WO2023221425A1 - 一种用于船舶过闸安全检测方法 - Google Patents

一种用于船舶过闸安全检测方法 Download PDF

Info

Publication number
WO2023221425A1
WO2023221425A1 PCT/CN2022/131860 CN2022131860W WO2023221425A1 WO 2023221425 A1 WO2023221425 A1 WO 2023221425A1 CN 2022131860 W CN2022131860 W CN 2022131860W WO 2023221425 A1 WO2023221425 A1 WO 2023221425A1
Authority
WO
WIPO (PCT)
Prior art keywords
ship
lock
detection
information
gate
Prior art date
Application number
PCT/CN2022/131860
Other languages
English (en)
French (fr)
Inventor
朱勇
杨喜
何兴华
张日民
王伟
张勉
李剑
周宇
Original Assignee
苏交科集团股份有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 苏交科集团股份有限公司 filed Critical 苏交科集团股份有限公司
Publication of WO2023221425A1 publication Critical patent/WO2023221425A1/zh

Links

Images

Classifications

    • E—FIXED CONSTRUCTIONS
    • E02—HYDRAULIC ENGINEERING; FOUNDATIONS; SOIL SHIFTING
    • E02C—SHIP-LIFTING DEVICES OR MECHANISMS
    • E02C1/00—Locks or dry-docks; Shaft locks, i.e. locks of which one front side is formed by a solid wall with an opening in the lower part through which the ships pass
    • E—FIXED CONSTRUCTIONS
    • E02—HYDRAULIC ENGINEERING; FOUNDATIONS; SOIL SHIFTING
    • E02C—SHIP-LIFTING DEVICES OR MECHANISMS
    • E02C1/00—Locks or dry-docks; Shaft locks, i.e. locks of which one front side is formed by a solid wall with an opening in the lower part through which the ships pass
    • E02C1/06—Devices for filling or emptying locks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06K—GRAPHICAL DATA READING; PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS
    • G06K17/00—Methods or arrangements for effecting co-operative working between equipments covered by two or more of main groups G06K1/00 - G06K15/00, e.g. automatic card files incorporating conveying and reading operations
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06K—GRAPHICAL DATA READING; PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS
    • G06K17/00—Methods or arrangements for effecting co-operative working between equipments covered by two or more of main groups G06K1/00 - G06K15/00, e.g. automatic card files incorporating conveying and reading operations
    • G06K17/0022—Methods or arrangements for effecting co-operative working between equipments covered by two or more of main groups G06K1/00 - G06K15/00, e.g. automatic card files incorporating conveying and reading operations arrangements or provisious for transferring data to distant stations, e.g. from a sensing device
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06K—GRAPHICAL DATA READING; PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS
    • G06K17/00—Methods or arrangements for effecting co-operative working between equipments covered by two or more of main groups G06K1/00 - G06K15/00, e.g. automatic card files incorporating conveying and reading operations
    • G06K17/0032—Apparatus for automatic testing and analysing marked record carriers, used for examinations of the multiple choice answer type
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N7/00—Television systems
    • H04N7/18—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast
    • H04N7/181—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast for receiving images from a plurality of remote sources
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
    • Y02A10/00—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE at coastal zones; at river basins
    • Y02A10/30—Flood prevention; Flood or storm water management, e.g. using flood barriers

