Cabin smoke and open fire identification alarm system based on AI video

Abstract

The invention relates to the technical field of marine transportation safety, and discloses an AI video-based cabin smoke and open fire identification alarm system, which comprises a shipborne system and a shore-based system, the shipborne system comprises an information acquisition unit used for acquiring video, temperature difference, audio and other information of a control area; the smoke open fire and temperature difference identification unit is used for identifying smoke open fire and temperature difference appearing in the control and management area; the fire development situation evaluation unit is used for evaluating the fire development situation of the control and management area based on a fire situation development evaluation model; the shore-based system and the information acquisition unit are used for training a ship smoke open fire identification standard model and a ship fire situation development evaluation standard model; and alarm management: generating a coping decision based on the shipborne fire asking-for-help information. The defect that an existing smoke and fire sensor is low in sensitivity can be overcome, fire hazards can be found as soon as possible without dead angles, ship-shore linkage scientific rescue can be carried out in time, and ship fire prevention and shore-based rescue efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the field of ship safety technology, and in particular to a cabin smoke and open fire recognition alarm system based on AI video. Background Art

[0002] During the voyage of a ship, the safety of the ship's navigation is directly related to the safety of the crew and cargo on board, as well as the protection of the marine environment. Among them, fire is one of the most serious safety hazards in the navigation of a ship. Timely identification and discovery of smoke and open flames in the ship's control area, early alarm, and allowing the crew to have more time to carry out pre-disaster emergency treatment are of great significance to reducing the occurrence of fires and reducing fire losses.

[0003] With the continuous expansion of the tonnage of modern transport ships, the continuous enhancement of transport capacity, and the continuous increase in the types of transported goods, modern ships are equipped with video monitoring systems, smoke detection alarm systems and fire-fighting facilities in the control area. However, the existing technology still has the following problems:

[0004] (1) Existing smoke and fire sensors are arranged at intervals, making it difficult to cover the entire control area without blind spots. In addition, they have low sensitivity to smoke and open flames generated in the early stages of a fire, making it difficult to detect fire hazards early.

[0005] (2) Existing video surveillance systems rely on the human eye to detect fires. However, due to factors such as eye fatigue, negligence of monitoring personnel, and the large number of cameras, it is difficult to detect fire hazards in the first place.

[0006] (3) After a fire occurs, it is difficult to promptly and clearly inform relevant personnel of what response decisions to take. When cabin personnel need to be evacuated, it is also difficult to provide a safe evacuation route in a timely manner.

[0007] (4) The ships involved in the rescue and the shore-based rescue commanders are unable to obtain the actual fire scene of the burning ship, which reduces the rescue efficiency. Summary of the invention

[0008] The purpose of the present invention is to provide a cabin smoke and open flame identification alarm system based on AI video, which aims to solve the problems in the prior art that it is difficult to discover fire hazards in the first time, promptly notify and help nearby crew members to take fire hazard elimination strategies in time, and cause hazardous fires; when the fire is out of control, it is difficult to provide safe evacuation channel prompts for evacuees, endangering the lives of crew members; rescue ships and shore-based command personnel cannot obtain visualized fire scene information of the fire ship, and then make scientific rescue decisions in time, reducing rescue efficiency and other problems.

[0009] An embodiment of the present invention provides a cabin smoke and open fire recognition alarm system based on AI video, comprising:

[0010] A cabin smoke and open fire identification alarm system based on AI video, characterized in that it comprises a ship-borne system arranged on a ship and a shore-based system arranged on land, wherein the ship-borne system and the shore-based system are connected via a wireless network;

[0011] The shipborne system is used to obtain the smoke and fire, temperature difference video image data of the controlled area in real time, perform abnormal detection on the smoke and fire video, temperature difference, and sound information based on the ship smoke and fire recognition standard model, generate fire danger alarm information based on the detection results, perform a comprehensive assessment of the fire development situation based on the ship fire situation development assessment standard model, and issue evacuation alarm and shore-based help alarm information based on the assessment results;

[0012] The evacuation alarm also provides guidance information on safe evacuation channels based on a safe channel algorithm for evacuees;

[0013] The shipborne system is also used to obtain smoke, open flames, temperature difference audio and video data, and fire scene data of the controlled area collected by video cameras and infrared temperature difference sensors, and send them to the shore-based system;

[0014] The shore-based system is used to receive fire help information sent by each ship, alarm and activate the shore-based rescue plan and conduct rescue dispatch command;

[0015] The shore-based system is also used to collect smoke, open flames, temperature difference audio and video data, and fire scene data sent by each ship, build and train standard models for ship smoke and open flame recognition, temperature difference recognition, and ship fire development situation assessment standard models, and send them to the ship-borne systems.

