Safety early warning system based on image recognition

Abstract

The invention relates to the technical field of safety early warning, and discloses a safety early warning system based on image recognition, and the system comprises an early warning image collection device which is responsible for collecting a real-time image file of a cultural relic monitoring sub-region, and transmitting the real-time image file to an early warning image processing device. And the early warning image processing device receives the image file, analyzes the fire risk factor type, the target cultural relic monitoring sub-region and the actual cultural relic type, and sends an analysis result to the early warning response device. And the early warning response equipment carries out processing according to the fire risk factor type, judges the smoke risk level according to the concentration value for smoke, determines proper fire extinguishing equipment and sends a first early warning instruction. And for the flame, analyzing a fire area proportion through an image segmentation model, determining a flame risk level, selecting fire extinguishing equipment and sending a second early warning instruction. The electromagnetic lock is used for responding to the early warning instruction, executing unlocking operation and releasing related fire extinguishing equipment. Therefore, safety early warning based on image recognition is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of security warning, and particularly relates to a security warning system based on image recognition. Background Art

[0002] With the continuous improvement of the awareness of cultural relics protection, the security monitoring and fire warning systems for cultural relics have become important technical means in places such as museums and cultural relics protection units. However, the existing fire monitoring and warning systems still face some technical challenges when dealing with complex cultural relic environments. Therefore, there is an urgent need for a security warning system based on image recognition to address this challenge.

[0003] In the existing security warning systems, the following technical problems often exist: First, the existing systems fail to comprehensively consider the differential factors of cultural relics, resulting in inaccurate fire risk assessment, untimely warning processing, and lack of personalization, leading to low warning efficiency and thus affecting the protection effect of cultural relics; Second, the existing security warning systems usually only focus on the warning of the fire itself, but do not consider the potential threats of environmental changes to cultural relics, and do not consider the specific requirements of cultural relics for fire extinguishing and do not automatically generate targeted repair plans, thus affecting the protection effect of cultural relics; Third, the existing warning systems do not consider the regional characteristics corresponding to the characteristics of cultural relics for undifferentiated fire extinguishing, which is prone to false alarms, resulting in inaccurate warnings, and untimely response in case of fire, leading to poor warning efficiency and thus poor protection effect of cultural relics. Summary of the Invention

[0004] This part of the summary of the invention is used to briefly introduce the concepts, which will be described in detail in the following detailed implementation part. This part of the summary of the invention is not intended to identify the key features or essential features of the claimed technical solution, nor is it intended to limit the scope of the claimed technical solution.

[0005] The present invention proposes a security warning system based on image recognition to solve one or more of the technical problems mentioned in the above background art part.

[0006] The present invention provides a security warning system based on image recognition, including: a plurality of warning image acquisition devices, which are used to send the acquired plurality of real-time image files to the warning image processing device; The warning image processing device is used to receive the plurality of real-time image files, and send the type of fire risk factors, the target cultural relic monitoring sub-region, and the actual cultural relic type determined according to the plurality of real-time image files to the warning response device; Early warning response device, which is used to receive the type of fire risk factors, the target cultural relic monitoring sub-region and the actual type of cultural relics. The type of fire risk factors is smoke or flame. If the type of fire risk factor is smoke, determine the smoke risk level of the target cultural relic monitoring sub-region. According to the actual type of cultural relics and the smoke risk level, determine the smoke extinguishing equipment information group from the smoke equipment deployment plan group and determine the first target electromagnetic lock. Send the first early warning instruction to the electromagnetic lock associated with the extinguishing equipment according to the equipment identifier corresponding to the first target electromagnetic lock. If the type of fire risk factor is flame, determine the flame risk level of the target cultural relic monitoring sub-region. According to the actual type of cultural relics and the flame risk level, determine the flame extinguishing equipment information group from the flame equipment deployment plan group and determine the second target electromagnetic lock. Send the second early warning instruction to the electromagnetic lock associated with the extinguishing equipment according to the equipment identifier corresponding to the second target electromagnetic lock. The electromagnetic lock associated with the extinguishing equipment, which is used to respond to the first early warning instruction or the second early warning instruction and perform the unlocking operation for the electromagnetic lock.

[0007] Optionally, a safety early warning system based on image recognition of the present invention further includes: Multiple light sensors are arranged in multiple cultural relic monitoring sub-regions. Each light sensor in the multiple light sensors is used to collect the light intensity of the corresponding cultural relic monitoring sub-region. Among them, each cultural relic monitoring sub-region corresponds to a region number. Send the light intensity and the region number to the early warning image processing device. Multiple color temperature sensors are arranged in multiple cultural relic monitoring sub-regions. Each color temperature sensor in the multiple color temperature sensors is used to collect the color temperature of the corresponding cultural relic monitoring sub-region. Send the color temperature and the region number to the early warning image processing device. The early warning image processing device is further used to receive the light intensity, color temperature and region number of each cultural relic monitoring sub-region. Determine the calibration parameters of the corresponding early warning image acquisition device according to the light intensity and color temperature. Send the calibration parameters of the early warning image acquisition device and the region number to the device management terminal so that the device management personnel can calibrate each early warning image acquisition device.

[0008] Optionally, the type of fire risk factors further includes no fire risk, and The early warning response device is further used to obtain the environmental stability index in the target cultural relic monitoring sub-region. The environmental stability index includes the current temperature, current humidity and current air flow velocity. Determine the suitable temperature range, suitable humidity range and suitable air flow velocity range of the cultural relics according to the actual type of cultural relics. Compare the current temperature with the suitable temperature range of the cultural relics to determine the degree of temperature deviation; determine the temperature risk level score according to the degree of temperature deviation; compare the current humidity with the suitable humidity range of the cultural relics to determine the degree of humidity deviation; determine the humidity risk level score according to the degree of humidity deviation; compare the current air flow velocity with the suitable air flow velocity range of the cultural relics to determine the degree of air flow velocity deviation; determine the air flow velocity risk level score according to the degree of air flow velocity deviation; Determine the comprehensive risk score according to the temperature risk level score, humidity risk level score and air flow velocity risk level score; generate an abnormal warning level according to the comprehensive risk score and the preset comprehensive risk score level; generate a corresponding abnormal warning instruction according to the abnormal warning level; send the abnormal warning instruction to the user terminal.

[0009] Optionally, the warning response device is also used to store a pre-configured cultural relic damage index table, and the cultural relic damage index table includes multiple cultural relic types, the cultural relic vulnerability levels corresponding to each cultural relic type, and damage indexes; The warning response device is also used to receive multiple post-disaster real-time image files of multiple cultural relic monitoring sub-areas. Each post-disaster real-time image file in the multiple post-disaster real-time image files includes a post-disaster real-time image and a corresponding image number; input the post-disaster real-time image in each post-disaster real-time image file into a pre-trained cultural relic recognition model to obtain a cultural relic recognition result, and the cultural relic recognition result includes cultural relics or no cultural relics; determine the post-disaster real-time image corresponding to the cultural relic recognition result as the target post-disaster real-time image; determine the corresponding target cultural relic type according to the target post-disaster real-time image; match the target cultural relic type with the cultural relic damage index table to obtain a matching result, and the matching result is used to represent successful or failed matching; if the matching result represents successful matching, analyze the damage index corresponding to the target cultural relic type to determine the degree of damage; If the matching result represents failed matching, determine the cultural relic vulnerability level corresponding to the target cultural relic type; compare the cultural relic vulnerability level corresponding to the target cultural relic type with the cultural relic vulnerability levels corresponding to each cultural relic type in the cultural relic damage index table to obtain multiple cultural relic vulnerability level differences; sort the multiple cultural relic types in ascending order according to the corresponding cultural relic vulnerability level differences to obtain a cultural relic type sequence; query the sorting numbers corresponding to each cultural relic type in the cultural relic type sequence in a preset damage index importance coefficient table to obtain the damage index importance coefficients corresponding to each cultural relic type; obtain the degree of damage corresponding to the target cultural relic type according to the damage index importance coefficients and damage indexes corresponding to each cultural relic type; Determine a target repair plan from a pre-stored repair plan set according to the degree of damage and the target cultural relic type; send the target repair plan to the repair terminal.

