CN111160270B - Bridge monitoring method based on intelligent video recognition - Google Patents
Bridge monitoring method based on intelligent video recognition Download PDFInfo
- Publication number
- CN111160270B CN111160270B CN201911402701.7A CN201911402701A CN111160270B CN 111160270 B CN111160270 B CN 111160270B CN 201911402701 A CN201911402701 A CN 201911402701A CN 111160270 B CN111160270 B CN 111160270B
- Authority
- CN
- China
- Prior art keywords
- vehicle
- image
- preset
- information
- bridge monitoring
- 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.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/08—Detecting or categorising vehicles
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Abstract
A bridge monitoring method based on intelligent video recognition relates to the field of bridge monitoring, and comprises the following steps: and intercepting a vehicle passing picture in the vehicle passing video, carrying out image recognition on the vehicle passing picture according to the target detection model so as to extract a real-time image and related information of a preset type of vehicle, carrying out image recognition on the real-time image according to the weight classification model so as to obtain weight classification information of the preset type of vehicle, and when the weight classification information is larger than a preset threshold value, sending alarm information containing the weight classification information to a preset bridge monitoring system and carrying out risk analysis by combining monitoring data. The invention has the beneficial effects that: the vehicle dynamic weighing system is not required to be additionally installed, and the overweight vehicle can be identified, recorded and analyzed based on video monitoring by utilizing the preset weighing system.
Description
Technical Field
The invention relates to the technical field of bridge monitoring, in particular to a bridge monitoring method based on intelligent video identification.
Background
The bridge is a key component of transportation, and a large number of bridges are raised along with the gradual promotion of the traffic country. However, some bridge engineering projects are damaged far from the design service life, and overload transportation is a main reason for the damage, and bridge ruts, network cracks, sudden increases of a dangerous bridge and the like are rarely generated by fatigue strength of vehicle loads, and most of the damage is caused by overload. Therefore, it is important to identify the overweight vehicle of the bridge and analyze the risk thereof for long-term healthy and stable operation of the bridge.
In the prior art, a vehicle dynamic weighing system (WIM) is generally utilized to obtain information such as passing time, passing image, license plate number, axle weight, total weight, vehicle speed, axle distance and the like of a running vehicle, a sensor is utilized to obtain structural response of a bridge under the action of external load, and the two data are combined to monitor an overweight vehicle passing through the bridge, however, the two data are relatively isolated and cannot exert maximum utilization value. Moreover, for bridges without a vehicle dynamic weighing system, overweight vehicles cannot be monitored in real time only through video monitoring of installation.
Disclosure of Invention
Aiming at the defects in the prior art, the invention provides a bridge monitoring method based on intelligent video recognition, which utilizes the existing system to realize real-time recognition, recording and analysis of overweight vehicles and improves the long-term operation safety of bridges.
In order to achieve the above purpose, the technical scheme adopted is as follows:
a bridge monitoring method based on intelligent video recognition, the bridge monitoring method comprising:
acquiring a plurality of first acquired information and a plurality of second acquired information by using a preset weighing system, wherein the first information acquired information comprises a first image and type labeling information related to a plurality of vehicles of preset types contained in the first image, the second acquired information comprises a second image and weight information related to a unique vehicle of preset types contained in the second image, a target detection model is obtained according to the processing of the plurality of first acquired information, and a weight classification model is obtained according to the processing of the plurality of second acquired information;
intercepting a vehicle passing image in a vehicle passing video, and carrying out image recognition on the vehicle passing image according to the target detection model so as to extract a real-time image of a preset type of vehicle from the vehicle passing image;
performing image recognition on the real-time image according to the weight classification model to obtain weight classification information of a preset type of vehicle;
judging whether the weight classification information is larger than a preset threshold value or not:
if yes, sending alarm information containing the weight classification information to a preset bridge monitoring system, and carrying out risk analysis by the preset bridge monitoring system according to the alarm information and monitoring data stored by the preset bridge monitoring system and related to the alarm information to obtain a bridge passing risk coefficient of a corresponding preset type of vehicle.
Preferably, the predetermined type of vehicle includes a predefined large vehicle.
Preferably, the first image includes a plurality of vehicles, and at least part of the plurality of vehicles is a preset type of vehicle;
the second image comprises a vehicle which is a preset type of vehicle.
