Fire prevention and alarm system based on multi-modal artificial intelligence model

Abstract

The invention discloses a fire prevention and alarm system based on a multi-modal artificial intelligence model. The fire prevention and alarm system comprises a data acquisition module, a multi-modal fusion module, an artificial intelligence analysis module, an alarm module and a background management module. The data acquisition module acquires multi-modal data such as video image data, smoke concentration data, temperature data, humidity data and combustible gas concentration data; the multi-modal fusion module carries out fusion processing on the data to generate a multi-modal feature vector; the artificial intelligence analysis module analyzes the feature vectors based on a multi-modal artificial intelligence model, and outputs a fire occurrence probability and a risk level; when the alarm condition is met, the alarm module is started and sends alarm information to related personnel; the background management module carries out all-around management on the system. According to the invention, various modal data are fused, the artificial intelligence technology is utilized to realize early warning and accurate alarm of a fire disaster, related personnel can be prompted to take measures in time, fire disaster loss is reduced, and the system is suitable for various places needing fire disaster monitoring and has important fire safety value.

Description

[0001] manual Technical Field

[0002] The present invention relates to the field of fire safety technology, and in particular to a fire prevention and alarm system based on a multimodal artificial intelligence model. Background Art

[0003] Fires are common safety incidents, posing a serious threat to life and property. Traditional fire alarm systems rely primarily on single-mode devices, such as smoke detectors and temperature sensors, to detect fires. This presents limitations. For example, smoke detectors may not detect fires in the early stages of a fire when smoke concentrations are low, and temperature sensors may not accurately reflect the extent of a fire in certain environments (such as well-ventilated areas). Furthermore, these traditional devices struggle to comprehensively analyze and provide early warning of potential risk factors before a fire breaks out.

[0004] With the development of artificial intelligence (AI), fire detection methods based on technologies such as video surveillance have emerged. However, most of these methods analyze a single video image and are susceptible to factors such as ambient lighting and obstructions, leading to false alarms and missed alarms. Therefore, a more intelligent and efficient fire prevention and alarm system that can integrate information from multiple modalities is needed to improve the accuracy and timeliness of fire detection, enabling effective fire prevention and rapid fire alarms. Summary of the Invention

[0005] The purpose of this invention is to provide a fire prevention and alarm system based on a multimodal artificial intelligence model, which uses artificial intelligence technology to integrate data from multiple modalities to achieve early warning and accurate alarm of fires.

[0006] The fire prevention and alarm system of the present invention includes the following modules:

[0007] Data acquisition module: used to collect data of various modes related to fire, including but not limited to video image data, smoke concentration data, temperature data, humidity data, and combustible gas concentration data. Video image data can be obtained through cameras installed in buildings to observe flames, smoke and other conditions at the scene; smoke concentration data, temperature data, humidity data and combustible gas concentration data are respectively monitored and collected in real time by corresponding sensors. Multimodal fusion module: performs fusion processing on the collected data of different modes. Due to the different types and characteristics of different data, various types of data are first pre-processed, such as data cleaning and normalization. Then, deep learning and other technologies are used to fuse the flame and smoke feature information in the video image with data such as smoke concentration, temperature, humidity and combustible gas concentration to form a multimodal feature vector that comprehensively reflects the fire risk situation.

[0008] Artificial intelligence analysis module: Constructs a fire identification algorithm based on a multimodal artificial intelligence model. Using the fused multimodal feature vector as input, the model is trained and learned on a large amount of multimodal data in scenes with and without fires, enabling it to accurately identify early signs of fire and ongoing fires, and output the probability and risk level of fire. Alarm module: When the artificial intelligence analysis module determines that the probability of fire has reached the preset alarm threshold and the risk level is high, the alarm module is quickly activated. Alarm methods include local sound and light alarms to alert people in the building; at the same time, alarm information is sent to preset management personnel, fire departments and other relevant personnel through the communication network. The alarm information contains key information such as the location and time of the fire, as well as related on-site video images and various sensor data, so that relevant personnel can take rescue and response measures in a timely manner.

