CN114049560A - Road surface multi-feature disease detection method and device based on combination of multiple neural networks - Google Patents
Road surface multi-feature disease detection method and device based on combination of multiple neural networks Download PDFInfo
- Publication number
- CN114049560A CN114049560A CN202111410199.1A CN202111410199A CN114049560A CN 114049560 A CN114049560 A CN 114049560A CN 202111410199 A CN202111410199 A CN 202111410199A CN 114049560 A CN114049560 A CN 114049560A
- Authority
- CN
- China
- Prior art keywords
- road surface
- pavement
- model
- disease
- training
- 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
- 201000010099 disease Diseases 0.000 title claims abstract description 116
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 title claims abstract description 116
- 238000001514 detection method Methods 0.000 title claims abstract description 95
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 24
- 238000012549 training Methods 0.000 claims abstract description 35
- 238000012544 monitoring process Methods 0.000 claims abstract description 11
- 238000010586 diagram Methods 0.000 claims description 8
- 230000000694 effects Effects 0.000 claims description 8
- 238000000034 method Methods 0.000 claims description 7
- 238000004364 calculation method Methods 0.000 claims description 4
- 230000005764 inhibitory process Effects 0.000 claims description 4
- 238000012216 screening Methods 0.000 claims description 4
- 230000005540 biological transmission Effects 0.000 claims description 2
- 230000003042 antagnostic effect Effects 0.000 abstract 1
- 230000006870 function Effects 0.000 description 6
- 238000007689 inspection Methods 0.000 description 4
- 238000007726 management method Methods 0.000 description 4
- 238000004891 communication Methods 0.000 description 3
- 238000013135 deep learning Methods 0.000 description 3
- 230000007246 mechanism Effects 0.000 description 3
- 238000005457 optimization Methods 0.000 description 3
- 238000007781 pre-processing Methods 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 238000010276 construction Methods 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 230000003631 expected effect Effects 0.000 description 2
- 238000002372 labelling Methods 0.000 description 2
- 238000012423 maintenance Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 241000283070 Equus zebra Species 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 238000005299 abrasion Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000003137 locomotive effect Effects 0.000 description 1
- 230000035515 penetration Effects 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/2431—Multiple classes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/253—Fusion techniques of extracted features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/088—Non-supervised learning, e.g. competitive learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/20—Administration of product repair or maintenance
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/08—Construction
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Business, Economics & Management (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- General Health & Medical Sciences (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Health & Medical Sciences (AREA)
- Human Resources & Organizations (AREA)
- Computational Linguistics (AREA)
- Marketing (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Economics (AREA)
- General Business, Economics & Management (AREA)
- Molecular Biology (AREA)
- Tourism & Hospitality (AREA)
- Strategic Management (AREA)
- Quality & Reliability (AREA)
- Operations Research (AREA)
- Entrepreneurship & Innovation (AREA)
- Primary Health Care (AREA)
- Traffic Control Systems (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a pavement multi-feature disease detection method based on combination of multiple neural networks, which comprises the following steps: generating an antagonistic network through training to generate a pavement disease picture, and enlarging a data set to train a pavement disease detection model; loading the trained optimal pavement disease detection model into a system board for pavement disease detection; the road surface detection information is sent to a monitoring information platform through MQTT; and finally, the user refers to the relevant road information through the platform. According to the invention, a large amount of time and manpower are not wasted to collect the labeled data set, the multi-characteristic diseases of the road surface can be rapidly detected, and a monitoring platform is established, so that a user can check the road surface information at any time and any place.
Description
Technical Field
The invention relates to the field of pavement disease detection, in particular to a pavement multi-feature disease detection method and device based on multi-neural network combination.