[0016] In some embodiments of the present application, the shipboard system includes:

[0017] The communication management unit is used to manage the working status of the ship's combined configuration of 700M 5G, 5G, satellite communication, Beidou satellite SMS communication, WiFi network and other communication equipment, and automatically select the communication link, communication method and communication time in the most economical and efficient way based on the preset communication level, and exchange information with the ship-borne client, mobile terminal and shore-based system;

[0018] The preset communication level is to initiate information interaction with the corresponding client, mobile terminal, and shore-based system based on the monitoring, identification, and evaluation results of the data acquisition unit, cabin smoke and fire identification unit, fire situation assessment unit, etc., such as equipment failure, smoke and fire warning, fire evacuation, and help alarm.

[0019] In some embodiments of the present application, the shipboard system includes:

[0020] A data acquisition unit is used to collect video from cameras deployed at multiple points in the control area, the working status of sensors such as audio and infrared temperature, and monitoring data, process, integrate, store and send them to the AI ​​video processing unit, and send them to the shore-based system according to instructions;

[0021] The AI ​​video processing unit is used to pre-process the video image data deployed at multiple points in the control area, and based on an adaptive noise elimination algorithm, denoise the pre-processed image data based on the cabin light noise under different light intensities, and use the denoised image data for storage and smoke and open flame identification and fire development situation assessment.

[0022] In some embodiments of the present application, the shipboard system includes:

[0023] A smoke and open fire identification unit is communicatively connected with the AI ​​video processing unit and the data acquisition unit, performs abnormality detection on the smoke information and the temperature difference information based on the ship smoke and open fire identification standard model and the SSD detection algorithm, obtains an abnormality detection result, and sends the abnormality detection result to the management unit;

[0024] The fire situation development assessment unit is connected to the abnormality detection unit and the data acquisition unit for communication, and generates a fire situation development assessment result based on the abnormality detection result, the fire point spread speed, the change in the number of fire points, the temperature distribution, and the number of deflagrations, and sends the abnormality detection result to the management unit;

[0025] The management unit is used to issue smoke, open fire, and temperature difference alarm information according to the identification result of the smoke and open fire identification unit, and the alarm information includes the scene video of the smoke and open fire and the location information of the cabin where the smoke and open fire occur;

[0026] The management module is also used to issue personnel evacuation alarm information and send help information to the shore-based system based on the fire development situation assessment results;

[0027] The personnel evacuation alarm information includes a recommended personnel evacuation channel plan generated based on the cabin channel map and the intelligent path planning algorithm;

[0028] The fire rescue information includes the name and location of the fire ship and the fire scene video information.

[0029] In some embodiments of the present application, the shore-based system comprises:

[0030] The communication management unit is used to manage the working status of the communication equipment such as 700M 5G, 5G, satellite communication, Beidou satellite SMS communication, and Internet configured in the shore-based system combination, and automatically select the communication link, communication method and communication time in the most economical and efficient way based on the preset communication level to exchange information with various shipborne systems;

[0031] The preset communication level is to activate the shore-based system emergency plan based on the help alarm of the shipborne system, and issue ship fire help alarm information to relevant clients and mobile clients. The help information includes the name of the burning ship, the real-time location of the ship, the name of the burning cabin and the fire scene video information.

[0032] In some embodiments of the present application, the shore-based system comprises:

[0033] A data acquisition unit connected to the communication management unit, used to receive fire help information issued by various shipborne systems in real time and send it to the management unit;

[0034] The data acquisition unit is also used to collect, process, integrate and store multi-point camera videos deployed in the control areas of each shipboard system, monitoring data of audio, infrared temperature and other sensors, and fire scene video data for training AI video's cabin smoke and open flame recognition and fire development situation assessment standard model.