Definitions

  • the invention relates to the field of information technology, and in particular to a method for safety detection of ships passing through locks.
  • Each link usually has at least one manual position for full-time monitoring.
  • the labor cost is high, the operating efficiency is low, and errors in manual operation and monitoring cannot be avoided. Therefore, in order to take into account the safety of ships passing through the lock, it is necessary to limit the overall speed of the ship passing through the lock, which seriously affects shipping efficiency.
  • Embodiments of the present invention provide a lock safety monitoring system and method for canal shipping, which can improve the digitalization of ships passing through locks, thereby improving the safety and efficiency of ships passing through locks.
  • the lock-waiting link includes at least: the ship's draft over-limit detection sub-link and other lock-waiting sub-links. If the ship's draft does not exceed the limit and has also passed the other sub-links, it will enter the lock-passing link. ;
  • the gate crossing link includes: lock chamber radar identification sub-link, mooring cable safety detection sub-link and pedestrian detection above the gate sub-link. If the ship passes the mooring cable safety detection sub-link and pedestrian detection above the gate, sub-link and other sub-links passing through the lock, it will be determined that the ship has passed the lock safely.
  • the lock-passing safety detection method for canal shipping uses computer data processing and computer network transmission technology to intelligently "collect, analyze, and process" the canal green shipping lock-passing safety information and centrally reflect it to the platform in real time.
  • the monitoring interface and dispatch control system enable ship lock managers to promptly and accurately obtain various safety risks that occur during the ship's passage through the lock, and adopt timely and effective management and processing of daily real-time issues at the lock, thereby eliminating hidden dangers and failures in lock dispatch. closed-loop processing. In this way, taking into account the safety of ships passing through locks, the overall speed of the ship passing through locks can be improved, thereby improving the digitalization of ships passing through locks and further improving the safety and efficiency of ships passing through locks.
  • Figure 1 is a schematic flow chart of a method provided by an embodiment of the present invention.
  • Figures 2 and 3 are schematic diagrams of possible ship lock scene scenarios provided by the embodiment of the present invention.
  • FIGS 4 and 5 are schematic diagrams of possible camera deployment methods provided by embodiments of the present invention.
  • FIGS 6 and 7 are schematic diagrams of specific examples provided by embodiments of the present invention.
  • Figure 8 is an alarm closed-loop processing flow chart provided by an embodiment of the present invention.
  • FIGS 9 and 10 are schematic diagrams of radar deployment methods provided by embodiments of the present invention.
  • An embodiment of the present invention provides a method for safety detection of ships passing through locks, as shown in Figure 1, including:
  • the lock-waiting link includes at least: the ship's draft over-limit detection sub-link and other lock-waiting sub-links. If the ship's draft does not exceed the limit and has also passed the other sub-links, it will enter the lock-passing link. .
  • the gate crossing link includes: lock chamber radar identification sub-link, mooring cable safety detection sub-link and pedestrian detection above the gate sub-link. If the ship passes the mooring cable safety detection sub-link and pedestrian detection above the gate, sub-link and other sub-links passing through the lock, it will be determined that the ship has passed the lock safely.
  • the design idea of this embodiment is to use computer data processing and computer network transmission technology to intelligently "collect, analyze, and process" the canal green shipping lock safety information and centrally reflect it to the platform's real-time monitoring interface and dispatch control system, so that the ship lock Managers can timely and accurately obtain various safety risks that occur when ships pass through the lock, and adopt timely and effective management and processing of daily real-time issues at the lock, so as to deal with hidden dangers and faults in the lock dispatch in a closed-loop manner.
  • the video surveillance above the lock chamber is integrated into a picture of the lock chamber, which can effectively avoid interference from pedestrian bridges and camera angles, and is conducive to real-time understanding of the behavior of ships in the lock chamber, and a more intuitive display of the lock chamber.
  • Ship docking situation when a ship passes through the lock, the video surveillance above the lock chamber is integrated into a picture of the lock chamber, which can effectively avoid interference from pedestrian bridges and camera angles, and is conducive to real-time understanding of the behavior of ships in the lock chamber, and a more intuitive display of the lock chamber.
  • the information collection equipment group includes at least two sets of ship lock photography equipment, which are respectively arranged at the downstream lock head of the ship lock and the upstream lock head of the ship lock.
  • at least two sets of ship lock photography equipment are arranged at the downstream lock head of the ship lock and the upstream lock head of the ship lock.
  • a pole can be erected on the east and west sides of the downstream remote control station (to avoid blind spots caused by too close distance). Capture cameras are installed on the poles to ensure that there are no blind spots to capture the ship.
  • infrared strobe light is used for fill light. Infrared light is invisible to the human eye, which enhances the clear capture effect at night without affecting the crew's driving.
  • a panoramic camera is installed on one side of the pole to capture the entire ship.
  • the camera equipment in each set of ship lock shooting equipment at least includes: a detection camera (also called a ship detection camera), a detail capture camera and a panoramic capture camera, where the shooting field of view of the panoramic capture camera includes Enter the lock room.
  • a detection camera also called a ship detection camera
  • a detail capture camera also called a detail capture camera
  • a panoramic capture camera where the shooting field of view of the panoramic capture camera includes Enter the lock room.
  • the information collection equipment group also includes at least one ship identity verification subsystem.
  • the ship identity verification subsystem includes: main control equipment, fill light, camera equipment (detection camera, detail capture camera, panoramic capture camera) and AIS
  • the receiver, the fill light and the camera device are arranged outside the gate, and the shooting field of view of the camera device each covers a part of the access channel of the gate.
  • the main control equipment includes special main control equipment, which can be arranged in the central computer room of the ship lock.
  • the existing video network link of the ship lock is used between the central computer room and the field equipment.
  • the specific model of the dedicated main control device can be IMX220.
  • one set of video capture equipment can be deployed at the downstream lock head of the XX ship lock (the downstream lock head of the three-line lock) and the upstream lock head of the XX ship lock (the upstream lock head of the three-line lock), for a total of 6 sets, including detection cameras and detailed capture Cameras, panoramic capture cameras and related auxiliary materials, etc., and field equipment are installed through customized brackets on vertical poles; the main control equipment is placed in the central computer room of the ship lock, and the existing video network link of the ship lock is used between the central computer room and the field equipment.
  • capture cameras are installed on the poles to ensure that there are no blind spots to capture the ship.
  • infrared strobe light is used for fill light. Infrared light is invisible to the human eye, which enhances the clear capture effect at night without affecting the crew's driving.
  • a panoramic camera is installed on one side of the pole to detect whether there are ships near the gate.
  • the identity of the ship is identified, including ship name, ship number, ship size, etc., and automatically compared with the ship information in the AIS system database to verify the ship's identity. For ships that are inconsistent with the registered information in the AIS system, the detection results are saved, and early warnings and information queries are performed to complete the verification of the ship's identity. That is, in this embodiment, not only the ship information can be entered, but the ship information can also be automatically verified.
  • the hardware architecture of the ship identity verification system is shown in Figure 5.
  • the hardware equipment mainly includes: main control equipment, fill light, detection camera, detail capture camera, panoramic capture camera and AIS receiver.
  • the field equipment passes through the pole. Customized bracket installation; the main control equipment is placed in the central computer room, and the existing video network link of the ship lock is used between the central computer room and the field equipment.
  • the operation process of the ship identity verification system is shown in Figure 6.
  • a ship passes through the gate, it first intelligently senses and automatically captures photos of the ship through the front-end field equipment, and then intelligently identifies the photos and videos.
  • the identification content includes the ship’s identity, ship size and AIS information, etc., compares the identified information with the information in the AIS system database, automatically alarms ships with abnormal sizes, inconsistent identities, etc., and notifies the staff for review.
  • the system can monitor ships passing through the lock every day, abnormal ships, etc. statistic and analysis.
  • the high-definition capture process shown in Figure 7 can be used.
  • the high-definition capture of the ship relies on the capture camera installed at the head of the lock.
  • the video stream of the ship detection camera is processed. Automatic analysis to accurately detect whether the ship is passing by and the precise position of the ship.
  • the video stream signal is input from a stationary camera.
  • the background modeling of the video stream is performed.
  • the model is first initialized, and then it is judged whether the pixel is a foreground point, and a certain update mechanism is set to update the background model.
  • the foreground points are obtained, and the connected domain analysis method is used to analyze the foreground points, and the obtained connected domain is used as the target candidate area. Based on prior knowledge, non-ship targets in the target candidate area are removed, and finally the ship target is obtained.
  • Multi-camera linkage uses image recognition and processing technology to monitor the gate and at the same time cooperate with the moving target detection algorithm to individually and continuously track and capture detailed feature information of any moving target within the monitoring range.
  • Multi-camera linkage technology uses the ship detection camera to detect the target ship and lock and track it. After confirming that the target is a ship, the detail capture camera and panoramic capture camera are triggered to take pictures. The captured panoramic photos and detailed photos will be presented in chronological order. in the screen.
  • the captured ship photos not only include the overall view of the ship, but also the details of the ship, and are multiple photos of one ship. In the detailed photos of the ship, the ship number, ship type, ship draft, cargo type and other information can be clearly seen.
  • S2 includes: using the camera device in the ship identity verification subsystem to capture the ship outside the gate and entering the channel, obtaining the image information of the ship, and based on the image information Obtain ship identity information.
  • the AIS information of the ship is obtained.
  • the ship identity information at least includes: ship name, ship number and ship type.
  • the detection of whether the identity of the ship is normal includes: using the AIS information to query the ship registration information in the AIS system database, and comparing the queried ship registration information with the ship identity information, if they are consistent This determines the ship's identity as normal.
  • the current technical means of ship identification mainly include: AIS, RFID and video images, etc.
  • RFID requires additional terminal equipment on ships and the construction of a large number of base stations. It is difficult and costly to promote and cannot be applied on a large scale.
  • AIS is an international standard ship automatic identification system, the activation rate of AIS for inland river ships is low. Even if AIS is activated, identity information is often tampered with, and the confidence level is low.
  • the fusion of AIS and image recognition is used for ship identity recognition.
  • the initial accuracy of the ship's identity is no less than 85%.
  • the theoretical accuracy can reach 95%.
  • the automatic review process of ship information specifically involves automatic comparison and review of ship identity information with AIS system and dispatch information data.
  • the way to automatically alarm abnormal ships is that the system will record the number of ships entering/exiting each lock and the identity of the ships entering and exiting the lock.
  • the ship identity and related information obtained by this system are compared with the AIS registration system.
  • the alarm linkage closed-loop processing method can be as follows:
  • the purpose of constructing the ship identity verification system is mainly to verify the identity of ships passing through the lock. This system captures high-definition images of ships through artificial intelligence technology, and automatically recognizes ship name characters based on ship images, and communicates with AIS The registration information is automatically reviewed, and ships with inconsistent identities are alerted and pushed to the staff for processing, thus forming a closed loop of alarm linkage.
  • the ship identity review and alarm linkage process is shown in Figure 8: the ship uses AIS information for self-service gate registration when passing through the lock. After the system confirms the ship's AIS information, it dispatches the ship. The ship's AIS information and dispatch information are stored in the database, while the ship waits to pass through the lock. , after passing through the ship identity verification system when entering the lock, the system first senses the appearance of the ship, triggers the capture camera to take high-definition snapshots of the ship, uses artificial intelligence technology to automatically identify the ship name, automatically compares the identified ship name with the database, and automatically determines the identity of the ship passing through the lock. Check whether the registered identities are consistent. If they are consistent, it will pass through the gate normally.
  • an alarm will be triggered and the alarm data will be pushed to the staff for processing.
  • the staff will confirm through verification and take action on the ship, such as re-registration, credit deduction, ban Pass the gate or pass the gate normally, forming a closed-loop process.
  • the ship draft over-limit detection sub-link includes: after identifying the identity of the ship, obtaining the size information corresponding to the ship. Through the information collection equipment group, the image information of the ship's draft position is obtained. According to the size information and the image information of the ship's draft position, the freeboard position of the ship and the position of the dividing line between the bottom and the water surface are determined, and then the distance between the bollard and the dividing line of the water surface is calculated to determine whether the ship is overloaded.
  • the image information of the ship's draft position includes: the image information of the bollard, the image information of the ship's freeboard position, and the image information of the water surface dividing line position.
  • the distance from the water surface boundary determines whether the ship is overloaded. For example: This embodiment provides a possible detection method for ship draft exceeding the limit.
  • a set of video capture equipment is deployed at the remote control station downstream of the Shiqiao ship lock, including a detection camera, a detail capture camera, a panoramic capture camera and related auxiliary materials. Field equipment is mounted via custom brackets on poles.
  • the remote control station downstream of Shiqiao Ship Lock has added a ship draft over-limit detection system to identify the actual draft of the ship through high-definition snapshots. At the same time, for the identified ships with draft exceeding the limit, the detection results are pushed to the dispatching system, which handles them and provides early warning and information query.
  • the hardware equipment of the ship draft over-limit detection system mainly includes: special main control equipment, strobe lights, detection cameras, detail capture cameras, panoramic capture cameras and AIS receivers, etc.
  • the field equipment is installed through customized brackets on poles; the main control equipment is placed In the central computer room, the existing video network link of the ship lock is used between the central computer room and the field equipment.
  • the process of ship draft over-limit detection includes: when the ship passes the video capture point, it first intelligently senses and automatically captures photos of the ship through the front-end field equipment, and then intelligently identifies the photos and videos.
  • the identification content includes the ship’s identity, load information and AIS information. etc., automatically alarm ships with draft exceeding the limit and push them to the dispatching system for processing.
