A method and system for detecting and extinguishing smoke in lockers

Abstract

The present invention provides a method and system for locker smoke detection and fire extinguishing. The method comprises: obtaining comprehensive descriptive data of the locker interior at different time points within a preset period; calculating a fire risk handling opinion for the locker within a target period based on the acquisition time of each piece of comprehensive descriptive data and the total comprehensive descriptive data, wherein the preset period includes a target period and the target period is shorter than the preset period; calculating a fire risk score for the locker interior at the current moment based on the comprehensive descriptive data of the locker interior at the current moment; determining a fire extinguishing opinion based on the fire risk handling opinion for the locker within the target period and the fire risk score for the locker interior at the current moment, and taking different fire extinguishing measures based on the fire extinguishing opinion. The present invention can improve the safety level of stored items and provide users with more reliable storage services.

Description

Technical Field

[0001] The present invention relates to the technical field of lockers, and in particular to a method and system for detecting and extinguishing smoke in lockers. Background Art

[0002] Currently, as an essential component of modern urban public service facilities, lockers are widely used in various public places such as airports, train stations, shopping malls, and libraries, providing users with convenient temporary storage services. However, traditional lockers still have significant technical shortcomings in terms of fire safety. Their fire protection systems generally use a single smoke sensor or a fixed temperature threshold alarm mechanism. This simple design model leads to a high false alarm rate in actual application and lacks fire-fighting advice and measures tailored to current environmental factors. Therefore, if technological innovation can be used to improve the detection accuracy and targeted response of locker fire-fighting systems, it will become a key breakthrough in protecting user property safety and improving the quality of public services. Summary of the Invention

[0003] The object of the present invention is to provide a method and system for detecting and extinguishing smoke in a locker to improve the above-mentioned problem.

[0004] In order to achieve the above objectives, the embodiments of the present application provide the following technical solutions:

[0005] In a first aspect, an embodiment of the present application provides a method for detecting and extinguishing smoke in a locker, the method comprising:

[0006] Obtaining comprehensive description data of the interior of the locker at different time points within a preset time period, where the end time of the preset time period is the current time, and the comprehensive description data is text data formed by combining temperature data, smoke particle concentration data, and carbon dioxide concentration data inside the locker;

[0007] Calculate, based on the acquisition time of each piece of comprehensive description data and all the comprehensive description data, a fire risk treatment opinion for the locker within a target period, where the preset period includes the target period and the target period is shorter than the preset period; and simultaneously calculate, based on the comprehensive description data of the locker interior at the current moment, a fire risk score for the locker interior at the current moment;

[0008] The fire extinguishing opinions are determined based on the fire risk handling opinions for the lockers during the target period and the fire risk score inside the lockers at the current moment, and different fire extinguishing measures are taken according to the fire extinguishing opinions.

[0009] In a second aspect, an embodiment of the present application provides a locker smoke detection and fire extinguishing system, comprising:

[0010] an acquisition module, configured to acquire comprehensive descriptive data of the interior of the locker at different time points within a preset time period, wherein the preset time period ends at the current time, and the comprehensive descriptive data is text data formed by combining temperature data, smoke particle concentration data, and carbon dioxide concentration data within the locker;

[0011] a calculation module for calculating, based on the acquisition time of each piece of comprehensive description data and all the comprehensive description data, a fire risk treatment opinion for the locker within a target time period, where the preset time period includes the target time period and the target time period is shorter than the preset time period; and calculating, based on the comprehensive description data of the locker interior at the current moment, a fire risk score for the locker interior at the current moment;

[0012] The fire extinguishing module is used to determine a fire extinguishing opinion based on the fire risk handling opinion for the locker within the target period and the fire risk score inside the locker at the current moment, and to take different fire extinguishing measures according to the fire extinguishing opinion.

[0013] In a third aspect, embodiments of the present application provide a device for detecting and extinguishing smoke in a locker, the device comprising a memory and a processor. The memory is configured to store a computer program, and the processor is configured to execute the computer program to implement the steps of the aforementioned method for detecting and extinguishing smoke in a locker.

[0014] In a fourth aspect, an embodiment of the present application provides a readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of the above-mentioned locker smoke detection and fire extinguishing method are implemented.

[0015] The beneficial effects of the present invention are:

[0016] 1. This invention comprehensively optimizes data collection. By simultaneously acquiring data on temperature, smoke particle concentration, and carbon dioxide concentration within lockers, it establishes a multi-dimensional environmental monitoring system. Compared to traditional single-parameter monitoring, this multi-parameter fusion acquisition method provides a more comprehensive and accurate reflection of the locker's true safety status.

[0017] 2. In the data processing link, the present invention adopts a dual screening mechanism: first, the collected data is preliminarily screened based on the time dimension, and then a secondary screening is performed through a preset data list. This hierarchical screening strategy effectively reduces the amount of data input into the recognition model and significantly improves the system's operating efficiency. The screened data is input into the preset recognition model, which can output accurate fire risk treatment opinions corresponding to each comprehensive description data. At the same time, the present invention innovatively introduces a dynamic fire risk scoring mechanism, which uses a preset algorithm formula to perform real-time calculations on the current environmental data to generate quantitative risk assessment results. Ultimately, a comprehensive analysis of the fire risk score and treatment opinions can help staff form the optimal fire extinguishing decision plan.

