CN117764904A - Cloud detection system and method for bridge inhaul cable detection robot - Google Patents
Cloud detection system and method for bridge inhaul cable detection robot Download PDFInfo
- Publication number
- CN117764904A CN117764904A CN202310837567.3A CN202310837567A CN117764904A CN 117764904 A CN117764904 A CN 117764904A CN 202310837567 A CN202310837567 A CN 202310837567A CN 117764904 A CN117764904 A CN 117764904A
- Authority
- CN
- China
- Prior art keywords
- detection
- cable
- defect
- training
- data set
- 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 117
- 238000000034 method Methods 0.000 title claims abstract description 14
- 230000007547 defect Effects 0.000 claims abstract description 57
- 238000012549 training Methods 0.000 claims abstract description 48
- 238000013136 deep learning model Methods 0.000 claims abstract description 21
- 230000004044 response Effects 0.000 claims abstract description 4
- 238000012545 processing Methods 0.000 claims description 13
- 238000004364 calculation method Methods 0.000 claims description 5
- 230000003321 amplification Effects 0.000 claims description 3
- 230000006835 compression Effects 0.000 claims description 3
- 238000007906 compression Methods 0.000 claims description 3
- 238000001914 filtration Methods 0.000 claims description 3
- 238000003199 nucleic acid amplification method Methods 0.000 claims description 3
- 238000007781 pre-processing Methods 0.000 claims description 3
- 230000011218 segmentation Effects 0.000 claims description 3
- 238000012360 testing method Methods 0.000 claims description 3
- 238000013519 translation Methods 0.000 claims description 3
- 230000006870 function Effects 0.000 claims 1
- 229910000831 Steel Inorganic materials 0.000 description 3
- 238000011897 real-time detection Methods 0.000 description 3
- 239000010959 steel Substances 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 230000002378 acidificating effect Effects 0.000 description 1
- 230000002411 adverse Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
Landscapes
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
Abstract
The invention provides a cloud detection system for a bridge cable detection robot, which is based on a flash frame and comprises a cable defect detection module and a deep learning model online training module, wherein the cable defect detection module comprises two sub-functional modules of cable surface image detection and real-time video detection, the deep learning model online training module comprises four sub-functional modules of data uploading, data set expansion, data set training and training weight downloading, the cloud detection system adopts a working mode of adopting browser requests and server responses, and the cloud detection method for the bridge cable detection robot can detect and process a cable surface image acquired by the detection robot.
Description
Technical Field
The invention belongs to the field of bridge detection robots, and particularly relates to a cloud detection system and method for a bridge cable detection robot.
Background
The cable-stayed bridge mainly comprises a bridge tower, a bridge deck and a stay cable, wherein one end of the stay cable is anchored at the bridge tower, and the other end of the stay cable is anchored at the bridge deck, and the stay cable is used as a main bearing component of the cable-stayed bridge to play a great role in the safety of the whole bridge. The inside of the stay cable is a multi-strand steel wire bundle, the outside of the stay cable is a PE sheath, and the upper end and the lower end of the stay cable are anchors. The stay cable can be damaged in different types and degrees due to the fact that the stay cable bears alternating stress caused by variable load on the bridge deck for a long time and is directly exposed to the air and is adversely affected by the environment. When the PE sheath is seriously damaged and the inner steel wire bundle is directly exposed to the air, the inner steel wire bundle can be corroded or even broken due to moisture and acidic substances in the air, so that the bridge safety is seriously critical.
The cable detection robot can effectively replace a worker to climb at the position of a high-altitude cable, the current detection mode mainly utilizes a local computer to cooperate with the robot to carry out detection and calculation processing, equipment requirements and preparation work are more, the detection operation flow and the hardware use cost are continuously increased, and cloud real-time detection cannot be carried out. The intelligent cloud real-time detection operation of the inhaul cable robot is still an important point and a difficult point in the current industrial application field.
Disclosure of Invention
The invention discloses a cloud detection system and a cloud detection method for a bridge cable detection robot, which solve the problem that the cable detection robot can only utilize a local computer to carry out detection and calculation processing in the field detection stage, but cannot carry out cloud detection, reduce the detection operation flow and the hardware use cost, simultaneously carry out data set update and deep learning model training conveniently, and promote the rapidity of the subsequent robot operation flow.
The cloud detection system comprises a cable defect detection module and a deep learning model online training module, wherein the cable defect detection module comprises two sub-functional modules of cable surface image detection and real-time video detection, the deep learning model online training module comprises four sub-functional modules of data uploading, data set expansion, data set training and training weight downloading, the cloud detection system adopts a working mode of adopting browser requests and server responses, and the cloud detection method for the bridge cable detection robot can detect and process cable surface images acquired by the detection robot.