  • the system can make statistics and queries on daily passing ships and alarming ships.
  • the process of ship overload identification can include: first, based on massive ship capture images, using deep learning and target detection technology to study the bollard positioning model, and then studying edge detection technology to analyze the ship's freeboard position and the position of the dividing line between the bottom and the water surface. Finally, it is determined whether the ship is overloaded by calculating the distance between the bollard and the water surface dividing line.
  • the system records ship photos, identity and load information of ships passing through the section. After system comparison, it is found that the ship is overloaded, an alarm is automatically issued, and the information is pushed to the dispatching system for processing, forming a closed loop.
  • the mooring cable safety detection sub-link includes: obtaining a panoramic view of the lock chamber through the information collection equipment group, and identifying the floating bollards at the head and tail of the lock in the panoramic view of the lock chamber. Surrounding images, using the acquired images around the floating bollard, detect whether the ship passes the mooring cable safety detection sub-link. In the process of mooring safety inspection, it includes: using the panoramic view of the lock chamber to identify the floating bollards at the head and stern of the lock.
  • the first is to identify whether the ship's moorings are fastened; the second is to identify whether there are boaters at the head and stern of the ship, and whether there are The action of throwing the mooring line, if any, will default to the ship being tied to the mooring line.
  • the recognition results are pushed to the zone broadcast.
  • the zone broadcast automatically announces to remind boaters to fasten the mooring.
  • the recognition accuracy of this function is required to reach more than 80%.
  • notification information is sent to safety managers for manual confirmation, and the confirmation results are saved to the database for next model learning and correction.
  • the video camera on the gantry of the lock chamber to perform image analysis on the position of the ship's bollards and mooring hooks, and identify the status of the ship's mooring cables, which includes data collection, data calibration, Model training link, model deployment link and mooring cable status identification link.
  • data collection collect images and videos of ships docking in the lock through images, and obtain cable videos through fixed video angles
  • data calibration manually calibrate the collected pictures to clarify whether the mooring cable is safe
  • model training use markers pictures, and train the model through deep learning algorithms
  • model deployment deploy the model to the lock video analysis server to analyze the real-time video
  • mooring status recognition detect the bollard position through real-time images and analyze the mooring status of the closing ship , obtain the status of the ship's mooring cable
  • the pedestrian detection sub-link above the gate includes: obtaining the image information above the gate through the information collection equipment group, and determining whether the ship passes through the gate based on the image information above the gate.
  • pedestrian recognition above the gate is based on the intrusion detection function of Hikvision surveillance cameras.
  • the camera uses deep learning hardware and algorithms to support cross-border detection, regional intrusion detection, area entry detection and exit area detection, and supports linked flash alarms. Lights, linked sound alarm.
  • the surveillance camera used to realize the automatic detection of pedestrians above the gate is implemented by reusing the black light camera above the gantry.
  • target recognition can be divided into two methods: motion-based recognition and shape-based recognition.
  • the motion-based recognition method refers to identifying pedestrians by analyzing the gait (Gait) characteristics of people when they move.
  • the gait of the human body has a specific periodicity. By analyzing the periodicity of the image sequence and then comparing it with the periodic pattern of the pedestrian's gait, the pedestrian can be identified.
  • the shape-based recognition method refers to identifying the target by analyzing the grayscale, edge and texture information of the target. Shape-based methods include: methods based on explicit human models, methods based on template matching, and methods based on statistical classification.
  • the method based on a clear human body model refers to constructing a clear 2D or 3D parameter model based on the knowledge of human body structure, and solving the model by extracting the underlying features of the image to identify pedestrians.
  • Methods based on template matching represent pedestrians by storing some grayscale or contour templates. During recognition, pedestrians can be identified by simply measuring the distance between the template and the input window.
  • the method based on statistical classification uses machine learning to learn a classifier from a series of training data, uses the classifier to represent pedestrians, and then uses the classifier to identify the input window.
  • Methods based on statistical models mainly include two steps: feature extraction and classifier design. The purpose of feature extraction is to reduce the dimensionality of the data and obtain features that can reflect the essential attributes of the pattern to facilitate subsequent classification; classifier design belongs to the field of machine learning category, the purpose is to obtain a classifier with low computational complexity and good generalization.
  • NN neural network-based
  • SVM support vector machine
  • Adaboost-based method pedestrian detection is divided into the following steps:
  • preprocessing stage first obtain the image information above the gate through video surveillance, and preprocess this information (such as noise reduction, enhancement, etc.);
  • classification and detection stage some image processing techniques such as image segmentation and model extraction are used to select some regions of interest (ROIs) in the image, that is, candidate areas for pedestrians, and then further analyze the ROIs.
  • ROIs regions of interest
  • technical methods such as classification are used to determine whether the candidate area contains pedestrians; in the decision-making and alarm stage, areas containing pedestrians are tracked.
  • the classification detection stage is the most important stage. Since the pedestrian detection system is a real-time system, the detection algorithm in the system should have high real-time performance, and those algorithms that use complex image processing are no longer applicable; and open detection scenarios, such as constantly changing road conditions, weather and lighting There are also random changes, and pedestrians' clothing and postures are changeable, etc., which makes the template matching method unable to be well applied to pedestrian detection problems. Due to the limitations of premise assumptions, the performance and speed of the scene 3D modeling method cannot meet practical requirements.
  • the main research method is to introduce various classifiers in pedestrian detection, mainly because the classification algorithm has good robustness, and reasonable selection of training samples and features, combined with a reasonably structured classification algorithm, can make it more efficient.
  • Classifiers commonly used for pedestrian detection include: support vector machines (SVM), various types of neural networks (NN), and other statistical-based learning classifiers (such as Adaboost, series classifiers), etc.
  • the information collection equipment group also includes ultra-height detection equipment.
  • the ultra-height detection equipment consists of at least one pair of alignment poles equipped with laser beam-shooting devices.
  • the other sub-links of waiting for the lock also include: a sub-link of ship ultra-height detection; using the ultra-height detection equipment to detect whether the height of the ship exceeds the height limit of the ship lock; if not, it is determined that the ship has passed all the height limits. Describes the sub-link of ship ultra-high detection. Among them, an alignment pole is installed on both sides of the ultra-high detection area, and the beams emitted by the laser alignment devices on each pair of alignment poles are aligned horizontally.
  • ultra-elevation detection equipment laser beam shooting
  • the specific installation height is selected according to the actual requirements of the ship lock. In principle Ensure that the installation height is higher than the top of the superelevation detection specified height so that the vertical laser beam array covers the specified elevation line.
  • the hardware equipment that needs to be deployed includes ultra-high detection equipment, back-end servers, high-pitched loudspeakers, and high-definition video surveillance for linked capture and evidence collection.
  • the ultra-high detection equipment is divided into a laser beam transmitting unit and a receiving unit, which are deployed on both sides of the ultra-high detection area to ensure that the endpoints on both sides of the beam array are aligned horizontally and maintain normal working conditions.
  • the receiving unit and the back-end server Through signal network bridging, it is used for ultra-high signal communication; the background server deploys ultra-high signal processing and push services, which are responsible for capturing ultra-high signals, linking the basic video surveillance network and tweeter alarms; the tweeter is responsible for receiving and broadcasting warning audio; basic video
  • the surveillance network is responsible for capturing and returning ultra-high video data.
  • the implementation of the ultra-high detection system needs to consider the following aspects: 1Height detection: quickly and accurately determine the ship's height and return the ultra-high signal, which is the key to this detection equipment Basic functions 2 Anti-sunlight interference: The working environment of ultra-high detection systems is mostly in the open air. The interference of sunlight on the system cannot be ignored and may cause the system to not work properly. Therefore, overcoming sunlight interference is an important prerequisite to ensure the normal operation of the system. As a light source with highly concentrated energy, laser was early used in aerospace and military facilities. The laser beam in the construction plan belongs to the category of active intrusion detectors, also called laser intrusion detectors. It consists of a laser transmitter and a laser receiver.
  • the laser transmitter consists of a laser transmitter, a modulated excitation power supply and a corresponding direction adjustment mechanism;
  • Laser receiver consists of laser receiver, photoelectric signal processor and corresponding support mechanism.
  • the beamer uses military-grade laser transmitting and receiving devices as the main components of the equipment, which improves the product's unique advantages in detection range, anti-interference, and stability.
  • the laser intrusion detector uses 808nm invisible laser as the light source and works according to the method of unilateral emission and unilateral reception. It is widely used in various complex environments and is an important detection equipment in the field of modern security and intelligence.
  • the laser transmitter emits a directional strong laser beam with good directionality, single frequency and consistent phase.
  • Laser intrusion detectors have the advantages of long detection range, high sensitivity, low false alarm rate, safety, reliability and concealment, easy maintenance and debugging, and adaptability to various harsh natural climate conditions.
  • the ultra-high detection based on laser in this embodiment has the advantages of long detection distance, excellent directionality of the equipment's laser beam, concentrated light energy that is not easily attenuated, high transmission efficiency, and accurate positioning.
  • the laser beam has high emission power density, small divergence angle, and good directivity; the receiver has high sensitivity and strong anti-interference ability.
  • the power density of the laser beam at the target receiving location is the infrared light-emitting diode beam power. Hundreds to thousands of times the density; strong environmental adaptability.
  • the transmission attenuation of the equipment's laser beam is much smaller than that of other similar detectors, and its ability to penetrate rain and fog is strong. It can ensure normal operation at long distances and reduce the risk of severe weather when the detection distance reaches hundreds of meters to several kilometers. False alarm rate.
  • the information collection equipment group also includes a millimeter wave radar subsystem, and the millimeter wave radar subsystem is composed of at least one air gate detection radar and at least one speed detection radar.
  • the empty lock detection radar is used to detect whether there are ships and other objects within the range of "150m-200m length, 20m-30m width" to ensure that there are no ships in the lock when the lock should be empty;
  • the speed detection radar is used to detect the presence of ships in the lock.
  • the minimum detection value should not be greater than 0.5Km/H, and the maximum detection value should not be less than 10Km/H; the speed measurement accuracy is required to be above 0.3Km/H.
  • a total of 7 millimeter wave radars are required, and the layout position of each radar is shown in Figure 9.
  • the lock chamber radar identification sub-link includes: determining whether there is a ship staying in the lock through the empty lock detection radar; if there is no ship staying in the lock, opening the lock gate and detecting the ship entering the lock through the speed detection radar If the speed of the ship entering the lock exceeds the speed limit, an alarm will be triggered.
  • “Empty Lock Detection Radar” and “Speed Detection Radar” are required to detect whether there are ships and other objects within the range of "150m-200m length, 20m-30m width" to ensure that there are no ships in the lock when the lock should be empty; it can detect
  • the minimum detection value should not be greater than 0.5Km/H, and the maximum detection value should not be less than 10Km/H; the speed measurement accuracy is required to be above 0.3Km/H; the "warning line” detection radar is required to be able to detect "15m length, "20m-30m width” range, whether there are ships and other targets.
  • the minimum detection value should not be greater than 0.5Km/H, and the maximum detection value should not be less than 10Km/H
  • the speed measurement accuracy is required to be above 0.3Km/H
  • the "warning line” detection radar is required to be able to detect "15m length, "20m-30m width” range, whether there are ships and other targets.
  • millimeter-wave radar Compared with other sensors, millimeter-wave radar has the characteristics of "all-weather, long range, and accurate speed measurement" and can work normally in various severe weather conditions (haze, snow, light rain, night, etc.). As shown in Figure 10, this embodiment detects the presence and speed information of ships in the lock through the linkage of multiple millimeter wave radars. After linking with the panoramic vision system, the information after dual system verification is output to the gate control system to realize the gate control. Automation control.
  • this embodiment also includes:
  • image information inside and outside the lock is obtained through at least one set of lock photography equipment; it is determined based on the image information inside the lock whether any ship in the lock stays in the area beyond the first virtual warning line , the first virtual cordon is set inside the ship lock, and the area outside the first virtual cordon includes the area within the ship lock from the first virtual cordon to the gate.
  • the image information in the ship lock can be used, but in practical applications, the image information and the empty lock detection radar can also be used to detect whether there are ships in the ship lock staying in the area outside the first virtual warning line.
  • the second virtual warning line is set outside the ship lock.
  • the area outside the second virtual warning line includes the area outside the ship lock.
  • ship identity verification can be performed. Specifically, through the integration of AIS and video, the identity of ships entering the surveillance area can be automatically verified. In the background of the system, the information registered in the AIS system of all ships entering the video surveillance area can be seen, but such as ship name, ship number, and size. It is not yet known whether the actual information of the ship is consistent with the information registered in the system. Traditional methods can only rely on manual verification of ship names and numbers. In order to solve the problem of difficult ship identity verification, high-definition snapshot cameras are additionally installed on the outside of the splayed walls downstream of Huai'an Lock and upstream of Shaobo Lock.
  • this embodiment can also perform over-limit detection of ship draft.
  • the detection of ship draft exceeding the limit is an important basis for judging the safety of ship navigation. Due to changes in water level, ships are often stranded. The traditional method of manual ship measurement through drafts is laborious and inefficient.
  • a set of ship draft over-limit detection devices is added to the remote regulating station downstream of the Shiqiao ship lock. Machine vision is used to identify the ship's draft depth, and is integrated with AIS for the identification of ships suspected to be over-draft. In conjunction with the pontoon type It's sonar, used to measure and review ships suspected of exceeding the draft limit.
  • radar technology is also used to realize ship detection in the gate area, lock chamber vacancy detection, ship overspeed detection, and virtual secondary warning line warning. The advantage of radar technology is its high accuracy.