Preferably, a plurality of first acquired information form a first data set, and the target detection model is obtained through training according to the first data set;
the first dataset comprises a total number of the first acquired information greater than 10000.
Preferably, the specific steps of the preset weighing system in acquiring the first acquired information are as follows:
acquiring the first image;
manually marking the regional position coordinates of a preset type of vehicle in all vehicles in the first image through rectangular frames to obtain a plurality of real marking frames and coordinate positions, wherein all the real marking frames and all the coordinate positions of the first image form the type marking information;
the first image and the type standard information constitute the first acquisition information.
Preferably, the coordinate positions are the position coordinates of the top left corner vertex and the position coordinates of the bottom right corner vertex of the real labeling frame.
Preferably, a plurality of second acquired information form a second data set, and the weight classification model is obtained through training according to the second data set;
the second data set includes a total number of the second acquired information greater than 10000.
Preferably, the real-time gravimetric classification model outputs results of (0, 10 t), (10 t,20 t), (20 t,30 t), (30 t,40 t), (40 t,50 t), (50 t,60 t) or (60 t, -).
Preferably, the target detection model adopts an algorithm of SSD, yolo, R-CNN, fast R-CNN or Fast R-CNN;
the weight classification model adopts an algorithm of VGG, resNet, mobileNet or acceptance.
Preferably, the monitoring data includes displacement data, deflection data and strain data associated with the preset type of vehicle.
The invention has the beneficial effects that: the method comprises the steps of acquiring data by using an existing preset weighing system, processing the data to obtain a target detection model and a weight classification model, identifying a vehicle passing image acquired in real time by video monitoring through the target detection model and the weight classification model to obtain weight classification information of a preset type of vehicle in the vehicle passing image, carrying out risk analysis by using the existing preset bridge monitoring system and combining the weight classification information exceeding a preset threshold value with monitoring data to obtain a bridge passing risk coefficient of the corresponding preset type of vehicle, wherein the preset type of vehicle can comprise an overweight vehicle, and accordingly identifying, recording and analyzing all road passing vehicles including the overweight vehicle.
Drawings
Fig. 1 is a flowchart of a bridge monitoring method based on intelligent video recognition in an embodiment of the invention.
Detailed Description
The present invention will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the specific examples described herein are intended to illustrate the invention and are not intended to limit the invention. Moreover, all other embodiments which can be made by a person of ordinary skill in the art based on the embodiments of the present invention without making any inventive effort are within the scope of the present invention.
The present invention will be described in further detail with reference to the accompanying drawings and examples.
As shown in fig. 1, a bridge monitoring method based on intelligent video recognition includes:
step S1, acquiring a plurality of first acquired information and a plurality of second acquired information by using a preset weighing system. The first information collection information comprises a first image and type labeling information associated with a plurality of vehicles of a preset type contained in the first image, and the second information collection information comprises a second image and weight information associated with a unique vehicle of a preset type contained in the second image. The first image comprises a plurality of vehicles, at least part of the vehicles are vehicles of preset types, and the type standard information corresponding to each first image comprises relevant information of all the vehicles of the preset types in the first image. The second acquired information comprises a second image and weight information associated with the only preset type of vehicle contained in the second image, namely the second image only comprises one preset type of vehicle, and the weight information is the weight information of the only preset type of vehicle in the second image. The preset type of vehicle is a predefined large vehicle (open type). In a first dataset made up of first collected information, the total number of first collected information is greater than 10000. And in the second data set formed by a plurality of second acquired information, the total number of the second acquired information is more than 10000 second acquired information.
And S2, processing according to the first acquired information to obtain a target detection model, and processing according to the second acquired information to obtain a weight classification model. The target detection model adopts an algorithm of SSD, yolo, R-CNN, fast R-CNN or Fast R-CNN. The SSD multi-target detection algorithm based on deep learning is preferred, and the SSD is high in speed and accuracy. The target detection model can also select other algorithms according to actual requirements. The weight classification model uses an algorithm of VGG, resNet, mobileNet or acceptance, preferably acceptance v3. The weight classification model can also select other algorithms according to actual requirements.
And S3, intercepting a vehicle passing image in the vehicle passing video, and carrying out image recognition on the vehicle passing image according to the target detection model.
S4, judging whether a large vehicle exists in the intercepted vehicle passing image, and if so, turning to S5; if not, go to step S6.