[0018] 3. Compared to traditional single-factor assessment methods, this invention offers significant advantages: First, multi-parameter fusion analysis significantly improves the accuracy of fire risk assessment; second, the system generates targeted action recommendations based on real-time environmental data; and finally, a dynamic scoring mechanism enables quantitative risk assessment. Through these three technological innovations, this invention enhances the security of stored items and provides users with more reliable storage services.

[0019] Other features and advantages of the present invention will be described in the following description, and in part will become apparent from the description, or understood by practicing the embodiments of the present invention. The purposes and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description, claims, and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained based on these drawings without paying any creative work.

[0021] Figure 1 1. It is a flowchart of a method for detecting and extinguishing smoke in a locker according to an embodiment of the present invention;

[0022] Figure 2 Schematic diagram of the system structure of locker smoke detection and fire extinguishing according to an embodiment of the present invention;

[0023] Figure 3 Schematic diagram of the equipment structure for locker smoke detection and fire extinguishing according to an embodiment of the present invention. DETAILED DESCRIPTION

[0024] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. The components of the embodiments of the present invention generally described and shown in the drawings herein can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0025] It should be noted that similar reference numerals or letters represent similar items in the following drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined or explained in subsequent drawings. At the same time, in the description of the present invention, the terms "first", "second", etc. are only used to distinguish the description and should not be understood as indicating or implying relative importance.

[0026] Example 1:

[0027] like Figure 1 As shown, this embodiment provides a method for detecting smoke and extinguishing fire in a locker, which includes step S1, step S2 and step S3.

[0028] Step S1: Obtaining comprehensive description data of the interior of the locker at different time points within a preset time period, where the end time of the preset time period is the current time, and the comprehensive description data is text data formed by combining temperature data, smoke particle concentration data, and carbon dioxide concentration data inside the locker;

[0029] In this step, the preset time period can be 10 minutes or 5 minutes before the current moment, and can be customized according to user needs. Comprehensive description data of the interior of the locker can be collected every 2 seconds within the preset time period. The comprehensive description data can be understood as: for example, the combined text data can be: the temperature inside the locker is 20°C, the smoke particle concentration inside the locker is 50 μg / m³, and the carbon dioxide concentration inside the locker is 450 ppm. In this step, by collecting multiple fire influencing factors to comprehensively determine the fire extinguishing measures, the accuracy of the fire extinguishing measures can be improved compared to a single factor.

[0030] Step S2: Calculate a fire risk management opinion for the locker within a target time period based on the acquisition time of each piece of comprehensive description data and all the comprehensive description data, where the preset time period includes the target time period and the target time period is shorter than the preset time period; and simultaneously calculate a fire risk score for the locker interior at the current moment based on the comprehensive description data for the locker interior at the current moment;

[0031] In this step, the end time of the target period is the current time. The length of the target period can be customized according to user needs. For example, if the preset time period is 5 minutes before the current time, the target period can be 2 minutes before the current time.

[0032] In this step, based on the acquisition time of each piece of comprehensive description data and all the comprehensive description data, the specific implementation steps for the fire risk treatment opinions for the lockers within the target period are calculated, including:

[0033] Step S21: cluster all the comprehensive description data using the KMEANS algorithm to obtain multiple first-class clusters, calculate the first number of comprehensive description data belonging to the target time period in each first-class cluster based on the acquisition time of each comprehensive description data, and calculate the second number of comprehensive description data contained in each first-class cluster;

[0034] Step S22, calculate the ratio of the first number to the second number; group the first clusters whose ratios are greater than a preset ratio threshold to form a first group; filter the comprehensive description data in the first group to obtain a second group, input each comprehensive description data in the second group into a preset recognition model, obtain the fire risk handling opinion corresponding to each comprehensive description data in the second group, and calculate the fire risk handling opinion for the lockers within the target time period based on all the fire risk handling opinions.

[0035] In this step, the comprehensive description data in the first set is screened to obtain the specific implementation steps of the second set, including:

[0036] Step S221, obtain a data list, which includes multiple comprehensive description data that do not need to be processed; calculate the keywords of each comprehensive description data in the first set and the keywords of each comprehensive description data that does not need to be processed in the data list, and vectorize the keywords; for a comprehensive description data, count the number of the same keywords in its corresponding keywords and the keywords of each comprehensive description data that does not need to be processed, take the maximum value of the number as the repetition value, and at the same time, calculate the cosine distance between the vector corresponding to its keyword and the vector corresponding to the keyword of each comprehensive description data that does not need to be processed, and take the maximum value of the cosine distance as the cosine value; delete the comprehensive description data whose repetition value is greater than the preset threshold from the first set, and at the same time, delete the comprehensive description data whose repetition value is less than or equal to the preset threshold and whose cosine value is greater than the preset cosine value threshold from the first set, and finally obtain the second set.

