CN116434533A - AI wisdom highway tunnel synthesizes monitoring platform based on 5G - Google Patents
AI wisdom highway tunnel synthesizes monitoring platform based on 5G Download PDFInfo
- Publication number
- CN116434533A CN116434533A CN202211615885.7A CN202211615885A CN116434533A CN 116434533 A CN116434533 A CN 116434533A CN 202211615885 A CN202211615885 A CN 202211615885A CN 116434533 A CN116434533 A CN 116434533A
- Authority
- CN
- China
- Prior art keywords
- tunnel
- monitoring
- image
- fire
- module
- 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
- 238000012544 monitoring process Methods 0.000 title claims abstract description 104
- 238000005516 engineering process Methods 0.000 claims abstract description 10
- 238000013135 deep learning Methods 0.000 claims abstract description 6
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims abstract description 6
- 238000011161 development Methods 0.000 claims abstract description 5
- 239000011435 rock Substances 0.000 claims abstract description 4
- 238000001514 detection method Methods 0.000 claims description 17
- 238000012549 training Methods 0.000 claims description 15
- 238000004891 communication Methods 0.000 claims description 14
- 238000000034 method Methods 0.000 claims description 13
- 238000011176 pooling Methods 0.000 claims description 12
- 238000013527 convolutional neural network Methods 0.000 claims description 10
- 230000005540 biological transmission Effects 0.000 claims description 9
- 230000036541 health Effects 0.000 claims description 9
- 239000003595 mist Substances 0.000 claims description 8
- 230000001052 transient effect Effects 0.000 claims description 8
- 238000007781 pre-processing Methods 0.000 claims description 7
- 230000002706 hydrostatic effect Effects 0.000 claims description 6
- 238000001228 spectrum Methods 0.000 claims description 6
- 230000000007 visual effect Effects 0.000 claims description 6
- 238000001914 filtration Methods 0.000 claims description 5
- 238000013528 artificial neural network Methods 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 claims description 3
- 238000006243 chemical reaction Methods 0.000 claims description 3
- 239000003086 colorant Substances 0.000 claims description 3
- 238000009826 distribution Methods 0.000 claims description 3
- 238000005315 distribution function Methods 0.000 claims description 3
- 230000002708 enhancing effect Effects 0.000 claims description 3
- 238000000605 extraction Methods 0.000 claims description 3
- 238000011010 flushing procedure Methods 0.000 claims description 3
- 230000006870 function Effects 0.000 claims description 3
- 230000007246 mechanism Effects 0.000 claims description 3
- 238000012545 processing Methods 0.000 claims description 3
- 230000002035 prolonged effect Effects 0.000 claims description 3
- 238000003384 imaging method Methods 0.000 abstract description 3
- 238000005457 optimization Methods 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 11
- 206010039203 Road traffic accident Diseases 0.000 description 4
- 230000015572 biosynthetic process Effects 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 238000003786 synthesis reaction Methods 0.000 description 4
- 230000009471 action Effects 0.000 description 3
- 230000018109 developmental process Effects 0.000 description 3
- 230000006399 behavior Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000002834 transmittance Methods 0.000 description 2
- 230000008033 biological extinction Effects 0.000 description 1
- 238000005336 cracking Methods 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000003760 hair shine Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 230000000737 periodic effect Effects 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 239000000779 smoke Substances 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 230000003313 weakening effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
-
- 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
- G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
- G01S13/88—Radar or analogous systems specially adapted for specific applications
- G01S13/91—Radar or analogous systems specially adapted for specific applications for traffic control
- G01S13/92—Radar or analogous systems specially adapted for specific applications for traffic control for velocity measurement
-
- 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
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- 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
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/41—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/46—Extracting features or characteristics from the video content, e.g. video fingerprints, representative shots or key frames
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B17/00—Fire alarms; Alarms responsive to explosion
- G08B17/12—Actuation by presence of radiation or particles, e.g. of infrared radiation or of ions
- G08B17/125—Actuation by presence of radiation or particles, e.g. of infrared radiation or of ions by using a video camera to detect fire or smoke
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B21/00—Alarms responsive to a single specified undesired or abnormal condition and not otherwise provided for
- G08B21/02—Alarms for ensuring the safety of persons
- G08B21/10—Alarms for ensuring the safety of persons responsive to calamitous events, e.g. tornados or earthquakes
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B25/00—Alarm systems in which the location of the alarm condition is signalled to a central station, e.g. fire or police telegraphic systems
- G08B25/01—Alarm systems in which the location of the alarm condition is signalled to a central station, e.g. fire or police telegraphic systems characterised by the transmission medium
- G08B25/10—Alarm systems in which the location of the alarm condition is signalled to a central station, e.g. fire or police telegraphic systems characterised by the transmission medium using wireless transmission systems
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B7/00—Signalling systems according to more than one of groups G08B3/00 - G08B6/00; Personal calling systems according to more than one of groups G08B3/00 - G08B6/00
- G08B7/06—Signalling systems according to more than one of groups G08B3/00 - G08B6/00; Personal calling systems according to more than one of groups G08B3/00 - G08B6/00 using electric transmission, e.g. involving audible and visible signalling through the use of sound and light sources
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/052—Detecting movement of traffic to be counted or controlled with provision for determining speed or overspeed
- G08G1/054—Detecting movement of traffic to be counted or controlled with provision for determining speed or overspeed photographing overspeeding vehicles
Abstract
The invention relates to the technical field of monitoring platforms, and discloses a 5G-based AI intelligent highway tunnel comprehensive monitoring platform, which comprises the following components: the tunnel congestion monitoring module is used for monitoring the congestion details of the tunnel in real time and displaying state information; the structure monitoring module is used for lifting out falling rocks in a tunnel hole, sliding down a slope at a hole opening, water rushing in the hole, falling blocks and collapse of lining concrete, rapid deformation and development of a lining structure or a road surface in the tunnel in a short time, falling of a cable bridge and rupture of a fire-fighting pipeline, and disasters endangering driving safety. The invention carries out deep optimization based on technologies such as 5G, wireless ad hoc network, wireless sensing, deep learning, image recognition, polarized light imaging and the like, and improves the safety and operation management level of the highway tunnel.