Background
In recent years, road construction in China is improved continuously, and transportation capacity is enhanced continuously, but with the increase of time, road facilities are damaged gradually, such as road structure damage, zebra line white line abrasion, pavement cracks and pits, pavement ruts and the like, which bring about more and less social influence and serious economic loss. Therefore, the planning construction of road maintenance is enhanced, the quality of the road can be effectively improved, and the economic loss caused by the damage of the road is avoided. At present, road disease inspection work is mainly manual detection due to technical limitation, only a small part of the road disease inspection work adopts intelligent inspection equipment, but the intelligent degree of the equipment function list is low. The road technical condition detection vehicle is high in cost, long in detection period, and incapable of meeting the requirement of disease normalized inspection on average once a year. How to rapidly and efficiently perform automatic detection on roads has become an important research topic in the transportation industry.
CN112215203A A road surface disease detection method and device based on deep learning, discloses a road surface disease detection method based on deep learning, including: acquiring a road surface image or a road surface video; acquiring a road surface image to be detected according to the road surface image or the road surface video; and identifying the pavement diseases in the pavement to-be-detected image by using a pavement disease detection model based on a deep learning network, and acquiring corresponding positioning data to form pavement disease comprehensive information. The road surface disease detection method can only detect the road surface disease, but the outside cannot acquire road information in real time, and the workload of the early training model is large.
Disclosure of Invention
The purpose of the invention is as follows: in order to solve the problems that the workload of an early training model of a general pavement disease detection method is large and the road information cannot be obtained in real time, the invention provides a pavement multi-feature disease detection method and a device based on combination of a multi-neural network.
The technical scheme is as follows: the pavement multi-feature disease detection method based on the combination of the multiple neural networks comprises the following steps:
(1) acquiring pictures of various road surface diseases by using a vehicle-mounted camera, and selecting a small number of images of various road surface diseases from the acquired images and labeling the images for classification;
(2) establishing and generating an confrontation network model and a pavement disease detection model;
(3) training to generate an confrontation network model, sending the classified road surface disease pictures into the generated confrontation network model in batches for training, adjusting the hyper-parameter gamma, and generating a plurality of data sets with different qualities and diversity;
(4) training a road surface disease detection model, forming a data set by a real road surface disease picture and a generated picture, inputting the data set into the road surface disease detection model for training, detecting a characteristic diagram to obtain a default frame, performing non-maximum value inhibition screening after calculation, judging road surface diseases, introducing an attention mechanism among Conv7, Conv8, Conv9, Conv10 and Conv11 of the road surface disease detection model, and improving the extraction efficiency of the disease characteristics.
(5) And returning the pavement disease data to the pavement information monitoring platform.
The confrontation network model generated in the step (2) comprises an encoder, a generator and a discriminator, the hidden variable value of the real image is obtained after the real image is sent to the encoder, the original random noise is replaced to train the confrontation network, the generated confrontation network model is used for expanding the data set quantity of the type aiming at the disease type with poor detection effect, the detection effect of the disease type is strengthened and trained, and the generated confrontation network model serves as the data set of the pavement disease detection model; the road surface disease detection model is an improved SSD model and is responsible for detecting the multi-characteristic diseases of the road surface.
The generated confrontation network model is obtained based on BEGAN network training, and the pavement disease detection model is obtained based on SSD network training.
The expected effect training judgment standard of the pavement disease detection model comprises recall rate and accuracy rate.
The pavement information monitoring platform adopts an MQTT protocol for data transmission, and realizes the management of pavement disease detection through docker and FRP tools.
The road surface multi-feature disease detection device based on the combination of the multi-neural network comprises a power supply module, a data acquisition module, a detection module and an external module;
the power supply module is used for supplying power to the acquisition module, the detection module and the external module;
the data acquisition module acquires pictures of various pavement diseases by using the vehicle-mounted camera;
the detection module configured to:
a confrontation network model and a pavement disease detection model are built and generated,
training to generate an confrontation network model, sending the classified pavement disease pictures into the generated confrontation network model in batches for training, adjusting the hyper-parameter gamma to generate a plurality of data sets with different quality diversity,
training a pavement disease detection model, forming a data set by a real pavement disease picture and the generated picture, inputting the data set into the pavement disease detection model for training, detecting the characteristic diagram to obtain a default frame, performing non-maximum inhibition screening after calculation, judging the pavement disease,
the detection module is connected with the data acquisition module;
the external module comprises a built-in display, a wireless network card, a sound box and a GPS, wherein the GPS is used for positioning the position of a road surface image, the sound box is used for alarming, the display screen is used for checking data, and the wireless network card is used for transmitting the data.