Landscapes

  • Engineering & Computer Science (AREA)
  • General Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Multimedia (AREA)
  • Signal Processing (AREA)
  • Ocean & Marine Engineering (AREA)
  • Mechanical Engineering (AREA)
  • Civil Engineering (AREA)
  • Structural Engineering (AREA)
  • Traffic Control Systems (AREA)

Abstract

本发明实施例公开了一种用于船舶过闸安全检测方法，涉及信息技术领域，能够提升船舶过闸的数字化程度，从而提升船舶过闸的安全性和效率。本发明包括：通过信息采集设备组，获取进入闸口的船舶的图像信息；利用所获取的图像信息和进入闸口的船舶的AIS信息，检测所述船舶的身份是否正常，若所述船舶的身份正常，则进入待闸环节；在所述待闸环节中，至少包括：船舶吃水超限检测和其它待闸，若所述船舶吃水未超限并且也通过了所述其它，则进入过闸环节；在所述过闸环节中，包括：闸室雷达识别、系缆安全检测和闸门上方行人检测，若所述船舶通过了系缆安全检测、闸门上方行人检测和其它过闸，则判定所述船舶安全过闸。

Description

一种用于船舶过闸安全检测方法 技术领域
本发明涉及信息技术领域，尤其涉及一种用于船舶过闸安全检测方法。
背景技术
目前应用在运河、水库等内河航运中的船舶过闸方式，依旧存在数字化成不高，人工操作误差大等问题，以至于在实际的生产经营中存在一些安全隐患。例如：针对闸室出空、船舶超高超限、进闸超速、不按档位停靠，未系缆绳等可能存在安装隐患的操作环节，缺乏信息化的监控手段，使得船闸过闸调度和闸阀门设备控制运行的过程通常都需要人工参与。
而每个环节通常都设置至少1个人工岗位进行专职监控，人力成本大、操作效率低并且依旧无法避免人工操作、监控中的误差问题。因此，为了兼顾船舶过闸的安全性，需要限制船舶过闸流程的整体速度，从而严重影响到了航运效率。
发明内容
本发明的实施例提供一种用于船闸安全监管系统及方法运河航运的过闸安全检测方法，能够提升船舶过闸的数字化程度，从而提升船舶过闸的安全性和效率。
为达到上述目的，本发明的实施例采用如下技术方案：
S1、通过信息采集设备组，获取进入闸口的船舶的图像信息；
S2、利用所获取的图像信息和进入闸口的船舶的AIS信息，检测所述船舶的身份是否正常，若所述船舶的身份正常，则进入待闸环节；
S3、在所述待闸环节中，至少包括：船舶吃水超限检测子环节和其它待闸 子环节，若所述船舶吃水未超限并且也通过了所述其它子环节，则进入过闸环节；
S4、在所述过闸环节中，包括：闸室雷达识别子环节、系缆安全检测子环节和闸门上方行人检测子环节，若所述船舶通过了系缆安全检测子环节、闸门上方行人检测子环节和其它过闸子环节，则判定所述船舶安全过闸。
在目前的船闸过闸调度以及相应的闸阀门设备控制运行过程中，暂无信息化手段对各环节的安全行为如闸室出空、船舶超高超限、进闸超速、不按档位停靠，未系缆绳等各项进行识别。并对接安全预警信号到控制系统中进行综合判断。目前主要依靠闸首闸尾的工作人员进行人工判断。具体来说，一方面，目前船舶过闸无法自动登记，另一方面，也缺少直观的数字化的导助航软件，这两个方面目前都需要大量的人工参与，效率低、安全性难以提升。
本发明实施例提供的用于运河航运的过闸安全检测方法，利用计算机数据处理和计算机网络传输技术，对运河绿色航运过闸安全信息进行智能“采集、分析、处理”并集中反映到平台实时监控界面和调度控制系统，使船闸管理人员能及时、准确地获得船舶过闸过程中出现的各类安全风险，并采取及时有效的船闸日常实时问题管理与处理，从而在过闸调度的隐患故障的闭环处理。从而兼顾船舶过闸的安全性，可以提升船舶过闸流程的整体速度，从而提升船舶过闸的数字化程度并进一步提升船舶过闸的安全性和效率。
附图说明
为了更清楚地说明本发明实施例中的技术方案，下面将对实施例中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本发明的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其它的附图。
图1为本发明实施例提供的方法流程示意图；
图2、3为本发明实施例提供的可能的船闸现场场景的示意图；
图4、5为本发明实施例提供的可能的相机部署方式的示意图；
图6、7为本发明实施例提供的具体实例的示意图；
图8为本发明实施例提供的告警闭环处理流程图；
图9、10为本发明实施例提供的雷达部署方式的示意图。
具体实施方式
为使本领域技术人员更好地理解本发明的技术方案，下面结合附图和具体实施方式对本发明作进一步详细描述。下文中将详细描述本发明的实施方式，所述实施方式的示例在附图中示出，其中自始至终相同或类似的标号表示相同或类似的元件或具有相同或类似功能的元件。下面通过参考附图描述的实施方式是示例性的，仅用于解释本发明，而不能解释为对本发明的限制。本技术领域技术人员可以理解，除非特意声明，这里使用的单数形式“一”、“一个”、“所述”和“该”也可包括复数形式。应该进一步理解的是，本发明的说明书中使用的措辞“包括”是指存在所述特征、整数、步骤、操作、元件和/或组件，但是并不排除存在或添加一个或多个其他特征、整数、步骤、操作、元件、组件和/或它们的组。应该理解，当我们称元件被“连接”或“耦接”到另一元件时，它可以直接连接或耦接到其他元件，或者也可以存在中间元件。此外，这里使用的“连接”或“耦接”可以包括无线连接或耦接。这里使用的措辞“和/或”包括一个或更多个相关联的列出项的任一单元和全部组合。本技术领域技术人员可以理解，除非另外定义，这里使用的所有术语(包括技术术语和科学术语)具有与本发明所属领域中的普通技术人员的一般理解相同的意义。还应该理解的是，诸如通用字典中定义的那些术语应该被理解为具有与现有技术的 上下文中的意义一致的意义，并且除非像这里一样定义，不会用理想化或过于正式的含义来解释。
本发明实施例提供一种用于船舶过闸安全检测方法，如图1所示，包括：
S1、通过信息采集设备组，获取进入闸口的船舶的图像信息。
S2、利用所获取的图像信息和进入闸口的船舶的AIS信息，检测所述船舶的身份是否正常，若所述船舶的身份正常，则进入待闸环节。
S3、在所述待闸环节中，至少包括：船舶吃水超限检测子环节和其它待闸子环节，若所述船舶吃水未超限并且也通过了所述其它子环节，则进入过闸环节。
S4、在所述过闸环节中，包括：闸室雷达识别子环节、系缆安全检测子环节和闸门上方行人检测子环节，若所述船舶通过了系缆安全检测子环节、闸门上方行人检测子环节和其它过闸子环节，则判定所述船舶安全过闸。
在目前的船闸过闸调度以及相应的闸阀门设备控制运行过程中，暂无信息化手段对各环节的安全行为如闸室出空、船舶超高超限、进闸超速、不按档位停靠，未系缆绳等各项进行识别。并对接安全预警信号到控制系统中进行综合判断。目前主要依靠闸首闸尾的工作人员进行人工判断。具体来说，一方面，目前船舶过闸无法自动登记，另一方面，也缺少直观的数字化的导助航软件，这两个方面目前都需要大量的人工参与，效率低、安全性难以提升。
本实施例的设计思路在于：利用计算机数据处理和计算机网络传输技术，对运河绿色航运过闸安全信息进行智能“采集、分析、处理”并集中反映到平台实时监控界面和调度控制系统，使船闸管理人员能及时、准确地获得船舶过闸过程中出现的各类安全风险，并采取及时有效的船闸日常实时问题管理与处理，从而在过闸调度的隐患故障的闭环处理。具体的，在船舶过闸时，通过闸 室上方视频监控，整合汇聚成闸室一张图，能够有效避免人行桥和摄像机角度干扰，有利于实时掌握闸室内船舶的行为，更直观展示闸室中船舶停靠情况。
本实施例中，所述信息采集设备组包括至少2套船闸拍摄设备，分别布置于船闸下游闸首和船闸上游闸首。例如图2所示的，至少2套船闸拍摄设备，分别布置于船闸下游闸首和船闸上游闸首。其中如图3所示的，可以在下游远调站东西两侧各立一杆(避免距离太近产生盲区)，杆上均安装抓拍摄像头，确保无死角对船舶进行抓拍。为保证夜间抓拍效果，采用红外爆闪灯进行补光。红外光人眼不可见，在不影响船员驾驶的同时，增强夜间清晰抓拍效果。此外，在一侧立杆上安装一台全景相机，用于拍摄船舶全貌。
本实施例中，每一套的船闸拍摄设备中的相机设备至少包括：检测相机(也可以称作船舶检测相机)、细节抓拍相机和全景抓拍相机，其中，所述全景抓拍相机的拍摄视野包括了闸室。其中的三种相机的具体数量和型号可以如表1所示的。
表1
序号 内容 单位 数量 型号
1 船舶检测相机 台 6 WH_5M0FGSNDS_G
2 全景抓拍相机 台 6 WH_5M0FGSNDS_BSS(联通、移动、电信4G，联通移动3G 2G)
3 细节抓拍相机 台 6 WH_HDFGVN_HSD_C
所述信息采集设备组还包括至少1套船舶身份核查子系统，所述船舶身份核查子系统包括：主控设备、补光灯、相机设备(检测相机、细节抓拍相机、全景抓拍相机)和AIS接收机，所述补光灯和相机设备布置于闸口的外面，并且相机设备的拍摄视野各自都覆盖了所述闸口的进入航道的一部分。其中，如图4所示的，主控设备包括专用主控设备，具体可以布置在船闸中心机房，中心机房 与外场设备间利用船闸现有的视频网络链路。专用主控设备的具体型号可以采用IMX220。在实际应用中，可以在XX船闸下游闸首(三线闸下游闸首)、XX船闸上游闸首(三线闸上游闸首)各布设1套视频抓拍设备，共计6套，包括检测相机、细节抓拍相机、全景抓拍相机及相关辅材等，外场设备通过立杆定制支架安装；主控设备放置于船闸中心机房，中心机房与外场设备间利用船闸现有的视频网络链路。通过在闸口东西两侧各立一杆(避免距离太近产生盲区)，杆上均安装抓拍摄像头，确保无死角对船舶进行抓拍。为保证夜间抓拍效果，采用红外爆闪灯进行补光。红外光人眼不可见，在不影响船员驾驶的同时，增强夜间清晰抓拍效果。此外，在一侧立杆上安装一台全景相机，用于检测闸门附近有无船舶。