[0037] By using the deduplication setting in this step, some comprehensive description data that does not need to be processed can be deleted. This method can improve the subsequent calculation speed;

[0038] At the same time, in step S22, each piece of comprehensive description data in the second set is input into a preset recognition model to obtain a fire risk handling opinion corresponding to each piece of comprehensive description data in the second set. Specific implementation steps for calculating the fire risk handling opinion for the locker during the target period based on all fire risk handling opinions include:

[0039] Step S222: Acquire multiple pieces of historical comprehensive description data, segment each piece of historical comprehensive description data to obtain multiple data segments, extract features from each data segment, perform local pooling on the features, obtain a first feature corresponding to each data segment, cluster all first features corresponding to each data segment, and construct a recognition model based on the clustering results;

[0040] In this step, the format of the historical comprehensive description data is the same as that of the comprehensive description data. When segmenting, a comma can be used as a segmentation mark. For example, if the historical comprehensive description data is that the temperature inside the locker is 20°C, the smoke particle concentration inside the locker is 50 μg / m³, and the carbon dioxide concentration data inside the locker is 400 ppm, then it can be segmented into three data segments: the temperature inside the locker is 20°C, the smoke particle concentration inside the locker is 50 μg / m³, and the carbon dioxide concentration data inside the locker is 400 ppm. Of course, other segmentation methods can also be used, which will not be repeated here.

[0041] At the same time, in this step, clustering is performed on all first features corresponding to each data segment, and the specific implementation steps of constructing a recognition model based on the clustering results include:

[0042] Step S2221: After clustering all the first features, for a cluster center, subtract the cluster center from each first feature to obtain multiple first vectors, and perform weighted summation on all the first vectors corresponding to the cluster center to obtain a second feature, wherein if the first feature belongs to the cluster center, the weight corresponding to the first feature is set to the first weight value, otherwise it is set to a second weight value different from the first weight value; the second features corresponding to each cluster center are spliced, and after splicing, global average pooling is performed to obtain feature information of each piece of historical comprehensive description data; clustering is performed according to the feature information of each piece of historical comprehensive description data to obtain multiple second clusters;

[0043] In this step, multiple cluster centers are obtained through clustering processing, and vector fusion processing is performed based on the distance vector between the first feature and the cluster center, that is, the first vector. This can erase the differences in the feature distribution of different clusters themselves, so that the fused distance vector can effectively represent the meaning of the corresponding cluster. Then, the fused distance vectors corresponding to the multiple cluster centers are combined into a fused feature, which can improve the accuracy of the fused feature. The feature information of each historical comprehensive description data finally obtained is the fused feature information. At the same time, clustering using the fused feature information can improve the accuracy of clustering;

[0044] Step S2222, performing a merge operation on all second clusters: selecting two different second clusters from all second clusters each time and recording them as the first processing cluster and the second processing cluster respectively; counting the number of historical comprehensive description data contained in the first processing cluster and the second processing cluster respectively, and recording them as the first value and the second value, and multiplying the first value and the second value to obtain a third value; calculating the similarity between each historical comprehensive description data in the first processing cluster and each historical comprehensive description data in the second processing cluster, and counting the number of historical comprehensive description data pairs whose similarity is greater than a preset similarity threshold, and recording it as a fourth value; dividing the fourth value by the third value to obtain a value as a fifth value, and judging whether the fifth value is greater than a preset category similarity threshold, if it is greater than, merging the historical comprehensive description data in the first processing cluster and the second processing cluster to form a third cluster; when the merge operation is completed, multiple third clusters are obtained;

[0045] Step S2223: perform data cleaning on the historical comprehensive description data contained in each third cluster and record the cleaned third cluster as the fourth cluster; merge all fourth clusters again, and when there are no clusters that can be merged, stop merging, and record the final cluster as the fifth cluster. Label each fifth cluster, and the labeling information is the fire risk handling opinion obtained by the staff based on the comprehensive analysis of the data in the fifth cluster. The labeling information of each historical comprehensive description data is the same as the labeling information of the fifth cluster to which it belongs; use the historical comprehensive description data with labeled information to train the convolutional neural network model to obtain the recognition model.

[0046] In this step, the data cleaning operation is to manually remove some abnormal historical comprehensive description data to improve the accuracy of subsequent annotations and thus improve the accuracy of recognition model training;

[0047] All fourth-class clusters are merged again, wherein the merging operation is the same as the method in step S2222. In this step, by merging clusters, the workload of subsequent labeling is reduced, the entire training cycle is shortened, and efficiency is improved;

[0048] Step S223: Input each comprehensive description data in the second set into the recognition model to obtain the fire risk handling opinion corresponding to each comprehensive description data in the second set, count the fire risk handling opinion that appears the most times and use it as the fire risk handling opinion for the lockers during the target period.

[0049] At the same time, in step S2, the specific implementation steps of calculating the fire risk score of the interior of the locker at the current moment based on the comprehensive description data of the interior of the locker at the current moment include:

[0050] Step S23: Calculate the fire risk score inside the locker at the current moment according to formula (1) and formula (2). Formula (1) and formula (2) are:

[0051] (1)

[0052] (2)

[0053] In formulas (1)-(2), H is the fire risk score; Detect the temperature at the current moment; is the current ambient reference temperature; To allow the maximum temperature rise, the value is 10℃; is the temperature weight coefficient, which is 0.5; is the concentration of smoke particles at the current moment, in μg / m³; is the standard fire particulate matter threshold, which is 500; is the particle weight, which is 0.3; is the rate of change of carbon dioxide concentration at the current moment, in ppm / s, is the gas weight, which is 0.2; is the material sensitivity coefficient; is the smoothing coefficient, which is 0.3; Detect the temperature at the last moment; It is the ambient reference temperature at the previous moment.