[0035] In some embodiments of the present application, the shore-based system comprises:

[0036] The model building unit, based on the monitoring scene data collected by the data collection unit, adopts the YOLO detection algorithm and the SSD detection algorithm to build a smoke feature map, generate a smoke ImageNet data set, and build and train a ship smoke open flame, temperature difference recognition, and fire development situation assessment standard model based on the smoke, temperature difference, and fire scene ImageNet data sets, and send it to the shipborne system to continuously improve the accuracy of the shipborne system's smoke open flame, temperature difference abnormal recognition rate, and fire development situation assessment;

[0037] The model building unit is also used to train the cabin passage map and intelligent path planning algorithm and send it to the shipboard system to continuously improve the accuracy of the fire evacuation recommendation plan.

[0038] In some embodiments of the present application, the shore-based system comprises:

[0039] A management unit, configured to receive the fire help information of the shipborne system in real time, and generate a recommended marine rescue plan based on the fire help information of the shipborne system and the fire scene video information and in accordance with the marine emergency plan;

[0040] The management unit is also used to establish and manage the maritime fire rescue emergency plan, and continuously improve the maritime rescue plan based on the rescue results.

[0041] In some embodiments of the present application, the shore-based system further comprises:

[0042] The maritime rescue command unit is used to issue the fire rescue alarm information of the ship to the client and mobile client of the relevant institutions and personnel through the communication management unit, 700M 5G, 5G, satellite communication, Beidou satellite SMS communication, Internet, etc. according to the determined maritime rescue plan;

[0043] It is also used to release real-time video information of fire scenes on rescued ships to ships and rescue agencies participating in maritime rescue through 700M 5G, 5G, satellite communications, and the Internet, and integrate and display real-time status information of rescue ships and rescue personnel to achieve flattening and visualization of ship fire rescue command.

[0044] The beneficial effects of the present invention are as follows: the shipboard system comprises: an information acquisition unit for acquiring video, temperature difference and audio information of the controlled area; a smoke and open flame and temperature difference recognition unit for identifying smoke and open flame and temperature difference appearing in the controlled area based on the smoke and open flame and temperature difference recognition model, obtaining the recognition result and giving an alarm; a fire development situation assessment unit for assessing the fire development situation in the controlled area based on the fire situation development assessment model, and deciding to initiate the emergency evacuation alarm and send rescue alarm information to the shore-based system according to the assessment result; the shore-based system, an information acquisition unit, for collecting the video, temperature difference, audio and other information of the controlled area sent by each shipboard system, the video information of the fire scene, training the ship smoke and open flame recognition standard model and the ship fire situation development assessment standard model, and updating the model of the shipboard system based on the training result; and an alarm management for generating a response decision based on the shipboard fire help information, so as to improve the rescue efficiency of marine ship fires. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art description. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0046] Figure 1 It is a structural schematic diagram of a cabin smoke and open fire identification alarm system provided by an embodiment of the present invention;

[0047] Figure 2 is a schematic structural diagram of a shipborne system provided by an embodiment of the present invention;

[0048] Figure 3 It is a schematic diagram of the structure of a shore-based system provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0049] Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments described herein. On the contrary, these embodiments are provided in order to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art. It should be noted that, in the absence of conflict, the embodiments of the present invention and the features described in the embodiments can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0050] The embodiment of the present invention provides a cabin smoke and open flame identification alarm system based on AI video, which is intended to solve the problem that the existing smoke and fire sensors are arranged at intervals, which makes it difficult to cover the entire control area without blind spots, and have low sensitivity to smoke and open flames generated in the early stage of a fire, making it difficult to detect fire hazards early. After a fire occurs, it is difficult to promptly and clearly notify relevant personnel of what response decisions to take, and it is difficult to provide a safe evacuation channel in time when cabin personnel need to be evacuated. In addition, the ships involved in the rescue and the shore-based rescue commanders cannot obtain the real fire scene of the fire ship, which reduces the rescue efficiency. Figure 1 and Figure 2 As shown, including:

[0051] A cabin smoke and open fire identification alarm system based on AI video, characterized in that it comprises a ship-borne system arranged on a ship and a shore-based system arranged on land, wherein the ship-borne system and the shore-based system are connected via a wireless network;