[0010] Optionally, multiple early warning image acquisition devices are arranged in multiple pre-divided cultural relic monitoring sub-areas. Each early warning image acquisition device among the multiple early warning image acquisition devices corresponds to a cultural relic monitoring sub-area. Each real-time image file among the multiple real-time image files includes a real-time image and an image number; and The early warning image processing device is further configured to input the real-time image in each real-time image file into a pre-deployed fire risk factor prediction model to obtain fire risk prediction information. The fire risk prediction information includes the type of fire risk factor and the prediction confidence level. The type of fire risk factor includes smoke or flame; determine the real-time image corresponding to the prediction confidence level greater than the preset prediction confidence level threshold as the target real-time image; determine the cultural relic monitoring sub-area corresponding to the target real-time image according to the image number corresponding to the target real-time image; determine the cultural relic monitoring sub-area corresponding to the target real-time image as the target cultural relic monitoring sub-area; and determine the corresponding actual cultural relic type according to the target cultural relic monitoring sub-area.

[0011] Optionally, the early warning response device is further configured to store a pre-configured set of fire extinguishing equipment deployment plans. Each fire extinguishing equipment deployment plan in the set of fire extinguishing equipment deployment plans includes fire risk factor information, cultural relic type, and fire extinguishing equipment information. The fire risk factor information includes the type of fire risk factor and the corresponding risk level. The fire extinguishing equipment information includes the electromagnetic lock information corresponding to the fire extinguishing equipment. The electromagnetic lock information includes the device identifier; group the set of fire extinguishing equipment deployment plans according to the type of fire risk factor to obtain a smoke equipment deployment plan group and a flame equipment deployment plan group.

[0012] The present invention has the following beneficial effects: 1. It improves the accuracy and efficiency of the early warning system. Specifically, the early warning image acquisition device is responsible for collecting real-time image files of the cultural relic monitoring sub-areas and sending them to the early warning image processing device. The early warning image processing device receives the image files, analyzes the type of fire risk factor, the target cultural relic monitoring sub-area, and the actual cultural relic type, and sends the analysis results to the early warning response device. The early warning response device processes according to the type of fire risk factor. For smoke, the smoke risk level is judged by the concentration value, the appropriate fire extinguishing equipment is determined, and the first early warning instruction is sent. For flame, the proportion of the fire area is analyzed by the image segmentation model, the flame risk level is determined, the fire extinguishing equipment is selected, and the second early warning instruction is sent. The electromagnetic lock is used to respond to the early warning instruction, perform the unlocking operation, and release the relevant fire extinguishing equipment; realizing the accurate identification of fire risk factors, efficient risk assessment, automatic fire extinguishing equipment deployment, and rapid response, thereby improving the accuracy and overall efficiency of the fire early warning system, and further enhancing the cultural relic protection effect; 2. Improve the reliability of the early warning system and the ability to protect cultural relics. Specifically, by comparing the environmental stability index with the suitable range of cultural relics, calculating the comprehensive risk score, generating the abnormal early warning level and instructions; by identifying and evaluating the damage degree of cultural relics through post-disaster images, combining with the vulnerability level of cultural relics and matching the damage index, sorting and determining the repair priority, and generating the target repair plan; realizing precise monitoring and precise repair of cultural relics, thereby improving the reliability of the early warning system and the ability to protect cultural relics; 3. Improve the accuracy of the early warning system and the response ability, reduce the false alarm rate, improve the efficiency of the early warning system, and thus enhance the effect of cultural relic protection. Specifically, through the early warning response device, according to the regional information and risk adjustment coefficient of the cultural relic monitoring area, adjust the smoke risk level or flame risk level of the target cultural relic monitoring sub-area. When the adjusted smoke risk level or the adjusted flame risk level reaches the corresponding preset risk level threshold, obtain the geographical coordinates of the target cultural relic monitoring area, screen the nearest fire station, and generate the optimal path from the fire station to the target area. Finally, send the optimal path and path details to the command terminal of the fire station for guiding fire response. According to the regional characteristics determined by the characteristics of cultural relics, protect cultural relics in a targeted manner, reduce the false alarm rate, and achieve accurate early warning; send the fire terminal in time in case of fire and respond quickly, thereby improving the accuracy of the early warning system and the response ability, improving the efficiency of the early warning system, and thus enhancing the effect of cultural relic protection. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Combined with the drawings and referring to the following specific embodiments, the above and other features, advantages and aspects of the embodiments of the present invention will become more obvious. Throughout the drawings, the same or similar reference numerals denote the same or similar elements. It should be understood that the drawings are schematic, and the elements and elements are not necessarily drawn to scale.

[0014] Figure 1 is an exemplary structural schematic diagram of a security early warning system based on image recognition of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0015] The present invention will be described in more detail below with reference to the drawings. Although some embodiments of the present invention are shown in the drawings, it should be understood that the present invention can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. On the contrary, these embodiments are provided to more thoroughly and completely understand the present invention. It should be understood that the drawings and embodiments of the present invention are only for exemplary purposes and are not used to limit the protection scope of the present invention.

[0016] In addition, it should be noted that for the sake of description, only the parts related to the relevant invention are shown in the drawings. Without conflict, the embodiments in the present invention and the features in the embodiments can be combined with each other.

[0017] It should be noted that the concepts such as "first" and "second" mentioned in the present invention are only used to distinguish different devices, modules or units, and are not used to limit the order of functions executed by these devices, modules or units or their interdependent relationships.

[0018] It should be noted that the modifications of "one" and "multiple" mentioned in the present invention are illustrative rather than restrictive. Those skilled in the art should understand that unless otherwise clearly specified in the context, it should be understood as "one or more".

[0019] The names of the messages or information exchanged between multiple devices of the present invention are only for illustrative purposes and are not used to limit the scope of these messages or information.

[0020] The present invention will be described in detail below with reference to the accompanying drawings and in conjunction with embodiments.

[0021] As Figure 1 shown, a security warning system based on image recognition according to the present invention is shown, including the following devices: a warning image acquisition device 101, a warning image processing device 102, a warning response device 103, and an electromagnetic lock 104.

[0022] Among them, multiple warning image acquisition devices 101 are used to send the acquired multiple real-time image files to the warning image processing device 102.

[0023] In some embodiments, the warning image acquisition device 101 may be a camera, which is responsible for acquiring real-time image files of the pre-divided cultural relic monitoring sub-areas. The multiple cultural relic monitoring sub-areas are multiple sub-areas obtained by dividing the overall monitoring area within the monitored cultural relic building, and each cultural relic monitoring sub-area corresponds to one camera. The real-time image file refers to the latest image data obtained by the camera, and usually these images are generated and updated in real time during the actual monitoring process. Among them, each real-time image file includes a real-time image and an image number. The image number is used to identify and track the image, distinguish different images, or distinguish the images acquired from different monitoring sub-areas. On this basis, the camera captures image data, performs preliminary processing through an image processing module, and stores the image together with the number as a file. The camera establishes a communication connection with the warning image processing device 102 and sends the real-time image file to the warning image processing device 102.

[0024] The warning image processing device 102 is used to receive multiple real-time image files and send the type of fire risk factors, the target cultural relic monitoring sub-area, and the actual cultural relic type determined according to the multiple real-time image files to the warning response device 103.

