Fire fighting access occupancy event detection and acousto-optic warning processing device and system

Abstract

The invention relates to the technical field of fire fighting management, and relates to a fire fighting access occupancy event detection and acousto-optic warning processing device and system. Technical problems are as follows: at present, professional and effective identification means for fire fighting access occupancy events are lacked, intelligent automatic detection modes for occupancy objects and vehicles on fire fighting accesses are lacked, and hierarchical linkage with a plurality of responsibility subjects and emergency response integration are not tight enough; according to the technical scheme, the fire fighting access occupancy event detection and acousto-optic warning processing device and system comprise a video access acquisition module, an event detection module and an intelligent linkage processing module; the fire fighting access occupancy event is intelligently identified by adopting an intelligent visual method, and the event is handled through a grading and classifying strategy in combination with the responsibility of a multi-responsibility subject and a rapid handling mechanism. The problems that fire fighting access occupancy events are passively responded and hidden danger risks are not found in advance are solved, potential hidden danger risks are pre-warned through intelligent recognition, a risk prevention system is perfected, and the prevention capacity is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of fire management, and is mainly directed to a fire passage occupancy event detection and sound and light warning disposal device and system. Background Art

[0002] Fire passages are the "life passages" for fire fighting and rescue, and should be kept open at all times to avoid delays in rescue due to occupation, which may lead to loss of life and property. No unit or individual may occupy, block or close evacuation passages, emergency exits or fire truck passages.

[0003] At present, there is a problem of insufficient technical means in the intelligent automatic detection of fire passage occupation; first, in image recognition technology, the existing algorithms have limited adaptability to complex scenes and changing environments, and it is difficult to accurately identify occupied objects and vehicles in fire passages; second, the application of sensors is limited; traditional sensors such as infrared sensors, pressure sensors, etc., are difficult to meet the needs of comprehensive and accurate detection of fire passage occupation in terms of detection range and accuracy; currently, except for the parking ball to monitor illegal parking behavior, there is no mature product to achieve a comprehensive and accurate judgment of fire passage occupation; finally, the existing intelligent detection system is not closely integrated with the emergency response mechanism of the responsible entity, public security, housing and construction, urban management, fire protection and other departments, resulting in the inability to quickly and effectively initiate corresponding disposal measures for occupation incidents.

[0004] Therefore, the existing intelligent automatic detection system for fire passage occupation has limited adaptability to complex scenes and changing environments, and it is difficult to accurately identify occupied objects and vehicles on the fire passage. In addition, the integration with the emergency response mechanisms of the responsible entities, public security, housing and construction, urban management, fire protection and other departments is not tight enough, resulting in the problem that the corresponding disposal measures cannot be initiated quickly and effectively for occupation incidents. A fire passage occupation event detection and sound and light warning disposal device and system solves the risk of occupation, blockage, and closure of fire truck passages that cannot be discovered in time and the risk of long disposal time for fire passage occupation incidents through intelligent fire passage occupation event discovery capabilities. Through hierarchical and classified disposal, the power of property and street grassroots units is mobilized to achieve a rapid disposal mechanism, solve the passive response to fire passage occupation incident disposal, and lack of prior hidden danger discovery risks. Through multi-source data research and analysis, potential hidden dangers are intelligently identified and warned, and disposed of in advance, the risk prevention system is improved, and the prior prevention capability is enhanced. Summary of the invention

[0005] In order to overcome the limited adaptability of the intelligent automatic detection system for fire passage occupancy under the existing technology to complex scenarios and changing environments, it is difficult to accurately identify occupied objects and vehicles on the fire passage, and the integration with the emergency response mechanisms of the responsible entities, public security, housing and construction, urban management, fire protection and other departments is not tight enough, resulting in the inability to quickly and effectively initiate corresponding disposal measures for occupation incidents.

[0006] The technical solution of the present invention is: a fire passage occupancy event detection and sound and light warning disposal device, including a panoramic video surveillance camera, a sound and light warning device is arranged on one side of the panoramic video surveillance camera, and a video capture device is arranged below the panoramic video surveillance camera.

[0007] Preferably, the video footage of the fire escape area captured by a panoramic video surveillance camera can provide a comprehensive perspective to ensure that all parts of the fire escape are effectively monitored so as to promptly discover and identify occupancy events. By real-time shooting and transmitting video streams, raw data is provided to the event detection system, so that the system can use intelligent visual algorithms to analyze the video and detect the occupancy of the fire escape. The video capture device is used in conjunction with the panoramic video surveillance camera to automatically or manually capture high-definition pictures or video clips of the scene when a fire escape occupancy event is detected. The video capture device can quickly record the specific circumstances of the occupancy event and on-site evidence, providing strong support for subsequent disposal and recording. The property center staff can more accurately understand the incident situation by viewing these on-site photos and make corresponding handling decisions. The sound and light warning device uses flashing warning lights and plays expulsion voices to persuade occupants to leave on the spot. This zero-delay intervention measure helps to quickly attract the attention of occupants and prompt them to evacuate the fire escape in time.

