A forest fire prevention monitoring and alarm method and device
Through drones, color and infrared images of forest fire areas are collected, combined with fire risk prediction models and feature recognition technology, the problem of delayed discovery of large-scale forest fires is solved, and fast and accurate fire alarms are achieved.
Patent Information
- Application Number
- CN202411267488.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-11
- Publication Date
- 2024-11-22
- Estimated Expiration
- 2044-09-11
AI Technical Summary
The existing technology is difficult to detect fires in large areas of original forests in a timely manner, which makes it difficult to delay the discovery and extinguish fires.
The drone is used to collect color images and infrared thermal images, combine fire risk prediction models and feature recognition technology to quickly identify smoke characteristics and temperature trends, and promptly alarms.
Accurate and rapid discovery of forest fires has been achieved, and the timeliness and accuracy of fire warnings has been improved.
Smart Images

Figure CN118803433B_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the field of fire prevention detection, in particular to a forest fire prevention detection alarm method and device. Background Art
[0002] For forest fire monitoring, the prior art generally sets various sensors in the forest to collect environmental parameters, or sets cameras to collect flame or smoke characteristics, so as to detect fire and issue an alarm in time.
[0003] However, for large areas of virgin forest, the existing detection difficulty lies in the huge forest area, so it is difficult to detect fires in time. Fires are often not discovered until some time after the fire occurs. At this time, the fire has already caused damage and is difficult to extinguish in time. Summary of the invention
[0004] In view of this, the purpose of the present invention is to provide a forest fire monitoring and alarm method and device to solve the problem that fires cannot be discovered in time in the prior art.
[0005] In order to achieve the above object, the present invention adopts the following technical solutions:
[0006] A forest fire prevention monitoring and alarm method of the present invention comprises the following steps:
[0007] Acquire current environmental data of a target block at a current time point, wherein the target forest area includes a plurality of blocks, the target block is one of the plurality of blocks, and the current environmental data includes values of a plurality of environmental parameters;
[0008] Determining the fire risk value of the target block at the current time point based on the current environmental data and a pre-constructed fire risk prediction model, wherein the fire risk prediction model is constructed based on historical fire data;
[0009] When the fire risk value is higher than a preset threshold, controlling the drone to collect color images and infrared thermal images of the target block at multiple time points;
[0010] Based on the color images at multiple time points, smoke characteristics of the target area are identified, and based on the infrared thermal images at multiple time points, temperature trends of the target area are identified;
[0011] When there are smoke features in the target block and there is a temperature increase trend in the target block, the location information of the target block and the fire alarm information are sent to the target object based on a pre-set signal station.
[0012] In one embodiment of the present application, the method for constructing the fire risk warning model includes:
[0013] Obtain multiple data samples of the target forest area or other forest areas , wherein the data samples are the values of various environmental parameters at multiple time points within the target time period before the fire occurs, The fire occurrence sequence number;
[0014] Calculate the mean of the values of each environmental parameter in each data sample , and calculate multiple data samples Medium environmental parameters The average Variance and average ;
[0015] Variance of environmental parameters Normalization is performed and based on the variance of multiple normalized environmental parameters Calculation environment parameters Risk weight , where the risk weight The mathematical expression is:
[0016] In the formula, is the number of environmental parameters;
[0017] Based on environmental parameters The average Variance and average Build environment parameters The risk reference range is as follows: The mathematical expression of the risk reference range is: ,in, is the range adjustment factor;
[0018] A fire risk warning model is constructed based on the risk reference range and weight of each environmental parameter.
[0019] In one embodiment of the present application, determining the fire risk value of the target block at the current time point based on the current environmental data and a pre-built fire risk prediction model includes:
[0020] Comparing the value of each environmental parameter in the current environmental data with the corresponding risk reference range to obtain a target environmental parameter, wherein the target environmental parameter is an environmental parameter whose value falls within the corresponding risk reference range;
[0021] The risk weights of all target environmental parameters are summed up to obtain the fire risk value.
[0022] In one embodiment of the present application, smoke feature recognition is performed on the target block based on color images at multiple time points, including:
[0023] Convert color images at multiple time points into grayscale images;
[0024] Preprocessing the multiple grayscale images to obtain preprocessed images, wherein the preprocessing method includes cropping and high-pass filtering;
[0025] Extracting contour features and texture features of the preprocessed image; and inputting the contour features and texture features of a plurality of preprocessed images into a pre-built recognition model to obtain a plurality of smoke recognition results;
[0026] When any smoke recognition result indicates that smoke exists, it is determined that smoke features exist in the target area.