[0052] The shipborne system is used to obtain smoke and fire, audio and video data of the controlled area in real time, perform abnormal detection on the smoke and fire video, temperature difference, and sound information based on the ship smoke and fire recognition standard model, generate fire danger alarm information based on the detection results, perform a comprehensive assessment of the fire development situation based on the ship fire situation development assessment standard model, and issue evacuation alarm and shore-based help alarm information based on the assessment results;

[0053] The evacuation alarm also provides guidance information on safe evacuation channels based on a safe channel algorithm for evacuees;

[0054] The shipborne system is also used to obtain smoke, open flames, temperature difference, audio and video data, and fire scene data of the controlled area collected by video cameras, infrared temperature difference and other sensors, and send them to the shore-based system;

[0055] The shore-based system is used to receive fire help information sent by each ship, alarm and activate the shore-based rescue plan and conduct rescue dispatch command;

[0056] The shore-based system is also used to collect smoke, open flames, temperature difference audio and video data, and fire scene data sent by each ship, build and train a standard model for ship smoke and open flame recognition, temperature difference recognition, and a standard model for ship fire development situation assessment, and send them to each ship-borne system;

[0057] In the embodiment of the present application, the controlled area refers to a pre-set area on the ship where fire is prone to occur and an area where cargo and people are densely populated;

[0058] In the embodiment of the present application, the training data refers to the image data and monitoring data collected by multiple cameras and sensors in the cabin control area of ​​each ship. This data can be current data and historical data. The current data and historical data are sent to the shore-based system for training the model.

[0059] In the application, when the shipboard system collects smoke information and temperature difference information in the controlled area, the system itself can use the ship smoke and open flame identification standard model and the ship fire damage assessment standard model to perform anomaly detection, and generate fire assessment information and fire rescue information based on the detection results according to its own algorithm, so as to further carry out fire registration assessment and personnel evacuation.

[0060] In the embodiment of the present application, a communication module, a data acquisition and processing module, an identification and evaluation module, and a management module are respectively provided in the shipborne system, and a communication connection is established between each module.

[0061] In the application, the communication module conducts a comprehensive assessment of the real-time monitoring information of the ship's communication equipment, the communication status of the shore-based system, and the preset communication level, selects the appropriate communication link, communication method, and communication time, and transmits data based on the determined communication link, communication method, and communication time. The data acquisition and processing module collects video data and monitoring data and processes the data, and sends the processed data to the shore-based system via the communication module for model training and construction.

[0062] In the application, the shore-based system will send the constructed ship smoke and open flame recognition standard model and ship fire damage assessment standard model to the shipborne system. The shipborne system can then use the ship smoke and open flame recognition standard model and ship fire damage assessment standard model to perform abnormal analysis, and generate fire assessment information and fire rescue information. Finally, the fire assessment information and fire rescue information will be sent to the shore-based system via the communication module for alarm and remote rescue.

[0063] In some embodiments of the present application, the shipboard system includes:

[0064] The communication management unit is used to manage the working status of the ship's combined configuration of 700M 5G, 5G, satellite communication, Beidou satellite SMS communication, WiFi network and other communication equipment, and automatically select the communication link, communication method and communication time in the most economical and efficient way based on the preset communication level, and exchange information with the ship-borne client, mobile terminal and shore-based system;

[0065] The preset communication level is to initiate information interaction with the corresponding client, mobile terminal, and shore-based system based on the monitoring, identification, and evaluation results of the data acquisition unit, cabin smoke and fire identification unit, fire situation assessment unit, etc., such as equipment failure, smoke and fire warning, fire evacuation, and help alarm.

[0066] In the embodiment of the present application, the communication management unit is used to construct a ship-borne marine broadband intelligent communication control mechanism, which performs weighted average analysis on the real-time monitoring information of the ship's communication equipment, the communication status of the shore-based system, and the preset communication level, and intelligently selects the communication link, method and time according to the analysis results to achieve timely, efficient and economical interconnection of ship-shore video and alarm information.

[0067] In the embodiment of the present application, the shipborne communication unit combines various communication modules such as 700M 5G, 5G, satellite communication, Beidou satellite SMS, etc. into one, and with the support of the communication management unit, forms a configurable, multi-channel, intelligent ship-shore communication system, which provides technical support for the full-time, efficient and economical interaction of information between ship and shore.

