948 results about "Fire behavior" patented technology

The invention discloses a large high-rise building indoor fire urgent evacuation indication escape method and system, and belongs to the field of high-rise building fire safety. The large high-rise building indoor fire urgent evacuation indication escape method comprises the following steps: step 1, a fire source is located; step 2, a dynamic module of fire spread is set up; step 3, according to the dynamic module of fire spread in combination with the structural information of a building, the spatial position of an ignition point and the spatial position of fire-fighting equipment in a large high-rise building are determined, predictions are made over fire spread paths, and evacuation escape routes are planned; step 4, according to the designed escape routes, site equipment is controlled in a selective mode for assisting people in evacuation. According to the large high-rise building indoor fire urgent evacuation indication escape method and system, the fire control escape routes are made based on specific fire behavior, so that evacuation time is shortened; the existing problems that when a fire breaks out, a fire control evacuation escape way is crowded , low in efficiency and disorderly and stampedes easily occur are solved fundamentally; the practical significance of improving the safe evacuation performance of the large high-rise building and reducing the property loss and casualties caused by the fire in the large high-rise building is quite high.

The invention discloses a forest fire detection, prevention and control method based on an unmanned aerial vehicle. According to the forest fire detection, prevention and control method, the mode that visible-light images and thermal infrared images are combined is adopted, a forest fire can be judged by observing thermal image difference values of high-brightness points of the images, and fire extinguishing bags are released so as to extinguish fire head points in time or control fire behaviors primarily; accurate analysis is further conducted through a receiving station on the ground, abnormal areas missed during rough judgment conducted by the unmanned aerial vehicle can be found in time, and particularly monitoring can be achieved more effectively by forming a percentage distribution diagram of fire behavior possibility coefficients; efficiency and accuracy are high.

The present invention discloses a forest fire status monitoring system based on infrared camera. Said monitoring system includes infrared camera placed on the monitoring tower and ground monitoring centre. Said invention utilizes the infrared camera placed on the monitoring tower to make circular scan so as to obtain topographical image, then according to the difference of fire point and background radiation intensity and quantification relationship of flame temperature and radiation intensity utilizes image identification technique to attain goal of judging forest fire status.

The invention relates to a fire extinguishing system for energy storage units of lithium batteries. The fire extinguishing system mainly comprises an automatic fire alarm system, a gas fire extinguishing system, a fire extinguishing linkage control device, a fire emergent broadcasting and fire extinguishing special telephone device. Fire extinguishing devices of the fire extinguishing system can be automatically or manually started, the automatic starting is triggered mainly by a fire alarm control unit, the fire alarm control unit can receive triggering signals input by a battery management system and a detection alarm, when fire alarm conditions are met, starting signals are output to a gas fire extinguishing controller, and a fire extinguishing device is started to spray a fire extinguishing agent to the energy storage units; when an operator finds fire behavior in a protective zone, a start button outside the protective zone or on the gas fire extinguishing controller can be manually pressed down, and the manual operation of the fire extinguishment system is realized. Specific to multiple fire spraying characteristics of the lithium batteries, the two sets of fire extinguishing devices of the fire extinguishing system share one fire extinguishing cabinet, and the fire extinguishment protection of the two energy storage units of the lithium batteries is realized. The two sets of fire extinguishing devices are independent of each other and spare to each other, and the whole fire extinguishing system has a secondary fire extinguishment function.

The invention discloses a forest fire identification method and a forest fire identification system. The forest fire identification method comprises the steps of analyzing forest zone monitoring images, extracting image features, judging whether the image features conform to features of fire behaviors, and judging that suspected fire points exist if the image features conform to the features of the fire behaviors so that the purpose of identifying forest fire is achieved. Moreover, alarm can be conducted. Monitoring personnel can visually identify images of suspected fire points so as to confirm whether forest fires occur. The method and the system have the forest fire automatic identification function, can automatically alarm the fire behaviors, combines a geographic information system (GIS) to accurately locate the fire points, form images clearly, avoid limitation of original manual observing fire behaviors, realize forest zone management digitization and signification, greatly reduce expenditure and management cost of forestry sectors, and improve efficiency of enterprises in forest areas.

