CN116087198B - Highway road surface situation data acquisition equipment and automatic rapid detection system thereof - Google Patents
Highway road surface situation data acquisition equipment and automatic rapid detection system thereof Download PDFInfo
- Publication number
- CN116087198B CN116087198B CN202211541287.XA CN202211541287A CN116087198B CN 116087198 B CN116087198 B CN 116087198B CN 202211541287 A CN202211541287 A CN 202211541287A CN 116087198 B CN116087198 B CN 116087198B
- Authority
- CN
- China
- Prior art keywords
- data
- module
- model
- disease
- pavement
- 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
- 238000001514 detection method Methods 0.000 title claims abstract description 111
- 201000010099 disease Diseases 0.000 claims abstract description 98
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 claims abstract description 98
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 40
- 238000012545 processing Methods 0.000 claims abstract description 33
- 238000000034 method Methods 0.000 claims abstract description 19
- 230000008569 process Effects 0.000 claims abstract description 9
- 238000003860 storage Methods 0.000 claims abstract description 9
- 238000012544 monitoring process Methods 0.000 claims abstract description 5
- 238000012549 training Methods 0.000 claims description 29
- 238000004458 analytical method Methods 0.000 claims description 24
- 238000011156 evaluation Methods 0.000 claims description 14
- 238000007726 management method Methods 0.000 claims description 13
- 238000012360 testing method Methods 0.000 claims description 12
- 238000012795 verification Methods 0.000 claims description 10
- 230000006870 function Effects 0.000 claims description 9
- 238000002372 labelling Methods 0.000 claims description 9
- 238000013473 artificial intelligence Methods 0.000 claims description 7
- 238000013135 deep learning Methods 0.000 claims description 6
- 238000011161 development Methods 0.000 claims description 6
- 230000000694 effects Effects 0.000 claims description 6
- 238000009877 rendering Methods 0.000 claims description 6
- 239000010426 asphalt Substances 0.000 claims description 4
- 239000004568 cement Substances 0.000 claims description 4
- 238000000605 extraction Methods 0.000 claims description 4
- 230000004927 fusion Effects 0.000 claims description 4
- 230000005540 biological transmission Effects 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 claims description 3
- 238000006243 chemical reaction Methods 0.000 claims description 3
- 238000004140 cleaning Methods 0.000 claims description 3
- 238000005520 cutting process Methods 0.000 claims description 3
- 230000008030 elimination Effects 0.000 claims description 3
- 238000003379 elimination reaction Methods 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 3
- 238000013515 script Methods 0.000 claims description 3
- 238000013519 translation Methods 0.000 claims description 3
- 239000013598 vector Substances 0.000 claims description 3
- 230000000007 visual effect Effects 0.000 claims description 3
- 238000013480 data collection Methods 0.000 claims description 2
- 230000003993 interaction Effects 0.000 claims description 2
- 238000012216 screening Methods 0.000 claims description 2
- 239000013589 supplement Substances 0.000 claims description 2
- 238000011426 transformation method Methods 0.000 claims description 2
- 238000005516 engineering process Methods 0.000 description 10
- 238000010586 diagram Methods 0.000 description 6
- 230000008901 benefit Effects 0.000 description 5
- 238000012423 maintenance Methods 0.000 description 5
- 101001074449 Crotalus durissus terrificus Phospholipase A2 inhibitor CNF Proteins 0.000 description 4
- 230000005484 gravity Effects 0.000 description 3
- 230000001360 synchronised effect Effects 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 239000000203 mixture Substances 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000006855 networking Effects 0.000 description 2
- 238000007619 statistical method Methods 0.000 description 2
- 238000012800 visualization Methods 0.000 description 2
- 230000003872 anastomosis Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000013145 classification model Methods 0.000 description 1
- 230000000052 comparative effect Effects 0.000 description 1
- DMBHHRLKUKUOEG-UHFFFAOYSA-N diphenylamine Chemical compound C=1C=CC=CC=1NC1=CC=CC=C1 DMBHHRLKUKUOEG-UHFFFAOYSA-N 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000012797 qualification Methods 0.000 description 1
- 230000008439 repair process Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012827 research and development Methods 0.000 description 1
- 238000000638 solvent extraction Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000001502 supplementing effect Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
-
- E—FIXED CONSTRUCTIONS
- E01—CONSTRUCTION OF ROADS, RAILWAYS, OR BRIDGES
- E01C—CONSTRUCTION OF, OR SURFACES FOR, ROADS, SPORTS GROUNDS, OR THE LIKE; MACHINES OR AUXILIARY TOOLS FOR CONSTRUCTION OR REPAIR
- E01C23/00—Auxiliary devices or arrangements for constructing, repairing, reconditioning, or taking-up road or like surfaces
- E01C23/01—Devices or auxiliary means for setting-out or checking the configuration of new surfacing, e.g. templates, screed or reference line supports; Applications of apparatus for measuring, indicating, or recording the surface configuration of existing surfacing, e.g. profilographs
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01D—MEASURING NOT SPECIALLY ADAPTED FOR A SPECIFIC VARIABLE; ARRANGEMENTS FOR MEASURING TWO OR MORE VARIABLES NOT COVERED IN A SINGLE OTHER SUBCLASS; TARIFF METERING APPARATUS; MEASURING OR TESTING NOT OTHERWISE PROVIDED FOR
- G01D21/00—Measuring or testing not otherwise provided for
- G01D21/02—Measuring two or more variables by means not covered by a single other subclass
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S17/00—Systems using the reflection or reradiation of electromagnetic waves other than radio waves, e.g. lidar systems
- G01S17/86—Combinations of lidar systems with systems other than lidar, radar or sonar, e.g. with direction finders
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S17/00—Systems using the reflection or reradiation of electromagnetic waves other than radio waves, e.g. lidar systems
- G01S17/88—Lidar systems specially adapted for specific applications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S19/00—Satellite radio beacon positioning systems; Determining position, velocity or attitude using signals transmitted by such systems
- G01S19/01—Satellite radio beacon positioning systems transmitting time-stamped messages, e.g. GPS [Global Positioning System], GLONASS [Global Orbiting Navigation Satellite System] or GALILEO
- G01S19/13—Receivers
- G01S19/14—Receivers specially adapted for specific applications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B19/00—Programme-control systems
- G05B19/02—Programme-control systems electric
- G05B19/04—Programme control other than numerical control, i.e. in sequence controllers or logic controllers
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- 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
- G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
- G06V20/58—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads
-
- 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
- G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
- G06V20/588—Recognition of the road, e.g. of lane markings; Recognition of the vehicle driving pattern in relation to the road
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/12—Protocols specially adapted for proprietary or special-purpose networking environments, e.g. medical networks, sensor networks, networks in vehicles or remote metering networks
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