[0009] Backend Management Module: This module is used to configure, manage, monitor, and maintain the entire system. It enables functions such as setting parameters for data acquisition equipment, adjusting multimodal fusion strategies, optimizing and updating AI models, and setting alarm thresholds. Furthermore, it enables querying, statistics, and analysis of historical fire data and alarm records, providing data support for fire risk assessment and fire safety management.

[0010] The beneficial effects of the present invention are:

[0011] It can integrate fire-related data from multiple modalities, overcome the shortcomings of single-modality detection, improve the accuracy and reliability of fire detection, and effectively reduce the occurrence of false alarms and missed alarms.

[0012] With the help of advanced artificial intelligence models, intelligent identification and early warning of early signs of fire can be achieved, which helps to detect and take measures in time at the early stages of fire and reduce fire losses.

[0013] A timely and comprehensive alarm mechanism can ensure that when a fire occurs, relevant personnel can obtain accurate information in the shortest possible time, quickly carry out rescue and response work, and protect the safety of people's lives and property.

[0014] The system has good scalability and maintainability, and is easy to flexibly configure and optimize and upgrade according to different scenarios and needs. DETAILED DESCRIPTION

[0015] The fire prevention and alarm system of the present invention is described in detail below.

[0016] The fire prevention and alarm system of the present invention is mainly composed of a data acquisition module, a multimodal fusion module, an artificial intelligence analysis module, an alarm module and a background management module.

[0017] In the data acquisition module, cameras are strategically installed in various areas of the building (such as corridors, halls, and rooms) to capture real-time video data to monitor for abnormalities such as flames and smoke. Smoke sensors, temperature sensors, humidity sensors, and combustible gas sensors are also deployed at various locations within the building to monitor smoke concentration, ambient temperature, humidity, and combustible gas concentration in real time. This data is then transmitted to the data acquisition module's central processing unit via wired or wireless communication.

[0018] After receiving various types of data, the multimodal fusion module first performs preprocessing operations on the data. For example, the video image data is subjected to noise removal and size normalization, and the sensor data is unified and standardized. Then, the feature extraction technology in the deep learning algorithm is used to extract characteristic information such as the shape, color, motion characteristics of the flame, as well as the concentration and diffusion range of the smoke from the video image. At the same time, the data collected by the smoke sensor, temperature sensor, humidity sensor and combustible gas sensor are combined and fused into a multimodal feature vector according to a certain fusion strategy (such as weighted fusion, feature splicing fusion, etc.). This feature vector can comprehensively reflect the fire risk status in the current environment and is input into the artificial intelligence analysis module.

[0019] The artificial intelligence analysis module pre-builds and trains a fire recognition model based on a multimodal artificial intelligence model. During the model training phase, a large amount of multimodal data, including fire scenes and normal scenes, is collected and annotated, and the annotated data set is divided into a training set, a validation set, and a test set. The model is trained using the training set, and the model parameters are adjusted through an optimization algorithm so that the model learns the characteristic patterns of various modal data when a fire occurs and the correlation between them, so that it can accurately distinguish between fire scenes and normal scenes. The model is verified and tuned on the validation set to determine the optimal model structure and hyperparameter settings. During actual operation, the fused multimodal feature vector is input into the trained model. After calculation and analysis, the model outputs the probability of fire occurrence at the current moment and the corresponding risk level (such as low risk, medium risk, high risk, etc.).

[0020] When the probability of fire output by the artificial intelligence analysis module reaches the preset alarm threshold and the risk level is high, the alarm module is immediately activated. The local sound and light alarm device will emit strong sound and light signals in the corresponding area within the building to alert on-site personnel to the fire hazard. At the same time, the alarm module sends detailed alarm information to preset management personnel (such as building security personnel, fire safety managers, etc.), fire departments and other relevant personnel through network communication (such as 4G, 5G networks or wired networks, etc.), including the specific location of the fire (accurate to the room number or area name in the building), the time of occurrence, on-site video images (a long video clip can be transmitted in real time so that relevant personnel can view the on-site situation) and data information such as smoke concentration, temperature, humidity and combustible gas concentration monitored by various sensors, so that relevant personnel can quickly grasp the fire situation and respond in a timely manner, such as organizing personnel evacuation and conducting fire fighting and rescue work.