The invention relates to an automatic target scoring method based on video analysis. The method comprises image correction, target surface parameter modeling, shooting determination and bullet hole positioning. Before shooting, a collected target surface image is corrected to obtain a corrected target surface image, and a target surface parameter model is established in the corrected target surface image; and in the shooting process, a collected target surface image sequence is analyzed in real time, shooting determination is carried out to determine whether shooting is carried out, and when a shooting behavior occurs, a bullet hole is positioned, and the score of the bullet hole is determined according to the target surface parameter model. According to the method, distortion of the target surface image is removed via geometric correct; the target surface parameter model is more accurate by improving a loop boundary point searching method; the bullet hole is positioned accurately via a classification method, the dynamic change feature of a bullet hole area is utilized fully, and the accuracy of automatic target scoring is improved; and the collected video sequence is analyzed in real time to implement shooting determination and bullet hole positioning, so that the automatic degree of target scoring is improved.

The invention provides a smoke thermal imaging fire automatic positioning detection system and method. All photoelectric smoke detectors are encoded through an encoder, the position serial number of each detector is made to correspond to a thermal imaging view field, and thus the specific position of a fire can be determined. A thermal imager is used for monitoring the field, the fire behavior can be discovered earlier compared with a common video detector, and the system and method are particularly suitable for hazardous places such as aircraft cargo spaces, nuclear power stations and high voltage transformer substations where people cannot enter easily. When a smoke alarm gives out a fire alarm, at the moment, if no excessive heat video areas are captured by the thermal imager, the fire alarm can be removed, and accordingly loss caused by a fake fire alarm can be lowered. The multi-working mode and multi-interface design is adopted, the practicability can be improved, and equipment transformation in the later period is facilitated. Smoke signal detection and thermal image detection are combined, an intelligent information processing method is used for detecting and positioning fires, and accordingly the intelligence of fire fighting equipment is promoted. The place where the smoke alarm has already been mounted can be upgraded and reconstructed, and the robustness of a detection result can also be improved.

The embodiment of the invention provides a mobile terminal alarm method and device. The mobile terminal alarm method comprises the steps that fire behavior information of one or more of the gas concentration, the temperature, the humidity and the gas pressure in an environment is acquired and obtained; the fire behavior level is judged according to the acquired fire behavior information; corresponding emergency operation information is output to a user according to the fire behavior level. According to the mobile terminal alarm method and device, alarms of disasters such as fires can be given through a mobile terminal.

The embodiment of the invention provides a fire disaster rescue system and an information processing method. The fire disaster rescue system comprises an information acquisition module, a server, at least one self-rescue terminal and a rescue terminal, wherein the information acquisition module is used for acquiring fire behavior information of a fire disaster area, and used for transmitting the fire behavior information to the server; any one self-rescue terminal in the fire disaster area is used for acquiring the position information of the self-rescue terminal and used for uploading the position information to the server; the server is used for generating a first fire ground map according to the fire behavior information and the position information of all self-rescue terminals in the fire disaster area; the rescue terminal is used for acquiring the first fire ground map generated from the server, and used for outputting the first fire ground map; the first fire ground map is used for prompting the fire behavior information of the fire disaster area and the positions and the number of beleaguered users. By adopting the embodiment of the invention, the fire disaster situation and the beleaguered people can be accurately confirmed in time, and rapid and accurate rescue can be realized.

The invention discloses an electric transmission line fire spreading predicating method and system based on a grid flow. The method includes the steps that firstly, fire point collecting equipment is used for obtaining fire point information; a GIS image of an affected zone is divided into a grid image, so that a grid terrain of the affected zone is obtained, the ground feature category of each grid is identified, grid-crossing fire spreading intensity is calculated under different meteorological conditions, and a fire spreading weighing grid chart of the affected zone is constructed; a fire spreading route from sources to sinks is searched for in the fire spreading weighing grid chart of the affected zone to establish a fire behavior spreading model, and coordinates of an affected pole tower, the zone area of spreading and the time for fire spreading to the pole tower are calculated; finally, corresponding alarms are given to different influence grades of fire spreading predicated according to the grid flow method. The electric transmission line fire spreading predicating method and system based on the grid flow have the advantages that the time for fire point spreading to an electric transmission line zone and fire behavior intensity can be rapidly calculated, fire behaviors are informed in time, and early warning is high in accuracy.