[0054] In this step, when calculating the ambient reference temperature, first install the temperature sensor in the cabinet, collect temperature data within 1 hour, and use the average of the temperature data collected within 1 hour as the ambient reference temperature at the initial moment; the last temperature data collected within this 1 hour is used as the detection temperature at the initial moment, and the subsequent ambient reference temperature calculation is based on this. Among them, the detection temperature is collected by the temperature sensor, and the temperature sensor is installed at any place in the cabinet; the smoke particle concentration is collected by the particle sensor, and the installation position is 1 / 3 of the upper part of the cabinet back panel; the carbon dioxide concentration is The sensor collects data and can be installed anywhere in the cabinet; The acquisition method is to obtain an image of the deposited object and input the image of the deposited object into a preset material sensitivity coefficient recognition model to obtain the material sensitivity coefficient of the current deposited object. The material sensitivity coefficient recognition model is used to characterize the one-to-one correspondence between the image of the deposited object and the material sensitivity coefficient;

[0055] In addition to the fire risk score calculation method described above, a fire risk score model can also be constructed. The fire risk score can be obtained by inputting the comprehensive description data of the locker interior at the current moment into the fire risk score model. The fire risk score model is used to represent the one-to-one correspondence between the comprehensive description data of the locker interior and the fire risk score. The construction method can be achieved by using conventional training data annotation and training convolutional neural network models. The details will not be repeated here.

[0056] In the above steps, when collecting the current detection temperature, multiple temperature sensors can be set at multiple locations inside the cabinet. After removing abnormal values ​​from the multiple temperatures collected, the average value is used as the current detection temperature. The specific implementation steps include:

[0057] Step S231: multiple temperature sensors are installed inside the locker, with the number of temperature sensors being an even number. The average of the multiple temperature data collected is used as the average temperature; the absolute value of the difference between the average temperature and each temperature data is used as the difference corresponding to each temperature data; the differences are sorted in ascending order, and after sorting, the differences with even numbers are selected and recorded as target differences;

[0058] In this step, multiple temperature sensors are evenly arranged on each panel in the cabinet. For example, two temperature sensors can be set on the front, back, left, right, top and bottom panels.

[0059] Step S232: For each target difference, all differences are compared with it, and the temperature data corresponding to differences greater than the target difference are aggregated to obtain a first data set corresponding to the target difference; and the temperature data corresponding to differences less than the target difference are aggregated to obtain a second data set corresponding to the target difference;

[0060] In the first data set, one temperature data is selected each time and recorded as the target temperature data, and the Euclidean distance between it and the remaining temperature data in the first data set is calculated respectively, and the average of all Euclidean distances is recorded as the first distance data corresponding to this target temperature data; at the same time, the Euclidean distance between it and each temperature data in the second data set is calculated respectively, and the average of all Euclidean distances is recorded as the second distance data corresponding to this target temperature data; the first distance data is subtracted from the second distance data to obtain difference data, the maximum value between the first distance data and the second distance data is screened out, and the ratio between the difference data and the maximum value is calculated; the calculation ends when the corresponding ratio is calculated for each temperature data in the first data set;

[0061] Then, in the second data set, one temperature data is selected each time and recorded as the target temperature data, and the Euclidean distance between it and the remaining temperature data in the second data set is calculated respectively, and the average of all Euclidean distances is recorded as the first distance data corresponding to this target temperature data; at the same time, the Euclidean distance between it and each temperature data in the first data set is calculated respectively, and the average of all Euclidean distances is recorded as the second distance data corresponding to this target temperature data; the first distance data is subtracted from the second distance data to obtain difference data, the maximum value between the first distance data and the second distance data is screened out, and the ratio between the difference data and the maximum value is calculated; the calculation ends when the corresponding ratio is calculated for each temperature data in the second data set;

[0062] Add the ratios corresponding to each temperature data in the first data set and the second data set to obtain a ranking value corresponding to each target difference;

[0063] Step S233: Filter out the target difference corresponding to the maximum ranking value, delete the temperature data with a difference greater than the target difference as abnormal data, and use the mean of the remaining temperature data as the detected temperature at the current moment.

[0064] Through the steps S231 to S233 above, abnormal temperature data can be dynamically deleted to ensure the accuracy of the temperature data provided each time;

[0065] Step S3: Determine a fire extinguishing opinion based on the fire risk handling opinion for the locker during the target period and the fire risk score inside the locker at the current moment, and take different fire extinguishing measures according to the fire extinguishing opinion.

[0066] The specific implementation steps of this step include:

[0067] Step S31: Determine whether the fire risk score is greater than a preset fire risk score threshold. If so, combine the fire risk handling opinions for the locker within the target period and the fire risk score inside the locker at the current moment to form a fire extinguishing opinion and send it to the staff to help them make fire extinguishing opinions and take different fire extinguishing measures based on the fire extinguishing opinions.

[0068] In this step, the fire risk handling opinions for the lockers within the target period and the fire risk score inside the lockers at the current moment are combined to form fire extinguishing opinions and sent to the staff. The staff can then adopt the fire risk handling opinions or further optimize them based on previous work experience to obtain the optimal fire extinguishing measures.