[0025] In some embodiments, the early warning image processing device 102 may be a background server. By establishing a communication connection with a camera, it can receive multiple real-time image files. The multiple real-time image files are a set of image files generated by the early warning image acquisition device 101. Each real-time image file includes a real-time image and an image number. On this basis, the pre-deployed fire risk factor prediction model is used to analyze the real-time image to obtain the fire risk factor type. The fire risk factor type represents the risk characteristics related to fire. According to the matching of the image number and the area mapping table, the cultural relics monitoring sub-area corresponding to the real-time image is located to obtain the target cultural relics monitoring sub-area. Among them, the area mapping table includes the image number and the area number. The target cultural relics monitoring sub-area refers to the specific monitoring sub-area where the fire risk occurs. According to the mapping relationship table between the area number, the area number, and the cultural relic type, the corresponding actual cultural relic type is determined. The actual cultural relic type refers to the specific cultural relic type monitored in this area, such as paper cultural relics, pottery cultural relics, or wooden cultural relics. The early warning image processing device 102 establishes a communication connection with the early warning response device 103, and thus sends the fire risk factor type, the target cultural relics monitoring sub-area, and the actual cultural relic type to the early warning response device 103.

[0026] The early warning response device 103 is used to receive the fire risk factor type, the target cultural relics monitoring sub-area, and the actual cultural relic type. The fire risk factor type is smoke or flame. If the fire risk factor type is smoke, the smoke risk level of the target cultural relics monitoring sub-area is determined; according to the actual cultural relic type and the smoke risk level, the smoke extinguishing equipment information group is determined from the smoke equipment deployment plan group and the first target electromagnetic lock is determined; a first warning instruction is sent to the electromagnetic lock associated with the extinguishing equipment according to the device identifier corresponding to the first target electromagnetic lock; if the fire risk factor type is flame, the flame risk level of the target cultural relics monitoring sub-area is determined; according to the actual cultural relic type and the flame risk level, the flame extinguishing equipment information group is determined from the flame equipment deployment plan group and the second target electromagnetic lock is determined; a second warning instruction is sent to the electromagnetic lock associated with the extinguishing equipment according to the device identifier corresponding to the second target electromagnetic lock.

[0027] In some embodiments, the early warning response device 103 may be a background server. The early warning response device 103 establishes a communication connection with the early warning image processing device 102 to receive the type of fire risk factors, the target cultural relic monitoring sub-region, and the actual cultural relic type. Among them, the type of fire risk factors includes smoke or flame. Among them, smoke or flame is the type of fire risk factor. Smoke usually represents the gas or particulate matter generated at the beginning of a fire, while flame is an obvious sign of a fire. If the type of fire risk factor is smoke, the early warning response device 103 obtains the smoke concentration in the target monitoring sub-region through a connected sensor. The early warning response device 103 compares the obtained smoke concentration with a preset concentration threshold range to determine whether the smoke concentration exceeds the preset concentration threshold. If it exceeds, the smoke risk level is determined according to the level of the exceeded smoke concentration. Among them, the smoke concentration refers to the concentration of smoke detected in the target cultural relic monitoring sub-region. The smoke concentration is measured by a sensor, such as a gas sensor or an optical sensor, and reflects the fire risk of the target cultural relic monitoring sub-region. The preset concentration threshold range refers to a pre-set concentration range used to determine whether the smoke concentration reaches a dangerous level. When the smoke concentration exceeds the preset concentration threshold range, it indicates that the fire risk in this area is relatively high and early warning measures need to be taken. The smoke risk level is the smoke risk level of the target cultural relic monitoring sub-region obtained based on the comparison between the smoke concentration and the preset concentration threshold range. The smoke risk level may be divided into multiple levels, such as level 1, level 2, level 3, etc. On this basis, the determined actual cultural relic type and smoke risk level are used as query conditions to search for the corresponding smoke extinguishing equipment information group in the smoke equipment allocation plan group. Among them, the smoke extinguishing equipment information group refers to the information set of available smoke extinguishing equipment, such as the type, quantity of fire extinguishers, and relevant information of the corresponding electromagnetic locks, etc. On this basis, according to the location of the target cultural relic monitoring area, the distance between devices is calculated using a geographic information system or a device location database, and the electromagnetic lock corresponding to the shortest distance is determined to obtain the first target electromagnetic lock. According to the device identifier corresponding to the first target electromagnetic lock, a communication connection is established with the electromagnetic lock 104 to send a first early warning instruction to the electromagnetic lock associated with the fire extinguishing equipment. Among them, the device identifier is an identifier used to uniquely identify the electromagnetic lock. The first early warning instruction is the unlocking instruction corresponding to the first target electromagnetic lock.

[0028] If the type of fire risk factor is flame, the corresponding target real-time image is used as input data and input into a pre-trained fire risk factor segmentation model. The fire risk factor segmentation model analyzes the real-time image, identifies and segments the flame area, and generates a segmentation image containing the flame area. Among them, the fire risk factor segmentation model is a trained deep learning model, such as a convolutional neural network, which aims to segment fire risk factors from images, can identify the fire area in the image, and perform pixel-level segmentation on fire risk factors. The fire risk factor segmentation image is the image obtained after being processed by the fire risk factor segmentation model, and some areas in the image will be marked as fire risk factors. Then, the flame area in the fire risk factor segmentation image is separated from the background through image binarization, and the shape of the flame is identified through contour detection. Then, the number of pixels in the flame area is calculated and compared with the total number of pixels in the entire image to obtain the area ratio of the flame area. Among them, image binarization is an image processing technology that converts a color or grayscale image into an image containing only two pixel values (such as black and white). In the fire risk factor segmentation image, the purpose of binarization is to separate the flame area from the background, making the flame area white and the non-flame area black. Among them, the area ratio of the risk factor image refers to the proportion of the pixels marked as the flame area in the fire risk factor segmentation image to the entire image. The preset image area ratio is a preset ratio used to distinguish the flame risk level. On this basis, the background server compares the area ratio of the risk factor image with the preset image area ratio to determine the flame risk level. For example, the flame risk level can be divided into level one, level two, or level three, and the flame risk level reflects the threat degree of the flame to the cultural relic monitoring sub-region. For example, if the area ratio of the flame area exceeds 50%, the flame risk level is level three; if the area ratio of the flame area is between 20% and 50%, the flame risk level is level two; if the area ratio of the flame area is less than 20%, the flame risk level is level one. On this basis, the determined actual cultural relic type and flame risk level are used as query conditions to search for the corresponding flame extinguishing equipment information group in the flame equipment deployment plan group. Among them, the flame extinguishing equipment information group refers to the information set of available flame extinguishing equipment, such as the type and quantity of fire extinguishers and the relevant information of the corresponding electromagnetic locks, etc. On this basis, according to the location of the target cultural relic monitoring area, the distance between devices is calculated using a geographic information system or device location database, and the electromagnetic lock corresponding to the shortest distance is determined to obtain the second target electromagnetic lock. According to the device identifier corresponding to the second target electromagnetic lock, a communication connection is established with the electromagnetic lock 104, so as to send a second warning instruction to the electromagnetic lock associated with the fire extinguishing equipment. The second warning instruction is the unlocking instruction corresponding to the second target electromagnetic lock.

[0029] The electromagnetic lock 104 associated with the fire extinguishing equipment is used to respond to the first warning instruction or the second warning instruction and perform the unlocking operation for the electromagnetic lock.

[0030] In some embodiments, the electromagnetic lock 104 is an electronic device for locking or releasing the fire extinguishing equipment, which can establish a communication connection with the warning response device 103 to receive instructions for unlocking operations and release the corresponding fire extinguishing equipment.

