
Coarse positioning of images
The brake beam strut components can be roughly located by using the wheel base information, the position of the components and other prior knowledge, and then the brake beam strut regions to be identified are cut out from a large map of the whole train (as shown in figure 2).
The image is reduced and then binarized (as shown in fig. 3), and screening is performed according to the area of the black communication area, so that the position and the direction of the pillar can be positioned.
The third concrete implementation mode: this embodiment mode is a further description of the second embodiment mode, and the difference between this embodiment mode and the second embodiment mode is that the specific step of step seven is:
step seven one: calculating affine correction parameters between two brake beam supports;
step seven and two: obtaining candidate positions of the shielding images according to the positions of the shielding masks in the shielding images and the affine correction parameters, and then carrying out geometric transformation according to the candidate positions of the shielding images;
step seven and three: and expanding the candidate position of the shielding image after the geometric conversion to the upper side, the lower side, the left side and the right side to obtain an expanded image area, wherein the expanded image area is larger than the shielding image, and finally obtaining a repairing image according to the expanded image area.
Image restoration
Rivet pin collar positioning
From the shaded areas near the brake beam strut and the shaded areas of the brake beam bodies on both sides (as shown in fig. 3), the center position of the strut, i.e., the position information of the rivet pin collar, can be calculated.
② generating a shading image
Comparing the normal image with the image of the lost rivet pin collar (as shown in fig. 4 and 5), it can be seen that the difference between the central areas of the rivet pin collar before and after the loss is the largest, so we choose to generate a blocking image in the central area of the rivet pin collar (as shown in fig. 6 and 7) and repair it.
Image restoration
Image restoration methods typically infer unknown information using known image data, and therefore, the reliability of the information source is critical to the effectiveness of the restoration. Because the brake beam strut rivet pin lantern ring is lost and lacks a fixed part, the angle iron is stressed to pull to generate certain displacement and plane rotation in the running process of a train, and therefore, the left and right strut images may have certain perspective transformation. The traditional image restoration is generally based on the same plane restoration, namely, the redundancy of the image is utilized, and the information of the known part of the image is used for completing the unknown part; or a large amount of calculation is needed to obtain possible plane information, and then searching and selecting are carried out in a plurality of planes. For the situation, structural information in the spliced image is fully utilized, and the process is improved:
first, affine correction parameters between the left and right struts need to be calculated. When the traditional method for extracting SIFT or ORB features is used, after the rivet pin lantern ring is lost, the feature points are difficult to match due to large difference. Therefore, we first perform the vanishing point estimation, as shown in fig. 8 and 9. And performing feature matching by using GMS (Gaussian filtered minimum mean Square) by using the line segments as features, and calculating affine correction parameters as constraint conditions for image restoration. The information can be used as middle layer structure analysis to guide a low-level synthesis algorithm, and the image restoration effect is greatly improved.
Secondly, for the image of the occlusion part, the candidate position and the geometric transformation can be directly calculated according to the position of the occlusion mask and the affine correction parameter calculated in the previous step. Thus, the regular sampling scheme greatly improves the completion quality and repair speed in the presence of repetitive structures in the image.
Finally, the repaired image is synthesized through multiple iterations within a small neighborhood of the candidate location, as shown in fig. 10 and 11.
The fourth concrete implementation mode: the present embodiment is a further description of the third embodiment, and the difference between the present embodiment and the third embodiment is that the candidate positions of the occlusion image after the geometric transformation in the seventh embodiment are expanded to the following upper, lower, left and right directions: and expanding the candidate positions of the shielding images after the geometric transformation to the neighborhood of 3 pixels respectively from top to bottom, left to right.
The fifth concrete implementation mode: the third embodiment is further described, and the difference between the third embodiment and the fourth embodiment is that the affine correction parameters are obtained by performing feature matching on features between line segments by using GMS,
extracting characteristics among the line segments according to the line segments to obtain the characteristics;
the line segments are obtained by estimating vanishing points of the shielded images.
The sixth specific implementation mode: the fifth embodiment is further described, and the difference between the fifth embodiment and the fifth embodiment is that vanishing point estimation is performed by LSD line segment detection and J-Linkage set clustering.