[0069] This embodiment has been comprehensively optimized in terms of data collection. By synchronously acquiring temperature data, smoke particle concentration data, and carbon dioxide concentration data inside the lockers, a multi-dimensional environmental monitoring system has been established. Compared with traditional single-parameter monitoring, this multi-parameter fusion collection method can more comprehensively and accurately reflect the true safety status of the lockers. In the data processing link, this embodiment adopts a double screening mechanism: first, the collected data is preliminarily screened based on the time dimension, and then a secondary screening is performed through a preset data list. This hierarchical screening strategy effectively reduces the amount of data input into the recognition model and significantly improves the system's operating efficiency. The screened data is input into the preset recognition model, which can output accurate fire risk treatment opinions corresponding to each comprehensive description data. At the same time, this embodiment innovatively introduces a dynamic fire risk scoring mechanism, which uses a preset algorithm formula to perform real-time calculations on the current environmental data to generate quantitative risk assessment results. The system ultimately forms the optimal fire extinguishing decision plan through a comprehensive analysis of the fire risk score and treatment opinions. Compared to traditional single-factor assessment methods, this embodiment offers significant advantages: First, multi-parameter fusion analysis significantly improves the accuracy of fire risk assessment; second, the system generates targeted action recommendations based on real-time environmental data; and finally, a dynamic scoring mechanism enables quantitative risk assessment. Through these three technological innovations, this embodiment enhances the security of stored items and provides users with more reliable storage services.

[0070] Example 2:

[0071] like Figure 2As shown, this embodiment provides a locker smoke detection and fire extinguishing system, which includes an acquisition module 1, a calculation module 2 and a fire extinguishing module 3.

[0072] Acquisition module 1, for acquiring comprehensive description data of the interior of the locker at different time points within a preset time period, where the end time of the preset time period is the current time, and the comprehensive description data is text data formed by combining temperature data, smoke particle concentration data, and carbon dioxide concentration data inside the locker;

[0073] Calculation module 2 is configured to calculate a fire risk management opinion for the locker within a target time period based on the acquisition time of each piece of comprehensive description data and all the comprehensive description data, where the preset time period includes the target time period and the target time period is shorter than the preset time period; and to calculate a fire risk score for the locker at the current moment based on the comprehensive description data of the locker interior at the current moment;

[0074] The fire extinguishing module 3 is configured to determine a fire extinguishing opinion based on the fire risk handling opinion for the locker within the target period and the fire risk score inside the locker at the current moment, and to take different fire extinguishing measures according to the fire extinguishing opinion.

[0075] In a specific embodiment of the present disclosure, the calculation module 2 further includes a first clustering unit 21 and a calculation unit 22 .

[0076] A first clustering unit 21 is configured to cluster all the comprehensive description data using a KMEANS algorithm to obtain a plurality of first clusters, calculate a first number of pieces of comprehensive description data belonging to a target time period in each first cluster based on an acquisition time of each piece of comprehensive description data, and calculate a second number of pieces of comprehensive description data contained in each first cluster;

[0077] The calculation unit 22 is used to calculate the ratio of the first number to the second number; group the first clusters whose ratios are greater than a preset ratio threshold to form a first set; filter the comprehensive description data in the first set to obtain a second set, input each comprehensive description data in the second set into a preset recognition model, obtain the fire risk handling opinion corresponding to each comprehensive description data in the second set, and calculate the fire risk handling opinion for the lockers within the target time period based on all the fire risk handling opinions.

[0078] In a specific implementation of the present disclosure, the calculation unit 22 further includes a first acquisition unit 221 .

[0079] The first acquisition unit 221 is used to obtain a data list, which includes multiple comprehensive description data that do not need to be processed; calculate the keywords of each comprehensive description data in the first set and the keywords of each comprehensive description data that does not need to be processed in the data list, and vectorize the keywords; for a comprehensive description data, count the number of the same keywords in its corresponding keywords and the keywords of each comprehensive description data that does not need to be processed, take the maximum value of the number as the repetition value, and at the same time, calculate the cosine distance between the vector corresponding to its keyword and the vector corresponding to the keyword of each comprehensive description data that does not need to be processed, and take the maximum value of the cosine distance as the cosine value; delete the comprehensive description data with a repetition value greater than a preset threshold from the first set, and at the same time, delete the comprehensive description data with a repetition value less than or equal to the preset threshold and a cosine value greater than the preset cosine value threshold from the first set, and finally obtain the second set.

[0080] In a specific embodiment of the present disclosure, the calculation unit 22 further includes a second acquisition unit 222 and an identification unit 223 .

[0081] A second acquisition unit 222 is configured to acquire multiple pieces of historical comprehensive description data, segment each piece of historical comprehensive description data to obtain multiple data segments, extract features from each data segment, perform local pooling processing on the features, obtain a first feature corresponding to each data segment, cluster all the first features corresponding to each data segment, and construct a recognition model based on the clustering results;

[0082] The identification unit 223 is used to input each comprehensive description data in the second set into the identification model, obtain the fire risk handling opinion corresponding to each comprehensive description data in the second set, count the fire risk handling opinion that appears the most times and use it as the fire risk handling opinion for the locker during the target period.

[0083] In a specific embodiment of the present disclosure, the second acquiring unit 222 further includes a second clustering unit 2221 , a third clustering unit 2222 and a training unit 2223 .