- G01N2021/8887—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges based on image processing techniques
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Remote Sensing (AREA)
- Radar, Positioning & Navigation (AREA)
- Theoretical Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Multimedia (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Electromagnetism (AREA)
- Signal Processing (AREA)
- Medical Informatics (AREA)
- Computing Systems (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Architecture (AREA)
- Civil Engineering (AREA)
- Structural Engineering (AREA)
- Life Sciences & Earth Sciences (AREA)
- Immunology (AREA)
- Automation & Control Theory (AREA)
- Artificial Intelligence (AREA)
- Biochemistry (AREA)
- Databases & Information Systems (AREA)
- Pathology (AREA)
- Software Systems (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a highway pavement condition data acquisition device and an automatic rapid detection system thereof, wherein the highway pavement condition data acquisition device consists of a core board, a depressurization module, a switch module, a Beidou module, a 4G module and a Wifi module; the core board adopts a 64-bit four-core X86 processor, acquires two paths of video data through a switch module, and then classifies, packages and stores the two paths of video data on an expansion storage; and in the storage process, the real-time data is configured in a wireless network through the WIFI module or the 4G module, so that the remote state is checked. The system collects original video data of road surface diseases through a high-definition camera; recording original data of road surface flatness through a flatness meter and collecting mileage data; identifying disease information through a pavement disease identification algorithm deployed by the data processing system; disease space positioning information is provided through the Beidou module; the 4G module is used for realizing interconnection between the acquisition equipment and the data processing system, monitoring the state of the control equipment and planning and checking track information; and summarizing and displaying the rapid detection results through a result display platform.
Description
Technical Field
The invention relates to road surface condition data acquisition equipment and a road surface disease detection and identification method, in particular to an automatic rapid detection system for a road surface.
Background
The rural highway is a pilot, basic and public facility for serving the economic and social development of vast rural areas, is an important component of a national highway network, and has an important basic supporting function for implementing a rural plain. In the face of complex conditions of rural highway points, line length and wide range, the traditional manual assessment mode has the problems of low efficiency, insufficient data accuracy and traceability and the like. The rural highway body is huge, and automatic detection of the road surface condition is a necessary trend of industry development. The existing automatic detection technology is high in cost and large in fund requirement, and is difficult to popularize and apply to rural highways.
Therefore, the intelligent detection meets the development requirements of the industry by fully utilizing advanced technologies such as big data, artificial intelligence and the like in the road condition detection assessment work of the road industry, and has important significance for comprehensively improving the fine and intelligent level of the road management and maintenance.
Intelligent detection of pavement is studied in the early 80 s of the 20 th century in the European and American countries, and is mature at present. There are many applications such as PCES system developed by Earth corporation of united states, HARRIS system of united kingdom, and automatic road analysis system-ARAN series of the company Fugro Roadware, canadian. In the aspect of China, the first intelligent detection vehicle of China is researched and developed by university of Nanjing university in 2002; the vehicle-mounted intelligent pavement automatic detection system which is independently developed by the university of Wuhan can detect pavement cracks, ruts and pavement evenness at the speed of 100 km/h; the road condition rapid detection vehicle (CiCS) developed by Highway maintenance technology and technology stock company in the sub-company of the road science institute of transportation department can be used for rapidly (0-100 km/h) detecting the damaged condition of the road surface, the flatness of the road surface, the depth of construction, ruts, front images and other condition detection, and is one of three matched rapid detection equipment of national province trunk line and expressway asset management system (CPMS) which are popularized in large scale in the current national department of transportation. The equipment has the advantages of complete functions, high price, high maintenance difficulty and large vehicle body, and is mainly applied to high-grade highways at present and difficult to popularize and apply on a large scale.
Under the background, in order to comprehensively improve the fine and intelligent level of rural highway management and maintenance, the development of pavement condition data acquisition equipment and a pavement disease detection and identification method/automatic rapid detection system suitable for rural highway pavement conditions is necessary for realizing rapid automatic analysis and comprehensive evaluation of pavement diseases.
Disclosure of Invention
Aiming at the defects of the prior art, the invention provides highway pavement condition data acquisition equipment and an automatic rapid pavement disease detection system. The method adopts the image acquisition and laser radar technology and adopts the pavement disease recognition algorithm based on artificial intelligence, thereby realizing the rapid and automatic analysis and comprehensive evaluation of pavement diseases, and being light in weight, rapid, easy to operate and low in cost.
The invention adopts the technical scheme that:
the main current road disease detection methods mainly comprise three categories: the detection method based on the monocular vision has the advantages of wide measurement range, low cost, high flexibility, good precision and the like, and becomes the focus of research. Detection methods based on monocular vision can be largely classified into two types of methods, namely traditional image processing and deep learning. The invention adopts the image acquisition and laser radar technology, develops light, rapid, easy-to-operate and low-cost rural highway pavement condition data acquisition equipment; based on an artificial intelligence pavement disease recognition algorithm and a server cloud computing technology, a pavement disease recognition management system with complete functions and strong visualization is built, and rapid automatic analysis and comprehensive evaluation of pavement diseases are realized.
System composition
The system consists of field acquisition equipment, a data processing system and a result display platform.
The field acquisition equipment is a source of system data (comprising high-definition video, flatness and GPS), and is uploaded to the data processing system through a webpage in a mobile storage mode; the data processing system is a core of data processing, and disease information is automatically identified through a deployed pavement disease identification algorithm; the achievement display platform is a view window for processing results and visually displays statistical achievement. The three parts are tightly connected when the system works, so that an integrated system with complete functions, reasonable labor division and up-down linkage is formed. The technical route of the pavement rapid system is shown in figure 1.