通过不同角度、不同位置的抓拍图像，识别船舶身份，包括船名、船号、船舶尺寸等，与AIS系统库中的船舶信息进行自动比对，对船舶身份进行核查。针对与AIS系统中备案信息不一致的船舶，保存检测结果，并预警和信息查询，从而完成船舶身份的核查，即本实施例中不仅可以做船舶信息录入，并且可以进行船舶信息自动核对。具体的，船舶身份核查系统的硬件架构如图5所示，硬件设备主要包括：主控设备、补光灯、检测相机、细节抓拍相机、全景抓拍相机和AIS接收机等，外场设备通过立杆定制支架安装；主控设备放置于中心机房，中心机房与外场设备间利用船闸现有的视频网网络链路。
船舶身份核查系统的运行流程如图6所示，船舶经过闸口时，首先通过前端外场设备智能感知并自动抓拍船舶照片，随后对照片和视频进行智能识别，识别内容包括，船舶身份、船舶尺寸和AIS信息等，将识别的信息与AIS系统库的信息进行比对，对尺寸异常、身份不符等船舶自动报警并通知工作人员进行复核处理，最后系统可对每日过闸船舶、异常船舶等进行统计和分析。具体的， 实际应用中可以采用如图7所示的高清抓拍流程，船舶高清抓拍依托于安装在闸首的抓拍相机，基于背景建模、边缘检测等视频分析技术，对船舶检测相机视频流进行自动分析，精准检测船舶是否经过及船舶精准位置。视频流信号由固定不动的摄像头输入。随后对视频流进行背景建模，建模时先对模型进行初始化，之后判断像素点是否为前景点，设定一定的更新机制更新背景模型。背景建模后得到前景点，用连通域分析法分析前景点，得到的连通域作为目标候选区。根据先验知识对目标候选区域中的非船舶目标进行去除，最后得到船舶目标。通过多相机联动技术，实现7×24小时高清抓拍。多相机联动采用图像识别和处理技术，在实现对闸口监控的同时，配合移动目标检测算法，对监控范围内的任何运动目标依次进行单个地、持续地特写跟踪和捕捉目标细节特征信息。多相机联动技术，通过船舶检测相机，检测到目标船舶，并锁定跟踪，确认目标为船舶后，触发细节抓拍相机和全景抓拍相机拍照，抓拍的全景照片和细节照片会依照时间顺序依此呈现在画面中。抓拍的船舶照片不仅包含船舶全貌，又包含船舶细节，且为一船多照，在船舶细节照片上能看清船舶船号、船舶类型、船舶吃水、货物类型等信息。
本实施例中，在S2中包括：通过所述船舶身份核查子系统中的相机设备，抓拍在所述闸口外的且驶入进入航道的船舶，获取船舶的图像信息，并根据所述图像信息中获取船舶身份信息。通过所述AIS接收机，获取所述船舶的AIS信息。其中，所述船舶身份信息至少包括：船舶名称、船舶船号和船舶类型。所述检测所述船舶的身份是否正常，包括：利用所述AIS信息查询AIS系统库中的船舶备案信息，并将查询到的船舶备案信息与所述船舶身份信息进行比对，若二者一致这判定所述船舶的身份正常。
目前的船舶身份识别的技术手段主要有：AIS、RFID和视频图像等。RFID需要在船舶上额外增加终端设备，需要建设大量基站，推广难度大成本高，无法大规模应用。AIS虽然为国际标准船舶自动识别系统，但是内河船舶AIS开机率低，而且即使已开AIS，身份信息也经常被篡改，置信度低。
本实施例中采用AIS和图像识别融合的方式进行船舶身份识别。船舶身份初始准确率不低于85％，随着数据库的完善和深度学习，理论上的准确率可达95％。将定位摄像机获得的目标位置数据及船名信息数据，与AIS信息中的船舶位置信息进行数据融合，可以得出当前拍摄到的船只目标位置和船舶的AIS船名。基于高清抓拍照片采用图像分析和深度学习等技术，进一步分析船舶身份、吃水和船舶类型等信息。通过对大数据的学习和训练，训练出船舶船名的精定位模型，采用目标检测算法精准定位船舶船名号区域，利用图像分割技术将船名号各字符分割，再利用机器学习识别每个字符，进而识别出字符，再根据船名库进行匹配纠错，得到最终识别结果。其中，船舶信息自动复核过程，具体是将船舶身份信息与AIS系统和调度信息数据进行自动比对复核。异常船舶自动报警的方式，具体是系统会记录每个闸次进/出闸的船舶数量、进出闸船舶身份。将本系统得到的船舶身份和相关信息与AIS登记系统比对，若实际通过船舶的身份与登记船舶身份不符，则发出预警。报表统计分析的方式，具体可按时间段/地点/航向/闸次等参数进行筛选统计日/月船舶信息复核情况和流量情况。告警联动闭环处理的方式，可以是：船舶身份核查系统的建设目的主要在于核查验证过闸船舶身份，本系统通过人工智能技术抓拍船舶高清图像，并基于船舶图像自动识别船名字符，并与AIS登记信息自动复核，对身份不符的船舶报警并推送工作人员处理，从而形成告警联动闭环。
船舶身份复核告警联动流程如图8所示：船舶过闸采用AIS信息进行自助过 闸登记，系统确认船舶AIS信息后，调度船舶，船舶的AIS信息和调度信息存入数据库，同时船舶等待过闸，进闸时经过船舶身份核查系统，系统首先感知到船舶出现，触发抓拍相机进行船舶高清抓拍，利用人工智能技术自动识别船名，自动比对识别的船名与数据库，自动判断过闸船舶与登记身份是否一致，若一致则正常过闸，若不一致，触发报警，并将报警数据推送给工作人员处置，工作人员通过核查确认，并对船舶做出处置，例如重新登记、信用扣分、禁止过闸或者正常过闸，形成流程闭环处理。
在本实施例的S3中，所述船舶吃水超限检测子环节包括：在识别了所述船舶的身份后，获取对应所述船舶的尺寸信息。通过所述信息采集设备组，获取所述船舶吃水位置的图像信息。根据所述尺寸信息和所述船舶吃水位置的图像信息，确定所述船舶的干舷位置和船底与水面分界线位置，之后计算系船柱与水面分界线距离来判断所述船舶是否超载。
其中，所述船舶吃水位置的图像信息包括了：系船柱的图像信息、船舶干舷位置的图像信息和水面分界线位置的图像信息。实际应用中，可以基于海量船舶抓拍图像，利用深度学习和目标检测技术，研究系船柱定位模型，然后研究边缘检测技术分析船舶干舷位置、船底与水面分界线位置，最后通过计算系船柱与水面分界线距离判断船舶是否超载。例如：本实施例中提供一种可能的船舶吃水超限检测方式，在施桥船闸下游远调站布设1套视频抓拍设备，包括检测相机、细节抓拍相机、全景抓拍相机及相关辅材等，外场设备通过立杆定制支架安装。施桥船闸下游远调站增设一套船舶吃水超限检测系统，通过高清抓拍，识别船舶的实际吃水深度。同时，针对识别到的吃水超限船舶，将检测结果推送至调度系统，由调度系统处置，提供预警和信息查询。
船舶吃水超限检测系统的硬件设备主要包括：专用主控设备、爆闪灯、检测相机、细节抓拍相机、全景抓拍相机和AIS接收机等，外场设备通过立杆定制支架安装；主控设备放置于中心机房，中心机房与外场设备间利用船闸现有的视频网网络链路。船舶吃水超限检测的过程包括：船舶经过视频抓拍点时，首先通过前端外场设备智能感知并自动抓拍船舶照片，随后对照片和视频进行智能识别，识别内容包括，船舶身份、载重信息和AIS信息等，对吃水超限的船舶自动报警并推送至调度系统进行处置，系统可对每日经过船舶、报警船舶等进行统计和查询。其中，船舶超载判别的过程可以包括：首先基于海量船舶抓拍图像，利用深度学习和目标检测技术，研究系船柱定位模型，然后研究边缘检测技术分析船舶干舷位置、船底与水面分界线位置，最后通过计算系船柱与水面分界线距离判断船舶是否超载。系统记录经过断面船舶的船舶照片、身份和载重信息。经系统比对发现船舶超载，自动报警，并将信息推送至调度系统处理，形成闭环。
本实施例中，所述系缆安全检测子环节，包括：通过所述信息采集设备组获取闸室全景图，并在所述闸室全景图中识别闸首和闸尾的浮式系船柱周围的图像，利用所获取的浮式系船柱周围的图像，检测所述船舶是否通过所述系缆安全检测子环节。在系缆安全检测的过程中，包括：利用闸室的全景图，识别闸首闸尾的浮式系船柱，一是识别船舶缆绳是否系好；二是识别船舶首尾是否有船民、是否有甩缆绳的动作，如果有，则默认船舶系好缆绳。同时，将识别结果推送至分区广播，针对未系缆的船舶，分区广播自动喊话，提醒船民系好缆绳。该功能识别准确率要求达到80％以上，对于未识别到系缆的船舶，放通知信息发送给安全管理人员进行人工确认，确认结果保存到数据库供下次模型学习 修正。其中，实现系缆识别功能，可以利用闸室龙门架上面的视频摄像机，对船舶系船柱、系船钩位置进行图像分析，识别船舶系缆状态，其中包括了数据采集环节、数据标定环节、模型训练环节、模型部署环节和系缆状态识别环节。具体的，数据采集：通过图像方式采集闸室内船舶停靠的图像视频，通过固定视频角度获取缆绳视频；数据标定：针对采集的图片进行人工标定，明确是否系缆安全的状态；模型训练：利用标记的图片，通过深度学习算法训练模型；模型部署：将模型部署到船闸视频分析服务器，对实时视频进行分析；系缆状态识别：通过实时图像检测系船柱位置处，分析收尾船舶的系缆状态，获得船舶系缆状态；把系缆状态传递给控制系统和大屏显示系统进行状态判断；定期采集缆绳识别的图片素材，用以模型训练和强化。
本实施例中，所述闸门上方行人检测子环节，包括：通过所述信息采集设备组获取闸门上方的图像信息，并根据所述闸门上方的图像信息确定，检测所述船舶是否通过所述闸门上方行人检测子环节。实际应用中，闸门上方行人识别基于海康监控摄像机入侵检测功能实现，摄像机采用深度学习硬件及算法，支持越界侦测，区域入侵侦测，进入区域侦测和离开区域侦测，支持联动闪光报警灯，联动声音报警。具体的，实现闸门上方行人自动检测所采用的监控摄像机复用龙门架上方黑光相机实现。其中，根据利用的信息的不同，目标识别可以分为基于运动的识别和基于形状的识别两种方法。基于运动的识别方法指通过分析人运动时的步态(Gait)特征来识别行人。人体的步态具有特定的周期性，通过分析图像序列的周期性，然后与行人步态的周期性的模式相比较，就可以识别出行人。基于形状的识别方法指通过分析目标的灰度、边缘和纹理信息来对目标进行识别。基于形状的方法包括：基于明确人体模型的方法，基于模板匹配的方法，基于统计分类的方法。基于明确人体模型的方法指根据人 体结构的知识，构造一个明确的2D或3D参数模型，通过提取图像的底层特征来求解模型，从而识别行人。