Step S5, extracting a real-time image of the preset type of vehicle from the vehicle passing image, and then turning to step S7.
And S6, discarding the intercepted vehicle passing image, and turning to step S3.
And S7, carrying out image recognition on the real-time image according to the weight classification model to obtain weight classification information of the vehicle of the preset type.
Step S8, judging whether a weight value corresponding to the weight classification information is larger than a preset threshold value, and if so, turning to step S9; if not, go to step S10. The preset threshold value limits the traffic weight of the bridge through which the vehicle passes at the time.
And S9, storing and recording the weight classification information, sending alarm information to a preset bridge monitoring system, wherein the alarm information comprises the weight classification information, and then turning to step S11. The weight classification model for vehicle weight prediction adopts a classification algorithm, vehicle data is a sequence, the vehicle data is artificially classified into 7 categories of (0, 10 t), (10 t,20 t), (20 t,30 t), (30 t,40 t), (40 t,50 t), (50 t,60 t) and (60 t), a large vehicle weight prediction model is trained by using pictures and weight category information, and the weight prediction result is one of 7 categories.
And step S10, the weight classification information is not stored or recorded, and the process is finished.
And S11, after receiving the alarm information, the preset bridge monitoring system performs risk analysis by combining the monitoring data of the preset type of vehicles corresponding to the alarm information stored by the preset bridge monitoring system to obtain the bridge traffic risk coefficient of the preset type of vehicles. The monitoring data includes displacement data, deflection data, and strain data associated with a preset type of vehicle. The monitored data is compared with a set threshold value to determine the risk of passing vehicles.
In the embodiment, the identification, recording and analysis of the overweight vehicle can be realized based on the existing preset weighing system without additionally installing a vehicle dynamic weighing system, the identification of the overweight vehicle is realized by utilizing a computer vision technology, meanwhile, the vehicle dynamic weighing system is linked with bridge structure monitoring data of a preset bridge monitoring system, the risk is judged in real time, and the long-term healthy and stable operation of the bridge is ensured.
When a data set is collected and a target detection model and a weight classification model are obtained according to the data set training, a rectangular frame is used for manually marking the position coordinates of a large vehicle (open type) area in a vehicle image, a real marking frame of the position of the vehicle area and 4 coordinate values of the real marking frame in each image are obtained, the 4 coordinate values are the position coordinates of the top left corner vertex and the position coordinates of the bottom right corner vertex of the vehicle area respectively, the information of the real marking frame of each image is stored as an xml file (namely type marking information), all the large vehicle images and the corresponding xml marking result file form a large vehicle marking data set (namely a first data set), and the first data set is used for training to obtain a target detector for identifying the large vehicle. And training by using the second data set to obtain a weight classification model for identifying weight classifications of large vehicles.
The invention is not limited to the embodiments, and it will be apparent to those skilled in the art that modifications and variations can be made without departing from the principle of the invention, and these modifications and variations are also considered to be within the scope of the invention. What is not described in detail in this specification is prior art known to those skilled in the art.
Claims (6)
1. A bridge monitoring method based on intelligent video identification is characterized in that: the bridge monitoring method comprises the following steps:
acquiring a plurality of first acquired information and a plurality of second acquired information by using a preset weighing system, wherein the first acquired information comprises a first image and type marking information related to a plurality of vehicles of preset types contained in the first image, the second acquired information comprises a second image and weight information related to a unique vehicle of preset types contained in the second image, a target detection model is obtained according to the plurality of first acquired information, and a weight classification model is obtained according to the plurality of second acquired information; the first image comprises a plurality of vehicles, and at least part of the vehicles are vehicles of a preset type; the second image comprises a vehicle which is a preset type of vehicle; the preset type of vehicle comprises a predefined large vehicle;
intercepting a vehicle passing image in a vehicle passing video, carrying out image recognition on the vehicle passing image according to the target detection model, judging whether a large vehicle exists in the intercepted vehicle passing image, and if so, extracting a real-time image of a preset type of vehicle from the vehicle passing image;
performing image recognition on the real-time image according to the weight classification model to obtain weight classification information of a preset type of vehicle;
judging whether the weight value corresponding to the weight classification information is larger than a preset threshold value or not:
if yes, sending alarm information containing the weight classification information to a preset bridge monitoring system, and carrying out risk analysis by the preset bridge monitoring system according to the alarm information and monitoring data stored by the preset bridge monitoring system and related to the alarm information to obtain a bridge traffic risk coefficient of a corresponding preset type of vehicle;
the monitoring data includes displacement data, deflection data, and strain data associated with the preset type of vehicle;
the weight classification model outputs results of [0,10t ], [10t,20t ], [20t,30t ], [30t,40t ], [40t,50t ], [50t,60 t), or [60t, + ].