Field collection equipment
The field collection equipment comprises a high-definition camera, a planeness meter, an integrated chip integrated machine, a collection control terminal and the like. The high-definition camera is used for collecting original video data of road surface diseases, is a source of DR index of the road surface damage rate, can be used for recording road landscape image data, and is used for tracing road diseases and checking along line facilities manually; the planeness meter records the original data of the road surface flatness and collects mileage data, and is a source of a flatness index IRI. The built-in SIM card of the edge all-in-one machine receives the 4G signal, realizes interconnection of the acquisition equipment and the back-end server, is used for monitoring and controlling the state of the equipment and recording the data acquired by the whole system; the GPS antenna provides space positioning information for matching disease positioning. And the on-vehicle configuration tablet personal computer is provided with an acquisition APP for recording and planning and checking track information, integrally controlling hardware equipment and the like.
The device can be mounted on various SUV vehicle bodies, and is convenient to mount and dismount. Compared with the traditional detection vehicle, the detection vehicle has the characteristics of low cost and high cost performance. The integral acquisition equipment is shown in the outline of figure 2.
Data processing system
The data processing system mainly comprises a cloud server and a disease identification management system. The cloud server mainly comprises a service server and a GPU server, wherein the service server is used for storing and deploying a service system and a database, storing GPS data and video data, performing frame cutting, duplicate removal, data cleaning, data fusion, data display, deploying a data interface and the like on the video data; the intelligent pavement disease recognition algorithm model based on artificial intelligence is deployed, intelligent recognition of main diseases such as network cracks, longitudinal cracks, transverse cracks, pits, crushing plates, pits and the like of asphalt and cement concrete pavement can be realized, rapid automatic analysis and comprehensive evaluation are performed on pavement disease characteristics, and then disease positions and disease pictures are accurately given based on GPS and mileage encoder information.
The disease identification management system is mainly used for automatic analysis, evaluation and summarization of pavement diseases. The method can check the field collection condition and the assessment result, upload video, GPS, IRI and other original data, check field collection record information and data processing progress, check the processed disease details, statistical analysis results and the like. Meanwhile, a fixed format form can be output according to business requirements of the country, province, city and county, and conventional edition evaluation data can be automatically generated.
Achievement display platform
The rapid detection results are summarized and displayed according to the Henan province common highway and the waterway management platform. The disease recognition system is communicated and shared with the common highway and the waterway management platform, automatically processed data and road surface disease images are pushed to the platform in real time, and are visually displayed on a GIS map, so that basic data can be provided for rewarding and supplementing, planning maintenance and the like.
Overview of disease detection algorithm
2.1 Algorithm construction
The disease detection algorithm mainly comprises three stages of data processing, model training and application deployment. The data processing part mainly comprises sample collection, sample labeling, sample training/verification/test set division and enhancement and expansion of data used for training. The model training part mainly comprises model selection, parameter adjustment training and model output of the target detection model and the classification discrimination model. The application deployment mainly comprises a background auxiliary module and an algorithm model deployment, wherein the background auxiliary module is used for receiving requests, analyzing images, rendering effects, summarizing result forms, sending feedback and the like, a target detection model in the algorithm model is used for positioning diseases in the images and primarily classifying the diseases, and a classification discrimination model is used for assisting the target detection to conduct classification subdivision and interference elimination.
The principle of the disease detection algorithm is shown in fig. 3, and the establishment flow is shown in fig. 4.
Data processing
(1) Sample labeling and sample partitioning
And (3) defining the disease type and the common confusing interferent type, collecting a large number of data sets, marking a target detection frame by adopting a colabeler marking tool, and finishing to generate a target detection data set. And on the basis of target detection labeling, selecting frame, matting and classifying by means of scripts, and sorting to generate a classified data set. The two types of data sets are respectively divided into training samples, verification samples and test samples according to the ratio of 6:2:2.
(2) Image data enhancement
The image data enhancement is used for further enriching a training sample set and increasing the robustness of the trained model. Using conventional enhancements, including translation, flipping, rotation, HSV transformation, etc., for expanding the richness of the sample; using distortion disturbance enhancement to increase the robustness of the model to picture distortion caused by the camera lens; and the dynamic blurring enhancement is used for improving the accuracy of the model on identifying diseases in a high-speed motion scene. The Mosaic enhancement is used for enriching the background of the detected object, and data of a plurality of images are calculated at one time during BN calculation, so that the model learning efficiency is improved. The various enhancement effects are fused as shown in figure 5.
Model training
(1) Target detection model
The target detection selects a target detection algorithm model (Yolo v 5) based on deep learning, adopts a backbone feature extraction network CSPdark net, and the detection head is a convolution combination with three steps of 8, 16 and 32 respectively.
The Yolo series undergoes evolution of the previous four generations, and the target detection framework Yolov5 of the latest version is realized. Yolov5 is slightly stronger in accuracy performance than Yolov4, but far exceeds Yolov4 in flexibility and speed, and has extremely strong advantages in rapid deployment of models. The Yolov5 framework is realized based on Pytorch, and the bottom code is more simplified and is easy to develop secondarily. Accordingly, yolov5 is used as a core algorithm for disease target detection. The whole Yolov5 structure can be divided into the following four parts: input, backbone, neck, prediction. The Yolov5 structure is shown in fig. 6.
The main parameters in training are model depth, scaling factor with, anchors size, learning rate, loss balance parameters, data enhancement corresponding parameters and the like. In training, three-fold poor verification is adopted, three models are fused for forward reasoning during prediction, and NMS is performed by integrating pre-selected frames of the three models to obtain a final result.
(2) Classification discrimination model
As shown in FIG. 7, the classification model selects the densnet 169 model, and all layers in the network are connected by two, so that each layer in the network receives the characteristics of all layers in front of the layer as input, and the fine characteristics of pavement diseases are extracted and reserved. The main parameters are loss balance parameters, learning rate, data enhancement corresponding parameters and the like during training.