The invention discloses a reinforced concrete shell fire resistance testing system. The system comprises a test furnace, a loading device and a control and data acquisition device; one or more tension-compression sensors (3.1.2) electrically connected to the strain acquisition circuit of the control and data acquisition device are distributed on bottom of a loading beam (2.3); each tension-compression sensor (3.1.2) is connected with a jack (2.4); each jack (2.4) is fixedly connected with a force transfer rod (2.5) which extends into the reinforced concrete shell (4). According to the reinforced concrete shell fire resistance testing system, the jacks are adopted for performing vertical loading so as to simulate the structural stress in the reality; the fire behaviors of the reinforced concrete shell and the influence rule of vertical uniform distribution or integrated tension-compression load effect on the fire resistance of the reinforced concrete shell can be deeply studied, and then the fire resisting design of the reinforced concrete shell can be guided according to the test data and the fire resistance of the reinforced concrete shell can be improved.

The invention relates to the technical field of fire control, in particular to an intelligent fire fighting system based on the extensive electric power internet of things. The intelligent fire fighting system is applicable to an extensive internet of things electric power system comprising a data server. A plurality of fire extinguishing devices used for fire extinguishing and arranged nearby anelectric power building or electric power equipment, a plurality of monitoring sensing devices, a collector, a cloud server and a monitoring terminal are included. The monitoring sensing devices monitor the electric power building fire behavior or the state of the fire extinguishing devices and arranged on the electric power building or the fire extinguishing devices. The collector collects the sensor data and is in communication connection with the monitoring sensing devices. The cloud server analyzes the sensor data, monitors the fire control state of the electric power system and is in communication connection with the collector and the data server. The monitoring terminal displays the fire control state of the electric power system, interacts with a user and is connected with the cloudserver. The intelligent fire fighting system has the substantive effects that the cloud server is used for conducting fire behavior early warning and fire extinguishing strategy and accident reason study and judgment, the advantages of the internet of things are brought into full play, and the fire control level is improved.

The invention provides a UAV firefighting scheduling system which comprises a UAV investigation system, a cloud-end processor, an information center and a handheld terminal; the UAV investigation system collects sky data of the forest, and transmits the data to the cloud-end processor; the cloud-end processor gathers, classifies and analyzes the received data, and transmits the data to the handheld terminal for display; the handheld terminal collects position and vital sign information of a holder of the handheld terminal, and transmits the information to the cloud-end processor; and the cloud-end processor gathers the data collected by inspection UAV and the information collected by the handheld terminal, transmits the data and information to the information center, and information center schedules the inspection UAV according to the data. The UAV firefighting scheduling system can gather and analyze the data collected by UAV and the information collected by the handheld terminal automatically via the cloud-end processor, the trend of fire behaviors can be analyzed intelligently, and the UAV can be scheduled to assist people on a fire site.

The invention discloses a fire-fighting robot capable of achieving automatic patrolling, and belongs to the field of automation. The fire-fighting robot can travel on a closed operating track throughfour traveling wheels installed on the fire-fighting robot and can rotate by 360 degrees. The fire-fighting robot is driven by a rotatable rotating module to rotate. A telescopic mechanism is fixedlyinstalled over the rotating module, and a fire monitor is fixedly supported on the telescopic mechanism through a supporting module. The telescopic mechanism drives the fire monitor to rise and fall within a permissible range, so that the distance between fire monitor and a fire source is decreased. The supporting module is hinged to a rotary motor, and a pitching angle of the fire monitor is adjusted through gear transmission. An infrared temperature detection module and a camera are installed at the front end of the fire monitor. During fire detection, the camera is used for video surveillance image analysis, and the infrared temperature detection module is used for detecting a fire point. The fire-fighting robot is easy to mount, can achieve automatic patrolling, automatic fire source detection and automatic fire extinguishing; and the fire-fighting range and the fire-fighting condition can be sent to a control center in real time, and a basis is provided for fire behavior analysisand prevention in the later stage.

The present invention relates a shooting-type mixed flame detector. The detector is provided with a MCU main control unit (10), an ultraviolet flame detection system, an image collection system, an infrared flame detection system and a power supply management system. The method comprises a step 1, a step 2, a step 3 and a step 4. The power consumption is low, field environment application anti-interference capability and adaptability are high, a mode of mutual awakening between fire behavior nodes can effectively determine fire behavior development speed and range; and moreover, wireless networking transmission sensing data, compression image data and high-definition image data can perform close-range point-to-point transmission so as to solve the problems of forest fire hazard preventionand evidence obtaining.