基于模板匹配的方法通过存储一些灰度或者轮廓模板来表示行人，识别的时候只需要度量模板与输入窗口的距离就可以识别行人。基于统计分类的方法通过机器学习从一系列训练数据中学习得到一个分类器，用该分类器来表示行人，然后利用该分类器对输入窗口进行识别。基于统计模型的方法主要包括两个步骤：特征提取和分类器设计.特征提取的目的是降低数据的维数，得到能反映模式本质属性的特征，方便后面的分类；分类器设计属于机器学习领域的范畴，其目的是得到一个计算复杂度较低，并且推广性较好的分类器.针对行人识别问题，可根据分类器的设计方法将现有的基于统计分类的方法分为基于神经网络(NN)的方法，基于支持向量机(SVM)的方法和基于Adaboost的方法的行人检测分为以下几个步骤：预处理阶段，首先通过视频监控获得闸门上方图像信息，对这些信息做预处理(如降噪、增强等)；分类检测阶段，用图像分割、模型提取等一些图像处理技术在图像中选取一些感兴趣的区域(RegionsofInterest，ROIs)，即行人的候选区域，然后对ROIs进行进一步的验证，用分类等技术方法判断候选区域中是否包含行人；决策报警阶段，对含有行人的区域进行跟踪。
在行人检测系统中，分类检测阶段是最为重要的一个阶段。由于行人检测系统是一个实时系统，因此系统中的检测算法应具有很高的实时性，那些使用复杂图像处理的算法便不再适用；而开放的检测场景，如道路状况不断变换、天气以及光照也随机变化，行人的服饰和姿态多变等，使得模板匹配的方法无法很好的应用于行人检测问题中。场景3D建模的方法由于前提假设的限制，其性能和速度无法达到实用的要求。如今主要的研究方法还是在行人检测中引入各 种各样的分类器，主要是因为分类算法具有较好的鲁棒性，而且合理的选择训练样本和特征，结合结构合理的分类算法，可以较好地克服许多不利条件，如行人多样性、场景多样性、光照环境多样性等的影响。因此，在当前情况下，分类检测是行人检测技术研究中的一种主流的方法。常用于行人检测的分类器有：支持向量机(SVM)、各种类型的神经网络(NN)以及其他基于统计的学习分类器(如Adaboost、串联分类器)等。
本实施例中，所述信息采集设备组还包括超高检测设备，所述超高检测设备由至少一对安装激光对射装置的对位立杆组成。所述其它待闸子环节，还包括：船舶超高检测子环节；通过所述超高检测设备检测所述船舶的高度是否超过船闸的限高高度，若没有超过则判定所述船舶通过了所述船舶超高检测子环节。其中，在超高检测区域的两岸分别安装一根对位立杆，每一对对位立杆上的激光对射装置发出的光束水平对齐。在实际应用中，在超高检测区域两岸安装对位立杆，将超高检测设备(激光对射)安装到立杆保持光束两端水平对齐，具体安装高度视根据船闸实地要求选定，原则上保证安装高度高于超高检测指定高度的顶部，以便垂向激光束阵列覆盖指定高程线。需要部署的硬件设备包括超高检测设备、后台服务器、高音喊话扬声器以及用于联动抓拍取证的高清视频监控。其中，超高检测设备分为激光对射发射单元和接收单元，分别部署在超高检测区域的两岸，保证对射光束阵列两侧端点水平对齐，保持正常工况供电状态；接收单元与后台服务器通过信号网络桥接，用于超高信号通信；后台服务器部署超高信号处理与推送服务，负责捕捉超高信号、联动基础视频监控网和高音扬声器警报；高音扬声器负责接收并播报警告音频；基础视频监控网络负责抓拍与返回超高视频数据。
实现超高检测功能，超高检测系统的(激光对射超高检测)的实施需要考虑以下几个方面：①高度检测：迅速且准确判断出船舶高度，返回超高信号，是该检测设备的基本功能②抗日光干扰：超高检测系统的工作环境多为露天工作，日光对系统的干扰是不容忽视的，可能导致系统不能正常工作，因此克服日光干扰是保证系统正常工作的重要前提。激光作为一种能量高度集中的光源，早期应用于航空航天及军工设施中。建设方案中的激光对射属于主动入侵探测器类，又叫激光入侵探测器，由激光发射机和激光接收机组成，激光发射机由激光发射器、调制激励电源及相应的方向调整机构组成；激光接收机由激光接收器、光电信号处理器以及相应的支撑机构组成。对射器采用军工级的激光发射和接收器件作为设备的主要部件，提高了产品在探测距离，抗干扰，稳定性方面的独有优势。激光入侵探测器采用808nm不可见激光作为光源，按照单边发射，单边接收的方式工作，广泛应用于各种复杂环境，是现代安防及智能化领域重要的探测设备。激光发射机将其发射出的定向强激光束，方向性好、频率单一、相位一致，以不可见调制激光光束形成警戒线，采用遮挡报警的方式对周界、平面和立体空间进行封闭布防的激光入侵方案系统。主动激光入侵探测器具有探测距离远，灵敏度高，误报率低，安全可靠隐蔽性好，检修调试方便，适应各种恶劣自然气候情况等优点。
本实施例中基于激光进行的超高检测，具有的优势在于：探测距离远，设备激光束的方向性极好，光能集中不易衰减，传输效率高；定位准确。激光束发射功率密度大，发散角小，方向性好；接收器的灵敏性高，抗干扰能力强，在使用同等功率器件的条件下，目标接收处激光束的功率密度是红外发光二极管光束功率密度的数百至几千倍；环境适应力强。在同样气候条件下，设备激光束的传输衰减远小于其他同类探测器，穿透雨雾能力强，在探测距离达数百 米至几公里的同时能够保证远距离的正常工作并减少恶劣天气时的误报率。
本实施例中，所述信息采集设备组还包括毫米波雷达子系统，所述毫米波雷达子系统由至少1个空闸检测雷达和至少1个速度检测雷达组成。其中，空闸检测雷达用于检测“150m-200m长度，20m-30m宽度”范围内是否存在船舶等物体，确保船闸应该空仓时，没有船只在闸内；速度检测雷达用于检测闸内船只的速度，最小检测值不应大于0.5Km/H，最大检测值不应小于10Km/H；测速精度要求在0.3Km/H以上。本实施例中一共需要7颗毫米波雷达，每个雷达布局位置分别如图9所示。
在所述过闸环节中先执行所述闸室雷达识别子环节。所述闸室雷达识别子环节中，包括：通过所述空闸检测雷达判定是否有船舶停留在闸内；若没有船舶停留在闸内，则开启闸门并通过所述速度检测雷达检测驶入船闸内的船舶速度，若驶入船闸内的船舶速度超过限速则触发报警。例如：“空闸检测雷达”和“速度检测雷达”要求可以检测“150m-200m长度，20m-30m宽度”范围内是否存在船舶等物体，确保船闸应该空仓时，没有船只在闸内；可以检测闸内船只的速度，最小检测值不应大于0.5Km/H，最大检测值不应小于10Km/H；测速精度要求在0.3Km/H以上；“警戒线”检测雷达要求能检测“15m长度，20m-30m宽度”范围内，是否有船舶等目标存在。船闸外侧近处区域，要求距离闸门20m范围内可检测是否有船舶存在。
相比较于其他传感器，毫米波雷达具有“全天候，测距远，测速准确”等特点，可以在各种恶劣天气(雾霾天，雪天，中小雨，黑夜等各种场景下)正常工作。如图10所示的，本实施例通过多颗毫米波雷达联动，检测闸内船舶有无及速度信息，与全景视觉系统联动后，输出双系统验证后的信息给到闸门控制 系统，实现闸门自动化控制。
进一步的，本实施例中还包括：
在船舶通过闸门驶入闸室后，通过至少一套船闸拍摄设备获取船闸内和船闸外的图像信息；根据船闸内的图像信息判定是否有船闸内的船舶停留在第一虚拟警戒线以外的区域，所述第一虚拟警戒线设置在船闸内，所述第一虚拟警戒线以外的区域包括了在船闸内的从所述第一虚拟警戒线到闸门之间的区域。其中，不仅可以根据船闸内的图像信息，在实际应用中还可以同时利用图像信息和所述空闸检测雷达进行检测是否有船闸内的船舶停留在第一虚拟警戒线以外的区域。
根据船闸外的图像信息判定是否有船闸内的船舶停留在第二虚拟警戒线以外的区域，所述第二虚拟警戒线设置在船闸外，所述第二虚拟警戒线以外的区域包括了在船闸外的从所述第二虚拟警戒线到闸门之间的区域；若第一虚拟警戒线以外的区域和第二虚拟警戒线以外的区域都没有船舶停留，则关闭闸门。
本实施例中，可以进行船舶身份核查。具体通过采用AIS与视频融合，可以对进入到监控区域的船舶身份自动核查，在系统后台可以看到所有进入视频监控区域内船舶在AIS系统中登记的信息，但诸如船名、船号、尺寸等船舶实际的信息与系统中登记的信息是否一致尚未可知。传统手段只能依靠人工进行船名船号的确认，为了解决船舶身份核查难的问题，在淮安闸下游、邵伯闸上游八字墙外侧补充安装高清抓拍相机，一方面对闸首闸尾八字墙外侧视频监控进行补盲；另一方面通过抓拍的船舶图片，进行船名、船号、船舶尺寸识别，通过与AIS系统中的船舶尺寸进行对比分析，将疑似船证不符的情况和相关证据推送 给海事执法部门，由海事执法部门进行处理。
同时，本实施例还可以进行船舶吃水超限检测。船舶吃水超限检测是船舶航行安全的重要判断依据，由于水位变化导致在船舶搁浅的情况屡见不鲜，传统方式通过水尺进行船舶手工丈量费事费力，效率偏低。本实施例在施桥船闸下游远调站增设一组船舶吃水超限检测装置，采用机器视觉的方式识别船舶吃水深度，并与AIS进行融合，用于疑似吃水超限船舶识别，并配合浮筒式是声纳，对疑似吃水超限船舶进行丈量复核。进一步的，本实施例中还利用雷达技术实现闸门区域船舶检测、闸室出空检测，船舶超速检测，虚拟二级警戒线预警。雷达技术其优点是准确率高。
本说明书中的各个实施例均采用递进的方式描述，各个实施例之间相同相似的部分互相参见即可，每个实施例重点说明的都是与其他实施例的不同之处。尤其，对于设备实施例而言，由于其基本相似于方法实施例，所以描述得比较简单，相关之处参见方法实施例的部分说明即可。以上所述，仅为本发明的具体实施方式，但本发明的保护范围并不局限于此，任何熟悉本技术领域的技术人员在本发明揭露的技术范围内，可轻易想到的变化或替换，都应涵盖在本发明的保护范围之内。因此，本发明的保护范围应该以权利要求的保护范围为准。