Model deployment
The application deployment mainly comprises a background auxiliary module and an algorithm model deployment.
The background auxiliary module is mainly realized by a flash framework and combines with an opencv and other method libraries to realize the functions of request receiving, format analysis, image base64 coding analysis, image size conversion, image normalization processing and the like.
The algorithm model deployment is mainly realized based on a torch and onnx framework, the target detection model performs forward reasoning by inputting a normalized image, the feature is extracted by a backbone network, and the detection head regresses to obtain frame selection positioning information and frame selection classification information as a preselected frame. And then selecting the frames according to the frame selection confidence level and the IOU through the NMS module. The classification discrimination model performs picture matting through the frame selection, inputs the model for analysis, generates classification feature vectors, and obtains a frame selection refined classification result as an output result of the final frame selection. And finally, the background auxiliary module performs image rendering, generates a result form and completes result feedback according to the frame selection result generated by the original image and the model.
The invention has the beneficial effects that:
1. the highway pavement condition data acquisition equipment can be mounted on various SUV vehicle bodies, is convenient to mount and dismount, is light in weight, quick and easy to operate, and has the characteristics of low cost and high cost performance. And uploading system data (comprising high-definition video, flatness and GPS) to a data processing system through a webpage in a mobile storage mode by adopting an image acquisition and laser radar technology and a unit module.
2. The automatic rapid detection system for the pavement diseases realizes rapid automatic analysis and comprehensive assessment of the pavement diseases based on an artificial intelligence pavement disease identification algorithm and a server cloud computing technology. A pavement disease identification management system platform with complete functions and strong visualization is built, and through a rural highway vehicle following comparison test with main stream multifunctional road condition rapid detection equipment (CICS) in Henan province, which is close to 30000km, the result shows that the system can be well used for working in rural highway pavement detection and evaluation, has good consistency in accuracy, and can completely meet the requirement of rural road pavement automatic detection.
3. The automatic rapid detection system for the pavement diseases adopts the target detection algorithm model Yolo v5 and densnet 169 classification and discrimination model based on deep learning, and is beneficial to realizing rapid automatic analysis and comprehensive evaluation of pavement disease characteristics. The target detection algorithm model Yolo v5 is slightly stronger than Yolo v4 in accuracy performance, but is greatly improved in flexibility and speed, and has extremely strong advantages in rapid deployment of the model. The classification discrimination model selects the densnet 169 model, all layers in the network are connected with each other, so that each layer in the network receives the characteristics of all layers in front of the network as input, and the extraction and the reservation of the fine characteristics of pavement diseases are facilitated. The comparison shows that the accuracy is better in consistency with the main stream multifunctional road condition rapid detection equipment, and the requirements of rural road surface detection are met.
Drawings
FIG. 1 is a schematic diagram showing the components of an automated rapid detection system for road surface defects;
FIG. 2 is a schematic block diagram of a road surface condition data acquisition device (field acquisition device);
FIG. 3 shows the principle of a disease detection algorithm;
FIG. 4 shows a disease recognition algorithm setup flow;
FIG. 5 is a schematic diagram showing the fusion of multiple enhancement effects;
FIG. 6 shows a diagram of the structure of Yolov 5;
FIG. 7 is a diagram showing a structure of a densnet;
FIG. 8 shows results of a synchronous acquisition and comparison test for Y029 line in Nanyang city;
fig. 9 shows the results of the X002 line car-following synchronous acquisition and comparison test in south yang city.
Detailed Description
In order to make the technical conception and advantages of the invention to achieve the objects of the invention more apparent, the technical scheme of the invention is further described in detail below with reference to the accompanying drawings. It is to be understood that the following examples are intended to illustrate and describe preferred embodiments of the invention and should not be construed as limiting the scope of the invention as claimed.
Example 1
As shown in fig. 2, the invention provides highway pavement condition data acquisition equipment, which consists of a core board with a depressurization module, a switch module, a Beidou module, a 4G module and a Wifi module, wherein the depressurization module provides working power for each component module of the equipment, and the core board adopts a 64-bit four-core X86 processor;
the switch module divides 4 network ports for intranet data transmission, wherein 1 network port is connected with the core board, two network ports are connected with the high-definition camera and the flatness instrument, and 1 network port is led to an external terminal;
the Beidou module is connected with the core board through a serial port, collects Beidou position data and provides real-time self coordinates and a real-time clock for the collection equipment;
the 4G module adopts an all-network version, and a data output interface is connected with the core board to provide external network connection for the acquisition equipment;
the Wifi module data input port is connected with the core board and used for interaction between the handheld terminal and the acquisition equipment;
the acquisition equipment acquires two paths of video data through the switch module, and then classifies, packages and stores the two paths of video data on the expansion storage; and in the storage process, the real-time data is configured in a wireless network through the WIFI module or the 4G module, so that the remote state is checked.
The core board adopts an on-board Intel Atom x5-Z8300 'slung' four-core processor, integrates Intel Gen8 HD core display (500 MHz), is provided with a double-channel 1GB/2GB/4GB LPDDR3-1866MHz memory, and the conventional expansion comprises: 1 USB 3.0, 3 USB 2.0, 1 USB OTG Type-C interface, gigabit LAN, HDMI 1.4 (4K@30 Hz), eDP and MIPI DSI, and 3.5mm audio interface; development extensions include: camera MIPI CSI I/F, 2xADC, 2xPWM, 2 xl 2C and 40 pin rasberry Pi compatible plugs. The original 4-core network port is changed into an RJ45 terminal.
The switch module chip uses IP175G, 4-way network port as switch function, and when in use, 2-way camera is connected, (camera network cable connection and core board network cable connection are provided). The model of the 4G module is AIR724, the Wifi module adopts an RTL8188FTV module, the TL8188FQA WiFi module provides ultra-low power consumption for networking equipment according to a more simplified standard, and the Wifi module scheme is developed based on an RTL8188FTV-VQ1 main chip. The wireless transmission rate is up to 72M, and the system can adapt to different working environments, so that a desktop or notebook computer user and other equipment needing to realize wireless networking can be conveniently accessed into a wireless network, and the TL8188FQA module developed based on the RTL8188FTV-VQ1 main chip can adapt to different working environments.