[0008] Preferably, a fire passage occupancy event detection and sound and light warning disposal system comprises:

[0009] The video acquisition module is responsible for acquiring the video stream of the fire passage area and providing real-time video data to the event detection module for intelligent analysis;

[0010] The event monitoring module is used to detect the occupancy of fire passages in the video stream based on the deep learning intelligent visual algorithm of the neural source network, automatically identify the type and urgency of the occupancy event, and provide decision-making basis for the intelligent linkage disposal module;

[0011] The intelligent linkage disposal module is used to automatically drive the sound and light warning equipment on site according to the results of the event detection module, achieving zero-delay intervention and persuasion.

[0012] Preferably, the video acquisition module uses high-definition, wide-angle cameras, including fixed gun cameras and rotatable ball cameras, to cover every corner of the fire passage. The camera has a night vision function and can shoot clearly at night or in low light conditions. By placing cameras according to the layout of the fire passage, deploying cameras at regular monitoring points and monitoring blind spots to achieve full coverage of the fire passage, the video acquisition module transmits the real-time captured video stream wirelessly to the central processing unit of the system.

[0013] Preferably, the video acquisition module comprises the following steps when working:

[0014] S401: Deploy fixed gun cameras and dome cameras at key locations of fire passages;

[0015] S402: The camera is connected to the system network via wireless mode;

[0016] S403: The camera starts to continuously capture video images of the fire passage area;

[0017] S404: The image sensor built into the camera converts the optical signal into an electrical signal to generate a digital video stream;

[0018] S405: Encode the video stream using a video encoding algorithm to remove redundant information and reduce the amount of data;

[0019] S406: The encoded video data is transmitted to the back-end server of the system via the network;

[0020] S407: The backend server receives and decodes the video data, performs real-time monitoring and display, and saves the video data to a storage device.

[0021] Preferably, the event detection module performs intelligent analysis on the video stream transmitted by the video acquisition module to identify occupancy events on the fire escape. The event detection module is based on an intelligent visual algorithm based on deep learning of a neural network. After training, it can identify various occupancy situations such as parking vehicles, stacking of debris, setting up obstacles, setting up stalls, constructing temporary buildings, and installing facilities that affect traffic. The event detection module classifies the type and urgency of occupancy events and generates warning signals of different levels.

[0022] Preferably, the event detection module includes the following steps when performing event detection and identification:

[0023] S601: pre-processing the collected video data, including cropping, adjusting resolution, and unifying frame rates;

[0024] S602: Determine a marking method for the category and location of the occupancy event according to a marking specification;

[0025] S603: Use the labeling tool LabelImg to perform fine labeling on the video data;

[0026] S604: Verify and correct the annotation results;

[0027] S605: Select TensorFlow deep learning framework and Python programming language;

[0028] S606: Construct a neural network structure, including multiple convolutional layers, pooling layers, and fully connected layers;

[0029] S607: Configure network parameters, including learning rate and batch size;

[0030] S608: Convert the annotated video data into a format acceptable to the model, including frame image sequences and video clips;

[0031] S609: Divide the converted data into a training set and a validation set;

[0032] S610: Train the model using the training set data, adjust the model parameters through the back propagation algorithm, and minimize the loss function;

[0033] S611: During training, regularly evaluate model performance using validation set data to monitor overfitting and underfitting;

[0034] S612: Adjust the model structure and training parameters based on the performance feedback of the validation set;

[0035] S613: Obtain the video stream of the fire passage from the surveillance camera in real time;

[0036] S614: Decoding and preprocessing the video stream, including adjusting the resolution and unifying the frame rate;

[0037] S615: Input the preprocessed video stream into the trained deep learning model;

[0038] S616: extracting features from the video frame using the convolution layer and the pooling layer in the model to obtain feature information of the occupancy event;

[0039] S617; input the extracted feature information into the fully connected layer of the model;

[0040] S618: Classify occupancy events using the output of the fully connected layer;

[0041] S619: According to the classification result, the location and category of the occupancy event are marked in the video frame.