[0031] In these embodiments, the accuracy and efficiency of the warning system are improved. Specifically, the warning image acquisition device is responsible for acquiring real-time image files of the cultural relic monitoring sub-region and sending them to the warning image processing device. The warning image processing device receives the image files, analyzes the types of fire risk factors, the target cultural relic monitoring sub-region and the actual cultural relic types, and sends the analysis results to the warning response device. The warning response device processes according to the types of fire risk factors. For smoke, the smoke risk level is judged through the concentration value, the appropriate fire extinguishing equipment is determined and the first warning instruction is sent. For flames, the proportion of the fire area is analyzed through the image segmentation model, the flame risk level is determined, the fire extinguishing equipment is selected and the second warning instruction is sent. The electromagnetic lock is used to respond to the warning instruction, perform the unlocking operation and release the relevant fire extinguishing equipment; realizing the accurate identification of fire risk factors, efficient risk assessment, automatic allocation of fire extinguishing equipment and rapid response, thus improving the accuracy and overall efficiency of the fire warning system, and further enhancing the cultural relic protection effect.

[0032] In some embodiments, in order to further solve the second technical problem described in the background art section, that is, "existing safety warning systems usually only focus on the warning of the fire itself, but do not consider the potential threats of environmental changes to cultural relics, and do not consider the needs of cultural relic characteristics for fire extinguishing and do not automatically generate targeted repair plans, thus affecting the cultural relic protection effect", in some embodiments of the present invention, the types of fire risk factors also include no fire risk, and The warning response device 103 is also used to obtain the environmental stability indicators in the target cultural relic monitoring sub-region, and the environmental stability indicators include the current temperature, the current humidity and the current air flow velocity; according to the actual cultural relic types, determine the appropriate temperature range, humidity range and air flow velocity range for the cultural relics.

[0033] In some embodiments, the fire risk factor type further includes no fire risk, where no fire risk indicates that there are no obvious fire risk factors in the target cultural relic monitoring sub-region. The background server establishes a communication connection with sensors pre-deployed in the target cultural relic monitoring sub-region to collect environmental data in the target cultural relic monitoring sub-region. Among them, the environmental stability index is a parameter used to describe whether the environmental conditions in the cultural relic monitoring sub-region are stable, and specifically includes the current temperature, the current humidity, and the current air flow velocity. Among them, the current temperature refers to the current environmental temperature in the corresponding target cultural relic monitoring sub-region, usually in degrees Celsius. The current humidity refers to the relative humidity in the target cultural relic monitoring sub-region, usually expressed as a percentage. The current air flow velocity represents the speed of air flow in the target cultural relic monitoring sub-region, usually in meters per second. According to the cultural relic type number corresponding to the actual cultural relic type in the target cultural relic monitoring sub-region, the suitable temperature range, the suitable humidity range, and the suitable air flow velocity range for the actual cultural relic type are retrieved from the preset cultural relic protection database.

[0034] Compare the current temperature with the suitable temperature range for the cultural relic to determine the temperature deviation degree; according to the temperature deviation degree, determine the temperature risk level score; compare the current humidity with the suitable humidity range for the cultural relic to determine the humidity deviation degree; according to the humidity deviation degree, determine the humidity risk level score; compare the current air flow velocity with the suitable air flow velocity range for the cultural relic to determine the air flow velocity deviation degree; according to the air flow velocity deviation degree, determine the air flow velocity risk level score.

[0035] In some embodiments, it is determined whether the current temperature is within the suitable temperature range for cultural relics. If the current temperature is lower than the lower limit of the suitable temperature range for cultural relics, the first temperature offset is the lower limit of the suitable temperature range for cultural relics minus the current temperature. According to the first temperature offset, the corresponding temperature deviation degree is queried. If the current temperature is higher than the upper limit of the suitable temperature range for cultural relics, the second temperature offset is the current temperature minus the upper limit of the suitable temperature range for cultural relics. According to the second temperature offset, the corresponding temperature deviation degree is queried. The temperature deviation degree refers to the degree of difference between the current temperature and the suitable temperature range for cultural relics, usually expressed as an absolute value. Then, by looking up the preset temperature risk level mapping table, the temperature risk level score corresponding to the temperature deviation degree is obtained. Among them, the preset temperature risk level mapping table includes the temperature deviation degree and the corresponding temperature risk level score. It is determined whether the current humidity is within the suitable humidity range for cultural relics. If the current humidity is lower than the lower limit of the suitable humidity range for cultural relics, the first humidity offset is the lower limit of the suitable humidity range for cultural relics minus the current humidity. According to the first humidity offset, the corresponding humidity deviation degree is queried. If the current humidity is higher than the upper limit of the suitable humidity range for cultural relics, the second humidity offset is the current humidity minus the upper limit of the suitable humidity range for cultural relics. According to the second humidity offset, the corresponding humidity deviation degree is queried. The humidity deviation degree refers to the degree of difference between the current humidity and the suitable humidity range for cultural relics, usually expressed as a percentage. Then, by querying the preset humidity risk level mapping table, the humidity risk level score corresponding to the humidity deviation degree is obtained. Among them, the preset humidity risk level mapping table includes the humidity deviation degree and the corresponding humidity risk level score. It is judged whether the current air flow velocity is within the suitable air flow velocity range for cultural relics. If the current air flow velocity is lower than the lower limit of the suitable air flow velocity range for cultural relics, the first air flow velocity offset is the lower limit of the suitable air flow velocity range for cultural relics minus the current air flow velocity. According to the first air flow velocity offset, the corresponding air flow velocity deviation degree is queried. If the current air flow velocity is higher than the upper limit of the suitable air flow velocity range for cultural relics, the second air flow velocity offset is the current air flow velocity minus the upper limit of the suitable air flow velocity range for cultural relics. According to the second air flow velocity offset, the corresponding air flow velocity deviation degree is queried. The air flow velocity deviation degree refers to the degree of difference between the current air flow velocity and the suitable air flow velocity range for cultural relics. Then, by looking up the preset air flow velocity risk level mapping table, the air flow velocity risk level score corresponding to the air flow velocity deviation degree is obtained. Among them, the preset air flow velocity risk level mapping table includes the air flow velocity deviation degree and the corresponding air flow velocity risk level score. Among them, the risk level score is usually expressed as a quantitative value.

[0036] Determine the comprehensive risk score based on the temperature risk level score, humidity risk level score, and air velocity risk level score; generate an abnormal warning level according to the comprehensive risk score and the preset comprehensive risk score level; generate a corresponding abnormal warning instruction according to the abnormal warning level; and send the abnormal warning instruction to the user terminal.

[0037] In some embodiments, corresponding weights are assigned to the temperature risk level score, humidity risk level score, and air velocity risk level score respectively to calculate the comprehensive risk score. Among them, the comprehensive risk score is used to evaluate the overall environmental risk level of the cultural relic monitoring sub-region. The preset comprehensive risk score level is a standard for classifying the comprehensive risk score, usually divided into multiple levels and corresponding to specific score ranges. On this basis, the comprehensive risk score is classified using the score range corresponding to the preset comprehensive risk score level, and the corresponding abnormal warning level is marked according to the range to which the comprehensive risk score belongs. Among them, the abnormal warning level is the severity of the warning determined based on the matching result of the comprehensive risk score and the preset comprehensive risk score level. According to the abnormal warning level, the corresponding abnormal warning instruction is found. The abnormal warning instruction is a ventilation equipment start signal. The warning response device 103 establishes a communication connection with the user terminal and sends the abnormal warning instruction to the user terminal. Among them, the user terminal is a device that interacts with the warning response device 103, receives and responds to the abnormal warning instruction generated by the comprehensive risk score. The user can view the risk information on the terminal and make a response. The user terminal can be a laptop computer or a tablet computer, etc.