A cloud detection method for a bridge cable detection robot comprises the following steps:
step (1): inputting a cloud detection system website through a browser, and inputting an account name and a password to log in;
step (2): after successful login, module selection is carried out, if a inhaul cable defect detection module is selected, the step (3) is carried out, if a deep learning model online training module is selected, the step (4) is carried out;
step (3): selecting a inhaul cable defect detection mode, detecting inhaul cable surface images or detecting real-time videos,
step (4): the deep learning model is trained on line,
step (5): and (3) after the cloud detection of the bridge cable detection robot is finished, returning to the step (2).
Further, the step 3 specifically includes the following steps:
step (3.1), cable defect detection starts to run, if cable surface image detection is selected, the step (3.3) is performed, and if real-time video detection is selected, the step (3.3) is performed;
sequentially performing image reading, image gray level processing, filtering processing, threshold segmentation and connected region calculation, outputting an identification result image, and performing defect discrimination in the step (3.4);
step (3.3), sequentially reading video stream, preprocessing video frames, loading a YOLOV3 model, reading weight, testing, continuously outputting single frame identification result images, and counting in step (3.4) to judge defects;
judging the generated inhaul cable defect identification result image, if the inhaul cable defect identification result image has a defect, entering the step (3.5), and if the inhaul cable defect identification result image has no defect, directly entering the step (3.6);
step (3.5), storing the inhaul cable surface defect image, and displaying the defect type and position;
and (3.6) finishing the cable defect detection, and entering the step (5).
Further, the step 4 specifically includes the following steps:
step (4.1) uploading the training data set in the form of compressed packets;
step (4.2), detecting whether uploading is successful, if so, entering step (4.3), otherwise, returning to step (4.9);
decompressing the data set compression package, and sequentially renaming names of the data set pictures and the tag files according to naming rules, wherein the specific naming rules are as follows: reading the existing training set data quantity n in the server, and renaming the data set pictures and the tag files according to numbers in sequence from n+1;
sequentially storing the picture and the tag file to the corresponding positions;
performing image translation, image rotation, mirroring and noise adding four data enhancement processes on the data set to finish training data set amplification;
step (4.6), inputting learning rate, batch_size and iteration times to set training parameters of the YOLOv3 model;
training and learning the inhaul cable image training data set based on a YOLOv3 model;
training is completed, and a weight file is saved;
and (4.9) finishing the online training of the deep learning model, and entering the step (5).
Further, the weight file read in the step (3.3) may automatically update the latest weight file generated for the step (4.8).
The invention has the beneficial effects that:
(1) The cloud detection system and the cloud detection method are used for detecting the surface defects of the inhaul cable and training the deep learning model, and compared with a conventional local computer operation method, the cloud detection system and the cloud detection method improve the convenience of operation and are not influenced by computer hardware.
(2) By using the cloud detection system and the method, the real-time detection of the defect by the inhaul cable detection robot is realized, the manual identification is effectively replaced, and the inhaul cable detection working intensity is reduced.
(3) By using the cloud detection system and the cloud detection method to judge the damage degree of the surface of the bridge inhaul cable, the remote statistic detection data can be directly obtained, and a basis is provided for further inhaul cable maintenance.
Drawings
Fig. 1 is a schematic diagram of a cloud detection system;
FIG. 2, a cloud cable defect detection flow chart;
FIG. 3 is a flow chart of online training of a deep learning model;
fig. 4 is a diagram of a surface image detection display interface of a dragline of the cloud detection system;
FIG. 5, a cloud detection system deep learning model online training parameter setting interface;
fig. 6 shows an original image and a defect identification result image of real-time video detection of the cloud detection system.
Detailed Description
The present invention is further illustrated in the following drawings and detailed description, which are to be understood as being merely illustrative of the invention and not limiting the scope of the invention. It should be noted that the words "front", "rear", "left", "right", "upper" and "lower" used in the following description refer to directions in the drawings, and the words "inner" and "outer" refer to directions toward or away from, respectively, the geometric center of a particular component.
As shown in fig. 1, the cloud detection system for a bridge cable detection robot of the present embodiment is based on a flash framework, and includes a cable defect detection module and a deep learning model online training module, where the cable defect detection module includes two sub-functional modules of cable surface image detection and real-time video detection, and the deep learning model online training module includes four sub-functional modules of data uploading, data set expansion, data set training and training weight downloading, and the cloud detection system adopts a working mode of adopting a browser request and a server response, and the cloud detection method for the bridge cable detection robot can perform cloud detection and processing on cable surface images collected by the detection robot.