2. The bridge monitoring method of claim 1, wherein: a plurality of first acquired information form a first data set, and the target detection model is obtained through training according to the first data set;
the first dataset comprises a total number of the first acquired information greater than 10000.
3. The bridge monitoring method of claim 1, wherein: the specific steps of the preset weighing system in acquiring the first acquired information are as follows:
acquiring the first image;
manually marking the regional position coordinates of a preset type of vehicle in all vehicles in the first image through rectangular frames to obtain a plurality of real marking frames and coordinate positions, wherein all the real marking frames and all the coordinate positions of the first image form the type marking information;
the first image and the type annotation information form the first acquisition information.
4. A bridge monitoring method according to claim 3, wherein: and the coordinate positions are the position coordinates of the top left corner vertex and the position coordinates of the bottom right corner vertex of the real annotation frame.
5. The bridge monitoring method of claim 1, wherein: the plurality of second acquired information form a second data set, and the weight classification model is obtained through training according to the second data set;
the second data set includes a total number of the second acquired information greater than 10000.
6. The bridge monitoring method of claim 1, wherein: the target detection model adopts an algorithm of SSD, yolo, R-CNN, fast R-CNN or Fast R-CNN;
the weight classification model adopts an algorithm of VGG, resNet, mobileNet or acceptance.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911402701.7A CN111160270B (en) | 2019-12-31 | 2019-12-31 | Bridge monitoring method based on intelligent video recognition |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911402701.7A CN111160270B (en) | 2019-12-31 | 2019-12-31 | Bridge monitoring method based on intelligent video recognition |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111160270A CN111160270A (en) | 2020-05-15 |
| CN111160270B true CN111160270B (en) | 2023-09-01 |
Family
ID=70559475
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911402701.7A Active CN111160270B (en) | 2019-12-31 | 2019-12-31 | Bridge monitoring method based on intelligent video recognition |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111160270B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111860201B (en) * | 2020-06-28 | 2023-07-25 | 中铁大桥科学研究院有限公司 | Ramp heavy vehicle identification method and system combining image identification and bridge monitoring |
| CN113033284B (en) * | 2020-12-22 | 2022-10-25 | 迪比（重庆）智能科技研究院有限公司 | Vehicle real-time overload detection method based on convolutional neural network |
| CN112906647B (en) * | 2021-03-24 | 2023-12-19 | 杭州鲁尔物联科技有限公司 | Method and device for monitoring load of small-span bridge, computer equipment and storage medium |
| CN114396877B (en) * | 2021-11-19 | 2023-09-26 | 重庆邮电大学 | Intelligent three-dimensional displacement field and strain field measurement method for mechanical properties of materials |
| CN114639061A (en) * | 2022-04-02 | 2022-06-17 | 山东博昂信息科技有限公司 | Vehicle detection method, system and storage medium |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR101564381B1 (en) * | 2014-09-01 | 2015-11-02 | 주식회사에스에이티 | System for controlling overloaded vehicle using axle-load weighting machine |
| CN109658711A (en) * | 2019-01-07 | 2019-04-19 | 合肥市规划设计研究院 | A kind of bridge overload early warning system |
| CN109815856A (en) * | 2019-01-08 | 2019-05-28 | 深圳中兴网信科技有限公司 | Status indication method, system and the computer readable storage medium of target vehicle |
| CN109827647A (en) * | 2019-01-17 | 2019-05-31 | 同济大学 | A kind of bridge dynamic weighing system |
| CN109903558A (en) * | 2019-03-07 | 2019-06-18 | 南京博瑞吉工程技术有限公司 | A kind of road and bridge vehicular load monitoring system and monitoring method |
| WO2019179809A1 (en) * | 2018-03-23 | 2019-09-26 | Bayerische Motoren Werke Aktiengesellschaft | Method for classifying a vehicle seat and classification system for classifying a vehicle seat |