The invention provides a clustered control fire fighting system. The system comprises an unmanned aerial vehicle in communication connection and a dispatcher control center used for containing and controlling the unmanned aerial vehicle. The unmanned aerial vehicle comprises at least one investigation unmanned plane and a plurality of fire fighting unmanned planes. The fire fighting unmanned planes are used for flying to be close to a target fire place to investigate and obtain a fire behavior of the target fire place according to a fire behavior investigation command sent by the dispatcher control center. The dispatcher control center is used for sending a dispatch command according to the fire behavior of the target fire place according to the investigation unmanned plane. The fire fighting unmanned planes are used for flying to an appointed place according to the dispatch command to conduct bomb putting fire fighting. The actual fire behavior of the target fire place is obtained through the system, corresponding bomb putting fire fighting is carried out according to the actual fire behavior, and therefore the problem that fire fighting cannot be carried out in time when a fire happens in a place where the fire is easily caught is solved.

The invention relates to the technical field of fire-fighting digitization, and provides a fire-fighting three-dimensional digital planning method fusing a fire behavior prediction model. The fire-fighting three-dimensional digital planning method comprises the steps: building a three-dimensional visualization model of a building structure according to a building structure drawing, and extractingthe internal layered structure information and fire-fighting information of the building structure; indoor panoramic image generation: collecting indoor information through a panoramic camera or a camera, and obtaining indoor material attributes and material parameters; and inputting the building structure internal layered structure information of the three-dimensional visualization model, the indoor material attribute and the material parameter information of the fire disaster and the weather information when the fire disaster occurs into the fire behavior prediction model to obtain a fire behavior spreading trend analysis result. By means of the three-dimensional visualization, indoor panorama and fire behavior prediction model technology, the fire-fighting three-dimensional digital planning method can be applied to various fire scenes in real time, can make corresponding adjustment of three-dimensional digitization and prediction models according to real-time fire scene dynamic changes, and can meet the fire fighting actual combat requirements. The fire-fighting three-dimensional digital planning method has the advantages of flexibility, convenience, accuracy, high efficiency, intelligent fire protection and scientific and technological fire protection.

The invention discloses a personnel characteristic-based fire evacuation path navigation method which comprises the following steps: 1, obtaining fire behavior information through a fire detector, calculating an effective time range capable of being used for escape of each path; 2, reading personnel label information by using an RFID (Radio Frequency Identification) reader-writer, locating personal position information, performing personnel evacuating speed classifying according to the obtained personnel label information and the obtained personnel position information; 3, for personnel with different speeds, calculating evacuating path information by using a shortest path precedence algorithm, wherein the evacuating path information comprises a shortest path length value from an initial position to an outlet node D and all nodes on the path; and 4, sending the shortest path length value and all nodes on the path to personnel to be evacuated. According to the personnel characteristic-based fire evacuation path navigation method, the utilization rate of the path can be effectively increased.

The invention relates to a method for applying a radio frequency identification label technology to indoor locating of a firefighter. The method for applying the radio frequency identification label technology to indoor locating of the firefighter comprises the steps that (1) information of radio frequency identification labels deployed inside a building nearby the position of the firefighter is read through a radio frequency identification label read head with variable power; (2) the information of the radio frequency identification labels is forwarded to a server through a 3G network; (3) the server carries out analysis processing on the information of the radio frequency identification labels to obtain the real-time indoor position of the firefighter. A commander outside the building can achieve accurate locating of personnel and command according to the real-time indoor position of the firefighter. A system implemented through the method for applying the radio frequency identification label technology to indoor locating of the firefighter is convenient to deploy and use, low in price and convenient to operate and control, and efficiency of search and rescue of people and efficiency of fire behavior surveys can be effectively improved.

The invention discloses a simulation experiment device of high-rise building three-dimensional fire behaviors under the effect of environment wind. The device comprises a high-rise building model, a rotary platform, a direction adjustable wind hole, a three-dimensional movable thermoelectric coupling frame and a fire source simulating device. The high-rise building model is a cuboid shell and is divided into four chambers, wherein the ignition experiment chamber is provided with the fire source simulating device and a size adjustable building window model. The bottom of the high-rise building model is connected with the rotary platform and matched with the external direction adjustable wind hole so that environment wind in any direction at any wind speed can be simulated. The three-dimensional movable thermoelectric coupling frame is of a combined rectangular structure, and the parameter measurement function of any position of an external wall surface can be achieved. The simulation experiment device for studying high-rise building three-dimensional fire spreading under the effect of environment wind is established for the first time, a wind direction adjusting system is safe, convenient to use and precise, the experiment device is provided with a complete parameter measuring system, and the complete process of a high-rise building from fire starting, flame spreading to final spreading towards the upper layer and adjacent side of the building can be studied.