Description
Technical Field
The invention relates to the field of monitoring platforms, in particular to a 5G-based AI intelligent highway tunnel comprehensive monitoring platform.
Background
With the continuous deep urban construction and the rapid increase of the automobile conservation quantity in China, the traffic problems such as urban congestion, traffic accidents and the like still stand out. Road tunnels are more prone to traffic accidents due to their special environment. Aiming at highway tunnel operation maintenance, traffic accident handling and violation check, the key problems to be solved at the present stage include:
(1) Traffic jam is often caused by vehicle faults, accidents, large traffic flow and the like in a highway tunnel, the on-site information of the tunnel can be obtained only by manually judging through video monitoring or dialing a help call by an accident vehicle driver, and then the rear vehicle is informed of slowing down through an electronic screen, so that the on-site information is obtained and early warning measures are carried out more slowly. During this time, there is a high possibility that a secondary or even more serious accident may be caused by the rear vehicle not being noticed or driving faster, or the like. Therefore, it is critical how to acquire information on site at the first time and early warn in time.
(2) The highway tunnel often exposes various comprehensive typical diseases such as concrete cracking, water leakage, deformation and the like in the operation process. The traditional periodic detection method is non-real-time and has hidden danger. The wired sensing monitoring method is inconvenient to install and deploy, and the construction difficulty of the sensing equipment is high in the follow-up increase. In addition, the structural health evaluation is mainly judged by experience, the workload is large, and the intelligent degree is low.
(3) Tunnel fires are one of the most common and most damaging disasters in tunnels. The reasons for the fire disaster are mainly as follows: a small probability event such as a vehicle circuit failure, an engine burning vehicle, a crash, a tunnel equipment failure, etc. Because the tunnel environment is sealed, escape and rescue are very difficult, once the escape and rescue happen, the escape and rescue not only can cause casualties, vehicle damage and traffic interruption, but also can damage the tunnel structure to influence the service life of the escape and rescue, and even can cause production pause in partial areas to cause order confusion, thereby causing great social influence and economic loss. The occurrence process of the road tunnel fire disaster can be divided into four stages of fire occurrence, fire behavior increase, full development, fire behavior weakening, extinction and the like, and personnel evacuation generally goes through four stages of fire detection, action preparation, escape action and arrival at a safety zone. Therefore, when the fire is just started, the fire is discovered and early warning is sent out at the first time under the condition of small fire, and precious time can be striven for timely extinguishing the fire source and organizing rescue. The existing smoke monitoring and optical fiber temperature monitoring methods can only sense fire under the condition of larger fire, and delay the key time of fire fighting and escape.
(4) The traffic accident is probably caused by the actions of dark light and poor sight in the tunnel, such as overspeed, call receiving and the like, so that the traffic management in the tunnel is particularly important. The light inside the tunnel is dim, or the automobile headlight shines the influence of reflection of light for the image that the camera took is fuzzy unclear, and the misrecognition rate is higher, is difficult to find out and criticize the education to the vehicle that violating regulations, has the potential safety hazard. Therefore, how to clearly obtain the picture of the license plate number in the severe environment is of great importance for accurately identifying the license plate number.
In order to solve the problems, the application provides an AI intelligent highway tunnel comprehensive monitoring platform based on 5G.
Disclosure of Invention
Object of the invention
In order to solve the technical problems in the background technology, the invention provides a 5G-based AI intelligent highway tunnel comprehensive monitoring platform, which is developed based on 5G, wireless ad hoc network, wireless sensing, deep learning, image recognition, polarized light imaging extension optimization, and comprises a plurality of systems for detecting the transient speed of a tunnel vehicle, monitoring and early warning the health of a tunnel structure, monitoring and early warning a tunnel fire disaster, and remotely acquiring evidence from illegal vehicles in the tunnel without intervention, and a 5G+AI intelligent highway tunnel comprehensive monitoring system platform is developed based on the subsystems, so that the safety and operation management level of a highway tunnel are improved.