[0068] It can be understood that the preset communication levels are level one, two, and three. Level one, two, and three refer to fire rescue alarm, scene data collection, and standard model update, respectively. They are used to help shipborne system operators select appropriate links, time, and methods, upload alarm information, execute the control instructions of the shore-based system, and realize efficient interconnection of ship-shore information.

[0069] In some embodiments of the present application, the shipboard system includes:

[0070] A data acquisition unit is used to collect video from cameras deployed at multiple points in the control area, the working status of sensors such as audio and infrared temperature, and monitoring data, process, integrate, store and send them to the AI ​​video processing unit, and send them to the shore-based system according to instructions;

[0071] The AI ​​video processing unit is used to pre-process the video image data deployed at multiple points in the control area, and based on an adaptive noise elimination algorithm, denoise the pre-processed image data based on the cabin light noise under different light intensities, and use the denoised image data for storage and smoke and open flame identification and fire development situation assessment.

[0072] In an embodiment of the present application, the AI ​​video processing unit uses an adaptive noise elimination algorithm to denoise the cabin light noise under different illumination in the image data, provide high-quality recognition images, and store the denoised image data in real time, providing training materials for building a model for the cabin smoke and open flame recognition alarm system of the shore-based AI video.

[0073] In the application, the data acquisition unit controls the cameras and sensors in the control area to collect data, and sends the collected data to the monitoring data processing unit and the AI ​​video processing unit. The monitoring data processing unit and the AI ​​video processing unit pre-process the data and perform a secondary noise removal process on the pre-processed image data to further ensure the accuracy of the video data. Finally, the image data after noise removal and the pre-processed sensor monitoring data are sent to the shore-based system.

[0074] In some embodiments of the present application, the data acquisition and processing module includes:

[0075] Monitoring data processing unit, used to pre-process the monitoring data and send it to the shore-based system;

[0076] The AI ​​video processing unit is used to pre-process the image data and, based on an adaptive noise elimination algorithm, denoise the pre-processed image data based on the cabin light noise under different illuminations, and send the denoised image data to the shore-based system;

[0077] The data acquisition unit is communicatively connected with the monitoring data processing unit and the AI ​​video processing unit, and is used to collect image data captured by the camera and monitoring data of the sensor, and send them to the monitoring data processing unit and the AI ​​video processing unit.

[0078] In an embodiment of the present application, the monitoring data processing unit pre-processes the sensor data collected in real time based on AI video smoke and open flame pattern recognition, ship fire situation assessment, and data display service requirements, and screens out data that does not meet the requirements and data with too large a difference between the previous and next data, to meet the needs of model training and model construction.

[0079] In some embodiments of the present application, the identification and evaluation module includes:

[0080] The collection unit is used to collect smoke information and temperature difference information in the control area in real time and send it to the abnormality detection unit;

[0081] The anomaly detection unit is connected to the acquisition unit for communication, and performs anomaly detection on the smoke information and the temperature difference information based on the ship smoke and open flame identification standard model, the ship fire damage assessment standard model, and the SSD detection algorithm to obtain anomaly detection results, and sends the anomaly detection results to the fire assessment unit;

[0082] The fire assessment unit is communicated with the anomaly detection unit, and based on the anomaly detection results and a comprehensive analysis based on the fire spread speed, changes in the number of fire points, temperature distribution, and number of explosions, generates fire assessment information and fire rescue information, and sends the fire assessment information and fire rescue information to the cabin smoke and open flame identification alarm system of the shore-based AI video.

[0083] In the embodiment of the present application, the abnormality detection unit detects the smoke information and the temperature difference information respectively. In the application, the detection of the smoke information is based on the standard model for identifying smoke and open flames in ships. The SSD detection algorithm is used to perform abnormal detection on the smoke information, and the presence of smoke and open flames in the controlled area is detected in time. Once the possible presence of smoke and open flames is found, the corresponding fire assessment information and fire rescue information are immediately generated.

[0084] In the application, the anomaly detection unit detects temperature difference information based on the standard model for ship fire damage assessment, and performs weighted calculations on indicators such as the fire spread rate, changes in the number of fire points, temperature distribution, and the number of explosions to obtain anomaly detection results. It detects excessive temperature differences in the controlled area and immediately generates corresponding fire assessment information and fire rescue information.