[0021] The backend management module provides system administrators with a comprehensive interface for management functions. This module allows administrators to configure parameters for data acquisition devices such as cameras and various sensors, such as adjusting camera capture parameters (resolution, frame rate, etc.) and setting sensor data acquisition frequency and upload cycles. Furthermore, the algorithms and strategies used in multimodal fusion can be adjusted and optimized to meet the fire monitoring needs of different building environments and usage scenarios. It also supports further training and optimization of artificial intelligence models, such as incrementally training the models with new fire data, continuously improving their recognition accuracy and generalization capabilities. Different alarm thresholds can be set within the backend management module, allowing flexible adjustment of alarm sensitivity based on actual needs and experience. The system also automatically records all alarm events and related data. Administrators can conveniently query historical alarm records through the backend management module and conduct statistical analysis on fire frequency, temporal distribution, and location distribution. This provides powerful data support for building fire safety management and helps formulate more scientific and rational fire safety strategies and preventive measures.

[0022] To sum up, the fire prevention and alarm system based on the multimodal artificial intelligence model of the present invention can achieve early warning and accurate alarm of fire by fusing fire-related data of multiple modalities and using advanced artificial intelligence technology for analysis and identification. It is of great significance to improving the level of fire safety and protecting the lives and property of people. It has broad application prospects and can be widely used in various types of buildings, industrial sites, forest fire prevention and other fields.

Claims

1. A fire prevention and alarm system based on a multimodal artificial intelligence model, characterized in that: include: Data acquisition module, used to collect video image data, smoke concentration data, temperature data, humidity data and combustible gas concentration data; The multimodal fusion module is used to fuse the collected data of different modalities to form a multimodal feature vector. The artificial intelligence analysis module builds a fire identification algorithm based on the multimodal artificial intelligence model, takes the fused multimodal feature vector as input, and outputs the fire probability and risk level. The alarm module activates the alarm and sends an alarm message to relevant personnel when the fire probability reaches the preset alarm threshold and the risk level is high. The background management module is used to configure, manage, monitor and maintain the system.

2. A fire prevention and alarm system based on a multimodal artificial intelligence model according to claim 1, characterized in that: The data acquisition module includes: a camera installed in the building to capture video image data; a smoke sensor, a temperature sensor, a humidity sensor and a combustible gas sensor, which are used to monitor smoke concentration, ambient temperature, humidity and combustible gas concentration respectively, and transmit the monitoring data to the central processing unit of the data acquisition module.

3. The fire prevention and alarm system based on a multimodal artificial intelligence model according to claim 1 is characterized in that: The method for the multimodal fusion module to perform data fusion processing includes: preprocessing various types of data, such as data cleaning and normalization; extracting characteristic information of flames and smoke from video image data; and using a deep learning algorithm to fuse video image features with smoke concentration, temperature, humidity, and combustible gas concentration data to form a multimodal feature vector.

4. The fire prevention and alarm system based on a multimodal artificial intelligence model according to claim 1, characterized in that: The multimodal artificial intelligence model of the artificial intelligence analysis module adopts a deep neural network architecture and is trained with a large amount of multimodal data in fire and non-fire scenarios. It can accurately identify early signs of fire and ongoing fire situations.

5. The fire prevention and alarm system based on a multimodal artificial intelligence model according to claim 1 is characterized in that: The alarm module's alarm modes include local sound and light alarms and network alarms. Local sound and light alarms are used to alert people in the building, and network alarms send alarm information containing the fire location, time, on-site video images, and various sensor data to preset management personnel, fire departments, and other relevant personnel through the communication network.

6. The fire prevention and alarm system based on a multimodal artificial intelligence model according to claim 1, characterized in that: The background management module has the following functions: parameter setting of data acquisition equipment; adjustment of multimodal fusion algorithms and strategies; optimization and updating of artificial intelligence models; setting of alarm thresholds; query, statistics and analysis of historical fire data and alarm records.