(II) technical scheme
In order to solve the problems, the invention provides a 5G-based AI intelligent highway tunnel comprehensive monitoring platform, which comprises:
the tunnel congestion monitoring module is used for monitoring the congestion details of the tunnel in real time and displaying state information;
the structure monitoring module is used for lifting out falling rocks in a tunnel hole, sliding down a slope at a hole opening side, water flushing in the hole, falling blocks and collapse of lining concrete, rapid deformation and development of a lining structure or a road surface in the tunnel in a short time, falling of a cable bridge and rupture of a fire-fighting pipeline, and disasters endangering driving safety;
the fire image monitoring module is used for monitoring the fire information in the tunnel in real time; capturing a flame image at the initial stage of a fire disaster through video shooting, preprocessing the image, dividing the area, then performing deep learning on image data by using a convolutional neural network, extracting flame characteristics, and realizing accurate identification of the tunnel fire disaster image;
the vehicle violation detection module is used for monitoring the vehicle violation in real time;
the system comprises a tunnel congestion monitoring module, a structure monitoring module, a fire image monitoring module and a vehicle violation detection module, wherein the monitored information is transmitted to a cloud server through a 5G communication module, and the cloud server is in communication connection with platform software.
Preferably, the tunnel congestion monitoring module includes:
the audible and visual alarm is used for alarming the congestion of the tunnel;
the speed measuring radar is used for monitoring the speed of the vehicle;
the electronic warning screen is used for displaying information in the tunnel;
and a flat mesh network is formed among the audible and visual alarm, the speed measuring radar and the electronic warning screen by a Mist mesh wireless ad hoc network technology.
Preferably, the structure monitoring module includes:
a hydrostatic level, inclinometer, strain gauge and crack gauge; and a flat mesh network is formed among the hydrostatic level, the inclinometer, the strain gauge and the crack gauge independently by a Mist mesh wireless ad hoc network technology.
Preferably, the fire image monitoring module includes:
and a video camera.
Preferably, the vehicle violation detection module includes:
speed measuring radar and a camera.
Preferably, the platform software includes:
the system comprises a vehicle transient speed detection early warning system, a structural health monitoring early warning system, a tunnel fire monitoring early warning subsystem and an illegal vehicle remote intervention digital evidence obtaining and judging subsystem; the vehicle transient speed detection and early warning system, the structural health monitoring and early warning system, the tunnel fire monitoring and early warning subsystem and the illegal vehicle remote intervention digital evidence obtaining and judging subsystem all comprise an intelligent early warning module, an information inquiry module, a historical data module, an accident state module and a user management module.
In an alternative embodiment, the lens of the camera is divided into:
a linear polarizer, which is responsible for filtering out ground reflections;
the rotating mechanism is responsible for adjusting the angle of the transmission axis of the linear polaroid to make the linear polaroid more effectively filter ground reflection:
a lens array responsible for focusing the received light onto the photosensitive array of the camera.
It should be noted that, the core structure of the camera lens with high light filtering is a linear polarizer, and the lens can filter strong ground reflection by properly adjusting the transmission axis angle of the linear polarizer, and only allows other normal light to pass through the lens.
Images obtained by the camera: the incident light will first pass through a sheet of HDR filters that direct the incident light to four channels of different light transmission rates, the four channels having 100%, 75%, 50%, 25% light transmission rates, respectively, before entering the photosensitive array of the camera. The photosensitive array of the camera captures images of four channels simultaneously, and simultaneously based on the maximum unsaturated image synthesis theory proposed by Song Zhang, the four images are synthesized by using the built-in processor of the monitoring camera, so that a clear HDR image which can be directly used for the traditional license plate recognition algorithm is finally obtained.
In an alternative embodiment, the method for implementing fire identification by the fire image monitoring module is as follows:
s1, enhancing brightness information of an image
The fire disaster image is processed by adopting a gray balance method, a pair of images with relatively uniform gray scale distribution probability are generated by converting an accumulated distribution function into a histogram, the number of pixel points on each gray scale of the image after gray balance is relatively balanced, the height of each gray scale of the corresponding gray scale histogram is relatively even, the image after gray balance has larger information quantity, and the color information of flame in the fire disaster image is a very important characteristic, the transformation is performed in an HSV color space which consists of three level information of Hue (Hue), saturation (Saturation) and brightness (Value); the hexagonal boundary represents hue, represents the position of the spectrum color, the parameter is represented by angle, and three colors of RGB are separated by 120 degrees respectively; the horizontal axis represents saturation, the value of which is from 0 to 1, and the saturation is higher as the value is larger; the brightness is measured along the vertical axis, the value of the brightness is also from 0 to 1, and the brightness is not related to the light intensity; in the tunnel, the camera is influenced by light irradiation and shadow, gray balance processing is carried out on a brightness channel, namely a V channel, of an HSV color space, and the contrast ratio of an image is enhanced;
s2, after preprocessing, segmenting a fire disaster area in the image, sending the area into a network for training, if the undivided image is directly trained, a large amount of non-fire disaster characteristic information can be learned, the training time is prolonged, the recognition work is not great, and in order to accurately segment the fire disaster area, a K adjacent algorithm is adopted for segmenting the fire disaster area
Preferably, the K-nearest neighbor algorithm is as follows:
a sample data set exists, namely a training sample set is also called, each data in the sample set corresponds to one label, namely, the corresponding relation between each data in the sample set and the class to which each data belongs is known, in a K-nearest neighbor algorithm, new data are input, the distance between the characteristics of each data and each data characteristic in the sample set is calculated respectively, then the classification label corresponding to K data with the smallest distance is selected, and finally the label with the largest occurrence frequency in the K labels is selected to be used as the classification label of the new data; and (3) optimizing a K-neighbor algorithm, clustering flame and non-flame data by using a K-median algorithm, then respectively applying the K-neighbor algorithm to two types of samples, calculating the similarity between the pixels to be classified and the pixels of each known class by using a similarity function, and judging the class to which the pixels belong according to the similarity.