The data acquisition module is a vehicle-mounted camera and sends road surface pictures shot by the camera into the detection module after being processed.
The detection module generates an confrontation network model which comprises an encoder, a generator and a discriminator and serves as a data set of the pavement disease detection model; the road surface disease detection model is an improved SSD model and is responsible for detecting the multi-characteristic diseases of the road surface.
The road surface multi-feature disease detection device based on the combination of the multi-neural network further comprises an external detection platform, and the external detection platform is used for reading road section information, GPS coordinates and disease types by a user and transmitting data with the outside through a wireless network card of an external module.
Has the advantages that: compared with the prior art, the invention has the following remarkable advantages:
the pavement multi-feature disease detection method based on the combination of the multi-neural network does not need to consume a large amount of time for collecting and marking a data set, has a strong detection effect on the pavement multi-feature diseases, and can enable an external user to search road section information, GPS coordinates and disease types through a pavement information monitoring platform.
Drawings
FIG. 1 is a flow chart of a road surface multi-feature disease detection method based on multi-neural network combination;
FIG. 2 is a diagram of a generation of a pairing countermeasure network model;
FIG. 3 is a diagram of a road surface detection network model;
FIG. 4 is a design diagram of a road surface multi-feature disease detection device based on multi-neural network combination;
fig. 5 is a flowchart of the work of the road information monitoring platform.
Detailed Description
The technical scheme of the invention is further explained by combining the attached drawings.
Example 1:
the invention provides a pavement multi-feature disease detection method based on combination of a multi-neural network, which comprises the following steps of:
before pavement multi-feature disease detection, a neural network needs to be trained and loaded into a Jetson nano system board, a small number of pavement disease pictures are obtained through a camera loaded on a vehicle, the size of the pictures is adjusted to be 300 x 300, and labels are marked for various pavement diseases to classify.
Table 1 pavement disease category table:
as shown in fig. 2, a confrontation network model is built and generated, wherein an encoder is an encoder module of a VAE variational self-encoder, a mean value and a variance obeyed by a real road surface disease picture are obtained after the real road surface disease picture is input into the encoder, and random sampling is performed according to normal distribution to obtain a hidden variable value. Because the problems of non-convergence and collapse are easy to occur in the training of the generated confrontation network, the hidden variable is used for replacing an original random noise input generator for generating the confrontation network input, and then the generated confrontation network input generator and the original random noise input generator are sent to the discriminator together with the real pavement disease picture, wherein the generator and the discriminator are the generator and the discriminator of the BEGAN network, and the formula of the loss function of the t step during the training of the discriminator, the generator and the discriminator is as follows:
wherein the loss function of the generated network is:
LG＝L(G(zG))，
the penalty function for the discriminator is:
LD＝L(X)-kt(G(zD))，
during training, the loss function of the t step is as follows:
kt+1＝kt+λk(γL(x)-L(G(zG)))。
wherein, different gamma can change the generation result of the image, when the gamma value is lower, the obtained image is single, but a finer image can be generated; and generating fuzzy image data as the gamma increases, and performing optimization between the quality and the diversity of the generated image by using the hyper-parameter gamma, wherein the optimization method is used for judging the training effect of the road disease model.
And generating a picture for generating the pavement diseases by using the trained generation countermeasure network, wherein the size of the generated picture is 300 x 300, labeling and classifying are performed, and finally the generated picture and the picture of the real pavement diseases form a data set of the pavement detection model together.