The invention provides a clustered-control fire fighting method and device. Unmanned aerial vehicles comprise the at least one reconnaissance unmanned aerial vehicle and the multiple fire fighting unmanned aerial vehicles. The clustered-control fire fighting method comprises the steps that the reconnaissance unmanned aerial vehicle receives a fire behavior detection instruction emitted by a dispatching control center and then flies to a target ignition area to reconnoiter the fire behavior; the reconnaissance unmanned aerial vehicle transmits the reconnoitered fire behavior condition of the target ignition area to the dispatching control center; and the multiple fire fighting unmanned aerial vehicles receive a fire fighting dispatching instruction emitted by the dispatching control center based on the fire behavior condition and fly to the target fire point to put out the fire. When a fire breaks out, the reconnaissance unmanned aerial vehicle can accurately obtain the actual fire behavior of a fire site in time, the fire fighting unmanned aerial vehicles can conduct fire fighting on the fire site according to the actual fire behavior, thus, losses caused by the fire can be decreased, meanwhile, no fireman needs to appear on the fire site to conduct fire fighting operation, and the situation that the life safety of the fireman is threatened by the fire is avoided.

The invention discloses a multi-parameter mine laneway fire monitoring and alarming system. Mine laneway fires are characterized in that finding is difficult, and the development is rapid; a traditional laneway fire alarming system is long in reaction time, and misstatement and reporting missing are likely to occur, so that the fire behavior cannot be controlled in time, and a large number of losses of life and personal injury are caused. The system mainly comprises an information processing server, an alarming device, a communication network, a gas concentration remote sensing device, a temperature monitoring device and the like; and by means of the system, the characteristic of concentration of gas in the monitoring zone environment can be measured through the gas concentration remote sensing device, analysis of integrated monitoring data such as temperature and smoke can be carried out, and a fire alarm can be given out. By means of the system, laneway fire characteristic gas can be accurately monitored, the alarming accuracy rate is greatly improved, and important guarantees are provided for coal mine safety production.

The invention provides an intelligent fire-fighting system and method for a transformer substation. The intelligent fire-fighting system comprises a plurality of fire-fighting robots and fire-fightingmedium supply equipment as well as a control center, wherein the fire-fighting robots and the fire-fighting medium supply equipment are arranged on each area in the transformer substation in a spacedmode; the fire-fighting medium supply equipment can be fixed, and also can be movable; while a fire behavior occurs, ignition point information can be obtained according to a smoke / temperature sensorarranged on each area in the transformer substation, also can be obtained by sensors on the fire-fighting robots, and also can be detected by arranging an unmanned aerial vehicle. The control centerschedules the fire-fighting robot closest to the ignition point, so that the fire-fighting robot firstly goes to the ignition point to extinguish fire; and meanwhile, other fire-fighting robots closerto the ignition point or the fire-fighting medium supply equipment can be scheduled to go to the ignition point. The intelligent fire-fighting system realizes timely fire-extinguishing and adaptive fire-extinguishing.

The invention relates to the technical field of data processing, in particular to a fire alarm method and device, a storage medium and a terminal device. The method comprises the steps of obtaining environment monitoring data of a fire occurrence place; inputting the environment monitoring data into the trained machine learning model to obtain a first fire behavior analysis result output by the machine learning model; inputting the environment monitoring data into a preset fuzzy comprehensive evaluation model to obtain a second fire behavior analysis result output by the fuzzy comprehensive evaluation model; performing result fusion on the first fire behavior analysis result and the second fire behavior analysis result by utilizing an evidence theory to obtain a fire behavior level corresponding to the environment monitoring data; determining an alarm mode corresponding to a fire level according to an evidence theory, performing fire alarm according to the alarm mode, performing fire analysis by simultaneously utilizing a machine learning model and a fuzzy comprehensive evaluation model, and performing result fusion by utilizing the evidence theory, thereby accurately and efficiently determining the fire level, and determining the corresponding alarm mode to perform fire alarm.