[0085] In some embodiments of the present application, the shore-based system comprises:

[0086] A management unit, configured to receive the fire help information of the shipborne system in real time, and generate a recommended marine rescue plan based on the fire help information of the shipborne system and the fire scene video information and in accordance with the marine emergency plan;

[0087] The management unit is also used to establish and manage the maritime fire rescue emergency plan, and continuously improve the maritime rescue plan based on the rescue results.

[0088] In the embodiment of the present application, the alarm management unit sends alarm prompt information to the designated driver and passenger user terminal and shore-based system according to the preset alarm process, and applies the target tracking algorithm to highlight the occurrence point of smoke and open fire hazards on the display screen, thereby greatly improving the identification efficiency of the driver and passenger and timely avoiding navigation risks.

[0089] In the embodiment of the present application, the evacuation channel management unit provides a recommended plan for personnel fire evacuation corridors based on an intelligent path planning algorithm based on the results of smoke identification and fire situation assessment and on the basis of the ship's cabin channel map, and uses flashing channel signposts and mobile terminal display mode to indicate the best escape channel for the evacuation of ship passengers, thereby effectively reducing casualties on board.

[0090] In some embodiments of the present application, the shore-based system comprises:

[0091] The model building unit, based on the monitoring scene data collected by the data collection unit, adopts the YOLO detection algorithm and the SSD detection algorithm to build a smoke feature map, generate a smoke ImageNet data set, and build and train a ship smoke open flame, temperature difference recognition, and fire development situation assessment standard model based on the smoke, temperature difference, and fire scene ImageNet data sets, and send it to the shipborne system to continuously improve the accuracy of the shipborne system's smoke open flame, temperature difference abnormal recognition rate, and fire development situation assessment;

[0092] The model building unit is also used to train the cabin passage map and intelligent path planning algorithm and send it to the shipboard system to continuously improve the accuracy of the fire evacuation recommendation plan.

[0093] In some embodiments of the present application, the shore-based system comprises:

[0094] A management unit, configured to receive the fire help information of the shipborne system in real time, and generate a recommended marine rescue plan based on the fire help information of the shipborne system and the fire scene video information and in accordance with the marine emergency plan;

[0095] The management unit is also used to establish and manage the maritime fire rescue emergency plan, and continuously improve the maritime rescue plan based on the rescue results.

[0096] In some embodiments of the present application, the shore-based system further comprises:

[0097] The maritime rescue command unit is used to issue the fire rescue alarm information of the ship to the client and mobile client of the relevant institutions and personnel through the communication management unit, 700M 5G, 5G, satellite communication, Beidou satellite SMS communication, Internet, etc. according to the determined maritime rescue plan;

[0098] It is also used to release real-time video information of fire scenes on rescued ships to ships and rescue agencies participating in maritime rescue through 700M 5G, 5G, satellite communications, and the Internet, and integrate and display real-time status information of rescue ships and rescue personnel to achieve flattening and visualization of ship fire rescue command.

[0099] In the embodiment of the present application, the ship smoke and open flame recognition standard model construction unit is based on the YOLO detection algorithm and the SSD detection algorithm, and is based on the image data of the ship fire control place. Through a large amount of detailed classification and positioning training, a large number of multi-scale feature maps of smoke and open flames and temperature differences under different backgrounds are constructed to form an ImageNet data set, which serves as a standard model for ship smoke and open flame recognition and temperature difference recognition, and provides technical support for improving the efficiency and accuracy of smoke and open flame recognition and temperature difference recognition in the cabin smoke and open flame recognition alarm system of the shipborne AI video.

[0100] In the embodiment of the present application, the ship fire damage assessment standard model construction unit is based on the YOLO detection algorithm and the SSD detection algorithm, based on the sensor monitoring data after the previous step of preprocessing, combined with the fire point spread speed, number, fire point center temperature, number of deflagrations, and cargo flammable parameters and other indicators, through a large number of detailed classification and positioning training, to build different types of ship fire situation assessment models, to improve the efficiency and accuracy of the shipboard AI video cabin smoke open fire recognition alarm system fire development situation assessment to provide technical support;

[0101] It can be understood that the detailed classification is to perform detailed classification on the video data shot at different angles and regions of the control area, and the positioning training is to perform separate positioning training on the data of each angle and region.