The seventh embodiment: this embodiment mode is a further description of a fifth embodiment mode, and is different from the fifth embodiment mode in that
The method for obtaining the restored image according to the expanded image area in the seventh step and the third step comprises the following specific steps:
step seven, three and one: establishing a multi-scale feature map according to the expanded image area;
step seven, three, two: calculating a matching error between the source Patch and the target Patch in a candidate region of the feature map of each scale;
step seven, step three: searching a plurality of scale spaces, repeatedly executing the step seven, the step three and the step two, when the matching error value is lower than a target value, selecting a compensation position from the rivet pin collar image according to the matching error value at the moment, and then solving a fusion boundary by utilizing a graph-cut algorithm;
step seven, three and four: and obtaining a restored image after the fusion boundary is processed by standard Poisson fusion.
And establishing a multi-scale feature map.
Introduction to PatchMatch: the method is mainly used for searching the patch with the highest similarity in the nearest field in the image. The method is mainly based on a random sampling idea, and provides a mechanism for rapidly spreading in the whole image area to improve the efficiency of searching and matching according to the similarity of the image area.
Within the feature map candidate region of each scale, a matching error between the source Patch and the target Patch is calculated.
When the matched patch is spliced to the original image, the graph-cut algorithm is used for solving the optimal fusion boundary;
the final cost is composed of two parts: match cost + graph-cut cost, i.e. cost ═ match cost + graph-cut cost.
The cost minimum is selected as the result of the final repair.
Searching a plurality of scale spaces, and selecting the best completion position according to the matching error; and then solving the optimal fusion boundary by using a graph-cut algorithm.
The fusion boundaries were processed using standard poisson fusion.
The specific implementation mode is eight: this embodiment mode is a further description of a seventh embodiment mode, and is different from the seventh embodiment mode in that the target value is 0.2
The specific implementation method nine: the present embodiment is a further description of a seventh embodiment, and the difference between the present embodiment and the seventh embodiment is that a multi-scale feature map is created by mask information of an expanded image and an occlusion image.
The detailed implementation mode is ten: the present embodiment is further described with reference to the first embodiment, and the difference between the present embodiment and the first embodiment is that the specific step of determining whether the blind rivet pin collar is faulty or not according to the area of the white region in the image after the difference image binarization processing in the ninth step is: and if the area of the white area is larger than 100 pixel points, determining that the white area is a fault, otherwise, determining that the white area is a non-fault.
It should be noted that the detailed description is only for explaining and explaining the technical solution of the present invention, and the scope of protection of the claims is not limited thereby. It is intended that all such modifications and variations be included within the scope of the invention as defined in the following claims and the description.
Claims (10)
1. The method for detecting the loss of the pull rivet pin collar of the truck brake beam strut based on image restoration is characterized by comprising the following steps of:
the method comprises the following steps: acquiring a linear array image of the railway wagon;
step two: carrying out coarse positioning on a brake beam strut part aiming at the obtained linear array image of the railway wagon;
step three: and (3) carrying out binarization processing on the roughly positioned image, and positioning the position and the direction of the brake beam strut according to the area of a black area in the binarized image, wherein the specific steps are as follows:
step three, firstly: sequentially traversing the black areas, and selecting two areas with the largest areas as the edges of the brake beam strut;
step three: projecting the selected edge to obtain a convex image, wherein the position of the convex image is the position of the brake beam strut angle iron, and the opposite direction of the convex image is the direction of the rivet pin lantern ring;
step four: splicing the brake beam pillar image on one side as the upper part and the brake beam pillar image on the other side as the lower part;
step five: positioning the rivet pin lantern ring in the spliced image according to the direction of the rivet pin lantern ring and by combining the edge of the brake beam strut;
step six: generating a shielding image for shielding the rivet pin sleeve ring in the rivet pin sleeve ring image;
step seven: carrying out image restoration on the shielded rivet pin collar image;
step eight: calculating the absolute value of the image gray difference before and after image restoration to obtain a difference image, and performing binarization processing on the difference image;
step nine: and judging whether the blind rivet pin collar is in failure or not according to the area of the white area in the image after the binarization processing of the difference image.
2. The method for detecting the loss of the riveting pin sleeve of the truck brake beam strut based on the image restoration as claimed in claim 1, wherein in the second step, the rough positioning of the brake beam strut part is carried out by axle distance information and a priori position information of the brake beam strut.