[0038] Among them, the warning response device 103 is also used to store a pre-configured cultural relic damage index table, which includes multiple cultural relic types, the cultural relic vulnerability levels corresponding to each cultural relic type, and damage indexes.

[0039] In some embodiments, the cultural relic damage index table is a pre-configured table or database that records different cultural relic types and their cultural relic vulnerability levels and damage indexes. Among them, the cultural relic vulnerability level is a numerical or hierarchical representation of the anti-damage ability of cultural relics in a fire. The damage index is a quantitative index reflecting the damage state of cultural relics, such as crack length, damaged area, etc.

[0040] The early warning response device 103 is also used to receive multiple post-disaster real-time image files of multiple cultural relic monitoring sub-areas. Each post-disaster real-time image file among the multiple post-disaster real-time image files includes a post-disaster real-time image and a corresponding image number; input the post-disaster real-time image in each post-disaster real-time image file into a pre-trained cultural relic recognition model to obtain a cultural relic recognition result, where the cultural relic recognition result includes cultural relics or no cultural relics; determine the post-disaster real-time image corresponding to the cultural relic recognition result of cultural relics as the target post-disaster real-time image; determine the corresponding target cultural relic type according to the target post-disaster real-time image; match the target cultural relic type with the cultural relic damage index table to obtain a matching result, and the matching result is used to represent successful or failed matching; if the matching result represents successful matching, analyze the damage index corresponding to the target cultural relic type to determine the degree of damage.

[0041] In some embodiments, the early warning response device 103 establishes a communication connection with a camera to receive multiple post-disaster real-time image files. Among them, the post-disaster real-time image file is a real-time image file collected after a fire. Each post-disaster real-time image file includes a post-disaster real-time image and a post-disaster real-time image number. Among them, the post-disaster real-time image is an image recording the state of cultural relics after the fire. The post-disaster real-time image number is a unique number used to identify the corresponding post-disaster real-time image. The cultural relic recognition model is a machine learning model used to identify whether there are cultural relics in an image. Input the real-time image into the cultural relic recognition model to obtain a cultural relic recognition result. Among them, the cultural relic recognition result includes cultural relics or non-cultural relics. Determine the post-disaster real-time image corresponding to the cultural relic recognition result of cultural relics as the target post-disaster real-time image. Query the corresponding area number according to the image number corresponding to the target post-disaster real-time image, and find the corresponding target cultural relic type according to the area number. Traverse the cultural relic types in the cultural relic damage index table to find whether there is a cultural relic type that is the same as the target cultural relic type. If they are the same, it means successful matching. If they are not the same, it means failed matching. If the matching is successful, further determine the damage index of the target cultural relic type. Then specifically analyze the damage situation according to the matched damage index to determine the degree of damage. Among them, the degree of damage refers to the level of damage to cultural relics.

[0042] If the matching result represents failed matching, determine the vulnerability level of the target cultural relic type corresponding to the target cultural relic type; compare the vulnerability level of the target cultural relic type with the vulnerability levels of each cultural relic type in the cultural relic damage index table to obtain multiple cultural relic vulnerability level differences; sort the multiple cultural relic types in ascending order according to the corresponding cultural relic vulnerability level differences to obtain a cultural relic type sequence; query the sorting numbers corresponding to each cultural relic type in the cultural relic type sequence in a preset damage index importance coefficient table to obtain the damage index importance coefficients corresponding to each cultural relic type; obtain the degree of damage corresponding to the target cultural relic type according to the damage index importance coefficients and damage indexes corresponding to each cultural relic type.

[0043] In some embodiments, if the matching fails, according to the target cultural relic type, query the mapping relationship table between the cultural relic type and the cultural relic vulnerability level to obtain the cultural relic vulnerability level corresponding to the target cultural relic type. Perform a difference operation between the vulnerability level of the target cultural relic type and the vulnerability levels of each cultural relic type in the damage index table to obtain multiple cultural relic vulnerability level differences. Sort the damage index table of cultural relics in ascending order according to the corresponding cultural relic vulnerability level differences of multiple cultural relic types to obtain a sequence of cultural relic types. Query the sorting numbers corresponding to each cultural relic type in the sequence of cultural relic types in the preset damage index importance coefficient table to obtain the damage index importance coefficients corresponding to each cultural relic type. The damage index importance coefficient table includes the cultural relic type, the sorting number corresponding to the cultural relic type, and the damage index importance coefficient. For the target cultural relic type, analyze each damage index, and through the damage index importance coefficients corresponding to each damage index, obtain the comprehensive damage degree of the target cultural relic type. The damage degree is used to determine the restoration plan.

[0044] According to the damage degree and the target cultural relic type, determine the target restoration plan from the pre-stored set of restoration plans; send the target restoration plan to the restoration terminal.

[0045] In some embodiments, use the target cultural relic type and the corresponding damage degree as query conditions to search in the set of restoration plans to obtain the target restoration plan. The background server sends the target restoration plan to the restoration terminal through the network. The restoration terminal refers to the device used by the operator to receive the restoration plan and perform the restoration operation, usually a computer or a smart terminal.

[0046] Among them, multiple warning image acquisition devices 101 are set in multiple pre-divided cultural relic monitoring sub-areas. Each warning image acquisition device 101 among the multiple warning image acquisition devices 101 corresponds to a cultural relic monitoring sub-area. Each of the multiple real-time image files includes a real-time image and an image number; and The warning image processing device 102 is further configured to input the real-time image in each real-time image file into a pre-deployed fire risk factor prediction model to obtain fire risk prediction information. The fire risk prediction information includes the type of fire risk factor and the prediction confidence level. The type of fire risk factor includes smoke or flame; determine the real-time image corresponding to the prediction confidence level greater than the preset prediction confidence level threshold as the target real-time image; according to the image number corresponding to the target real-time image, determine the cultural relic monitoring sub-area corresponding to the target real-time image; determine the cultural relic monitoring sub-area corresponding to the target real-time image as the target cultural relic monitoring sub-area; according to the target cultural relic monitoring sub-area, determine the corresponding actual cultural relic type.

[0047] In some embodiments, the input pre-deployed fire risk factor prediction model is a model trained using machine learning or deep learning algorithms, aiming to analyze the content in images, especially for the identification of fire risks. The model can identify potential fire factors by inputting real-time images. The prediction process of the fire risk factor prediction model is to extract features and classify the input images through machine learning algorithms (such as convolutional neural networks) to identify specific fire risk factors in the images. The machine learning algorithm automatically extracts features from a large number of labeled image data and classifies them, and finally can accurately predict the fire risk factors existing in the real-time images. On this basis, the early warning image processing device 102 extracts real-time images from each received real-time image file, performs necessary preprocessing, and then inputs the real-time images into the pre-deployed fire risk factor prediction model to obtain fire risk prediction information. The fire risk prediction information includes the type of fire risk factors and the prediction confidence level. The type of fire risk factors includes smoke or flame. Among them, the fire risk prediction information refers to the output result obtained through the fire risk factor prediction model. It contains two parts of information: the type of fire risk factors and the corresponding prediction confidence level. The type of fire risk factors refers to the type of fire risk factors that may exist in the real-time image, such as smoke or flame. The prediction confidence level represents the confidence level of the model in the prediction result, usually a probability value, reflecting the reliability of the model's judgment on the existence of a certain fire risk factor. For example, the type of fire risk factor: smoke. Prediction confidence level: 85%.