[0042] Preferably, the event detection module includes the following steps when classifying the type and urgency of the occupancy event and generating warning signals of different levels:

[0043] S701: Classifying events according to the type of occupied objects;

[0044] S702: Identify special occupancy situations, including whether they involve key areas such as fire escape routes and emergency exits;

[0045] S703: Assess the occupied area, volume or quantity and determine the severity of the occupancy;

[0046] S704: Assess the length of the occupation, including whether it causes prolonged or sustained obstruction;

[0047] S705: Setting thresholds of different urgency levels, including based on occupancy type, occupancy level, and key area influencing factors;

[0048] S706: Classify the urgency into multiple levels, such as “low”, “medium”, “high”, and “urgent”;

[0049] S707: Determine the urgency level of the event according to the occupancy classification identification result;

[0050] S708: define warning signals corresponding to different emergency levels, including color, flashing frequency, and sound;

[0051] S709: Select a corresponding warning signal according to the emergency level;

[0052] S710: Prepare the output format of the warning signal, including LED screen display, mobile phone APP push, SMS notification, and voice alarm;

[0053] S711: Determine the users and departments that need to receive the warning information based on the nature and location of the event;

[0054] S712: Push the warning information to target users and departments through the preset communication channels.

[0055] Preferably, through deep learning algorithms, the system can continuously learn and optimize recognition capabilities, improve adaptability to different scenarios and complex environments, and at the same time, the system can also classify the type and urgency of occupancy events and generate warning signals of different levels.

[0056] As a preferred option, the intelligent linkage disposal module is responsible for automatically triggering the corresponding disposal measures after detecting the fire passage occupation event, and automatically pushes warning information to the property center, public security and other departments according to the urgency of the event and the disposal results, so as to take further disposal measures in time. The general strategy is as follows: After the intelligent linkage disposal module detects the occupation event, it triggers the sound and light warning equipment on site to conduct preliminary persuasion; if the road is still occupied within 5 minutes, it will push the property department to handle it on site within 10 minutes, and send notification information to the car owner or relevant personnel through the public security APP and SMS system; conduct an occupation inspection every 30 minutes. If it is still not handled, the event will be pushed to the business system of the public security department, and the relevant personnel will be dispatched to the scene for disposal.

[0057] Preferably, the intelligent linkage disposal module comprises the following steps when working:

[0058] S901: Send a control instruction to the intelligent linkage disposal module according to the warning level to activate the corresponding sound and light warning device;

[0059] S902: At the scene of the occupation incident, the occupants are persuaded to leave through flashing warning lights, ground projection, and voice broadcasting. The warning content includes the illegality of the occupation behavior and the instruction to evacuate immediately;

[0060] S903: Continuously monitor the occupation event site to observe whether the occupation behavior has stopped;

[0061] S904: If the occupation behavior has not stopped, a reminder of illegal occupation of the fire passage will be sent to the vehicle owner or relevant personnel through the public safety SMS system, informing them to evacuate as soon as possible;

[0062] S905: If the property management cannot effectively handle the occupation incident, the incident information will be pushed to the business system of the public security department, requesting relevant personnel to go to the scene to persuade or tow away the vehicles or items occupying the fire passage;

[0063] S906: Record event information, and record the entire process of the event in the system, including the time, location, occupancy, disposal measures and results.

[0064] Preferably, a fire passage occupancy event detection and sound and light warning disposal device and system comprises the following steps when working:

[0065] S1: The surveillance camera continuously collects video streams from the fire escape area;

[0066] S2: The event detection module detects the occupancy of the fire passage location in the video stream and identifies the type and urgency of the occupancy event;

[0067] S3: Generate different levels of warnings according to the type and urgency of the occupancy event, and push the warning information to different users according to the level;

[0068] S4: At locations where conditions permit, immediately activate the on-site sound and light warning equipment to warn the occupants;

[0069] S5: If the occupant does not respond, the warning information will be pushed to the property management center, and the property management center staff will check the photos of the incident scene and guide relevant personnel to rush to the scene to evacuate;

[0070] S6: If the occupant has left after the relevant personnel arrive at the scene and cannot be removed on the spot, a text message reminder of illegal occupation of the fire passage will be sent to the vehicle owner or relevant personnel, informing them to evacuate as soon as possible;

[0071] S7: If the property management company is unable to handle the occupation incident, an early warning will be issued to the public security department, who will dispatch relevant personnel to the scene to persuade or tow away the vehicles or items occupying the fire escape.