The suspected flame area is segmented by utilizing the color features, and then the partial images are used as training sets to be sent to a CNN network for training, so that feature extraction can be more specifically carried out, and the recognition rate of fire images is effectively improved.
The fire image recognition algorithm flow based on the convolutional neural network comprises the following steps:
the convolution neural network based on the maximum pooling mode is adopted, and the calculation formula is as follows:
wherein F is a characteristic spectrum, F ij The pixel value at the position (i, j) of the characteristic map, the size of the pooling domain is s×s, and the pooling result obtained finally is denoted by P.
The technical scheme of the invention has the following beneficial technical effects:
the system is based on the technical improvements of 5G, wireless ad hoc network, wireless sensing, deep learning, image recognition, polarized light imaging and the like, and comprises a plurality of systems for detecting the transient speed of a tunnel vehicle, monitoring and early warning the health of a tunnel structure, monitoring and early warning the fire disaster of the tunnel, and judging the remote non-intervention digital evidence taking of illegal vehicles in the tunnel, and a 5G+AI intelligent highway tunnel comprehensive monitoring system platform is developed based on the subsystems, so that the safety and the operation management level of a highway tunnel are improved;
the on-site video/picture data in the tunnel is collected in real time through the high-definition camera, the on-site manual data is transmitted at high speed and low delay through 5G broadband communication, and under the condition that a flame just happens, the fire is identified at the first time through image identification and an alarm signal is sent out, so that precious rescue time is striven for emergency personnel.
Drawings
Fig. 1 is a schematic structural diagram of a 5G-based AI intelligent highway tunnel comprehensive monitoring platform according to the present invention.
Fig. 2 is a schematic diagram of a part of a structure of the 5G-based AI intelligent highway tunnel comprehensive monitoring platform according to the present invention.
Fig. 3 is a schematic diagram of a filtering lens structure of a camera in the 5G-based AI intelligent highway tunnel comprehensive monitoring platform.
Fig. 4 is a schematic structural diagram of an HSV color space model in the 5G-based AI intelligent highway tunnel comprehensive monitoring platform.
Fig. 5 is a flowchart of image preprocessing in the 5G-based AI intelligent highway tunnel comprehensive monitoring platform.
Fig. 6 is a schematic diagram of a decision process of a K nearest neighbor algorithm in the 5G-based AI intelligent highway tunnel comprehensive monitoring platform.
Fig. 7 is a flowchart of an algorithm in the 5G-based AI intelligent highway tunnel comprehensive monitoring platform provided by the invention.
Fig. 8 is a flowchart of a fire image recognition algorithm based on a convolutional neural network in the 5G-based AI intelligent highway tunnel comprehensive monitoring platform.
Fig. 9 is a diagram showing the result from 4×4 feature mapping to maximum pooling in the 5G-based AI intelligent highway tunnel comprehensive monitoring platform according to the present invention.
Fig. 10 is a schematic structural diagram of an HDR filter in a 5G-based AI intelligent highway tunnel comprehensive monitoring platform according to the present invention.
Fig. 11 is a schematic structural diagram of a filter of the 5G-based AI intelligent highway tunnel comprehensive monitoring platform according to the present invention, wherein each four pixels in each group are numbered A, B, C, D.
Detailed Description
The present invention will be further described in detail with reference to the accompanying drawings 1 to 11 in conjunction with the detailed description of the invention, in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the description is only illustrative and is not intended to limit the scope of the invention. In addition, in the following description, descriptions of well-known structures and techniques are omitted so as not to unnecessarily obscure the present invention.
The invention provides a 5G-based AI intelligent highway tunnel comprehensive monitoring platform, which comprises the following steps:
the tunnel congestion monitoring module is used for monitoring the congestion details of the tunnel in real time and displaying state information;
the structure monitoring module is used for lifting out falling rocks in a tunnel hole, sliding down a slope at a hole opening side, water flushing in the hole, falling blocks and collapse of lining concrete, rapid deformation and development of a lining structure or a road surface in the tunnel in a short time, falling of a cable bridge and rupture of a fire-fighting pipeline, and disasters endangering driving safety;
the fire image monitoring module is used for monitoring the fire information in the tunnel in real time; capturing a flame image at the initial stage of a fire disaster through video shooting, preprocessing the image, dividing the area, then performing deep learning on image data by using a convolutional neural network, extracting flame characteristics, and realizing accurate identification of the tunnel fire disaster image;
the vehicle violation detection module is used for monitoring the vehicle violation in real time;
the system comprises a tunnel congestion monitoring module, a structure monitoring module, a fire image monitoring module and a vehicle violation detection module, wherein the monitored information is transmitted to a cloud server through a 5G communication module, and the cloud server is in communication connection with platform software.