The invention discloses an intelligent fire fighting system based on the Internet of things. The intelligent fire fighting system based on the Internet of things comprises a server, a fire behavior monitoring subsystem, an alarming subsystem, fire extinguishing equipment, an equipment monitoring subsystem and a monitoring terminal; the server is used for running fire fighting strategies to generates the subsystems; the fire behavior monitoring subsystem is arranged in a fire fighting area, used for detecting the fire behavior and connected with the server; the alarming subsystem is arranged inthe fire fighting area, used for giving out sound and light alarm information and connected with the server; the fire extinguishing equipment is arranged on a fire fighting site and used for extinguishing a fire; the equipment monitoring subsystem is arranged on fire fighting equipment, used for detecting the state and the position of the fire fighting equipment and connected with the server; andthe monitoring terminal is used for displaying the running state of a sensing and alarming subsystem and the equipment monitoring subsystem and connected with the server. The intelligent fire fighting system based on the Internet of things has the function of extinguishing the fire when the fire occurs; escape routes are intelligently formulated through the monitoring equipment and sent to residents and guardians thereof through messages, and the safety of the residents is guaranteed; and the fire behavior is controlled in time through a self-walking fire extinguishing robot, and all-area-covered intelligent automatic fire extinguishing is achieved.

The invention provides a fire-fighting control system and method of wind generating set. The fire-fighting control system comprises multiple fire hazard detectors, a fire-fighting controller, multiple fire extinguishing devices and a communication module, wherein the fire hazard detectors are used for detecting a fire hazard and emitting alarm information; the fire-fighting controller is used for receiving the alarm information emitted by the fire hazard detectors, and determining whether a fire behavior occurs in a fire-fighting area where certain fire hazard detector is placed according to the alarm information and preset fire behavior determining conditions; the fire extinguishing devices are used for extinguishing the fire under control of the fire-fighting controller; and the communication module is used for sending a fire alarm signal and receiving a fire-fighting control command. If the fire-fighting controller determines that the fire behavior occurs in the fire-fighting area where the fire hazard detector is placed, the fire alarm signal which indicates occurrence of the fire behavior of the wind generating set is sent to the communication module; and if the communication module receives the fire-fighting control command, the fire-fighting controller controls the corresponding fire extinguishing device to extinguish the fire according to the fire-fighting control command.

The invention discloses a halogen-free environmental protection flame-retardant asphalt for tunnel road surface, comprising asphalt, a modifier and a flame retardant, wherein the flame retardant is a halogen-free flame retardant, the weight of which is 3% to 20% of total weight of the product, and the flame retardant has the following components in percentage by weight: 50% to 75% of a nitrogen-series flame retardant, 20% to 40% of magnesium hydrate, 3% to 10% of a coupling agent, and 0.5% to 5% of a smoke suppressant. The asphalt does not contain halogen and produce a great deal of smoke and halogen hydride poisonous noxious gas (HCl or HBr, and the like) in combustion process, thereby buying more escape time for human, reducing the difficulty of emergency rescue work and guaranteeing the security of the lives and property to the utmost extent while improving the flame retardance of tunnel asphalt road, reducing the smoke produced with fire and delaying fire behavior.

The invention provides a ship accumulator cabin ventilation and automatic extinguishing method based on fuzzy control. The method comprises the steps that the environment temperature and the smoke density of a ship accumulator cabin are collected; a fuzzy control model is established; a collected signal is used as the input quantity of the fuzzy control model, and the input quantity is fuzzified according to a preset fuzzy set; the fuzzy control quantity is obtained through fuzzy reasoning according to a preset control rule list; the fuzzy control quantity is multiplied by a proportionality coefficient after being defuzzified to obtain the control quantity; the control quantity is output to an alarm device, an extinguishing device and a ventilation system; and whether the alarm device, the extinguishing device and the ventilation system are started or not is selected according to the control quantity. By adoption of the ship accumulator cabin ventilation and automatic extinguishing method, cabin ventilation and early warning can be achieved, it is ensured that a fire is prevented before occurring, or a fire is automatically extinguished after occurring, the fire behavior is weakened maximally, personnel are evacuated in advance, and casualties and property loss are reduced.

The invention discloses a fire image detection method based on a mixed color model and a neural network. The method comprises the following steps: firstly, performing an inter-frame difference methodon a monitoring video image to obtain a moving foreground region; extracting an area conforming to flame color characteristics through the mixed color model, and performing secondary color area expansion to obtain candidate areas filtered by the color model; taking a union set of the motion foreground area and the color candidate area to obtain a suspected flame area; and finally, constructing a dense connection convolutional neural network, and comprehensively judging whether a suspected flame region has a fire behavior or not through the trained network model and the secondary mixed color model. According to the method, the flame candidate area with high precision is obtained through motion detection and the mixed color model, the detection speed is high, the deep neural network guarantees the flame detection accuracy and the generalization capability, the method can be widely deployed in an actual intelligent security system, fire hazards are early warned, and losses caused by the fire hazards are reduced.