3. The method for detecting the loss of the riveting pin sleeve of the truck brake beam strut based on image restoration as claimed in claim 2, wherein the specific steps of the seventh step are as follows:
step seven one: calculating affine correction parameters between two brake beam supports;
step seven and two: obtaining candidate positions of the shielding images according to the positions of the shielding masks in the shielding images and the affine correction parameters, and then carrying out geometric transformation according to the candidate positions of the shielding images;
step seven and three: and expanding the candidate position of the shielding image after the geometric conversion to the upper side, the lower side, the left side and the right side to obtain an expanded image area, wherein the expanded image area is larger than the shielding image, and finally obtaining a repairing image according to the expanded image area.
4. The method for detecting the loss of the riveting pin sleeve of the truck brake beam strut based on image restoration as claimed in claim 3, wherein the candidate positions of the shielding images after the geometric transformation in the seventh step and the third step are expanded to the following positions: and expanding the candidate positions of the shielding images after the geometric transformation to the neighborhood of 3 pixels respectively from top to bottom, left to right.
5. The method for detecting the loss of the riveting pin sleeve of the truck brake beam strut based on the image restoration as claimed in claim 3, wherein the affine correction parameters are obtained by performing feature matching on the features between the line segments by using GMS,
extracting characteristics among the line segments according to the line segments to obtain the characteristics;
and the line segment is obtained by estimating a vanishing point of the shielded image.
6. The method for detecting the loss of the riveting pin sleeve of the truck brake beam strut based on the image restoration as claimed in claim 5, wherein the vanishing point estimation is performed by LSD line segment detection and J-Linkage collective clustering.
7. The method for detecting the loss of the riveting pin sleeve of the truck brake beam strut based on image restoration as claimed in claim 5, wherein the concrete steps of obtaining the restoration image according to the expanded image area in the seventh step and the third step are as follows:
step seven, three and one: establishing a multi-scale feature map according to the expanded image area;
step seven, three, two: calculating a matching error between the source Patch and the target Patch in a candidate region of the feature map of each scale;
step seven, step three: searching a plurality of scale spaces, repeatedly executing the step seven, the step three and the step two, when the matching error value is lower than a target value, selecting a compensation position from the rivet pin collar image according to the matching error value at the moment, and then solving a fusion boundary by utilizing a graph-cut algorithm;
step seven, three and four: and obtaining a restored image after the fusion boundary is processed by standard Poisson fusion.
8. The image restoration-based detection method for detecting the loss of the riveting pin collar of the brake beam strut of the wagon as claimed in claim 7, wherein the target value is 0.2.
9. The method for detecting the loss of the riveting pin sleeve of the truck brake beam strut based on the image restoration as claimed in claim 7, wherein the multi-scale feature map is established by mask information of the expanded image and the shielding image.
10. The method for detecting the loss of the riveting pin collar of the truck brake beam strut based on image restoration as claimed in claim 1, wherein the specific steps of judging whether the riveting pin collar is in failure or not according to the area of the white area in the image after the binarization processing of the difference image in the ninth step are as follows: and if the area of the white area is larger than 100 pixel points, determining that the white area is a fault, otherwise, determining that the white area is a non-fault.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011217314.9A CN112329781B (en) | 2020-11-04 | 2020-11-04 | Method for detecting loss of pull rivet pin collar of truck brake beam strut based on image restoration |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011217314.9A CN112329781B (en) | 2020-11-04 | 2020-11-04 | Method for detecting loss of pull rivet pin collar of truck brake beam strut based on image restoration |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112329781A true CN112329781A (en) | 2021-02-05 |
| CN112329781B CN112329781B (en) | 2021-05-11 |
Family
ID=74323846