[0048] In some embodiments, the preset prediction confidence level threshold refers to a preset threshold used to distinguish which prediction results are sufficiently credible. Only real-time images with a prediction confidence level greater than the preset prediction confidence level threshold will be determined as target real-time images. Target real-time images refer to those real-time images with a prediction confidence level greater than the preset threshold. Target real-time images are considered to have potential fire risks and need to be further processed and analyzed. The target cultural relic monitoring sub-region refers to when a certain target real-time image is identified, the cultural relic monitoring sub-region corresponding to the target real-time image is determined as the target cultural relic monitoring sub-region. On this basis, each image file has a unique image number, and the image number corresponds to the cultural relic monitoring sub-region. By looking up the mapping relationship table between the image numbers and the cultural relic monitoring sub-regions established in the database through the image number, the cultural relic monitoring sub-region corresponding to the target real-time image is obtained. The cultural relic monitoring sub-region corresponding to the target real-time image is determined as the target cultural relic monitoring sub-region. According to the target cultural relic monitoring sub-region, the corresponding actual cultural relic type is determined.

[0049] Among them, the early warning response device 103 is also used to store a pre-configured set of fire-fighting equipment deployment plans. Each fire-fighting equipment deployment plan in the set of fire-fighting equipment deployment plans includes fire risk factor information, cultural relic type, and fire-fighting equipment information. The fire risk factor information includes the fire risk factor type and the corresponding risk level. The fire-fighting equipment information includes the electromagnetic lock information corresponding to the fire-fighting equipment. The electromagnetic lock information includes the device identifier. The set of fire-fighting equipment deployment plans is grouped according to the fire risk factor type to obtain a smoke equipment deployment plan group and a flame equipment deployment plan group.

[0050] In some embodiments, the early warning response device 103 stores the pre-configured set of fire-fighting equipment deployment plans in the form of a database. The set of fire-fighting equipment deployment plans is a collection of pre-configured fire-fighting equipment usage and allocation plans, which is used to provide appropriate fire-fighting equipment suggestions for different fire risk factors and cultural relic types. Among them, each fire-fighting equipment deployment plan includes fire risk factor information, cultural relic type, and fire-fighting equipment information. The fire risk factor information includes the fire risk factor type and the corresponding risk level. The fire-fighting equipment information includes the electromagnetic lock information corresponding to the fire-fighting equipment. The electromagnetic lock information includes the device identifier. On this basis, each record in the set of fire-fighting equipment deployment plans is traversed and classified according to the "fire risk factor type" field. For the plans of the smoke type, they are added to the smoke equipment deployment plan group. For the plans of the flame type, they are added to the flame equipment deployment plan group.

[0051] Among them, a safety early warning system based on image recognition of the present invention further includes: A plurality of light sensors are arranged in a plurality of cultural relic monitoring sub-regions. Each light sensor in the plurality of light sensors is used to collect the light intensity of the corresponding cultural relic monitoring sub-region. Among them, each cultural relic monitoring sub-region corresponds to a region number; the light intensity and the region number are sent to the early warning image processing device 102.

[0052] In some embodiments, the light sensor is a device for sensing and measuring the light intensity within the cultural relic monitoring sub-region. It can generally measure the light intensity in the environment and convert it into an electrical signal for transmission to the background server. Each cultural relic monitoring sub-region corresponds to one light sensor. The light intensity refers to the illumination level within a specific cultural relic monitoring sub-region, which helps to judge the potential impact of ambient light on the cultural relics. The region number is a unique identifier for each cultural relic monitoring sub-region, used to identify different cultural relic monitoring sub-regions and ensure accurate correspondence with the light intensity of each sub-region. Based on this, when each light sensor is installed in the corresponding cultural relic monitoring sub-region, it continuously or periodically collects the ambient light in that region and then converts the ambient light into an electrical signal representing the light intensity of the corresponding cultural relic monitoring sub-region. Among them, the light intensity is usually a digital signal and can be transmitted via wireless communication or a wired network. Each cultural relic monitoring sub-region has a uniquely pre-set region number. After each light sensor collects the light intensity and matches it with the corresponding region number, it establishes a communication connection with the background server to transmit the light intensity and the corresponding region number to the warning image processing device 102.

[0053] Multiple color temperature sensors are arranged in multiple cultural relic monitoring sub-regions. Each color temperature sensor among the multiple color temperature sensors is used to collect the color temperature of the corresponding cultural relic monitoring sub-region; the color temperature and the region number are sent to the warning image processing device 102.

[0054] In some embodiments, the color temperature sensor can be an RGB color temperature sensor, where RGB represents the three color channels of red, green, and blue. The color temperature sensor is used to measure the color temperature of the ambient light in each cultural relic monitoring sub-region. Color temperature is the hue of the light emitted by a light source, reflecting the degree of warmth or coldness of the light. Each color temperature sensor corresponds to a specific cultural relic monitoring sub-region, and each cultural relic monitoring sub-region has a unique region number. The color temperature sensor will be automatically associated with a certain region number, thus ensuring that the collected color temperature data can correspond to a specific cultural relic monitoring sub-region. Based on this, the color temperature sensor captures the spectral characteristics of the ambient light through optical elements and a color temperature measurement module (such as a photoelectric sensor or a filter) and calculates the color temperature according to a specific algorithm. After each color temperature sensor collects the color temperature and its corresponding region number, it establishes a communication connection with the background server to send the color temperature and the corresponding region number to the warning image processing device 102.

[0055] The warning image processing device 102 is further configured to receive the light intensity, color temperature, and region number of each cultural relic monitoring sub-region; determine the calibration parameters of the corresponding warning image acquisition device according to the light intensity and color temperature; and send the warning image acquisition device calibration parameters and the region number to the device management terminal so that the device management personnel can calibrate each warning image acquisition device.

[0056] In some embodiments, the early warning image processing device 102 establishes communication connections with the light sensor and the color temperature sensor respectively, so as to receive the light intensity, color temperature and area number of each cultural relic monitoring sub-region. The calibration parameters of the early warning image acquisition device are used to calibrate the camera, so that it can accurately and stably acquire images under different lighting conditions. The calibration parameters generally include exposure time, gain, white balance, brightness and contrast, etc. On this basis, the background server calculates the corresponding calibration parameters according to the light intensity and color temperature of each cultural relic monitoring sub-region. Adjust the exposure time of the image acquisition device according to the light intensity to ensure the image quality. Adjust the white balance according to the color temperature to make the color of the image more accurate and natural. Under low light conditions, it may be necessary to adjust the gain and contrast to improve the clarity and visibility of the image. The background server sends the calculated calibration parameters of the early warning image acquisition device and the area number to the device management terminal together. The sending process can be realized by various communication means. The calibration parameters of the early warning image acquisition device are provided to the device management personnel for their subsequent device calibration operations. Among them, the device management terminal is a system or interface for managing and configuring devices, and device management personnel can use it to adjust, calibrate and maintain the devices.

[0057] In these embodiments, the reliability of the early warning system and the cultural relic protection ability are improved. Specifically, by comparing the environmental stability index with the suitable range of cultural relics, calculating the comprehensive risk score, generating the abnormal early warning level and instructions; by identifying the damaged degree of cultural relics through post-disaster images, combining with the vulnerability level of cultural relics and the matching of damage indicators, sorting and determining the repair priority, and generating the target repair plan; realizing precise monitoring and precise repair of cultural relics, thus improving the reliability of the early warning system and the cultural relic protection ability.

[0058] In some embodiments, in order to further solve Technical Problem 3 described in the background art part, that is, "the existing early warning system does not consider the regional characteristics corresponding to the characteristics of cultural relics for undifferentiated fire extinguishing, which is prone to false alarms, resulting in inaccurate early warnings, and the response is not timely during a fire, resulting in poor early warning efficiency and further poor cultural relic protection effect", in some embodiments of a safety early warning system based on image recognition of the present invention, the early warning response device 103 stores the area information within the target cultural relic monitoring area, wherein the area information includes the area information of each cultural relic monitoring sub-region among a plurality of cultural relic monitoring sub-regions, and the area information includes the risk adjustment coefficient corresponding to each cultural relic monitoring sub-region, the area number corresponding to each cultural relic monitoring sub-region, and the area type corresponding to each cultural relic monitoring sub-region, and the area type is area prohibited or non-area prohibited; on this basis, the early warning response device 103 is further configured to perform the following steps: Step 1: Determine the area type according to the area number corresponding to the target cultural relic monitoring sub-area; if the area type is area prohibition, generate an area prohibition strategy; if the area type is non-area prohibition, obtain the adjusted risk level according to the fire risk factor type and risk adjustment coefficient corresponding to the target cultural relic monitoring sub-area, and the adjusted risk level is the adjusted flame risk level or the adjusted smoke risk level.