In an alternative embodiment, the tunnel congestion monitoring module includes:
the audible and visual alarm is used for alarming the congestion of the tunnel;
the speed measuring radar is used for monitoring the speed of the vehicle;
the electronic warning screen is used for displaying information in the tunnel;
and a flat mesh network is formed among the audible and visual alarm, the speed measuring radar and the electronic warning screen by a Mist mesh wireless ad hoc network technology.
The monitoring of structural state parameters such as tunnel deformation, cracks, water leakage and the like is realized, and simultaneously, the sensors such as temperature, humidity and gas are utilized to monitor the internal environment of the tunnel in real time. The sensing device integrates a wireless ad hoc network communication module, all sensors are connected and communicated through self-networking, data are converged to the intelligent gateway uniformly and then are transmitted in a wireless communication remote way through 5G, and the data are uploaded to the server.
In an alternative embodiment, the structure monitoring module includes:
a hydrostatic level, inclinometer, strain gauge and crack gauge; and a flat mesh network is formed among the hydrostatic level, the inclinometer, the strain gauge and the crack gauge independently by a Mist mesh wireless ad hoc network technology.
The fire image monitoring module comprises: and a video camera. The vehicle violation detection module includes: speed measuring radar and a camera.
Because the tunnel structure is complex, the arch and bend structures of the tunnel structure easily enable electromagnetic wave signals of wireless communication to generate multipath propagation and attenuation, and the transmission stability is affected. The Mist mesh wireless ad hoc network has self-adaption and self-healing capabilities, can automatically select the most communication link, can reconstruct a new network after the nodes are damaged, and can furthest ensure the reliability of sensing communication.
The Mist mesh wireless ad hoc network technology is a planar mesh network technology, has no primary and secondary parts, is simpler in network architecture, enables nodes to form a flat mesh network independently, does not depend on a gateway or any special node to form a network, and can communicate with any other node and any gateway. When nodes form a planar mesh network, clusters are formed within the network to coordinate the communication cycles of the individual nodes and facilitate communication between the nodes.
In an alternative embodiment, the platform software includes:
the system comprises a vehicle transient speed detection early warning system, a structural health monitoring early warning system, a tunnel fire monitoring early warning subsystem and an illegal vehicle remote intervention digital evidence obtaining and judging subsystem; the vehicle transient speed detection and early warning system, the structural health monitoring and early warning system, the tunnel fire monitoring and early warning subsystem and the illegal vehicle remote intervention digital evidence obtaining and judging subsystem all comprise an intelligent early warning module, an information inquiry module, a historical data module, an accident state module and a user management module.
It should be noted that, the system in this embodiment detects and analyzes the detected data respectively to make corresponding decisions, and by providing an intelligent early warning module, an information query module, a historical data module, an accident state module and a user management module, the system is beneficial to subsequent query and intelligent early warning.
In an alternative embodiment, the lens of the camera is divided into:
a linear polarizer, which is responsible for filtering out ground reflections;
the rotating mechanism is responsible for adjusting the angle of the transmission axis of the linear polaroid to make the linear polaroid more effectively filter ground reflection:
a lens array responsible for focusing the received light onto the photosensitive array of the camera.
The dynamic range of the monitoring camera is greatly expanded by specially preparing a piece of filter with different light passing quantity pixel by pixel in front of the photosensitive array of the camera, so that the purpose of enabling the monitoring camera to clearly capture darker license plates in a strong light environment is achieved.
The camera photosensitive array HDR filter specifically includes an image filter module and an HDR image synthesis module, as shown in fig. 10:
before entering the photosensitive array of the monitoring camera, the incident light firstly passes through an HDR filter, and the HDR filter divides the incident light into four channels with different light transmittance, wherein the light transmittance of the four channels is 100%, 75%, 50% and 25% respectively. The photosensitive array of the monitoring camera captures four-channel images simultaneously, and the four images are synthesized by using the built-in processor of the camera based on the maximum unsaturated image synthesis theory proposed by Song Zhang, so that a clear HDR image which can be directly used for the traditional license plate recognition algorithm is finally obtained.
The filter consists of 2 layers of linear polaroids, every four pixels on the filter are in a group, and the four pixels in each group are respectively numbered A, B, C, D, as shown in fig. 11. The included angles of the transmission axes of the two polarizers at the corresponding positions on the pixel A, B, C, D are respectively 0 degree, 41.40 degrees, 60.00 degrees and 75.52 degrees.
According to Ma Lvsai law, the light intensity ratios of the light entering the camera photosensitive array at the four pixel positions of A, B, C, D are respectively 0%, 25%, 50% and 75%. Then, all A, B, C, D pixels are recombined into a new picture respectively, and the four pictures are recombined into a new HDR picture which can be used for license plate recognition through an HDR synthesis algorithm.