As shown in fig. 3, a road surface multi-feature detection model is built, the size of an input picture is 300 × 300, a data set is divided into a training set and a testing set according to the ratio of 8: 2, an attention mechanism is introduced among Conv7, Conv8, Conv9, Conv10 and Conv11, and a selected attention module is SE-NET. The attention mechanism can increase the weight coefficient of the disease position, improve the detection efficiency and loss accuracy, set the batch processing number to be 20, the learning rate to be 0.0004, the momentum optimization value to be 0.8, the iteration number to be 10000, and select L2 regularization according to the regularization standard.
The discrimination criteria of the expected effect of the road surface multi-feature detection model training are recall rate and precision rate. In the training and testing process, multiple tests are performed to optimize the quality and diversity generated by the generated countermeasure networks of different gamma, the model with the best training effect is selected, and the generated countermeasure networks are used for increasing the corresponding pavement disease data sets to perform reinforced training aiming at the disease types with weaker detection effects, so that the model training effect is further improved.
And loading the optimal pavement disease multi-feature detection model into a Jetson nano system board to complete the algorithm part of the pavement multi-feature disease detection device based on the combination of the multiple neural networks.
Example 2:
the invention discloses a road surface multi-feature disease detection device based on combination of a multi-neural network, which is a hardware design diagram as shown in fig. 4 and comprises a power supply module, an acquisition module, a detection module and an external module.
The power supply module supplies power for the acquisition module, the detection module and the external module.
The collection module comprises the USB camera, and the camera passes through in the USB inserts Jetson nano system board, installs the camera at locomotive or rear of a vehicle and shoots collection road surface information, detects in transmitting the image information to road surface disease detection model.
The detection module is internally provided with a road surface disease characteristic detection model and a picture preprocessing module which are trained, wherein the picture preprocessing module is mainly used for preprocessing pictures, and comprises the steps of scaling the size of the pictures to 300 x 300, carrying out Gaussian denoising and mean denoising, and improving the contrast of the pictures. And then the data are sent into a detection model for pavement disease detection, are connected with an external module and send the processed data to the outside.
The external module comprises a GPS, a sound box, a display screen and a wireless network card. The GPS is used for acquiring position information and knowing the quality condition of the road surface in time. The audio device is used for alarming or giving out prompt sound, and the sound device can feed back to the client in the case that the system is wrong or a user needs to be reminded. The display screen is convenient for a user to operate the system and a developer to debug the system. The wireless network card uses a USB network card, so that the network communication of the system is ensured, and the detection data can be transmitted to the road surface detection information platform in time.
As shown in fig. 5, which is a working flow chart of a road information monitoring platform, initialization is performed first and whether MQTT works normally is judged; then judging whether the camera works normally or not; if the camera works abnormally, the abnormal information is sent to the MQTT for the user to subscribe, if the camera works normally, the road surface diseases are collected, the road surface diseases are transmitted in real time through the MQTT and wait for subscription, and meanwhile, the effectiveness of the disease data is judged according to whether the vehicle is static or not. And the monitoring information platform subscribes pavement disease topic information to the MQTT and performs graphical processing and display.
The wireless communication system needs to complete the function of mutual communication between each part of the pavement detection system and the pavement monitoring information platform, and the management of the detection system is realized through tools such as docker, FRP and the like.
Through the FRP tool, local web items can be provided for external network access, and the FRP supports domain name binding and internal network penetration performance with high performance. In the invention
In the road surface detection system, the FRP is configured among the servers, so that the remote management and the daily maintenance of the detection system can be realized.
And a dc (Docker-composition) tool is used for realizing the management and convenient use of Docker.
In the method and the device for detecting the road surface multi-feature diseases based on the combination of the multi-neural network, the road surface data are collected by a camera on a running vehicle, then the road surface data are sent to a development board for road surface detection and identification, and then a background and a platform are uploaded through an MQTT, so that a user can look up the related information of the road surface on the platform.