The highway road surface condition data acquisition equipment, when specifically using, can adopt vehicle sensor, vehicle sensor passes through GPIO and is connected with the core board for gather vehicle sensor data when big dipper positioning is not punctual, provide a driving speed's supplement for acquisition equipment.
The highway pavement condition data acquisition equipment (field acquisition equipment) comprises a high-definition camera, a planeness meter, an integrated chip all-in-one machine, an acquisition control terminal and the like. The high-definition camera is used for collecting original video data of road surface diseases, is a source of DR index of the road surface damage rate, can be used for recording road landscape image data, and is used for tracing road diseases and checking along line facilities manually; the planeness meter records the original data of the road surface flatness and collects mileage data, and is a source of a flatness index IRI. The 4G module is internally provided with a SIM card for receiving 4G signals, so that the acquisition equipment is interconnected with the back-end server and is used for monitoring and controlling the state of the equipment and recording the data acquired by the whole system; the Beidou module provides space positioning information and is used for matching disease positioning.
The field collection equipment can be mounted on various SUV car bodies, is convenient to mount and dismount, and has the characteristics of low cost and high cost performance. When the system is used, the tablet personal computer is configured on the vehicle to install and collect the APP, so that track information is recorded and planned to be checked, and the system is used for overall control of hardware equipment and the like. The device composition schematic block diagram is shown in fig. 2, and system data (including high-definition video, flatness and GPS) is uploaded to a data processing system through a webpage in a mobile storage mode.
Example 2
As shown in fig. 1, the automatic rapid detection system for the road surface comprises field collection equipment, a data processing system and a result display platform, wherein the field collection equipment uses the road surface condition data collection equipment, and the system collects original video data of road surface diseases through a high-definition camera; recording original data of road surface flatness through a flatness meter and collecting mileage data; space positioning information is provided through the Beidou module and used for matching disease positioning; the 4G module is internally provided with a SIM card to receive the 4G signal, so that the acquisition equipment is interconnected with the data processing system, and the acquisition equipment is used for monitoring the state of the control equipment, planning and checking track information and recording data acquired by the whole system;
the data processing system mainly comprises a cloud server and a disease identification management system, and disease information is automatically identified through a deployed pavement disease identification algorithm; the disease recognition system is mainly used for automatic analysis, evaluation and summarization of pavement diseases, and automatically processed data and pavement disease images are pushed to a result display platform in real time; the cloud server mainly comprises a service server and a GPU server, wherein the service server is used for storing and deploying a service system and a database, storing GPS data and video data, performing frame cutting, de-duplication, data cleaning, data fusion and data showing on the video data, and deploying a data interface; the GPU server is deployed with a pavement disease recognition algorithm model based on artificial intelligence, so that intelligent recognition of main diseases such as network cracks, longitudinal cracks, transverse cracks, pits, crushing plates, pits and the like of asphalt and cement concrete pavement is realized, rapid automatic analysis and comprehensive evaluation are performed on pavement disease characteristics, and then disease positions and disease pictures are accurately given based on GPS and mileage encoder information;
and the achievement display platform is used for summarizing and displaying the rapid detection achievement and providing a visual window for displaying the statistical achievement.
The disease identification management system is shared by intercommunication with the common highway and the waterway management platform, and can check the field collection condition and assessment result, upload video, GPS, IRI and other original data, check field collection record information and data processing progress, check the processed disease details and statistical analysis result; and outputting a fixed format form according to the service requirement, and automatically generating conventional edition evaluation data.
Example 3
The highway pavement automatic rapid detection system of this embodiment, unlike the embodiment, further: the progress discloses a pavement disease recognition algorithm which is specifically adopted. The principle of the disease detection algorithm is shown in fig. 4, and the establishment flow is shown in fig. 5.
The pavement disease recognition algorithm comprises three stages of data processing, model training and application deployment; the data processing stage mainly comprises sample collection, sample labeling, sample division and enhancement and expansion of data used for training;
the model training stage mainly comprises the selection of a target detection model and a classification discrimination model, parameter adjustment training and model output;
the application deployment stage mainly comprises a background auxiliary module and an algorithm model deployment part, wherein the background auxiliary module is used for receiving a request, analyzing an image, rendering an effect, summarizing a result form and sending feedback; the algorithm model comprises a target detection model and a classification discrimination model, wherein the target detection model is used for disease positioning and preliminary classification in the image, and the classification discrimination model is used for assisting target detection to conduct classification subdivision and interference elimination.
In the data processing process, the sample labeling and sample dividing processes are as follows:
the disease type and the common confusing interferent type are clarified, a large number of data sets are collected, a target detection frame is marked by adopting a colabeler marking tool, and a target detection data set is generated by arrangement;
on the basis of target detection labeling, selecting frame, matting and classifying by means of scripts, and sorting to generate a classified data set; dividing the two types of data sets into training samples, verification samples and test samples according to the proportion of 6:2:2;
image data enhancement, namely enriching a training sample set by using translation, overturning, rotation and HSV transformation methods; using distortion disturbance enhancement to increase the robustness of the model to picture distortion caused by the camera lens; the dynamic blur enhancement is used for improving the accuracy of the model for identifying diseases in a high-speed motion scene; the Mosaic enhancement is used for enriching the background of the detected object, and data of a plurality of images are calculated at one time during BN calculation, so that the model learning efficiency is improved. The various enhancement effects are fused as shown in figure 5.
The model training process includes (1) building a target detection model:
selecting a target detection algorithm model based on deep learning, extracting a network CSPdark by adopting backbone characteristics, wherein the detection head is a convolution combination with three steps of 8, 16 and 32 respectively;
the parameter adjustment is model depth, scaling factor with, anchors size, learning rate, loss balance parameter and data enhancement corresponding parameter, three-fold poor verification is adopted in training, three models are fused for forward reasoning during prediction, and NMS is conducted by integrating pre-selected frames of the three models to obtain a final result; the Yolov5 structure is shown in figure 6;
(2) Establishing a classification discrimination model
As shown in fig. 7, the densnet 169 model is selected, all layers in the network are connected in pairs, and each layer receives the characteristics of all the layers in front of the densnet as input, so as to extract and retain the subtle characteristics of road surface diseases; the main parameters are loss balance parameters, learning rate and data enhancement corresponding parameters during training.