In an alternative embodiment, the method for implementing fire identification by the fire image monitoring module is as follows:
s1, enhancing brightness information of an image
The fire image is processed by adopting a gray balance method, a pair of images with relatively uniform gray level distribution probability are generated by converting an accumulated distribution function into a histogram, the number of pixel points on each gray level of the image after gray balance is relatively balanced, the gray level of each gray level of the corresponding gray level histogram is relatively even, the image after gray balance has larger information quantity, the color information of flame in the fire image is a very important characteristic, if the original color image is directly subjected to gray level conversion, the original information of flame pixels can be destroyed, the conversion is performed in an HSV color space, and the HSV color space consists of three level information of Hue (Hue), saturation (Saturation) and brightness (Value) (the model structure is shown in figure 4); the hexagonal boundary represents hue, represents the position of the spectrum color, the parameter is represented by angle, and three colors of RGB are separated by 120 degrees respectively; the horizontal axis represents saturation, the value of which is from 0 to 1, and the saturation is higher as the value is larger; the brightness is measured along the vertical axis, the value of the brightness is also from 0 to 1, and the brightness is not related to the light intensity; in tunnels, cameras are affected by light irradiation and shadows, and the photographed pictures have the problem of low brightness. In order to solve the problem, the subsequent fire region segmentation is convenient, gray balance processing is carried out on a brightness channel, namely a V channel, of the HSV color space, and the contrast ratio of the image is enhanced;
s2, referring to fig. 5, in the tunnel, the camera of the camera is affected by light irradiation, shadow, and the like, and the shot photo has a problem of low brightness. In order to solve the problem, after preprocessing, a fire region in an image is segmented, the region is sent to a network for training, if the undivided image is directly trained, a large amount of non-fire characteristic information can be learned, the training time is prolonged, the recognition work is not very useful, and in order to accurately segment the fire region, a K adjacent algorithm is adopted to segment the fire region.
In an alternative embodiment, the K-nearest neighbor algorithm is as follows:
a sample data set exists, namely a training sample set is also called, each data in the sample set corresponds to one label, namely, the corresponding relation between each data in the sample set and the class to which each data belongs is known, in a K-nearest neighbor algorithm, new data are input, the distance between the characteristics of each data and each data characteristic in the sample set is calculated respectively, then the classification label corresponding to K data with the smallest distance is selected, and finally the label with the largest occurrence frequency in the K labels is selected to be used as the classification label of the new data; the decision diagram of the K-nearest neighbor algorithm is shown in fig. 6: and (3) optimizing a K-neighbor algorithm, clustering flame and non-flame data by using a K-median algorithm, then respectively applying the K-neighbor algorithm to two types of samples, calculating the similarity between the pixels to be classified and the pixels of each known class by using a similarity function, and judging the class to which the pixels belong according to the similarity. The algorithm flow is shown in figure 7.
The suspected flame area is segmented by utilizing the color features, and then the partial images are used as training sets to be sent to a CNN network for training, so that feature extraction can be more specifically carried out, and the recognition rate of fire images is effectively improved. The flow is as in fig. 8.
The fire image recognition algorithm flow based on the convolutional neural network comprises the following steps:
the convolution neural network based on the maximum pooling mode is adopted, and the calculation formula is as follows:
wherein F is a characteristic spectrum, F ij The pixel value at the position (i, j) of the characteristic map, the size of the pooling domain is s×s, and the pooling result obtained finally is denoted by P. The process of maximum pooling is shown in fig. 9 (the leftmost diagram in fig. 9 is a 4 x 4 feature map, and the right is the maximum pooling result diagram).
It is to be understood that the above-described embodiments of the present invention are merely illustrative of or explanation of the principles of the present invention and are in no way limiting of the invention. Accordingly, any modification, equivalent replacement, improvement, etc. made without departing from the spirit and scope of the present invention should be included in the scope of the present invention. Furthermore, the appended claims are intended to cover all such changes and modifications that fall within the scope and boundary of the appended claims, or equivalents of such scope and boundary.
Claims (10)
1. AI wisdom highway tunnel synthesizes monitoring platform based on 5G, its characterized in that includes:
the tunnel congestion monitoring module is used for monitoring the congestion details of the tunnel in real time and displaying state information;
the structure monitoring module is used for lifting out falling rocks in a tunnel hole, sliding down a slope at a hole opening side, water flushing in the hole, falling blocks and collapse of lining concrete, rapid deformation and development of a lining structure or a road surface in the tunnel in a short time, falling of a cable bridge and rupture of a fire-fighting pipeline, and disasters endangering driving safety;
the fire image monitoring module is used for monitoring the fire information in the tunnel in real time; capturing a flame image at the initial stage of a fire disaster through video shooting, preprocessing the image, dividing the area, then performing deep learning on image data by using a convolutional neural network, extracting flame characteristics, and realizing accurate identification of the tunnel fire disaster image;
the vehicle violation detection module is used for monitoring the vehicle violation in real time;
the system comprises a tunnel congestion monitoring module, a structure monitoring module, a fire image monitoring module and a vehicle violation detection module, wherein the monitored information is transmitted to a cloud server through a 5G communication module, and the cloud server is in communication connection with platform software.