Claims (8)
1. The pavement multi-feature disease detection method based on the combination of the multiple neural networks is characterized by comprising the following steps of:
(1) acquiring pictures of various pavement diseases by using a vehicle-mounted camera;
(2) establishing and generating an confrontation network model and a pavement disease detection model;
(3) training to generate an confrontation network model, sending the classified road surface disease pictures into the generated confrontation network model in batches for training, adjusting the hyper-parameter gamma, and generating a plurality of data sets with different quality diversity;
(4) training a pavement disease detection model, forming a data set by a real pavement disease picture and the generated picture, inputting the data set into the pavement disease detection model for training, detecting the characteristic diagram to obtain a default frame, performing non-maximum inhibition screening after calculation, and judging the pavement disease.
(5) And returning the pavement disease data to the pavement information monitoring platform.
2. The method for detecting the road surface multi-feature diseases based on the combination of the multi-neural networks as claimed in claim 1, wherein the generation of the confrontation network model in the step (2) comprises an encoder, a generator and a discriminator; the road surface disease detection model is an improved SSD model and is responsible for detecting the multi-characteristic diseases of the road surface.
3. The method for detecting the road surface multi-feature diseases based on the combination of the multi-neural networks as claimed in claim 1, wherein the generated confrontation network model is obtained based on BEGAN network training, and the road surface disease detection model is obtained based on SSD network training.
4. The method for detecting the road surface multi-feature diseases based on the combination of the multi-neural networks is characterized in that the judgment criteria of the expected training effect of the road surface disease detection model comprise recall rate and accuracy rate.
5. The method for detecting the pavement multi-feature diseases based on the combination of the multi-neural network as claimed in claim 1, wherein the pavement information monitoring platform adopts an MQTT protocol for data transmission, and management of pavement disease detection is realized through docker and FRP tools.
6. The road surface multi-feature disease detection device based on the combination of the multi-neural network is characterized by comprising a power supply module, a data acquisition module, a detection module and an external module;
the power supply module is used for supplying power to the acquisition module, the detection module and the external module;
the data acquisition module acquires pictures of various pavement diseases by using the vehicle-mounted camera;
the detection module configured to:
a confrontation network model and a pavement disease detection model are built and generated,
training to generate an confrontation network model, sending the classified pavement disease pictures into the generated confrontation network model in batches for training, adjusting the hyper-parameter gamma to generate a plurality of data sets with different quality diversity,
training a pavement disease detection model, forming a data set by a real pavement disease picture and the generated picture, inputting the data set into the pavement disease detection model for training, detecting the characteristic diagram to obtain a default frame, performing non-maximum inhibition screening after calculation, judging the pavement disease,
the detection module is connected with the data acquisition module;
the external module comprises a built-in display, a wireless network card, a sound box and a positioning system.
7. The device for detecting the multi-characteristic diseases on the road surface based on the combination of the multi-neural network as claimed in claim 5, wherein the data acquisition module is a vehicle-mounted camera, and road surface pictures shot by the camera are processed and then sent to the detection module.