| CN110633690A (en) * | 2019-09-24 | 2019-12-31 | 北京邮电大学 | Vehicle feature identification method and system based on bridge monitoring |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10459615B2 (en) * | 2014-12-11 | 2019-10-29 | Rdi Technologies, Inc. | Apparatus and method for analyzing periodic motions in machinery |
| US10049286B2 (en) * | 2015-12-15 | 2018-08-14 | International Business Machines Corporation | Image-based risk estimation |
-
2019
- 2019-12-31 CN CN201911402701.7A patent/CN111160270B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR101564381B1 (en) * | 2014-09-01 | 2015-11-02 | 주식회사에스에이티 | System for controlling overloaded vehicle using axle-load weighting machine |
| WO2019179809A1 (en) * | 2018-03-23 | 2019-09-26 | Bayerische Motoren Werke Aktiengesellschaft | Method for classifying a vehicle seat and classification system for classifying a vehicle seat |
| CN109658711A (en) * | 2019-01-07 | 2019-04-19 | 合肥市规划设计研究院 | A kind of bridge overload early warning system |
| CN109815856A (en) * | 2019-01-08 | 2019-05-28 | 深圳中兴网信科技有限公司 | Status indication method, system and the computer readable storage medium of target vehicle |
| CN109827647A (en) * | 2019-01-17 | 2019-05-31 | 同济大学 | A kind of bridge dynamic weighing system |
| CN109903558A (en) * | 2019-03-07 | 2019-06-18 | 南京博瑞吉工程技术有限公司 | A kind of road and bridge vehicular load monitoring system and monitoring method |
| CN110633690A (en) * | 2019-09-24 | 2019-12-31 | 北京邮电大学 | Vehicle feature identification method and system based on bridge monitoring |
Non-Patent Citations (1)
| Title |
|---|
| 区域高速公路网疲劳驾驶行为分布特征研究;钱圣隆;《中国优秀硕士学位论文全文数据库 工程科技辑》;全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111160270A (en) | 2020-05-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111160270B (en) | Bridge monitoring method based on intelligent video recognition | |
| CN111310645A (en) | Overflow bin early warning method, device, equipment and storage medium for cargo accumulation amount | |
| CN111881730A (en) | Wearing detection method for on-site safety helmet of thermal power plant | |
| CN108537154A (en) | Transmission line of electricity Bird's Nest recognition methods based on HOG features and machine learning | |
| CN111626277B (en) | Vehicle tracking method and device based on over-station inter-modulation index analysis | |
| CN109935080B (en) | Monitoring system and method for real-time calculation of traffic flow on traffic line | |
| CN102867183B (en) | Method and device for detecting littered objects of vehicle and intelligent traffic monitoring system | |
| CN114998852A (en) | Intelligent detection method for road pavement diseases based on deep learning | |
| CN111260629A (en) | Pantograph structure abnormity detection algorithm based on image processing | |
| CN109948455B (en) | Detection method and device for left-behind object | |
| CN111626170B (en) | Image recognition method for railway side slope falling stone intrusion detection | |
| CN110120155A (en) | A kind of chemical industry plant area vehicle overload overload intelligent monitoring and alarming system | |
| CN112070135A (en) | Power equipment image detection method and device, power equipment and storage medium | |
| CN114511718B (en) | Intelligent management method and system for materials for building construction | |
| CN108508023B (en) | Defect detection system for contact end jacking bolt in railway contact network | |
| CN108664875A (en) | Underground belt-conveying monitoring method based on image recognition | |
| CN111626169A (en) | Image-based railway dangerous falling rock size judgment method | |
| CN116665011A (en) | Coal flow foreign matter identification method for coal mine belt conveyor based on machine vision | |
| CN115995056A (en) | Automatic bridge disease identification method based on deep learning | |
| CN111724604A (en) | Highway non-stop speed measurement weighing system with license plate recognition function and method | |
| CN106910334B (en) | Method and device for predicting road section conditions based on big data | |
| CN114627286A (en) | Method for detecting wagon staff invasion based on PSPNet and improved YOLOv4 | |
| CN115810161A (en) | Transformer substation fire identification method and system | |
| CN111667655A (en) | Infrared image-based high-speed railway safety area intrusion alarm device and method | |
| CN110765900A (en) | DSSD-based automatic illegal building detection method and system |
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 |