2. The 5G-based AI intelligent highway tunnel comprehensive monitoring platform of claim 1, wherein the tunnel congestion monitoring module comprises:
the audible and visual alarm is used for alarming the congestion of the tunnel;
the speed measuring radar is used for monitoring the speed of the vehicle;
the electronic warning screen is used for displaying information in the tunnel;
and a flat mesh network is formed among the audible and visual alarm, the speed measuring radar and the electronic warning screen by a Mist mesh wireless ad hoc network technology.
3. The 5G-based AI intelligent highway tunnel comprehensive monitoring platform of claim 2, wherein the structural monitoring module comprises:
a hydrostatic level, inclinometer, strain gauge and crack gauge; and a flat mesh network is formed among the hydrostatic level, the inclinometer, the strain gauge and the crack gauge independently by a Mist mesh wireless ad hoc network technology.
4. The 5G-based AI intelligent highway tunnel comprehensive monitoring platform of claim 3, wherein the fire image monitoring module comprises:
and a video camera.
5. The 5G-based AI intelligent highway tunnel comprehensive monitoring platform of claim 4, wherein the vehicle violation detection module comprises:
speed measuring radar and a camera.
6. The 5G-based AI intelligent highway tunnel comprehensive monitoring platform of claim 5, wherein the platform software comprises:
the system comprises a vehicle transient speed detection early warning system, a structural health monitoring early warning system, a tunnel fire monitoring early warning subsystem and an illegal vehicle remote intervention digital evidence obtaining and judging subsystem; the vehicle transient speed detection and early warning system, the structural health monitoring and early warning system, the tunnel fire monitoring and early warning subsystem and the illegal vehicle remote intervention digital evidence obtaining and judging subsystem all comprise an intelligent early warning module, an information inquiry module, a historical data module, an accident state module and a user management module.
7. The 5G-based AI intelligent highway tunnel comprehensive monitoring platform of claim 5, wherein the lens of the camera is divided into:
a linear polarizer, which is responsible for filtering out ground reflections;
the rotating mechanism is responsible for adjusting the angle of the transmission axis of the linear polaroid to make the linear polaroid more effectively filter ground reflection:
a lens array responsible for focusing the received light onto the photosensitive array of the camera.
8. The 5G-based AI intelligent highway tunnel comprehensive monitoring platform as claimed in claim 7, wherein the method for realizing fire disaster identification by the fire disaster image monitoring module is as follows:
s1, enhancing brightness information of an image
The fire disaster image is processed by adopting a gray balance method, a pair of images with relatively uniform gray level distribution probability are generated by converting an accumulated distribution function into a histogram, the number of pixel points on each gray level of the image after gray balance is relatively balanced, the gray level of each gray level of the corresponding gray level histogram is relatively even, the image after gray balance has larger information quantity, and the color information of flame in the fire disaster image is a very important characteristic, the conversion is performed in an HSV color space which consists of three layers of information of tone, saturation and brightness; the hexagonal boundary represents hue, represents the position of the spectrum color, the parameter is represented by angle, and three colors of RGB are separated by 120 degrees respectively; the horizontal axis represents saturation, the value of which is from 0 to 1, and the saturation is higher as the value is larger; the brightness is measured along the vertical axis, the value of the brightness is also from 0 to 1, and the brightness is not related to the light intensity; in the tunnel, the camera is influenced by light irradiation and shadow, gray balance processing is carried out on a brightness channel, namely a V channel, of an HSV color space, and the contrast ratio of an image is enhanced;
s2, after preprocessing, segmenting a fire disaster area in the image, sending the area into a network for training, if the undivided image is directly trained, a large amount of non-fire disaster characteristic information can be learned, the training time is prolonged, the recognition work is not great, and in order to accurately segment the fire disaster area, a K adjacent algorithm is adopted for segmenting the fire disaster area
9. The 5G-based AI intelligent highway tunnel comprehensive monitoring platform of claim 8, wherein the step of the K-nearest neighbor algorithm is as follows:
a sample data set exists, namely a training sample set is also called, each data in the sample set corresponds to one label, namely, the corresponding relation between each data in the sample set and the class to which each data belongs is known, in a K-nearest neighbor algorithm, new data are input, the distance between the characteristics of each data and each data characteristic in the sample set is calculated respectively, then the classification label corresponding to K data with the smallest distance is selected, and finally the label with the largest occurrence frequency in the K labels is selected to be used as the classification label of the new data; optimizing a K-neighbor algorithm, clustering flame and non-flame data by using a K-median algorithm, then respectively applying the K-neighbor algorithm to two types of samples, calculating the similarity between the pixels to be classified and the pixels of each known class by using a similarity function, and judging the class to which the pixels belong according to the similarity;
the suspected flame area is segmented by utilizing the color characteristics, and then the partial pictures are used as training sets to be sent to a CNN network for training, so that characteristic extraction can be more specifically carried out, and the recognition rate of fire images is improved;
the fire image recognition algorithm flow based on the convolutional neural network comprises the following steps:
the convolution neural network based on the maximum pooling mode is adopted, and the calculation formula is as follows:
wherein F is a characteristic spectrum, F ij The pixel value at the position (i, j) of the characteristic map, the size of the pooling domain is s×s, and the pooling result obtained finally is denoted by P.