[0072] Beneficial effects of the present invention:

[0073] 1. Compared with the existing technology, the intelligent automatic detection system for fire passage occupation has limited adaptability to complex scenes and changing environments, and it is difficult to accurately identify occupied objects and vehicles on the fire passage. In addition, the integration with the emergency response mechanism of the responsible entity, public security, housing and construction, urban management, fire protection and other departments is not tight enough, resulting in the inability to quickly and effectively initiate corresponding disposal measures for occupation events; the system solves the risk of not being able to timely discover the occupation, blockage, and closure of fire truck passages and the risk of long disposal time for fire passage occupation events through the intelligent fire passage occupation event discovery capability. Through hierarchical and classified disposal, the power of property and street grassroots units is mobilized to realize a rapid disposal mechanism, solve the passive response to the disposal of fire passage occupation events, and the lack of prior hidden danger discovery risks. Through multi-source data research and analysis, potential hidden dangers are intelligently identified and warned, and disposed in advance, the risk prevention system is improved, and the pre-prevention capability is enhanced;

[0074] 2. Preprocess the collected video data through the event detection module, and use the labeling tool LabelImg to finely label the video data; select the TensorFlow deep learning framework and Python programming language to build a neural network structure, and convert the labeled video data into a format acceptable to the model, including frame image sequences and video clips; divide the converted data into a training set and a validation set; use the training set data to train the model, adjust the model parameters through the back propagation algorithm, and minimize the loss function; during the training process, regularly use the validation set data to evaluate the model performance to monitor overfitting and underfitting; adjust the model structure and training parameters based on the performance feedback of the validation set; obtain the video stream of the fire escape from the surveillance camera in real time; decode and preprocess the video stream; input the preprocessed video stream into the trained deep learning model; use the convolutional layer and pooling layer in the model to extract features from the video frame to obtain the feature information of the occupancy event; input the extracted feature information into the fully connected layer of the model; use the output of the fully connected layer to classify the occupancy event; based on the classification result, mark the location and category of the occupancy event in the video frame. Classify events according to the type of occupied objects; identify special occupancy situations; assess the occupied area, volume or quantity to determine the severity of the occupancy; assess the duration of occupancy, including whether it causes long-term or continuous obstruction; set thresholds for different levels of urgency; divide the urgency into multiple levels, such as "low", "medium", "high" and "urgent"; determine the urgency level of the event based on the occupancy classification and identification results; define warning signals corresponding to different levels of urgency, including color, flashing frequency, and sound; select the corresponding warning signal according to the urgency level; prepare the output format of the warning signal; determine the users and departments who need to receive the warning information based on the nature and location of the event; push the warning information to the target users and departments through the preset communication channels; thereby solving the risk of not being able to timely discover the occupation, blockage, and closure of fire truck passages and the risk of long handling time for fire passage occupation events through the intelligent fire passage occupation event detection capability;

[0075] 3. According to the warning level, the intelligent linkage disposal module sends control instructions to the intelligent linkage disposal module to start the corresponding sound and light warning equipment; at the scene of the occupation incident, the occupation behavior is persuaded to leave on the spot through flashing warning lights, ground projection, and voice broadcasting. The warning content includes the illegality of the occupation behavior and the instruction to evacuate immediately; the occupation incident scene is continuously monitored to observe whether the occupation behavior has stopped; if the occupation behavior has not stopped, the public safety SMS system will send a reminder of illegal occupation of the fire passage to the car owner or relevant personnel, informing them to evacuate as soon as possible; if the property cannot effectively deal with the occupation incident, the incident message will be sent to the owner. The information is pushed to the business system of the public security department, requesting relevant personnel to go to the scene to persuade or tow away the vehicles or items occupying the fire passage; record the event information, and record the whole process information of the event in the system, including the time, location, occupancy, handling measures and results; thus, through hierarchical and classified handling, mobilize the power of property and street-level grassroots units, realize a rapid handling mechanism, solve the passive response to the handling of fire passage occupation incidents, lack of prior hidden danger discovery risks, through multi-source data research and analysis, intelligent identification and warning of potential hidden dangers, handling in advance, improving the risk prevention system, and enhancing pre-prevention capabilities. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] Figure 1 Shown is a schematic diagram of the three-dimensional structure of the fire passage occupancy event detection and sound and light warning disposal device of the present invention;

[0077] Figure 2 Shown is a schematic diagram of a fire passage occupancy event detection and sound and light warning device and system workflow of the present invention;

[0078] Explanation of the accompanying drawings: 1. Panoramic video surveillance camera; 2. Sound and light warning device; 3. Video capture equipment. DETAILED DESCRIPTION

[0079] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0080] See also Figure 1-2 The present invention provides an embodiment: a fire passage occupancy event detection and sound and light warning disposal device, comprising a panoramic video surveillance camera 1, a sound and light warning device 2 and a video capture device 3; the sound and light warning device 2 is arranged on one side of the panoramic video surveillance camera 1, and the video capture device 3 is arranged below the panoramic video surveillance camera 1.