In the application deployment stage, a background auxiliary module is mainly realized by a flash framework, and functions of request receiving, format analysis, image base64 coding analysis, image size conversion, image normalization processing and the like are realized by combining an opencv method library;
the algorithm model deployment is mainly realized based on a torch and onnx framework, a target detection model carries out forward reasoning through inputting a normalized image, has the characteristics of main network extraction, and a detection head regresses to obtain frame selection positioning information and frame selection classification information, and is used as a pre-selected frame, and then frame selection screening is carried out through an NMS module according to frame selection confidence and an IOU;
the classification discrimination model performs picture matting through the selection frame, inputs the model for analysis, generates classification feature vectors, and obtains a result of frame selection refinement classification as an output result of the final selection frame; and finally, the background auxiliary module performs image rendering, generates a result form and completes result feedback according to the frame selection result generated by the original image and the model.
Application analysis
In order to optimize an algorithm model and improve disease recognition accuracy, the technical research and development team sequentially selects and tests typical asphalt and cement pavements of nearly 200km on plain and rural roads in mountain areas, extracts disease pictures aiming at insufficient disease recognition or false recognition and the like, takes the extracted disease pictures as samples to continuously perfect a disease sample database after manual calibration, strengthens algorithm deep learning, and further improves algorithm performance and accuracy. The result also proves that the invention has the characteristics of high efficiency, low cost, light weight, general easy operation and the like, is suitable for the rapid detection of low-grade rural highways, has better consistency with the main stream multifunctional road condition rapid detection equipment in accuracy through comparison verification, and meets the requirements of rural road surface detection.
1. Car-following synchronous acquisition and comparison analysis for X002 and Y029 lines in Nanyang city
In 2021, on roads X002 and Y029 (total 74.519 km) in south yang city, vehicles using the rapid detection system of the present invention and CICS road condition detection vehicles start simultaneously on the same road, and data are collected simultaneously for detection, and PCI (road damage condition index) of mileage of each road section is shown in fig. 8 and 9.
As can be seen from fig. 8 and 9: the change fluctuation conditions of the PCI values are basically consistent, and the real condition of road damage can be reflected. Meanwhile, the PCI anastomosis degree is higher for the 'excellent' and 'good' road sections with PCI more than or equal to 80; there is a certain deviation in the "bad", "secondary" and "medium" road sections where the road PCI is determined to be less than 80.
The method is characterized by further counting different disease types: the identification accuracy rate of all crack types is 85%, and the recall rate is 92%; the identification accuracy rate of all pits is 78%, and the recall rate is 85%; the identification accuracy rate of all the repairs is 70%, the recall rate is 80%, and the detection requirements of rural roads are basically met.
2. Analysis of test point detection results of Lanc county
The rapid detection system is adopted for collection of 635 kilometers in the county road of Lane in 2021 and evaluation of test point road sections: and meanwhile, entrusts four professional qualification units to carry out spot check verification on the road sections with 138 kilometers of partial routes. The comparative deviation statistics obtained are shown in Table 1.
Table 1: detecting system and professional detecting mechanism detect result contrast deviation condition
It can be seen that the road section with the PCI deviation value error within 5% is about 73 km and the specific gravity is about 52.90%; the error is about 30 km at a road section of 5% -10%, and the specific gravity is about 21.74%; the road section with the error of more than 10 percent is about 35 kilometers and the specific gravity is 25.36 percent. Meanwhile, the road sections with the error of more than 10% are mainly X053 and Y007 road sections, the road technical condition grade of the road surface is mainly a secondary difference road section, and the algorithm can perform further optimization learning.
3. Inspection-saving vehicle following acquisition comparison
The road detection vehicle provided with the rapid detection system of the invention follows CICS vehicles of two professional detection institutions for detecting the road detection vehicle following the road detection vehicle in the general state of the road detection vehicle for 1949.97 kilometers in total in 7-8 months in 2021, and further the reliability of the system is verified. The following comparison results are shown in table 2.
Table 2: vehicle following comparison PCI deviation mileage and duty ratio condition
The comparison of the data comparison automation processing result and the CICS detection vehicle can show that the ratio of the road surface damage condition index (PCI) deviation smaller than 15% reaches more than 94%, and the requirements of related standards are met.
Through comparison and verification, the pavement automatic rapid detection equipment developed by the image acquisition and laser radar technology can realize rapid automatic analysis and comprehensive evaluation of pavement disease characteristics. The accuracy is better than that of main stream multifunctional road condition rapid detection equipment, and the requirements of rural road surface detection are met.
By the year 2021, the road surface detection work of all county and rural roads in Zheng Zhou, zhaoma shop and nan yang test point areas is finished, the road surface automatic rapid detection system operates well in the road surface detection and assessment work of rural roads in Henan province at present, and the test application and popularization can be carried out in high-grade trunk roads in the later period.
The above description is only a preferred embodiment of the present invention, and is not intended to limit the present invention. Other modifications of the practice of the invention will be apparent to those skilled in the art from consideration of the specification and practice of the invention without the need for inventive faculty, and any modification or substitution of equivalents which fall within the spirit and principles of the invention, or which are obvious to those skilled in the art, are intended to be encompassed within the scope of the invention.