Claims (10)

  1. 一种用于船舶过闸安全检测方法，其特征在于，包括：
    S1、通过信息采集设备组，获取进入闸口的船舶的图像信息；
    S2、利用所获取的图像信息和进入闸口的船舶的AIS信息，检测所述船舶的身份是否正常，若所述船舶的身份正常，则进入待闸环节；
    S3、在所述待闸环节中，至少包括：船舶吃水超限检测子环节和其它待闸子环节，若所述船舶吃水未超限并且也通过了所述其它子环节，则进入过闸环节；
    S4、在所述过闸环节中，包括：闸室雷达识别子环节、系缆安全检测子环节和闸门上方行人检测子环节，若所述船舶通过了系缆安全检测子环节、闸门上方行人检测子环节和其它过闸子环节，则判定所述船舶安全过闸。
  2. 根据权利要求1所述的方法，其特征在于，所述信息采集设备组包括至少2套船闸拍摄设备，分别布置于船闸下游闸首和船闸上游闸首，每一套的船闸拍摄设备中的相机设备至少包括：检测相机、细节抓拍相机和全景抓拍相机，其中，所述全景抓拍相机的拍摄视野包括了闸室；
    所述信息采集设备组还包括至少1套船舶身份核查子系统，所述船舶身份核查子系统包括：主控设备、补光灯、相机设备和AIS接收机，所述补光灯和相机设备布置于闸口的外面，并且相机设备的拍摄视野各自都覆盖了所述闸口的进入航道的一部分。
  3. 根据权利要求2所述的方法，其特征在于，在S2中，包括：
    通过所述船舶身份核查子系统中的相机设备，抓拍在所述闸口外的且驶入进入航道的船舶，获取船舶的图像信息，并根据所述图像信息中获取船舶身份信息，所述船舶身份信息至少包括：船舶名称、船舶船号和船舶类型；
    通过所述AIS接收机，获取所述船舶的AIS信息；
    所述检测所述船舶的身份是否正常，包括：利用所述AIS信息查询AIS系统库中的船舶备案信息，并将查询到的船舶备案信息与所述船舶身份信息进行比对，若二者一致这判定所述船舶的身份正常。
  4. 根据权利要求1所述的方法，其特征在于，在S3中，所述船舶吃水超限检测子环节包括：
    在识别了所述船舶的身份后，获取对应所述船舶的尺寸信息；
    通过所述信息采集设备组，获取所述船舶吃水位置的图像信息，所述船舶吃水位置的图像信息包括了：系船柱的图像信息、船舶干舷位置的图像信息和水面分界线位置的图像信息；
    根据所述尺寸信息和所述船舶吃水位置的图像信息，确定所述船舶的干舷位置和船底与水面分界线位置，之后计算系船柱与水面分界线距离来判断所述船舶是否超载。
  5. 根据权利要求1或2所述的方法，其特征在于，所述系缆安全检测子环节，包括：通过所述信息采集设备组获取闸室全景图，并在所述闸室全景图中识别闸首和闸尾的浮式系船柱周围的图像，利用所获取的浮式系船柱周围的图像，检测所述船舶是否通过所述系缆安全检测子环节；
    所述闸门上方行人检测子环节，包括：通过所述信息采集设备组获取闸门上方的图像信息，并根据所述闸门上方的图像信息确定，检测所述船舶是否通过所述闸门上方行人检测子环节。
  6. 根据权利要求1所述的方法，其特征在于，所述信息采集设备组还包括超高检测设备，所述超高检测设备由至少一对安装激光对射装置的对位立杆组成，其中，在超高检测区域的两岸分别安装一根对位立杆，每一对对位立杆上的激光对射装置发出的光束水平对齐；
    所述其它待闸子环节，还包括：船舶超高检测子环节；
    通过所述超高检测设备检测所述船舶的高度是否超过船闸的限高高度，若没有超过则判定所述船舶通过了所述船舶超高检测子环节。
  7. 根据权利要求2所述的方法，其特征在于，所述信息采集设备组还包括毫米波雷达子系统，所述毫米波雷达子系统由至少1个空闸检测雷达和至少1个速度检测雷达组成。
  8. 根据权利要求7所述的方法，其特征在于，在所述过闸环节中先执行所述闸室雷达识别子环节；
    所述闸室雷达识别子环节中，包括：
    通过所述空闸检测雷达判定是否有船舶停留在闸内；
    若没有船舶停留在闸内，则开启闸门并通过所述速度检测雷达检测驶入船闸内的船舶速度，若驶入船闸内的船舶速度超过限速则触发报警。
  9. 根据权利要求8所述的方法，其特征在于，还包括：
    在船舶通过闸门驶入闸室后，通过至少一套船闸拍摄设备获取船闸内和船闸外的图像信息；
    根据船闸内的图像信息判定是否有船闸内的船舶停留在第一虚拟警戒线以外的区域，所述第一虚拟警戒线设置在船闸内，所述第一虚拟警戒线以外的区域包括了在船闸内的从所述第一虚拟警戒线到闸门之间的区域。
  10. 根据权利要求9所述的方法，其特征在于，还包括：
    在判定出船闸内的船舶没有停留在第一虚拟警戒线以外的区域之后，根据船闸外的图像信息判定是否有船闸内的船舶停留在第二虚拟警戒线以外的区域，所述第二虚拟警戒线设置在船闸外，所述第二虚拟警戒线以外的区域包括了在船闸外的从所述第二虚拟警戒线到闸门之间的区域；
    若第一虚拟警戒线以外的区域和第二虚拟警戒线以外的区域都没有船舶停留，则关闭闸门。
PCT/CN2022/131860 2022-05-18 2022-11-15 一种用于船舶过闸安全检测方法 WO2023221425A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202210547744.XA CN114875877A (zh) 2022-05-18 2022-05-18 一种用于船舶过闸安全检测方法
CN202210547744.X 2022-05-18