[0081] Preferably, the panoramic video surveillance camera 1 is used to capture and monitor the video footage of the fire passage area. The panoramic video surveillance camera 1 can provide a comprehensive viewing angle to ensure that all parts of the fire passage are effectively monitored so as to timely discover and identify occupation events. By real-time shooting and transmitting video streams, the panoramic video surveillance camera 1 provides raw data for the event detection system, so that the system can use intelligent visual algorithms to analyze the video and thus detect the occupation of the fire passage. The video capture device 3 is used in conjunction with the panoramic video surveillance camera 1 to automatically or manually capture high-definition pictures or video clips of the scene when a fire passage occupation event is detected. The video capture device 3 can quickly record the specific circumstances of the occupation event and on-site evidence, providing strong support for subsequent disposal and recording. The property center personnel can more accurately understand the event situation by viewing these on-site photos and make corresponding processing decisions. The sound and light warning device 2 uses flashing warning lights and plays expulsion voices to persuade the occupants to leave on the spot. This zero-delay intervention measure helps to quickly attract the attention of the occupants and prompt them to evacuate the fire passage in time.

[0082] Preferably, a fire passage occupancy event detection and sound and light warning disposal system comprises:

[0083] The video acquisition module is responsible for acquiring the video stream of the fire passage area and providing real-time video data to the event detection module for intelligent analysis;

[0084] The event monitoring module is used to detect the occupancy of fire passages in the video stream based on the deep learning intelligent visual algorithm of the neural source network, automatically identify the type and urgency of the occupancy event, and provide decision-making basis for the intelligent linkage disposal module;

[0085] The intelligent linkage disposal module is used to automatically drive the sound and light warning equipment on site according to the results of the event detection module, achieving zero-delay intervention and persuasion.

[0086] Preferably, the video acquisition module uses high-definition, wide-angle cameras, including fixed gun cameras and rotatable ball cameras, to cover every corner of the fire passage. The camera has a night vision function and can shoot clearly at night or in low light conditions. By placing cameras according to the layout of the fire passage, deploying cameras at regular monitoring points and monitoring blind spots to achieve full coverage of the fire passage, the video acquisition module transmits the real-time captured video stream wirelessly to the central processing unit of the system.

[0087] Preferably, the video acquisition module comprises the following steps when working:

[0088] S401: Deploy fixed gun cameras and dome cameras at key locations of fire passages;

[0089] S402: The camera is connected to the system network via wireless mode;

[0090] S403: The camera starts to continuously capture video images of the fire passage area;

[0091] S404: The image sensor built into the camera converts the optical signal into an electrical signal to generate a digital video stream;

[0092] S405: Encode the video stream using a video encoding algorithm to remove redundant information and reduce the amount of data;

[0093] S406: The encoded video data is transmitted to the back-end server of the system via the network;

[0094] S407: The backend server receives and decodes the video data, performs real-time monitoring and display, and saves the video data to a storage device.

[0095] Preferably, the event detection module performs intelligent analysis on the video stream transmitted by the video acquisition module to identify occupancy events on the fire escape. The event detection module is based on an intelligent visual algorithm based on deep learning of a neural network. After training, it can identify various occupancy situations such as parking vehicles, stacking of debris, setting up obstacles, setting up stalls, constructing temporary buildings, and installing facilities that affect traffic. The event detection module classifies the type and urgency of occupancy events and generates warning signals of different levels.

[0096] Preferably, the event detection module includes the following steps when performing event detection and identification:

[0097] S601: pre-processing the collected video data, including cropping, adjusting resolution, and unifying frame rates;

[0098] S602: Determine a marking method for the category and location of the occupancy event according to a marking specification;

[0099] S603: Use the labeling tool LabelImg to perform fine labeling on the video data;

[0100] S604: Verify and correct the annotation results;

[0101] S605: Select TensorFlow deep learning framework and Python programming language;

[0102] S606: Construct a neural network structure, including multiple convolutional layers, pooling layers, and fully connected layers;

[0103] S607: Configure network parameters, including learning rate and batch size;

[0104] S608: Convert the annotated video data into a format acceptable to the model, including frame image sequences and video clips;

[0105] S609: Divide the converted data into a training set and a validation set;

[0106] S610: Train the model using the training set data, adjust the model parameters through the back propagation algorithm, and minimize the loss function;

[0107] S611: During training, regularly evaluate model performance using validation set data to monitor overfitting and underfitting;