[0059] In some embodiments, the area information can be stored in the database of the early warning response device 103. The area information refers to the specific division of each cultural relic monitoring sub-area. The area information includes multiple cultural relic monitoring sub-areas, the risk adjustment coefficient corresponding to each cultural relic monitoring sub-area, and the area type corresponding to each cultural relic monitoring sub-area, and the area type is area prohibition or non-area prohibition. Among them, the risk adjustment coefficient is a correction parameter assigned to each cultural relic monitoring sub-area according to the location and environment of the cultural relics in the area, and is used to reflect the impact of the importance and vulnerability of the cultural relics in the area on risk assessment. According to the area number corresponding to the target cultural relic monitoring sub-area, extract the area type and risk adjustment coefficient corresponding to each cultural relic monitoring sub-area from the area information of each cultural relic monitoring sub-area, and the area type is area prohibition or non-area prohibition. If the area type is area prohibition, generate specific protection measures for this area according to the preset area prohibition strategy template. Among them, area prohibition means that in certain specific areas, the risk adjustment coefficient corresponding to area prohibition is 0. The area prohibition strategy is a special protection measure formulated for the area corresponding to area prohibition. For example, increase the patrol frequency and the number of patrol personnel, etc. In practice, in a large cultural relic exhibition hall, there are activities such as worship in a certain cultural relic monitoring sub-area, and items such as incense and candles appear during the activities. When these items burn, there will also be risk factors such as smoke and flame. To avoid misidentifying the smoke or flame in this area, this area is set as area prohibition. If the area type is non-area prohibition, obtain the adjusted risk level according to the fire risk factor type and risk adjustment coefficient corresponding to the target cultural relic monitoring sub-area, and the adjusted risk level is the adjusted flame risk level or the adjusted smoke risk level. Among them, the wind direction adjustment coefficient of the area corresponding to non-area prohibition is greater than 0. Specifically, the risk adjustment coefficient is lower when the target cultural relic monitoring sub-area is near the fire extinguishing equipment or in a position with poor ventilation conditions. Extract the risk adjustment coefficient of the target cultural relic monitoring sub-area from the area information. If the fire risk factor type corresponding to the target cultural relic monitoring sub-area is smoke, use the risk adjustment coefficient to correct the smoke risk level corresponding to the target cultural relic monitoring sub-area to obtain the updated and adjusted smoke risk level. If the fire risk factor type corresponding to the target cultural relic monitoring sub-area is flame, use the risk adjustment coefficient to correct the flame risk level corresponding to the target cultural relic monitoring sub-area to obtain the updated and adjusted flame risk level. Store the adjusted adjusted smoke risk level or adjusted flame risk level for subsequent early warning and response decisions.

[0060] Step 2: If the adjusted smoke risk level is greater than or equal to the preset smoke risk level threshold, or the adjusted fire risk level is greater than or equal to the preset fire risk level threshold, obtain the geographical coordinate information of the target cultural relic monitoring area; extract the geographical coordinate information of the fire stations from the pre-stored fire station database, and screen the nearest fire station; determine the nearest fire station as the target fire station; obtain obstacle data based on the geographical coordinate information of the target cultural relic monitoring area and the geographical coordinate information of the target fire station; generate the optimal path and the corresponding path details according to the pre-stored path network diagram and the obstacle data.

[0061] In some embodiments, the preset smoke risk level threshold is a preset value, representing the demarcation point for whether further emergency response is required for the risk. The preset fire risk level threshold is also a preset value, representing the demarcation point for whether further emergency response is required for the risk. On this basis, if the adjusted smoke risk level is greater than or equal to the preset smoke risk level threshold, or the adjusted fire risk level is greater than or equal to the preset fire risk level threshold, then query the mapping relationship table between the cultural relic monitoring area and the geographic coordinate information to obtain the geographic coordinate information of the target cultural relic monitoring area. The fire station database is a pre-stored database that contains information on all fire stations, including names, addresses, geographic coordinates, etc. The geographic coordinates of each fire station are also represented by standard GPS (Global Positioning System) coordinates. The GPS coordinates consist of latitude and longitude. On this basis, extract the geographic coordinate fields of all fire stations from the fire station database to obtain the geographic coordinate information of each fire station. Use a distance calculation method. For example, use the Euclidean distance formula to calculate the distance between the target cultural relic monitoring area and each fire station, compare the distances of all fire stations, and select the fire station with the shortest distance as the target fire station. Among them, the Euclidean distance is the length of the shortest straight line between two points measured on a plane or in space. Through the geographic information system, based on the geographic coordinate information of the target cultural relic monitoring area and the geographic coordinate information of the target fire station, analyze and identify the obstacle data between the two. Among them, the obstacle data is a description of the obstacle information on the path, including dynamic or static obstacles such as construction areas, traffic congestion, and road closures. On this basis, load the pre-stored path network diagram and obstacle data, and use the shortest path algorithm to calculate the optimal path from the nearest fire station to the target cultural relic monitoring area and the corresponding path details. Among them, the path network diagram is a map model based on nodes and edges, reflecting the road system in the target area and its surroundings, including each road node and connection relationship. Usually generated in combination with map data. The optimal path is the driving route with the shortest time and the fewest obstacles from the fire station to the target cultural relic monitoring area. The optimal path is calculated dynamically and may vary due to real-time road conditions. The path details include the detailed information of the optimal path, such as the names of the sections passed along the way, turning points, mileage, estimated arrival time, etc.

[0062] Step 3, send the optimal path and the corresponding path details to the command terminal corresponding to the target fire station.

[0063] In some embodiments, the early warning response device 103 establishes a communication connection with the command terminal corresponding to the target fire station, so as to send the optimal path and the corresponding path details to the command terminal corresponding to the target fire station. Among them, the command terminal can be a computer, a tablet or a special display device, connected to the command and dispatch system. The command terminal is a dedicated device in the fire station, used to receive and display rescue instructions, path planning and other information.

[0064] In these embodiments, the accuracy and response ability of the early warning system are improved, the false alarm rate is reduced, the efficiency of the early warning system is improved, and thus the cultural relic protection effect is enhanced. Specifically, the early warning response device adjusts the smoke risk level or the flame risk level of the target cultural relic monitoring sub-region according to the regional information and the risk adjustment coefficient of the cultural relic monitoring area. When the adjusted smoke risk level or the adjusted flame risk level reaches the corresponding preset risk level threshold, the geographical coordinates of the target cultural relic monitoring area are obtained, the nearest fire station is selected, and the optimal path from the fire station to the target area is generated. Finally, the optimal path and the path details are sent to the command terminal of the fire station for guiding the fire response. According to the regional characteristics determined by the cultural relic characteristics, the cultural relics are protected in a targeted manner, the false alarm rate is reduced, and accurate early warning is achieved; when a fire occurs, the fire terminal is sent in time for rapid response, so as to improve the accuracy and response ability of the early warning system, improve the efficiency of the early warning system, and thus enhance the cultural relic protection effect.