8. The device for detecting the multi-characteristic road surface diseases based on the combination of the multi-neural networks as claimed in claim 5, wherein the generation of the confrontation network model in the detection module comprises an encoder, a generator and a discriminator, and the generation of the confrontation network model serves for a data set of the road surface disease detection model; the road surface disease detection model is an improved SSD model and is responsible for detecting the multi-characteristic diseases of the road surface.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111410199.1A CN114049560A (en) | 2021-11-25 | 2021-11-25 | Road surface multi-feature disease detection method and device based on combination of multiple neural networks |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111410199.1A CN114049560A (en) | 2021-11-25 | 2021-11-25 | Road surface multi-feature disease detection method and device based on combination of multiple neural networks |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114049560A true CN114049560A (en) | 2022-02-15 |
Family
ID=80210892
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111410199.1A Pending CN114049560A (en) | 2021-11-25 | 2021-11-25 | Road surface multi-feature disease detection method and device based on combination of multiple neural networks |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114049560A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114943693A (en) * | 2022-05-09 | 2022-08-26 | 盐城工学院 | Jetson Nano bridge crack detection method and system |
| CN115638831A (en) * | 2022-12-21 | 2023-01-24 | 四川九通智路科技有限公司 | Highway facility risk monitoring method and system based on MEMS sensor |
| CN117237925A (en) * | 2023-11-16 | 2023-12-15 | 南京萨利智能科技有限公司 | Intelligent road disease inspection method and system based on computer vision |
-
2021
- 2021-11-25 CN CN202111410199.1A patent/CN114049560A/en active Pending
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114943693A (en) * | 2022-05-09 | 2022-08-26 | 盐城工学院 | Jetson Nano bridge crack detection method and system |
| CN115638831A (en) * | 2022-12-21 | 2023-01-24 | 四川九通智路科技有限公司 | Highway facility risk monitoring method and system based on MEMS sensor |
| CN115638831B (en) * | 2022-12-21 | 2023-04-25 | 四川九通智路科技有限公司 | Highway facility risk monitoring method and system based on MEMS sensor |
| CN117237925A (en) * | 2023-11-16 | 2023-12-15 | 南京萨利智能科技有限公司 | Intelligent road disease inspection method and system based on computer vision |
| CN117237925B (en) * | 2023-11-16 | 2024-02-13 | 南京萨利智能科技有限公司 | Intelligent road disease inspection method and system based on computer vision |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114049560A (en) | Road surface multi-feature disease detection method and device based on combination of multiple neural networks | |
| CN109201514B (en) | Waste sorting recycle method, garbage classification device and classified-refuse recovery system | |
| CN111339882B (en) | Power transmission line hidden danger detection method based on example segmentation | |
| CN109977921B (en) | Method for detecting hidden danger of power transmission line | |
| CN111126238B (en) | X-ray security inspection system and method based on convolutional neural network | |
| Cord et al. | Automatic road defect detection by textural pattern recognition based on AdaBoost | |
| CN106570467B (en) | A kind of personnel based on convolutional neural networks leave the post detection method | |
| Nguyen et al. | Pavement crack detection using convolutional neural network | |
| CN110390691A (en) | A kind of ore scale measurement method and application system based on deep learning | |
| CN109815998A (en) | A kind of AI dress dimension method for inspecting and system based on YOLO algorithm | |
| Bo et al. | Particle pollution estimation from images using convolutional neural network and weather features | |
| CN110618134A (en) | Steel plate surface quality defect detection and rating system and method | |
| CN1806501B (en) | Marine phytoplankton automatic distinguishing method and apparatus | |
| CN109685761B (en) | Power cable defect detection method and detection system based on cloud platform | |
| CN110163278A (en) | A kind of flame holding monitoring method based on image recognition | |
| CN108664875A (en) | Underground belt-conveying monitoring method based on image recognition | |
| CN102236925B (en) | System and method for offline secondary detection and checking of machine detected data of large-piece checker | |
| CN112967255A (en) | Shield segment defect type identification and positioning system and method based on deep learning | |
| CN102915432A (en) | Method and device for extracting vehicle-bone microcomputer image video data | |
| Zhang et al. | Deep learning‐based automatic detection of muck types for earth pressure balance shield tunneling in soft ground | |
| CN117197591B (en) | Data classification method based on machine learning | |
| CN110175517A (en) | A kind of Coal-rock identification method based on transform domain numerical characteristic | |
| CN105334224A (en) | Automatic quality detection cloud platform | |
| CN111814873A (en) | Method for distinguishing drainage pipeline defect types and automatically identifying defect grades | |
| CN110196886A (en) | The multi-source heterogeneous big data correlating method of agricultural non-point source pollution and the big data supervising platform for using this method |
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 |