[0108] S612: Adjust the model structure and training parameters based on the performance feedback of the validation set;

[0109] S613: Obtain the video stream of the fire passage from the surveillance camera in real time;

[0110] S614: Decoding and preprocessing the video stream, including adjusting the resolution and unifying the frame rate;

[0111] S615: Input the preprocessed video stream into the trained deep learning model;

[0112] S616: extracting features from the video frame using the convolution layer and the pooling layer in the model to obtain feature information of the occupancy event;

[0113] S617; input the extracted feature information into the fully connected layer of the model;

[0114] S618: Classify occupancy events using the output of the fully connected layer;

[0115] S619: According to the classification result, the location and category of the occupancy event are marked in the video frame.

[0116] Preferably, the event detection module includes the following steps when classifying the type and urgency of the occupancy event and generating warning signals of different levels:

[0117] S701: Classifying events according to the type of occupied objects;

[0118] S702: Identify special occupancy situations, including whether they involve key areas such as fire escape routes and emergency exits;

[0119] S703: Assess the occupied area, volume or quantity and determine the severity of the occupancy;

[0120] S704: Assess the length of the occupation, including whether it causes prolonged or sustained obstruction;

[0121] S705: Setting thresholds of different urgency levels, including based on occupancy type, occupancy level, and key area influencing factors;

[0122] S706: Classify the urgency into multiple levels, such as “low”, “medium”, “high”, and “urgent”;

[0123] S707: Determine the urgency level of the event according to the occupancy classification identification result;

[0124] S708: define warning signals corresponding to different emergency levels, including color, flashing frequency, and sound;

[0125] S709: Select a corresponding warning signal according to the emergency level;

[0126] S710: Prepare the output format of the warning signal, including LED screen display, mobile phone APP push, SMS notification, and voice alarm;

[0127] S711: Determine the users and departments that need to receive the warning information based on the nature and location of the event;

[0128] S712: Push the warning information to target users and departments through the preset communication channels.

[0129] Preferably, through deep learning algorithms, the system can continuously learn and optimize recognition capabilities, improve adaptability to different scenarios and complex environments, and at the same time, the system can also classify the type and urgency of occupancy events and generate warning signals of different levels.

[0130] Preferably, the intelligent linkage disposal module is responsible for automatically triggering corresponding disposal measures after detecting the fire passage occupation incident. The intelligent linkage disposal module realizes zero-delay on-site persuasion by controlling the on-site sound and light warning equipment, and automatically pushes warning information to the property center, public security and other departments according to the urgency of the incident and the disposal results, so as to take further disposal measures in time; after the system detects the occupation incident, the intelligent linkage disposal module triggers the on-site sound and light warning equipment for preliminary persuasion; if unsuccessful, a notification message is sent to the car owner or relevant personnel through the public security SMS system; if it still cannot be handled, the event will be pushed to the public security department's business system.

[0131] Preferably, the intelligent linkage disposal module comprises the following steps when working:

[0132] S901: Send a control instruction to the intelligent linkage disposal module according to the warning level to activate the corresponding sound and light warning device;

[0133] S902: At the scene of the occupation incident, the occupants are persuaded to leave through flashing warning lights, ground projection, and voice broadcasting. The warning content includes the illegality of the occupation behavior and the instruction to evacuate immediately;

[0134] S903: Continuously monitor the occupation event site to observe whether the occupation behavior has stopped;

[0135] S904: If the occupation behavior has not stopped, a reminder of illegal occupation of the fire passage will be sent to the vehicle owner or relevant personnel through the public safety SMS system, informing them to evacuate as soon as possible;

[0136] S905: If the property management cannot effectively handle the occupation incident, the incident information will be pushed to the business system of the public security department, requesting relevant personnel to go to the scene to persuade or tow away the vehicles or items occupying the fire passage;

[0137] S906: Record event information, and record the entire process of the event in the system, including the time, location, occupancy, disposal measures and results.

[0138] Preferably, the fire passage occupancy event detection and sound and light warning disposal system comprises the following steps when operating:

[0139] S1: The surveillance camera continuously collects video streams from the fire escape area;

[0140] S2: The event detection module detects the occupancy of the fire passage location in the video stream and identifies the type and urgency of the occupancy event;

[0141] S3: Generate different levels of warnings according to the type and urgency of the occupancy event, and push the warning information to different users according to the level;

[0142] S4: At locations where conditions permit, immediately activate the on-site sound and light warning equipment to warn the occupants;

[0143] S5: If the occupant does not respond, the warning information will be pushed to the property management center, and the property management center staff will check the photos of the incident scene and guide relevant personnel to rush to the scene to evacuate;

[0144] S6: If the occupant has left after the relevant personnel arrive at the scene and cannot be removed on the spot, a text message reminder of illegal occupation of the fire passage will be sent to the vehicle owner or relevant personnel, informing them to evacuate as soon as possible;

[0145] S7: If the property management company is unable to handle the occupation incident, an early warning will be issued to the public security department, who will dispatch relevant personnel to the scene to persuade or tow away the vehicles or items occupying the fire escape.