Publications (1)

Publication Number Publication Date
WO2023221425A1 true WO2023221425A1 (zh) 2023-11-23

Family

ID=82678458

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/131860 WO2023221425A1 (zh) 2022-05-18 2022-11-15 一种用于船舶过闸安全检测方法

Country Status (2)

Country Link
CN (1) CN114875877A (zh)
WO (1) WO2023221425A1 (zh)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114875877A (zh) * 2022-05-18 2022-08-09 苏交科集团股份有限公司 一种用于船舶过闸安全检测方法
CN117058599B (zh) * 2023-10-12 2023-12-15 南京苏润科技发展有限公司 船闸运行数据处理方法及系统
CN117593703B (zh) * 2024-01-19 2024-04-16 深圳市创达电子有限公司 一种停车场道闸的控制管理方法及系统

Citations (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP3882025B1 (ja) * 2006-06-30 2007-02-14 国土交通省国土技術政策総合研究所長 広域用船舶動静監視方法及びシステム
JP2014163067A (ja) * 2013-02-22 2014-09-08 Toshiba Corp 水門制御システム及び水門制御方法
CN205992306U (zh) * 2016-08-30 2017-03-01 浙江华是科技股份有限公司 航道智能截面管理系统
CN107067821A (zh) * 2017-05-12 2017-08-18 山东海诺港务有限公司 一种船舶通过船闸的安全导航辅助系统及方法
CN110299029A (zh) * 2019-06-25 2019-10-01 中山市聚云软件科技有限公司 船舶特征检测系统和方法以及桥梁超高预警系统和方法
CN111862684A (zh) * 2020-07-31 2020-10-30 长江三峡通航管理局 船闸闸室禁停区安全监控系统及防撞方法
CN111967409A (zh) * 2020-08-20 2020-11-20 中国长江三峡集团有限公司 船舶过闸系缆图像识别装置及识别方法
CN112132385A (zh) * 2020-08-11 2020-12-25 长江三峡通航管理局 一种内河船舶监测系统及船舶管理信息系统
CN113465521A (zh) * 2021-09-03 2021-10-01 武汉理工大学 一种快速测量船舶高度的装置、船舶超高预警系统及方法
CN113762141A (zh) * 2021-09-03 2021-12-07 广州海事科技有限公司 智能视频识别预警方法、系统、计算机设备及存储介质
CN114401354A (zh) * 2021-12-30 2022-04-26 中船重工(浙江自贸区)海洋科学研究院有限公司 一种面向海上船舶的超视距监控的智能化控制方法及系统
CN114875877A (zh) * 2022-05-18 2022-08-09 苏交科集团股份有限公司 一种用于船舶过闸安全检测方法

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN204463471U (zh) * 2015-03-05 2015-07-08 长江三峡通航管理局 一种船舶过闸船速检测预警装置
CN111139767A (zh) * 2019-06-14 2020-05-12 深圳市博思高科技有限公司 防尾随无人值守的人行通道设备
CN112085642A (zh) * 2020-09-27 2020-12-15 广州海事科技有限公司 船港管理方法、系统、计算机设备、及存储介质
KR102374711B1 (ko) * 2020-10-14 2022-03-15 한국항공우주연구원 선박자동식별정보를 이용한 위성 촬영임무 재계획 방법 및 장치
CN112212804B (zh) * 2020-12-11 2021-03-05 思创数码科技股份有限公司 船闸系缆柱系缆状态检测方法、系统及可读存储介质

Patent Citations (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP3882025B1 (ja) * 2006-06-30 2007-02-14 国土交通省国土技術政策総合研究所長 広域用船舶動静監視方法及びシステム
JP2014163067A (ja) * 2013-02-22 2014-09-08 Toshiba Corp 水門制御システム及び水門制御方法
CN205992306U (zh) * 2016-08-30 2017-03-01 浙江华是科技股份有限公司 航道智能截面管理系统
CN107067821A (zh) * 2017-05-12 2017-08-18 山东海诺港务有限公司 一种船舶通过船闸的安全导航辅助系统及方法
CN110299029A (zh) * 2019-06-25 2019-10-01 中山市聚云软件科技有限公司 船舶特征检测系统和方法以及桥梁超高预警系统和方法
CN111862684A (zh) * 2020-07-31 2020-10-30 长江三峡通航管理局 船闸闸室禁停区安全监控系统及防撞方法
CN112132385A (zh) * 2020-08-11 2020-12-25 长江三峡通航管理局 一种内河船舶监测系统及船舶管理信息系统
CN111967409A (zh) * 2020-08-20 2020-11-20 中国长江三峡集团有限公司 船舶过闸系缆图像识别装置及识别方法
CN113465521A (zh) * 2021-09-03 2021-10-01 武汉理工大学 一种快速测量船舶高度的装置、船舶超高预警系统及方法
CN113762141A (zh) * 2021-09-03 2021-12-07 广州海事科技有限公司 智能视频识别预警方法、系统、计算机设备及存储介质
CN114401354A (zh) * 2021-12-30 2022-04-26 中船重工(浙江自贸区)海洋科学研究院有限公司 一种面向海上船舶的超视距监控的智能化控制方法及系统
CN114875877A (zh) * 2022-05-18 2022-08-09 苏交科集团股份有限公司 一种用于船舶过闸安全检测方法

Also Published As

Publication number Publication date
CN114875877A (zh) 2022-08-09

Similar Documents

Publication Publication Date Title
WO2023221425A1 (zh) 一种用于船舶过闸安全检测方法
CN108062349B (zh) 基于视频结构化数据及深度学习的视频监控方法和系统
CN103676829B (zh) 基于视频的智能城市综合管理系统及其方法
CN109164443A (zh) 基于雷达及图像分析的铁路线路异物检测方法及系统
US9154741B2 (en) Apparatus and method for processing data of heterogeneous sensors in integrated manner to classify objects on road and detect locations of objects
CN102509090B (zh) 一种基于天网工程中公共安全视频图像的车辆特征识别装置
US7489802B2 (en) Miniature autonomous agents for scene interpretation
CN103366506A (zh) 一种驾驶员行车途中接打手机行为的自动监控装置及方法
CN114419825B (zh) 基于毫米波雷达和摄像头的高铁周界入侵监测装置及方法
CN101964145A (zh) 一种自动车牌识别方法及其系统
CN105611244A (zh) 一种基于球机监控视频的机场外来异物检测方法
CN204990743U (zh) 一种超高车辆预警系统
CN102646309A (zh) 一种智能视频周界围栏系统及其控制方法
CN102867417A (zh) 一种出租车防伪系统及方法
CN103646550A (zh) 一种智能车牌识别系统
CN103236181A (zh) 一种基于机器视觉的交通信号灯状态监控系统及方法
CN113593250A (zh) 一种基于视觉识别的违章停车检测系统
CN106845359A (zh) 基于红外发射的隧道入口行车提示装置及方法
CN112287823A (zh) 一种基于视频监控的面部口罩识别方法
US11580659B2 (en) Method for size estimation by image recognition of specific target using given scale
CN103152558B (zh) 基于场景识别的入侵检测方法
CN113743260A (zh) 一种地铁站台密集人流情况下的行人跟踪方法
CN114248819A (zh) 基于深度学习的铁路侵限异物无人机检测方法、装置及系统
CN116846059A (zh) 一种用于电网巡检和监控的边缘检测系统
KR20210158037A (ko) 영상감시시스템에서의 고속 이동물체의 위치 및 모션 캡쳐 방법

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22942446

Country of ref document: EP

Kind code of ref document: A1