As shown in fig. 2-5, a cloud detection method for a bridge cable detection robot includes the following steps:
step (1): inputting a cloud detection system website through a browser, and inputting an account name and a password to log in;
step (2): after successful login, module selection is performed, and if a cable defect detection module is selected, the step (3) is entered: if the deep learning model online training module is selected, the step (4) is entered;
step (3): selecting a inhaul cable defect detection mode, detecting inhaul cable surface images or detecting real-time videos, wherein the specific processing process comprises the following steps of:
step (3.1): the cable defect detection starts to run, if cable surface image detection is selected, the step (3.3) is carried out, and if real-time video detection is selected, the step (3.3) is carried out;
step (3.2): sequentially performing image reading, image gray processing, filtering processing, threshold segmentation and connected region calculation, outputting an identification result image, and performing defect discrimination in the step (3.4);
step (3.3): sequentially reading video stream, preprocessing video frames, loading a YOLOV3 model, reading weight, testing, continuously outputting single-frame identification result images, and counting in the step (3.4) to judge defects;
step (3.4): judging the generated inhaul cable defect identification result image, if the inhaul cable defect identification result image has a defect, entering the step (3.5), and if the inhaul cable defect identification result image has no defect, directly entering the step (3.6);
step (3.5): storing the surface defect image of the inhaul cable, and displaying the defect type and position;
step (3.6): the cable defect detection is finished, and the step (5) is carried out;
step (4): the deep learning model is trained on line, and the specific processing process comprises the following steps:
step (4.1): uploading the training data set in the form of compressed packets;
step (4.2): detecting whether uploading is successful, if so, entering a step (4.3), otherwise, returning to the step (4.9);
step (4.3): decompressing a data set compression package, and sequentially renaming names of a data set picture and a tag file according to a naming rule, wherein the specific naming rule is as follows: reading the existing training set data quantity n in the server, and renaming the data set pictures and the tag files according to numbers in sequence from n+1;
step (4.4): sequentially storing the pictures and the tag files to corresponding positions;
step (4.5): performing image translation, image rotation, mirroring and noise adding four data enhancement processes on the data set to finish training data set amplification;
step (4.6): inputting learning rate, batch_size and iteration times to set training parameters of the YOLOv3 model;
step (4.7): training and learning based on a YOLOv3 model are carried out on the inhaul cable image training data set;
step (4.8): after training is completed, saving a weight file;
step (4.9), finishing the online training of the deep learning model, and entering step (5); step (5): and (3) after the cloud detection of the bridge cable detection robot is finished, returning to the step (2).
The weight file read in the step (3.3) in this embodiment may automatically update the latest weight file generated for the step (4.8).
As shown in fig. 6; the cloud detection system detects an original image and a defect identification result image in real time; the real-time processing from the original picture to the defect identification can be realized through the cloud detection system, the position, the category and the probability of the defect can be clearly seen from the defect identification result image, the defect detection result statistical processing and analysis are convenient, and the defect state is convenient to observe by workers in real time.
The technical means disclosed by the scheme of the invention is not limited to the technical means disclosed by the embodiment, and also comprises the technical scheme formed by any combination of the technical features.
Claims (5)
1. A high in clouds detecting system for bridge cable detection robot, its characterized in that: the system is based on a flash frame and comprises a cable defect detection module and a deep learning model online training module, wherein the cable defect detection module comprises two sub-functional modules of cable surface image detection and real-time video detection, the deep learning model online training module comprises four sub-functional modules of data uploading, data set expansion, data set training and training weight downloading, and the cloud detection system adopts a working mode of browser request and server response, so that the cloud detection method for the bridge cable detection robot can detect and process cable surface images acquired by the detection robot.
2. The cloud detection method for the bridge cable detection robot is characterized by comprising the following steps of:
step (1): inputting a cloud detection system website through a browser, and inputting an account name and a password to log in;
step (2): after successful login, module function selection is carried out, if a inhaul cable defect detection module is selected, the step (3) is carried out, and if an online training module of a deep learning model is selected, the step (4) is carried out;
step (3): selecting a inhaul cable defect detection mode, and detecting inhaul cable surface images or real-time video;
step (4): performing online training of the deep learning model;
step (5): and (3) after the cloud detection of the bridge cable detection robot is finished, returning to the step (2).
3. The cloud detection method for the bridge cable detection robot according to claim 2, wherein the step 3 specifically includes the following steps:
step (3.1): the cable defect detection starts to run, if cable surface image detection is selected, the step (3.3) is carried out, and if real-time video detection is selected, the step (3.3) is carried out;
step (3.2): sequentially performing image reading, image gray processing, filtering processing, threshold segmentation and connected region calculation, outputting an identification result image, and performing defect discrimination in the step (3.4);
step (3.3): sequentially performing video stream reading, video frame preprocessing, YOLOV3 model loading, weight reading, recognition testing, continuously outputting single-frame recognition result images, and counting in step (3.4) to perform defect judgment;
step (3.4): judging the generated inhaul cable defect identification result image, if the inhaul cable defect identification result image has a defect, entering the step (3.5), and if the inhaul cable defect identification result image has no defect, directly entering the step (3.6);
step (3.5): storing the surface defect image of the inhaul cable, and displaying the defect type and position; step (3.6): and (5) outputting the identification result, finishing the cable defect detection, and entering the step (5).