[0102] The beneficial effects of the present invention are as follows: the shipboard system includes: an information acquisition unit, which is used to acquire video and temperature difference information of the controlled area; a smoke and open flame, temperature difference identification unit, which is used to identify the smoke and open flame, and temperature difference appearing in the controlled area based on the smoke and open flame, temperature difference identification model, obtain the identification result and alarm; a fire development situation assessment unit, which is used to evaluate the fire development situation of the controlled area based on the fire situation development assessment model, and decide to start the emergency evacuation alarm and send rescue alarm information to the shore-based system according to the assessment result; the shore-based system, the information acquisition unit, is used to collect the video and temperature difference information of the controlled area and the fire scene video information sent by each shipboard system, train the ship smoke and open flame identification standard model and the ship fire situation development assessment standard model, and update the model of the shipboard system based on the training result; the alarm management is used to generate a response decision based on the shipboard fire help information to improve the efficiency of marine ship fire rescue.

[0103] The working principle of the present invention may be: the cameras and sensors located in the control area inside the ship collect image information, smoke information and temperature difference information in real time, and analyze and judge through the ship smoke and open flame recognition standard model and the ship fire damage assessment standard model to judge whether a fire has occurred and determine the fire level, and send the current judgment result to the cockpit display screen, which displays the alarm information and highlights the location of the fire. Based on the location of the fire and the cabin passage map, an evacuation plan is generated, and the crew can organize the evacuation of personnel in an orderly manner, and notify the shore-based system to sound the alarm and provide remote assistance and rescue.

[0104] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A cabin smoke and open fire recognition alarm system based on AI video, characterized in that: It includes a ship-borne system arranged on a ship and a shore-based system arranged on land, wherein the ship-borne system and the shore-based system are connected via a wireless network; The shipborne system is used to obtain the smoke and fire, temperature difference video image data of the controlled area in real time, perform abnormal detection on the smoke and fire video, temperature difference, and sound information based on the ship smoke and fire recognition standard model, generate fire danger alarm information based on the detection results, perform a comprehensive assessment of the fire development situation based on the ship fire situation development assessment standard model, and issue evacuation alarm and shore-based help alarm information based on the assessment results; The evacuation alarm also provides guidance information on safe evacuation channels based on a safe channel algorithm for evacuees; The shipborne system is also used to obtain smoke, open flames, temperature difference audio and video data, and fire scene data of the controlled area collected by video cameras and infrared temperature difference sensors, and send them to the shore-based system; The shore-based system is used to receive fire help information sent by each ship, alarm and activate the shore-based rescue plan and conduct rescue dispatch command; The shore-based system is also used to collect smoke and open flames, temperature differences, audio and video data, and fire scene data sent by each ship, build and train standard models for ship smoke and open flame recognition, temperature difference recognition, and ship fire development situation assessment standard models, and send them to the shipborne systems.

2. According to claim 1, a cabin smoke and open fire identification alarm system based on AI video is characterized in that: The shipboard system comprises: The communication management unit is used to manage the working status of the ship's combined configuration of 700M 5G, 5G, satellite communication, Beidou satellite SMS communication, WiFi network and other communication equipment, and automatically select the communication link, communication method and communication time in the most economical and efficient way based on the preset communication level, and exchange information with the ship-borne client, mobile terminal and shore-based system; The preset communication level is to initiate information interaction with the corresponding client, mobile terminal, and shore-based system based on the monitoring, identification, and evaluation results of the data acquisition unit, cabin smoke and fire identification unit, fire situation assessment unit, etc., such as equipment failure, smoke and fire warning, fire evacuation, and help alarm.

3. According to claim 1, a cabin smoke and open fire identification alarm system based on AI video is characterized in that: The shipboard system comprises: A data acquisition unit is used to collect video from cameras deployed at multiple points in the control area, the working status of sensors such as audio and infrared temperature, and monitoring data, process, integrate, store and send them to the AI ​​video processing unit, and send them to the shore-based system according to instructions; The AI ​​video processing unit is used to pre-process the video image data deployed at multiple points in the control area, and based on an adaptive noise elimination algorithm, denoise the pre-processed image data based on the cabin light noise under different light intensities, and use the denoised image data for storage and smoke and open flame identification and fire development situation assessment.