[0084] The second clustering unit 2221 is configured to cluster all the first features, subtract each first feature from the cluster center to obtain a plurality of first vectors, and perform weighted summation on all the first vectors corresponding to the cluster center to obtain a second feature, wherein if the first feature belongs to the cluster center, the weight corresponding to the first feature is set to a first weight value; otherwise, the weight is set to a second weight value different from the first weight value; concatenate the second features corresponding to each cluster center, and perform global average pooling processing after concatenation to obtain feature information of each piece of historical comprehensive description data; and perform clustering based on the feature information of each piece of historical comprehensive description data to obtain a plurality of second clusters;

[0085] The third clustering unit 2222 is used to perform a merging operation on all the second clusters: each time, two different second clusters are selected from all the second clusters and recorded as the first processing cluster and the second processing cluster respectively; the number of historical comprehensive description data contained in the first processing cluster and the second processing cluster is counted respectively, and recorded as the first value and the second value, and the first value and the second value are multiplied to obtain a third value; the similarity between each historical comprehensive description data in the first processing cluster and each historical comprehensive description data in the second processing cluster is calculated, and the number of historical comprehensive description data pairs whose similarity is greater than a preset similarity threshold is counted and recorded as a fourth value; the value of the fourth value divided by the third value is recorded as a fifth value, and it is determined whether the fifth value is greater than the preset category similarity threshold. If it is greater than, the historical comprehensive description data in the first processing cluster and the second processing cluster are merged to form a third cluster; when the merging operation is completed, multiple third clusters are obtained;

[0086] The training unit 2223 is used to perform data cleaning operations on the historical comprehensive description data contained in each third cluster and record the cleaned third cluster as the fourth cluster; merge all the fourth clusters again, and when there are no clusters that can be merged, stop merging, and record the final cluster as the fifth cluster. Each fifth cluster is labeled, and the labeling information is the fire risk handling opinion obtained by the staff based on the comprehensive analysis of the data in the fifth cluster. The labeling of each historical comprehensive description data is the same as the labeling of the fifth cluster to which it belongs; use the historical comprehensive description data with labeled information to train the convolutional neural network model to obtain a recognition model.

[0087] It should be noted that, regarding the system in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated on here.

[0088] Example 3:

[0089] Corresponding to the above method embodiment, the embodiment of the present disclosure also provides a locker smoke detection and fire extinguishing device. The locker smoke detection and fire extinguishing device described below and the locker smoke detection and fire extinguishing method described above can be referenced to each other.

[0090] Figure 3 FIG. 1 is a block diagram of a locker smoke detection and fire extinguishing device 300 according to an exemplary embodiment. Figure 3 As shown, the locker smoke detection and fire extinguishing device 300 may include: a processor 301, a memory 302. The locker smoke detection and fire extinguishing device 300 may also include one or more of a multimedia component 303, an I / O interface 304, and a communication component 305.

[0091] The processor 301 is used to control the overall operation of the locker smoke detection and fire extinguishing device 300 to complete all or part of the steps in the above-mentioned locker smoke detection and fire extinguishing method. The memory 302 is used to store various types of data to support the operation of the locker smoke detection and fire extinguishing device 300. This data may include, for example, instructions for any application or method operating on the locker smoke detection and fire extinguishing device 300, as well as application-related data such as contact information, sent and received messages, images, audio, video, etc. The memory 302 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk. The multimedia component 303 may include a screen and an audio component. The screen may be, for example, a touch screen, and the audio component is used to output and / or input audio signals. For example, the audio component may include a microphone for receiving external audio signals. The received audio signal may be further stored in the memory 302 or transmitted via the communication component 305. The audio component also includes at least one speaker for outputting audio signals. The I / O interface 304 provides an interface between the processor 301 and other interface modules, which may be a keyboard, a mouse, buttons, etc. These buttons may be virtual buttons or physical buttons. The communication component 305 is used for wired or wireless communication between the locker smoke detection and fire extinguishing device 300 and other devices. Wireless communication, such as Wi-Fi, Bluetooth, Near Field Communication (NFC), 2G, 3G or 4G, or a combination of one or more thereof, therefore, the corresponding communication component 305 may include: a Wi-Fi module, a Bluetooth module, an NFC module.

[0092] In an exemplary embodiment, the locker smoke detection and fire extinguishing device 300 can be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors or other electronic components to perform the above-mentioned locker smoke detection and fire extinguishing method.

[0093] In another exemplary embodiment, a computer-readable storage medium including program instructions is also provided. When executed by a processor, the program instructions implement the steps of the aforementioned locker smoke detection and fire extinguishing method. For example, the computer-readable storage medium may be the aforementioned memory 302 including the program instructions. The program instructions may be executed by the processor 301 of the locker smoke detection and fire extinguishing device 300 to perform the aforementioned locker smoke detection and fire extinguishing method.

[0094] Example 4:

[0095] Corresponding to the above method embodiment, the embodiment of the present disclosure further provides a readable storage medium. The readable storage medium described below and the locker smoke detection and fire extinguishing method described above can be referenced to each other.

[0096] A readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the locker smoke detection and fire extinguishing method of the above method embodiment.

[0097] The readable storage medium may specifically be any readable storage medium that can store program code, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0098] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention are intended to be within the scope of protection of the present invention.