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011217314.9A Active CN112329781B (en) | 2020-11-04 | 2020-11-04 | Method for detecting loss of pull rivet pin collar of truck brake beam strut based on image restoration |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112329781B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2005185383A (en) * | 2003-12-25 | 2005-07-14 | Fuji Photo Film Co Ltd | Abnormal shade detection method, apparatus, and program |
| CN104268588A (en) * | 2014-06-19 | 2015-01-07 | 江苏大学 | Automatic detection method for brake shoe borer loss fault of railway wagon |
| US20180174330A1 (en) * | 2016-12-20 | 2018-06-21 | International Business Machines Corporation | Mobile user interface design testing tool |
| CN109785286A (en) * | 2018-12-12 | 2019-05-21 | 中国科学院深圳先进技术研究院 | A kind of image repair detection method based on Texture Feature Fusion |
| CN110717477A (en) * | 2019-10-22 | 2020-01-21 | 深圳市微度数字技术有限公司 | Intelligent early warning and analyzing device for track |
-
2020
- 2020-11-04 CN CN202011217314.9A patent/CN112329781B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2005185383A (en) * | 2003-12-25 | 2005-07-14 | Fuji Photo Film Co Ltd | Abnormal shade detection method, apparatus, and program |
| CN104268588A (en) * | 2014-06-19 | 2015-01-07 | 江苏大学 | Automatic detection method for brake shoe borer loss fault of railway wagon |
| US20180174330A1 (en) * | 2016-12-20 | 2018-06-21 | International Business Machines Corporation | Mobile user interface design testing tool |
| CN109785286A (en) * | 2018-12-12 | 2019-05-21 | 中国科学院深圳先进技术研究院 | A kind of image repair detection method based on Texture Feature Fusion |
| CN110717477A (en) * | 2019-10-22 | 2020-01-21 | 深圳市微度数字技术有限公司 | Intelligent early warning and analyzing device for track |
Non-Patent Citations (2)
| Title |
|---|
| ORHAN YAMAN ET AL.: "Image processing based fault detection approach for rail surface", 《2015 23ND SIGNAL PROCESSING AND COMMUNICATIONS APPLICATIONS CONFERENCE (SIU)》 * |
| 李卓: "动车组故障检测机器人***的设计与实现", 《铁路计算机应用》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112329781B (en) | 2021-05-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102663760B (en) | Location and segmentation method for windshield area of vehicle in images | |
| CN111079817B (en) | Method for identifying fault image of cross beam of railway wagon | |
| CN105741291A (en) | Method for detecting faults of equipotential lines of high-speed railway overhead line system suspension devices | |
| CN107273802A (en) | A kind of detection method and device of railroad train brake shoe drill ring failure | |
| CN111667462A (en) | Automatic detection method and system for small and weak targets in welding defects | |
| CN107967681B (en) | Elevator compensation chain impact defect detection method based on machine vision | |
| CN113221839B (en) | Automatic truck image identification method and system | |
| CN112801972A (en) | Bridge defect detection method, device, system and storage medium | |
| CN112101182B (en) | Railway wagon floor damage fault identification method based on improved SLIC method | |
| CN111080600A (en) | Fault identification method for split pin on spring supporting plate of railway wagon | |
| CN111080607B (en) | Rolling bearing oil slinging fault detection method based on image recognition | |
| CN116188459B (en) | Line laser rapid identification method and system for belt tearing detection | |
| CN111080613A (en) | Image recognition method for damage fault of wagon bathtub | |
| CN115527170A (en) | Method and system for identifying closing fault of door stopper handle of automatic freight car derailing brake device | |
| CN111832562B (en) | Spring supporting plate fault identification method based on image processing | |
| CN112329781B (en) | Method for detecting loss of pull rivet pin collar of truck brake beam strut based on image restoration | |
| CN111652228B (en) | Railway wagon sleeper beam hole foreign matter detection method | |
| CN111091557B (en) | Method and system for detecting breakage fault of flange of bearing saddle of railway wagon | |
| CN112132824B (en) | Automatic detection method for failure of freight car axle box spring | |
| CN112102293B (en) | Rapid detection method for foreign matters in triangular holes of railway wagon | |
| CN112365481B (en) | Method for detecting bolt loss of cross beam assembly based on image processing | |
| CN112150453B (en) | Automatic detection method for breakage fault of bolster spring of railway wagon | |
| CN115439737B (en) | Railway box car window fault image recognition method based on image restoration | |
| CN116580245B (en) | Rail wagon bearing saddle dislocation fault identification method | |
| CN118015000B (en) | Surface defect detection method for guide rail based on image processing |
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 |