4. According to claim 1, a cabin smoke and open fire identification alarm system based on AI video is characterized in that: The shipboard system comprises: A smoke and open fire identification unit is communicatively connected with the AI ​​video processing unit and the data acquisition unit, performs abnormality detection on the smoke information and the temperature difference information based on the ship smoke and open fire identification standard model and the SSD detection algorithm, obtains an abnormality detection result, and sends the abnormality detection result to the management unit; The fire situation development assessment unit is connected to the abnormality detection unit and the data acquisition unit for communication, and generates a fire situation development assessment result based on the abnormality detection result, the fire point spread speed, the change in the number of fire points, the temperature distribution, and the number of deflagrations, and sends the abnormality detection result to the management unit; The management unit is used to issue smoke, open fire, and temperature difference alarm information according to the identification result of the smoke and open fire identification unit, and the alarm information includes the scene video of the smoke and open fire and the location information of the cabin where the smoke and open fire occur; The management module is also used to issue personnel evacuation alarm information and send help information to the shore-based system based on the fire development situation assessment results; The personnel evacuation alarm information includes a recommended personnel evacuation channel plan generated based on the cabin channel map and the intelligent path planning algorithm; The fire rescue information includes the name and location of the fire ship and the fire scene video information.

5. According to the AI ​​video-based cabin smoke and fire identification alarm system of claim 1, it is characterized in that: The shore-based system includes: The communication management unit is used to manage the working status of the communication equipment such as 700M 5G, 5G, satellite communication, Beidou satellite SMS communication, and Internet configured in the shore-based system combination, and automatically select the communication link, communication method and communication time in the most economical and efficient way based on the preset communication level to exchange information with various shipborne systems; The preset communication level is to activate the shore-based system emergency plan based on the help alarm of the shipborne system, and issue ship fire help alarm information to relevant clients and mobile clients. The help information includes the name of the burning ship, the real-time location of the ship, the name of the burning cabin and the fire scene video information.

6. The AI ​​video-based cabin smoke and fire identification alarm system according to claim 1 is characterized in that: The shore-based system includes: A data acquisition unit connected to the communication management unit, used to receive fire help information issued by various shipborne systems in real time and send it to the management unit; The data acquisition unit is also used to collect, process, integrate and store multi-point camera videos deployed in the control areas of each shipboard system, monitoring data of audio, infrared temperature and other sensors, and fire scene video data for training AI video's cabin smoke and open flame recognition and fire development situation assessment standard model.

7. The AI ​​video-based cabin smoke and fire identification alarm system according to claim 1 is characterized in that: The shore-based system includes: The model building unit, based on the monitoring scene data collected by the data collection unit, adopts the YOLO detection algorithm and the SSD detection algorithm to build a smoke feature map, generate a smoke ImageNet data set, and build and train a ship smoke open flame, temperature difference recognition, and fire development situation assessment standard model based on the smoke, temperature difference, and fire scene ImageNet data sets, and send it to the shipborne system to continuously improve the accuracy of the shipborne system's smoke open flame, temperature difference abnormal recognition rate, and fire development situation assessment; The model building unit is also used to train the cabin passage map and intelligent path planning algorithm and send it to the shipboard system to continuously improve the accuracy of the fire evacuation recommendation plan.

8. The AI ​​video-based cabin smoke and fire identification alarm system according to claim 1 is characterized in that: The shore-based system includes: A management unit, configured to receive the fire help information of the shipborne system in real time, and generate a recommended marine rescue plan based on the fire help information of the shipborne system and the fire scene video information and in accordance with the marine emergency plan; The management unit is also used to establish and manage the maritime fire rescue emergency plan, and continuously improve the maritime rescue plan based on the rescue results.

9. The AI ​​video-based cabin smoke and fire identification alarm system according to claim 1 is characterized in that: The shore-based system also includes: The maritime rescue command unit is used to issue the fire rescue alarm information of the ship to the clients and mobile clients of the relevant institutions and personnel through the communication management unit, 700M5G, 5G, satellite communication, Beidou satellite SMS communication, Internet, etc. according to the determined maritime rescue plan; It is also used to release real-time video information of fire scenes on rescued ships to ships and rescue agencies participating in maritime rescue through 700M 5G, 5G, satellite communications, and the Internet, and integrate and display real-time status information of rescue ships and rescue personnel to achieve flattening and visualization of ship fire rescue command.

Images