Claims

1. A method for detecting and extinguishing smoke in a locker, characterized in that: include: Obtaining comprehensive description data of the interior of the locker at different time points within a preset time period, where the end time of the preset time period is the current time, and the comprehensive description data is text data formed by combining temperature data, smoke particle concentration data, and carbon dioxide concentration data inside the locker; Calculate, based on the acquisition time of each piece of comprehensive description data and all the comprehensive description data, a fire risk treatment opinion for the locker within a target period, where the preset period includes the target period and the target period is shorter than the preset period; and simultaneously calculate, based on the comprehensive description data of the locker interior at the current moment, a fire risk score for the locker interior at the current moment; Determine a fire extinguishing opinion based on the fire risk management opinion for the locker during the target period and the fire risk score inside the locker at the current moment, and take different fire extinguishing measures based on the fire extinguishing opinion; Among them, based on the acquisition time of each comprehensive description data and all comprehensive description data, the fire risk handling opinions for the lockers during the target period are calculated, including: The K-MEANS algorithm is used to cluster all the comprehensive description data to obtain multiple first-class clusters. The first number of comprehensive description data belonging to the target time period in each first-class cluster is calculated according to the acquisition time of each comprehensive description data, and the second number of comprehensive description data contained in each first-class cluster is calculated; Calculating the ratio of the first number to the second number; grouping the first clusters whose ratios are greater than a preset ratio threshold to form a first set; filtering the comprehensive description data in the first set to obtain a second set; inputting each comprehensive description data in the second set into a preset recognition model to obtain a fire risk treatment opinion corresponding to each comprehensive description data in the second set; and calculating the fire risk treatment opinion for the lockers within the target time period based on all the fire risk treatment opinions; The comprehensive description data in the first set is screened to obtain the second set, including: Obtain a data list, which includes multiple comprehensive description data that do not need to be processed; calculate the keywords of each comprehensive description data in the first set and the keywords of each comprehensive description data that does not need to be processed in the data list, and vectorize the keywords; for one comprehensive description data, count the number of the same keywords in its corresponding keywords and the keywords of each comprehensive description data that does not need to be processed, take the maximum value of the number as the repetition value, and at the same time, calculate the cosine distance between the vector corresponding to its keyword and the vector corresponding to the keyword of each comprehensive description data that does not need to be processed, and take the maximum value of the cosine distance as the cosine value; delete the comprehensive description data whose repetition value is greater than a preset threshold from the first set, and at the same time, delete the comprehensive description data whose repetition value is less than or equal to the preset threshold and whose cosine value is greater than the preset cosine value threshold from the first set, and finally obtain the second set.

2. The locker smoke detection and fire extinguishing method according to claim 1, characterized in that: Each piece of comprehensive description data in the second set is input into a preset recognition model to obtain a fire risk handling opinion corresponding to each piece of comprehensive description data in the second set. Based on all fire risk handling opinions, a fire risk handling opinion for the locker during the target period is calculated, including: Acquire multiple pieces of historical comprehensive description data, segment each piece of historical comprehensive description data to obtain multiple data segments, extract features of each data segment and perform local pooling processing on the features to obtain the first feature corresponding to each data segment, cluster all the first features corresponding to each data segment, and build a recognition model based on the clustering results; Each comprehensive description data in the second set is input into the recognition model to obtain the fire risk handling opinion corresponding to each comprehensive description data in the second set, and the fire risk handling opinion with the most occurrences is counted and used as the fire risk handling opinion for the lockers during the target period.

3. The locker smoke detection and fire extinguishing method according to claim 2, characterized in that: Clustering is performed on all first features corresponding to each data segment, and a recognition model is constructed based on the clustering results, including: After clustering all the first features, for a cluster center, subtract the cluster center from each first feature to obtain multiple first vectors, and perform weighted summation on all the first vectors corresponding to the cluster center to obtain the second feature, wherein if the first feature belongs to the cluster center, the weight corresponding to the first feature is set to the first weight value, otherwise it is set to a second weight value different from the first weight value; the second features corresponding to each cluster center are spliced, and after splicing, global average pooling is performed to obtain feature information of each historical comprehensive description data; clustering is performed according to the feature information of each historical comprehensive description data to obtain multiple second clusters; Perform a merging operation on all second clusters: select two different second clusters from all second clusters each time and record them as the first processing cluster and the second processing cluster respectively; count the number of historical comprehensive description data contained in the first processing cluster and the second processing cluster respectively, and record them as the first value and the second value, and multiply the first value and the second value to obtain a third value; calculate the similarity between each historical comprehensive description data in the first processing cluster and each historical comprehensive description data in the second processing cluster, and count the number of historical comprehensive description data pairs whose similarity is greater than a preset similarity threshold, and record it as a fourth value; divide the fourth value by the third value to obtain a fifth value, and judge whether the fifth value is greater than the preset category similarity threshold. If it is greater than, merge the historical comprehensive description data in the first processing cluster and the second processing cluster to form a third cluster; when the merging operation is completed, obtain multiple third clusters; A data cleaning operation is performed on the historical comprehensive description data contained in each third cluster, and the cleaned third cluster is recorded as the fourth cluster; all the fourth clusters are merged again. When there are no clusters that can be merged, the merging is stopped, and the final cluster is recorded as the fifth cluster. Each fifth cluster is labeled. The labeling information is the fire risk treatment opinion obtained by the staff based on the comprehensive analysis of the data in the fifth cluster. The labeling of each historical comprehensive description data is the same as the labeling of the fifth cluster to which it belongs; the convolutional neural network model is trained using the historical comprehensive description data with labeled information to obtain a recognition model.