Claims (6)
1. The automatic rapid detection system for the road surface comprises field acquisition equipment, a data processing system and a result display platform, wherein the system acquires original video data of road surface diseases through a high-definition camera; recording original data of road surface flatness through a flatness meter and collecting mileage data; space positioning information is provided through the Beidou module and used for matching disease positioning; the 4G module is internally provided with a SIM card to receive the 4G signal, so that the acquisition equipment is interconnected with the data processing system, and the acquisition equipment is used for monitoring the state of the control equipment, planning and checking track information and recording data acquired by the whole system; the data processing system mainly comprises a cloud server and a disease identification management system, and disease information is automatically identified through a deployed pavement disease identification algorithm; the disease recognition system is mainly used for automatic analysis, evaluation and summarization of pavement diseases, and automatically processed data and pavement disease images are pushed to a result display platform in real time; the achievement display platform is used for summarizing and displaying rapid detection achievement and providing a visual window for displaying statistical achievement; the method is characterized in that:
the pavement disease recognition algorithm comprises three stages of data processing, model training and application deployment;
the data processing stage mainly comprises sample collection, sample labeling, sample division and enhancement and expansion of data used for training; in the data processing process, the sample labeling and sample dividing processes are as follows:
the disease type and the common confusing interferent type are clarified, a large number of data sets are collected, a target detection frame is marked by adopting a colabeler marking tool, and a target detection data set is generated by arrangement;
on the basis of target detection labeling, selecting frame, matting and classifying by means of scripts, and sorting to generate a classified data set; dividing the two types of data sets into training samples, verification samples and test samples according to the proportion of 6:2:2;
image data enhancement, namely enriching a training sample set by using translation, overturning, rotation and HSV transformation methods; using distortion disturbance enhancement to increase the robustness of the model to picture distortion caused by the camera lens; the dynamic blur enhancement is used for improving the accuracy of the model for identifying diseases in a high-speed motion scene; the Mosaic enhancement is used for enriching the background of the detected object, and data of a plurality of images are calculated at one time during BN calculation, so that the model learning efficiency is improved;
the model training stage mainly comprises the selection of a target detection model and a classification discrimination model, parameter adjustment training and model output;
the model training process includes (1) building a target detection model:
selecting a target detection algorithm model based on deep learning, extracting a network CSPdark by adopting backbone characteristics, wherein the detection head is a convolution combination with three steps of 8, 16 and 32 respectively;
the parameter adjustment is model depth, scaling factor with, anchors size, learning rate, loss balance parameter and data enhancement corresponding parameter, three-fold poor verification is adopted in training, three models are fused for forward reasoning during prediction, and NMS is conducted by integrating pre-selected frames of the three models to obtain a final result;
(2) Establishing a classification discrimination model
Selecting a densnet 169 model, connecting all layers in a network in pairs, and extracting and retaining fine features such as pavement diseases by taking the features of all layers in front of the densnet as input by each layer; the main parameter adjustment is a loss balance parameter, a learning rate and a data enhancement corresponding parameter during training;
the application deployment stage mainly comprises a background auxiliary module and an algorithm model deployment part, wherein the background auxiliary module is used for receiving a request, analyzing an image, rendering an effect, summarizing a result form and sending feedback; the algorithm model comprises a target detection model and a classification discrimination model, wherein the target detection model is used for disease positioning and preliminary classification in an image, and the classification discrimination model is used for assisting target detection to conduct classification subdivision and interference elimination;
in the application deployment stage, a background auxiliary module is mainly realized by a flash framework, and functions of request receiving, format analysis, image base64 coding analysis, image size conversion, image normalization processing and the like are realized by combining an opencv method library;
the algorithm model deployment is mainly realized based on a torch and onnx framework, a target detection model carries out forward reasoning through inputting a normalized image, has the characteristics of main network extraction, and a detection head regresses to obtain frame selection positioning information and frame selection classification information, and is used as a pre-selected frame, and then frame selection screening is carried out through an NMS module according to frame selection confidence and an IOU;
the classification discrimination model performs picture matting through the selection frame, inputs the model for analysis, generates classification feature vectors, and obtains a result of frame selection refinement classification as an output result of the final selection frame; the final background auxiliary module performs image rendering, generates a result form and completes result feedback according to the frame selection result generated by the original image and the model
And the achievement display platform is used for summarizing and displaying the rapid detection achievement and providing a visual window for displaying the statistical achievement.
2. The automated highway pavement rapid detection system of claim 1, wherein: the cloud server mainly comprises a service server and a GPU server, wherein the service server is used for storing and deploying a service system and a database, storing GPS data and video data, performing frame cutting, de-duplication, data cleaning, data fusion and data showing on the video data, and deploying a data interface; the GPU server is deployed with an artificial intelligence-based pavement disease recognition algorithm model to realize intelligent recognition of main diseases such as network cracks, longitudinal cracks, transverse cracks, pits, crushing plates, pits and the like of asphalt and cement concrete pavement, rapidly and automatically analyze and comprehensively evaluate pavement disease characteristics, and accurately give out disease positions and disease pictures based on GPS and mileage encoder information.
3. The automated highway pavement rapid detection system of claim 1 or 2, wherein: the field collection equipment uses highway pavement condition data collection equipment and consists of a core board with a depressurization module, a switch module, a Beidou module, a 4G module and a Wifi module, wherein the depressurization module provides working power for each component module of the equipment, and the core board adopts a 64-bit four-core X86 processor; the switch module divides 4 network ports for intranet data transmission, wherein 1 network port is connected with the core board, two network ports are connected with the high-definition camera and the flatness instrument, and 1 network port is led to an external terminal; the Beidou module is connected with the core board through a serial port, collects Beidou position data and provides real-time self coordinates and a real-time clock for the collection equipment; the 4G module adopts an all-network version, and a data output interface is connected with the core board to provide external network connection for the acquisition equipment; the Wifi module data input port is connected with the core board and used for interaction between the handheld terminal and the acquisition equipment; the acquisition equipment acquires two paths of video data through the switch module, and then classifies, packages and stores the two paths of video data on the expansion storage; and in the storage process, the real-time data is configured in a wireless network through the WIFI module or the 4G module, so that the remote state is checked.
4. The automated highway pavement rapid detection system of claim 3, wherein: the core board adopts an on-board Intel Atom x5-Z8300 'slung' four-core processor, integrates Intel Gen8 HD core display (500 MHz), is provided with a dual-channel 1GB/2GB/4GB LPDDR3-1866MHz memory, and conventional expansion comprises: 1 USB 3.0, 3 USB 2.0, 1 USB OTG Type-C interface, gigabit LAN, HDMI 1.4 (4K@30 Hz), eDP and MIPI DSI, and 3.5mm audio interface; development extensions include: camera MIPI CSI I/F, 2xADC, 2xPWM, 2 xl 2C and 40 pin rasberry Pi compatible plugs.