[0146] The embodiments of the present invention are described in detail above with reference to the accompanying drawings, but the present invention is not limited to the above embodiments, and various changes can be made within the knowledge scope of those skilled in the art without departing from the spirit of the present invention.

Claims

1. A fire passage occupancy event detection and sound and light warning device, comprising a panoramic video surveillance camera (1); characterized in that: An audible and visual warning device (2) is arranged on one side of the panoramic video surveillance camera (1), and a video capture device (3) is arranged below the panoramic video surveillance camera (1).

2. A fire passage occupancy event detection and sound and light warning disposal system, comprising: The video acquisition module is used to obtain the video stream captured by the front-end equipment in the fire channel area and provide real-time video data for the event detection module for intelligent analysis; The event monitoring module is used to detect the occupancy of fire passages in the video stream based on the deep learning intelligent visual algorithm of the neural source network, automatically identify the type and urgency of the occupancy event, and provide decision-making basis for the intelligent linkage disposal module; The linkage handling module is used to handle events using different handling strategies based on the results of the event detection module and to automatically review the results.

3. A fire passage occupancy event detection and sound and light warning handling system according to claim 1, characterized in that: The video acquisition module uses high-definition, wide-angle cameras, including fixed gun cameras and rotatable ball cameras, to cover every corner of the fire passage. The camera has night vision function and can shoot clearly at night or in low light conditions. By placing cameras according to the layout of the fire passage, deploying cameras at regular monitoring points and monitoring blind spots to achieve comprehensive coverage of the fire passage, the video acquisition module transmits the real-time captured video stream wirelessly to the system's central processing unit.

4. According to the fire passage occupancy event detection and sound and light warning disposal system of claim 3, when the video acquisition module is working, it includes the following steps: S401: Deploy fixed gun cameras and dome cameras at key locations of fire passages; S402: The camera is connected to the system network via wireless mode; S403: The camera starts to continuously capture video images of the fire passage area; S404: The image sensor built into the camera converts the optical signal into an electrical signal to generate a digital video stream; S405: Encode the video stream using a video encoding algorithm to remove redundant information and reduce the amount of data; S406: The encoded video data is transmitted to the back-end server of the system via the network; S407: The backend server receives and decodes the video data, performs real-time monitoring and display, and saves the video data to a storage device.

5. A fire passage occupancy event detection and sound and light warning handling system according to claim 4, characterized in that: The event detection module performs intelligent analysis on the video stream transmitted by the video acquisition module to identify occupancy events on the fire escape. The event detection module is based on an intelligent visual algorithm based on deep learning of a neural network. After training, it can identify various occupancy situations such as parking vehicles, stacking of debris, setting up obstacles, setting up stalls, constructing temporary buildings, and installing facilities that affect traffic. The event detection module classifies the type and urgency of occupancy events and generates warning signals of different levels.

6. A fire passage occupancy event detection and sound and light warning handling system according to claim 5, characterized in that: When the event detection module performs event detection and identification, the following steps are included: S601: pre-processing the collected video data, including cropping, adjusting resolution, and unifying frame rates; S602: Determine a marking method for the category and location of the occupancy event according to a marking specification; S603: Use the labeling tool LabelImg to perform fine labeling on the video data; S604: Verify and correct the annotation results; S605: Select the TensorFlow deep learning framework and Python programming language; S606: Construct a neural network structure, including multiple convolutional layers, pooling layers, and fully connected layers; S607: Configure network parameters, including learning rate and batch size; S608: Convert the annotated video data into a format acceptable to the model, including frame image sequences and video clips; S609: Divide the converted data into a training set and a validation set; S610: Train the model using the training set data, adjust the model parameters through the back propagation algorithm, and minimize the loss function; S611: During training, regularly evaluate model performance using validation set data to monitor overfitting and underfitting; S612: Adjust the model structure and training parameters based on the performance feedback of the validation set; S613: Obtain the video stream of the fire passage from the surveillance camera in real time; S614: decoding and preprocessing the video stream, including adjusting the resolution and unifying the frame rate; S615: Input the preprocessed video stream into the trained deep learning model; S616: extracting features from the video frame using the convolution layer and the pooling layer in the model to obtain feature information of the occupancy event; S617; input the extracted feature information into the fully connected layer of the model; S618: Classify occupancy events using the output of the fully connected layer; S619: According to the classification result, the location and category of the occupancy event are marked in the video frame.