4. A locker smoke detection and fire extinguishing system, characterized in that: include: an acquisition module, configured to acquire comprehensive descriptive data of the interior of the locker at different time points within a preset time period, wherein the preset time period ends at the current time, and the comprehensive descriptive data is text data formed by combining temperature data, smoke particle concentration data, and carbon dioxide concentration data within the locker; a calculation module for calculating, based on the acquisition time of each piece of comprehensive description data and all the comprehensive description data, a fire risk treatment opinion for the locker within a target time period, where the preset time period includes the target time period and the target time period is shorter than the preset time period; and calculating, based on the comprehensive description data of the locker interior at the current moment, a fire risk score for the locker interior at the current moment; A fire extinguishing module, configured to determine a fire extinguishing opinion based on the fire risk treatment opinion for the locker within the target period and the fire risk score inside the locker at the current moment, and to take different fire extinguishing measures according to the fire extinguishing opinion; The computing module includes: A first clustering unit is used to cluster all the comprehensive description data using a K-MEANS algorithm to obtain multiple first clusters, calculate the first number of comprehensive description data belonging to the target time period in each first cluster according to the acquisition time of each comprehensive description data, and calculate the second number of comprehensive description data included in each first cluster; a calculation unit configured to calculate a ratio of the first number of items to the second number of items; grouping first clusters whose ratios are greater than a preset ratio threshold to form a first set; filtering the comprehensive description data in the first set to obtain a second set; inputting each comprehensive description data item in the second set into a preset recognition model to obtain a fire risk handling opinion corresponding to each comprehensive description data item in the second set; and calculating a fire risk handling opinion for the locker within a target time period based on all the fire risk handling opinions; The computing unit includes: The first acquisition unit is used to obtain a data list, which includes multiple comprehensive description data that do not need to be processed; calculate the keywords of each comprehensive description data in the first set and the keywords of each comprehensive description data that does not need to be processed in the data list, and vectorize the keywords; for a comprehensive description data, count the number of the same keywords in its corresponding keywords and the keywords of each comprehensive description data that does not need to be processed, take the maximum value of the number as the repetition value, and at the same time, calculate the cosine distance between the vector corresponding to its keyword and the vector corresponding to the keyword of each comprehensive description data that does not need to be processed, and take the maximum value of the cosine distance as the cosine value; delete the comprehensive description data whose repetition value is greater than the preset threshold from the first set, and at the same time, delete the comprehensive description data whose repetition value is less than or equal to the preset threshold and whose cosine value is greater than the preset cosine value threshold from the first set, and finally obtain the second set.

5. The locker smoke detection and fire extinguishing system according to claim 4, characterized in that: Computing unit, including: a second acquisition unit, configured to acquire a plurality of pieces of historical comprehensive description data, segment each piece of historical comprehensive description data to obtain a plurality of data segments, extract features of each data segment and perform local pooling processing on the features to obtain a first feature corresponding to each data segment, cluster all the first features corresponding to each data segment, and construct a recognition model based on the clustering results; The recognition unit is used to input each comprehensive description data in the second set into the recognition model, obtain the fire risk handling opinion corresponding to each comprehensive description data in the second set, count the fire risk handling opinion that appears the most times and use it as the fire risk handling opinion for the locker during the target period.

6. The locker smoke detection and fire extinguishing system according to claim 5, characterized in that: The second acquisition unit includes: The second clustering unit is used to cluster all the first features, subtract the cluster center from each first feature to obtain multiple first vectors, and perform weighted summation on all the first vectors corresponding to the cluster center to obtain a second feature, wherein if the first feature belongs to the cluster center, the weight corresponding to the first feature is set to the first weight value, otherwise it is set to a second weight value different from the first weight value; the second features corresponding to each cluster center are spliced, and after splicing, global average pooling is performed to obtain feature information of each historical comprehensive description data; clustering is performed according to the feature information of each historical comprehensive description data to obtain multiple second clusters; The third clustering unit is used to perform a merging operation on all the second clusters: each time, two different second clusters are selected from all the second clusters and recorded as the first processing cluster and the second processing cluster respectively; the number of historical comprehensive description data contained in the first processing cluster and the second processing cluster is counted respectively, and recorded as the first value and the second value, and the first value and the second value are multiplied to obtain a third value; the similarity between each historical comprehensive description data in the first processing cluster and each historical comprehensive description data in the second processing cluster is calculated, and the number of historical comprehensive description data pairs whose similarity is greater than a preset similarity threshold is counted and recorded as a fourth value; the value of the fourth value divided by the third value is recorded as a fifth value, and it is judged whether the fifth value is greater than the preset category similarity threshold. If it is greater than, the historical comprehensive description data in the first processing cluster and the second processing cluster are merged to form a third cluster; when the merging operation is completed, multiple third clusters are obtained; The training unit is used to perform data cleaning operations on the historical comprehensive description data contained in each third cluster and record the cleaned third cluster as the fourth cluster; merge all the fourth clusters again, stop merging when there are no clusters that can be merged, record the final cluster as the fifth cluster, and label each fifth cluster. The labeling information is the fire risk treatment opinion obtained by the staff based on the comprehensive analysis of the data in the fifth cluster. The labeling of each historical comprehensive description data is the same as the labeling of the fifth cluster to which it belongs; use the historical comprehensive description data with labeled information to train the convolutional neural network model to obtain the recognition model.

Images