5. The automated highway pavement rapid detection system of claim 3, wherein: the switch module chip uses IP175G, the 4G module uses model AIR724, the Wifi module adopts TL8188FQA module developed based on RTL8188FTV-VQ1 main chip, and the switch module chip can adapt to different working environments.
6. The automated highway pavement rapid detection system of claim 4 or 5, wherein: the highway pavement condition data acquisition equipment comprises a vehicle sensor, wherein the vehicle sensor is connected with the core board through a GPIO (general purpose input/output) and is used for acquiring vehicle sensor data when the Beidou is positioned out of time, and supplying a driving speed supplement for the acquisition equipment.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211541287.XA CN116087198B (en) | 2022-12-02 | 2022-12-02 | Highway road surface situation data acquisition equipment and automatic rapid detection system thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211541287.XA CN116087198B (en) | 2022-12-02 | 2022-12-02 | Highway road surface situation data acquisition equipment and automatic rapid detection system thereof |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116087198A CN116087198A (en) | 2023-05-09 |
| CN116087198B true CN116087198B (en) | 2024-02-23 |
Family
ID=86209223
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211541287.XA Active CN116087198B (en) | 2022-12-02 | 2022-12-02 | Highway road surface situation data acquisition equipment and automatic rapid detection system thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116087198B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116758507B (en) * | 2023-07-03 | 2023-12-19 | 中铁建设集团有限公司 | Pavement quality analysis method, device and program based on disease image acquisition and segmentation |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109870456A (en) * | 2019-02-01 | 2019-06-11 | 上海智能交通有限公司 | A kind of road surface health status rapid detection system and method |
| CN112184625A (en) * | 2020-09-01 | 2021-01-05 | 燊赛(上海)智能科技有限公司 | Pavement defect identification method and system based on video deep learning |
| CN113256601A (en) * | 2021-06-10 | 2021-08-13 | 北方民族大学 | Pavement disease detection method and system |
| CN113449806A (en) * | 2021-07-12 | 2021-09-28 | 苏州大学 | Two-stage forestry pest identification and detection system and method based on hierarchical structure |
| CN114359695A (en) * | 2021-12-31 | 2022-04-15 | 三峡大学 | Insulator breakage identification method based on uncertainty estimation |
| CN115201218A (en) * | 2022-07-13 | 2022-10-18 | 鲁朗软件(北京)有限公司 | Vehicle-mounted pavement disease intelligent detection method and system |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112598672A (en) * | 2020-11-02 | 2021-04-02 | 坝道工程医院(平舆) | Pavement disease image segmentation method and system based on deep learning |
-
2022
- 2022-12-02 CN CN202211541287.XA patent/CN116087198B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109870456A (en) * | 2019-02-01 | 2019-06-11 | 上海智能交通有限公司 | A kind of road surface health status rapid detection system and method |
| CN112184625A (en) * | 2020-09-01 | 2021-01-05 | 燊赛(上海)智能科技有限公司 | Pavement defect identification method and system based on video deep learning |
| CN113256601A (en) * | 2021-06-10 | 2021-08-13 | 北方民族大学 | Pavement disease detection method and system |
| CN113449806A (en) * | 2021-07-12 | 2021-09-28 | 苏州大学 | Two-stage forestry pest identification and detection system and method based on hierarchical structure |
| CN114359695A (en) * | 2021-12-31 | 2022-04-15 | 三峡大学 | Insulator breakage identification method based on uncertainty estimation |
| CN115201218A (en) * | 2022-07-13 | 2022-10-18 | 鲁朗软件(北京)有限公司 | Vehicle-mounted pavement disease intelligent detection method and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116087198A (en) | 2023-05-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109870456B (en) | Rapid detection system and method for road surface health condition | |
| CN111855664B (en) | Adjustable three-dimensional tunnel defect detection system | |
| CN109716108B (en) | Bituminous paving disease detecting system based on two mesh image analysis | |
| CN103778681A (en) | Vehicle-mounted high-speed road inspection system and data acquisition and processing method | |
| Lei et al. | Automated pavement distress detection and deterioration analysis using street view map | |
| CN105956942A (en) | Detection method for quality of electric power pipe gallery reinforcing steel bar mesh based on machine vision and detection device | |
| CN102901444B (en) | Method for detecting component size based on matching pursuit (MP) wavelet filtering and detecting system thereof | |
| CN106296814B (en) | Highway maintenance detection and virtual interactive interface method and system | |
| CN112800913A (en) | Pavement damage data space-time analysis method based on multi-source feature fusion | |
| CN110503070A (en) | Traffic automation monitoring method based on Aerial Images object detection process technology | |
| CN116087198B (en) | Highway road surface situation data acquisition equipment and automatic rapid detection system thereof | |
| CN110021167A (en) | A kind of highway cloud inspection system and data processing method | |
| Luo et al. | Semantic Riverscapes: Perception and evaluation of linear landscapes from oblique imagery using computer vision | |
| CN112819066A (en) | Res-UNet single tree species classification technology | |
| CN109635629A (en) | A kind of bus platform crowd density detection method and device based on deep learning | |
| CN114627021A (en) | Point cloud and deep learning based defect detection method and system | |
| CN113326865A (en) | Highway road surface disease three-dimensional information detecting system based on deep learning | |
| Abou Chacra et al. | Municipal infrastructure anomaly and defect detection | |
| Zhang et al. | Research on Baidu street view road crack information extraction based on deep learning method | |
| CN114971166A (en) | Lightweight road pavement service quality inspection system | |
| CN113420109B (en) | Method for measuring permeability of street interface, computer and storage medium | |
| CN112446343B (en) | Vehicle-mounted point cloud road shaft-shaped object machine learning automatic extraction method integrating multi-scale features | |
| CN103063674A (en) | Detection method for copper grade of copper block, and detection system thereof | |
| CN103048257A (en) | Sandy soil type identification and particle component analysis device | |
| CN102955935A (en) | Sediment recognition method and sediment recognition system based on image fusion |
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 |