10. The 5G-based AI intelligent highway tunnel comprehensive monitoring platform of claim 7, wherein the camera lens comprises an HDR filter, the HDR filter comprises 2 layers of linear polarizers, each four pixels on the filter are in a group, and each four pixels in each group are respectively numbered A, B, C, D; the included angles of the transmission axes of the two polarizers at the corresponding positions on the pixel A, B, C, D are respectively 0 degree, 41.40 degrees, 60.00 degrees and 75.52 degrees.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211615885.7A CN116434533A (en) | 2022-12-15 | 2022-12-15 | AI wisdom highway tunnel synthesizes monitoring platform based on 5G |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211615885.7A CN116434533A (en) | 2022-12-15 | 2022-12-15 | AI wisdom highway tunnel synthesizes monitoring platform based on 5G |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116434533A true CN116434533A (en) | 2023-07-14 |
Family
ID=87080314
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211615885.7A Pending CN116434533A (en) | 2022-12-15 | 2022-12-15 | AI wisdom highway tunnel synthesizes monitoring platform based on 5G |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116434533A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117288269A (en) * | 2023-11-27 | 2023-12-26 | 四川交通职业技术学院 | Intelligent monitoring system and method for urban road landscape quality |
| CN117367411A (en) * | 2023-12-07 | 2024-01-09 | 深圳市拓安科技有限公司 | Safe navigation method and system for AI (advanced technology attachment) internet of things tunnel |
| CN117409341A (en) * | 2023-12-15 | 2024-01-16 | 深圳市光明顶技术有限公司 | Unmanned aerial vehicle illumination-based image analysis method and system |
-
2022
- 2022-12-15 CN CN202211615885.7A patent/CN116434533A/en active Pending
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117288269A (en) * | 2023-11-27 | 2023-12-26 | 四川交通职业技术学院 | Intelligent monitoring system and method for urban road landscape quality |
| CN117288269B (en) * | 2023-11-27 | 2024-01-30 | 四川交通职业技术学院 | Intelligent monitoring system and method for urban road landscape quality |
| CN117367411A (en) * | 2023-12-07 | 2024-01-09 | 深圳市拓安科技有限公司 | Safe navigation method and system for AI (advanced technology attachment) internet of things tunnel |
| CN117367411B (en) * | 2023-12-07 | 2024-04-16 | 深圳市拓安科技有限公司 | Safe navigation method and system for AI (advanced technology attachment) internet of things tunnel |
| CN117409341A (en) * | 2023-12-15 | 2024-01-16 | 深圳市光明顶技术有限公司 | Unmanned aerial vehicle illumination-based image analysis method and system |
| CN117409341B (en) * | 2023-12-15 | 2024-02-13 | 深圳市光明顶技术有限公司 | Unmanned aerial vehicle illumination-based image analysis method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN116434533A (en) | AI wisdom highway tunnel synthesizes monitoring platform based on 5G | |
| CN101334924B (en) | Fire hazard probe system and its fire hazard detection method | |
| KR101822924B1 (en) | Image based system, method, and program for detecting fire | |
| CN106650584B (en) | Flame detecting method and system | |
| CN104599427B (en) | A kind of intelligent image type fire alarm system for vcehicular tunnel | |
| CN108389359B (en) | Deep learning-based urban fire alarm method | |
| KR101095528B1 (en) | An outomatic sensing system for traffic accident and method thereof | |
| CN201936415U (en) | Automatic forest fire identification and alarm system | |
| CN107437318B (en) | Visible light intelligent recognition algorithm | |
| CN103106766A (en) | Forest fire identification method and forest fire identification system | |
| CN101908142A (en) | Feature analysis-based video flame detecting method | |
| CN210899299U (en) | Tunnel monitoring system | |
| CN103324910A (en) | Fire alarming method based on video detection | |
| CN112785809B (en) | Fire re-ignition prediction method and system based on AI image recognition | |
| CN111383429A (en) | Method, system, device and storage medium for detecting dress of workers in construction site | |
| CN201091014Y (en) | Fire detecting device | |
| CN110517441A (en) | Based on the frame-embedded smog of deep learning and flame video alarming system and method | |
| CN114724330A (en) | Implementation method of self-adaptive mode switching multi-channel video fire real-time alarm system | |
| CN109410497A (en) | A kind of monitoring of bridge opening space safety and alarm system based on deep learning | |
| CN111476964A (en) | Remote forest fire prevention monitoring system and method | |
| CN111783700B (en) | Automatic recognition and early warning method and system for pavement foreign matters | |
| CN116453278A (en) | Intrusion monitoring method combining deep learning intelligent detection and optical fiber vibration sensing | |
| KR20220072783A (en) | System and method for real-time flood detecting, and monitoring using CCTV image, and a recording medium recording a computer readable program for executing the method | |
| CN105046223A (en) | Device for detecting severity of ''black-hole effect'' at tunnel entrance and method thereof | |
| CN115457331A (en) | Intelligent inspection method and system for construction site |
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 |