[0065] The above description is only some preferred embodiments of the present invention and an explanation of the applied technical principles. Those skilled in the art should understand that the scope of the invention involved in the present invention is not limited to the technical solutions formed by the specific combination of the above technical features, and should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above inventive concept. For example, the technical solutions formed by mutually replacing the above features with the technical features (but not limited to) having similar functions disclosed in the present invention.

Claims

1. A safety warning system based on image recognition, characterized in that: include: A plurality of early warning image acquisition devices, wherein the plurality of early warning image acquisition devices are used to send a plurality of acquired real-time image files to the early warning image processing device; An early warning image processing device, the early warning image processing device is used to receive a plurality of real-time image files, and send the fire risk factor type, the target cultural relic monitoring sub-area and the actual cultural relic type determined according to the plurality of real-time image files to the early warning response device; An early warning response device, the early warning response device is used to receive the fire risk factor type, the target cultural relic monitoring sub-area and the actual cultural relic type, the fire risk factor type is smoke or flame; if the fire risk factor type is smoke, then determine the smoke risk level of the target cultural relic monitoring sub-area; According to the actual cultural relic type and smoke risk level, determine the smoke fire extinguishing equipment information group from the smoke equipment deployment plan group and determine the first target electromagnetic lock; send a first warning instruction to the electromagnetic lock associated with the fire extinguishing equipment according to the equipment identifier corresponding to the first target electromagnetic lock; if the fire risk factor type is flame, determine the flame risk level of the target cultural relic monitoring sub-area; According to the actual cultural relic type and the flame risk level, determine the flame extinguishing equipment information group from the flame equipment deployment plan group and determine the second target electromagnetic lock; Sending a second warning instruction to the electromagnetic lock associated with the fire extinguishing equipment according to the equipment identifier corresponding to the second target electromagnetic lock; An electromagnetic lock associated with the fire extinguishing equipment is used to respond to the first warning instruction or the second warning instruction and perform an unlocking operation on the electromagnetic lock.

2. The image recognition-based safety warning system according to claim 1, characterized in that: Also includes: A plurality of light sensors are arranged in a plurality of cultural relics monitoring sub-areas, each of the plurality of light sensors is used to collect the light intensity of the corresponding cultural relics monitoring sub-area, wherein each cultural relics monitoring sub-area corresponds to an area number; and the light intensity and the area number are sent to the early warning image processing device; A plurality of color temperature sensors are arranged in a plurality of cultural relics monitoring sub-areas; each of the plurality of color temperature sensors is used to collect the color temperature of the corresponding cultural relics monitoring sub-area; and the color temperature and the area number are sent to the early warning image processing device; The early warning image processing device is also used to receive the light intensity, color temperature and area number of each cultural relic monitoring sub-area; determine the corresponding early warning image acquisition device calibration parameters according to the light intensity and the color temperature; send the early warning image acquisition device calibration parameters and the area number to the equipment management terminal, so that the equipment management personnel can calibrate each early warning image acquisition device.

3. The image recognition-based safety warning system according to claim 2, characterized in that: The fire risk factor types also include no fire risk, and The early warning response device is also used to obtain an environmental stability index in the target cultural relic monitoring sub-area, wherein the environmental stability index includes current temperature, current humidity and current air flow velocity; and determine a suitable range of cultural relic temperature, a suitable range of cultural relic humidity and a suitable range of cultural relic air flow velocity according to the actual cultural relic type; Comparing the current temperature with the appropriate temperature range of the cultural relic to determine the degree of temperature deviation; Determine the temperature risk level score according to the temperature deviation degree; compare the current humidity with the suitable range of humidity for the cultural relics to determine the humidity deviation degree; determine the humidity risk level score according to the humidity deviation degree; compare the current airflow speed with the suitable range of airflow speed for the cultural relics to determine the airflow speed deviation degree; determine the airflow speed risk level score according to the airflow speed deviation degree; Determining a comprehensive risk score according to the temperature risk level score, the humidity risk level score and the airflow velocity risk level score; Generate an abnormal warning level according to the comprehensive risk score and the preset comprehensive risk score level; According to the abnormal warning level, a corresponding abnormal warning instruction is generated; and the abnormal warning instruction is sent to the user terminal.

4. The image recognition-based safety warning system according to claim 3, characterized in that: The early warning response device is also used to store a pre-configured cultural relic damage index table, which includes multiple cultural relic types, cultural relic vulnerability levels and damage indicators corresponding to each cultural relic type; The early warning response device is also used to receive multiple post-disaster real-time image files of multiple cultural relics monitoring sub-areas, each of the multiple post-disaster real-time image files including a post-disaster real-time image and a corresponding image number; input the post-disaster real-time image in each post-disaster real-time image file into a cultural relic recognition model to obtain a cultural relic recognition result, the cultural relic recognition result including cultural relics or no cultural relics; determine the post-disaster real-time image corresponding to the cultural relic with the cultural relic recognition result as the target post-disaster real-time image; determine the corresponding target cultural relic type according to the target post-disaster real-time image; match the target cultural relic type with the cultural relic damage index table to obtain a matching result, the matching result is used to indicate a successful match or a failed match; if the matching result indicates a successful match, analyze the damage index corresponding to the target cultural relic type to determine the degree of damage; If the matching result indicates a matching failure, determining the cultural relic vulnerability level corresponding to the target cultural relic type according to the target cultural relic type; Compare the cultural relic vulnerability level corresponding to the target cultural relic type with the cultural relic vulnerability level corresponding to each cultural relic type in the cultural relic damage index table to obtain a plurality of cultural relic vulnerability level differences; sort the plurality of cultural relic types in order from small to large according to the corresponding cultural relic vulnerability level differences to obtain a cultural relic type sequence; query the sorting number corresponding to each cultural relic type in the cultural relic type sequence in a preset damage index importance coefficient table to obtain the damage index importance coefficient corresponding to each cultural relic type; According to the damage index importance coefficient and damage index corresponding to each type of cultural relic, the damage degree corresponding to the target cultural relic type is obtained; Determining a target restoration plan from a set of pre-stored restoration plans according to the degree of damage and the type of the target cultural relic; The target repair solution is sent to the repair terminal.

5. The image recognition-based safety warning system according to claim 4, characterized in that: The plurality of early warning image acquisition devices are arranged in a plurality of pre-divided cultural relics monitoring sub-areas, each of the plurality of early warning image acquisition devices corresponds to a cultural relics monitoring sub-area, and each of the plurality of real-time image files includes a real-time image and an image number; as well as The early warning image processing device is also used to input the real-time image in each real-time image file into a pre-deployed fire risk factor prediction model to obtain fire risk prediction information, wherein the fire risk prediction information includes a fire risk factor type and a prediction confidence, wherein the fire risk factor type includes smoke or flame; the real-time image corresponding to the prediction confidence greater than a preset prediction confidence threshold is determined as the target real-time image; the cultural relics monitoring sub-area corresponding to the target real-time image is determined according to the image number corresponding to the target real-time image; the cultural relics monitoring sub-area corresponding to the target real-time image is determined as the target cultural relics monitoring sub-area; and the corresponding actual cultural relics type is determined according to the target cultural relics monitoring sub-area.

6. The image recognition-based safety warning system according to claim 5, characterized in that: The early warning response device is also used to store a pre-configured fire extinguishing equipment deployment plan set, each fire extinguishing equipment deployment plan in the fire extinguishing equipment deployment plan set includes fire risk factor information, cultural relic type and fire extinguishing equipment information, the fire risk factor information includes fire risk factor type and corresponding risk level, the fire extinguishing equipment information includes electromagnetic lock information corresponding to the fire extinguishing equipment, and the electromagnetic lock information includes equipment identification; the fire extinguishing equipment deployment plan set is grouped according to the fire risk factor type to obtain a smoke equipment deployment plan group and a flame equipment deployment plan group.

Images