4. The cloud detection method for the bridge cable detection robot according to claim 2, wherein the step 4 specifically includes the following steps:
step (4.1): uploading the training data set in the form of compressed packets;
step (4.2): detecting whether uploading is successful, if so, entering a step (4.3), otherwise, returning to the step (4.9);
step (4.3): decompressing a data set compression package, and sequentially renaming names of a data set picture and a tag file according to a naming rule, wherein the specific naming rule is as follows: reading the existing training set data quantity n in the server, and renaming the data set pictures and the tag files according to numbers in sequence from n+1;
step (4.4): sequentially storing the pictures and the tag files to corresponding positions;
step (4.5): performing image translation, image rotation, mirroring and noise adding four data enhancement processes on the data set to finish training data set amplification;
step (4.6): inputting learning rate, batch_size and iteration times to set training parameters of the YOLOv3 model;
step (4.7): training and learning based on a YOLOv3 model are carried out on the inhaul cable image training data set;
step (4.8): after training is completed, saving a weight file;
step (4.9): and (5) finishing online training of the deep learning model, and entering a step (5).
5. The cloud detection system and method for a bridge cable detection robot according to claim 3, wherein the weight file read in the step (3.3) can automatically update the latest weight file generated for the step (4.8).
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310837567.3A CN117764904A (en) | 2023-07-10 | 2023-07-10 | Cloud detection system and method for bridge inhaul cable detection robot |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310837567.3A CN117764904A (en) | 2023-07-10 | 2023-07-10 | Cloud detection system and method for bridge inhaul cable detection robot |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117764904A true CN117764904A (en) | 2024-03-26 |
Family
ID=90320661
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310837567.3A Pending CN117764904A (en) | 2023-07-10 | 2023-07-10 | Cloud detection system and method for bridge inhaul cable detection robot |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117764904A (en) |
-
2023
- 2023-07-10 CN CN202310837567.3A patent/CN117764904A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20220036533A1 (en) | Image defect detection method and apparatus, electronic device, storage medium and product | |
| CN108416294B (en) | Fan blade fault intelligent identification method based on deep learning | |
| CN111797890A (en) | Method and system for detecting defects of power transmission line equipment | |
| CN113283344A (en) | Mining conveying belt deviation detection method based on semantic segmentation network | |
| CN111489352A (en) | Tunnel gap detection and measurement method and device based on digital image processing | |
| CN112801972A (en) | Bridge defect detection method, device, system and storage medium | |
| CN111833328A (en) | Aircraft engine blade surface defect detection method based on deep learning | |
| CN110493574B (en) | Security monitoring visualization system based on streaming media and AI technology | |
| CN116485779B (en) | Adaptive wafer defect detection method and device, electronic equipment and storage medium | |
| CN113515655A (en) | Fault identification method and device based on image classification | |
| CN111127448A (en) | Model and method for detecting air spring fault based on isolated forest | |
| CN112101138A (en) | Bridge inhaul cable surface defect real-time identification system and method based on deep learning | |
| CN111899219A (en) | Image identification method and system for power transmission line machine patrol | |
| CN109521021A (en) | A kind of nuclear power plant containment shell appearance inspecting system and method | |
| CN115995056A (en) | Automatic bridge disease identification method based on deep learning | |
| CN115937101A (en) | Quality detection method, device, equipment and storage medium | |
| CN115753809A (en) | Insulator contamination detection method, device, equipment and storage medium | |
| CN115170988A (en) | Power grid line fault identification method and system based on deep learning | |
| CN112749735B (en) | Converter tapping steel flow identification method, system, medium and terminal based on deep learning | |
| CN110618129A (en) | Automatic power grid wire clamp detection and defect identification method and device | |
| CN114120086A (en) | Pavement disease recognition method, image processing model training method, device and electronic equipment | |
| CN117764904A (en) | Cloud detection system and method for bridge inhaul cable detection robot | |
| CN112801072B (en) | Elevator non-flat-layer door opening fault recognition device and method based on computer vision | |
| CN113989742A (en) | Nuclear power station plant pedestrian detection method based on multi-scale feature fusion | |
| CN112232307A (en) | Method for detecting wearing of safety helmet in night vision environment |
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 |