7. A fire passage occupancy event detection and sound and light warning handling system according to claim 6, characterized in that: The event detection module includes the following steps when classifying the type and urgency of occupancy events and generating warning signals of different levels: S701: Classifying events according to the type of occupied objects; S702: Identify special occupancy situations, including whether they involve key areas such as fire escapes and emergency exits; S703: Assess the occupied area, volume or quantity and determine the severity of the occupancy; S704: Assess the length of the occupation, including whether it causes prolonged or sustained obstruction; S705: Setting thresholds of different urgency levels, including based on occupancy type, occupancy level, and key area influencing factors; S706: Classify the urgency into multiple levels, such as "low", "medium", "high", and "urgent"; S707: Determine the urgency level of the event according to the occupancy classification identification result; S708: define warning signals corresponding to different emergency levels, including color, flashing frequency, and sound; S709: Select a corresponding warning signal according to the emergency level; S710: Prepare the output format of the warning signal, including LED screen display, mobile phone APP push, SMS notification, and voice alarm; S711: Determine the users and departments that need to receive the warning information based on the nature and location of the event; S712: Push the warning information to target users and departments through the preset communication channels.

8. A fire passage occupancy event detection and sound and light warning handling system according to claim 7, characterized in that: The intelligent linkage disposal module is responsible for automatically triggering the corresponding disposal measures after detecting the fire passage occupation event, and automatically pushes warning information to the property center, public security and other departments according to the urgency of the event and the disposal results, so that further disposal measures can be taken in time. The general strategy is as follows: After the intelligent linkage disposal module detects the occupation event, it triggers the sound and light warning equipment on site to conduct preliminary persuasion; if the road is still occupied within 5 minutes, it will push the property department to handle it on site within 10 minutes, and send notification information to the car owner or relevant personnel through the public security APP and SMS system; the road occupation inspection will be carried out after 30 minutes. If it is still not handled, the event will be pushed to the business system of the public security department, and the relevant personnel will be dispatched to the scene for disposal.

9. A fire passage occupancy event detection and sound and light warning handling system according to claim 8, characterized in that: When the intelligent linkage disposal module is working, it includes the following steps: S901: Send a control instruction to the intelligent linkage disposal module according to the warning level to activate the corresponding sound and light warning device; S902: At the scene of the occupation incident, the occupants are persuaded to leave through flashing warning lights, ground projection, and voice broadcasting. The warning content includes the illegality of the occupation behavior and the instruction to evacuate immediately; S903: Continuously monitor the occupation event site to observe whether the occupation behavior has stopped; S904: If the occupation behavior has not stopped, the property management department will be notified to arrive at the scene within five minutes for on-site disposal, and within ten minutes, a reminder of illegal occupation of the fire passage will be sent to the car owner or relevant personnel through the public safety APP or SMS system, informing them to evacuate as soon as possible; S905: If the property management cannot effectively handle the occupation incident within 30 minutes, the incident information will be pushed to the business system of the public security department, requesting relevant personnel to go to the scene to persuade or tow away the vehicles or items occupying the fire passage; S906: Record event information, and record the entire process of the event in the system, including the time, location, occupancy, disposal measures and results.

10. A fire passage occupancy event detection and sound and light warning handling system according to claim 9, characterized in that: The fire escape occupancy event detection and sound and light warning disposal system includes the following steps when working: S1: The surveillance camera continuously collects video streams from the fire escape area; S2: The event detection module detects the occupancy of the fire passage location in the video stream and identifies the type and urgency of the occupancy event; S3: Generate different levels of warnings according to the type and urgency of the occupancy event, and push the warning information to different users according to the level; S4: At locations where conditions permit, immediately activate the on-site sound and light warning equipment to warn the occupants; S5: If the occupant does not respond, the warning information will be pushed to the property management center, and the property management center staff will check the photos of the incident scene and guide relevant personnel to rush to the scene to evacuate; S6: If the occupant has left after the relevant personnel arrive at the scene and cannot be removed on the spot, a text message reminder of illegal occupation of the fire passage will be sent to the vehicle owner or relevant personnel, informing them to evacuate as soon as possible; S7: If the property management company is unable to handle the occupation incident, an early warning will be issued to the public security department, who will dispatch relevant personnel to the scene to persuade or tow away the vehicles or items occupying the fire escape.

Images