[0027] In one embodiment of the present application, the method for constructing the recognition model includes:
[0028] Acquire a positive sample image and a negative sample image, wherein the positive sample image contains smoke features, and the negative sample image does not contain smoke features;
[0029] Converting the positive sample image and the negative sample image into grayscale sample images;
[0030] Preprocessing a plurality of grayscale sample images to obtain preprocessed sample images;
[0031] Extracting contour features and texture features of the preprocessed sample image to obtain training data;
[0032] The artificial neural network is trained based on the training data to obtain a recognition model.
[0033] In one embodiment of the present application, the temperature trend of the target block is identified based on the infrared thermal images at multiple time points, including:
[0034] Convert pixel values in multiple infrared thermal images into temperature values based on the calibration parameters of the infrared camera;
[0035] Filter out target pixels whose temperature values in the infrared thermal image are higher than the temperature threshold, and obtain a high temperature region and the area of the high temperature region, wherein the area of the high temperature region is the number of corresponding target pixels;
[0036] Calculate the average temperature value corresponding to all target pixels in the high temperature area;
[0037] The areas and average temperature values of high-temperature areas of multiple infrared thermal images are mapped to a two-dimensional coordinate system in chronological order, and the high-temperature area spreading trend and the average temperature change trend are extracted in the two-dimensional coordinate system, and the temperature trend of the target block is determined based on the high-temperature area spreading trend and the average temperature change trend, wherein the horizontal axis of the two-dimensional coordinate system is time and the vertical axis is area.
[0038] In one embodiment of the present application, extracting the high temperature area spreading trend and the average temperature change trend in the two-dimensional coordinate system includes:
[0039] Use the pre-built sliding window to slide along the two-dimensional coordinate system, and calculate the The average area of the high temperature area of the infrared thermal image and average temperature ,in, is the sliding step number;
[0040] In the average area When the average temperature increases along the time axis, When the temperature is not decreasing in one direction, it is determined that the target block has a temperature increasing trend.
[0041] In one embodiment of the present application, the location information of the target block and the fire alarm information are sent to the target object based on the pre-set signal station, including:
[0042] The longitude and latitude information of the drone is used as the location information of the target block, and the location information of the target block and the fire alarm information are transmitted to the nearest signal station, and then forwarded to the target object by the nearest signal station.
[0043] In one embodiment of the present application, it also includes:
[0044] When there is no smoke feature in the target block, or when there is no temperature rising trend in the target block, the target block is marked as a high-risk block, and blocks with a fire risk value not higher than a preset threshold are marked as low-risk blocks;
[0045] The high-risk blocks are patrolled according to a preset first patrol frequency, and the low-risk blocks are patrolled according to a preset second patrol frequency, wherein the first patrol frequency is greater than the second patrol frequency.
[0046] The present application also provides a forest fire prevention monitoring and alarm device, comprising: an acquisition module, used to acquire current environmental data of a target block at a current time point, wherein the target forest area includes multiple blocks, the target block is one of the multiple blocks, and the current environmental data includes values of multiple environmental parameters;
[0047] A risk determination module, used to determine the fire risk value of the target block at the current time point based on the current environmental data and a pre-built fire risk prediction model, wherein the fire risk prediction model is built based on historical fire data;
[0048] An image acquisition module, used to control the drone to acquire color images and infrared thermal images of the target block at multiple time points when the fire risk value is higher than a preset threshold;
[0049] A feature recognition module, used to perform smoke feature recognition on the target block based on color images at multiple time points, and to perform temperature trend recognition on the target block based on infrared thermal images at multiple time points;
[0050] The alarm module is used to send the location information of the target block and the fire alarm information to the target object based on a pre-set signal station when there is a smoke feature in the target block and there is a temperature rising trend in the target block.
[0051] The beneficial effects of the present invention are as follows: a forest fire monitoring and alarm method and device of the present invention constructs a fire risk prediction model through historical data of a period of time before the forest fire, and then predicts the fire risk value of the target block at the current time based on the environmental data at the current time and the fire risk prediction model. If the fire risk value is large, the drone patrol is immediately triggered, and the drone hovers above the target block to take color images and infrared thermal images. If a fire has already occurred at this time, the smoke characteristics and temperature rise trends can be extracted from the color images and infrared thermal images. At this time, the drone returns the positioning information to the preset signal station, and the signal station forwards the alarm information. The present application performs preliminary fire prediction based on environmental information, and controls the drone to perform image acquisition and feature extraction after being triggered, so that fires or fire hazards can be discovered more accurately and quickly. BRIEF DESCRIPTION OF THE DRAWINGS
[0052] The present invention will be further described below in conjunction with the accompanying drawings and embodiments:
[0053] Figure 1 is a flow chart of a forest fire monitoring and alarm method shown in one embodiment of the present application;
[0054] Figure 2 This is a schematic diagram of the drone signal backhaul scenario in this application;
[0055] Figure 3 A schematic diagram of a signal transmission scenario of a signal station in this application;
[0056] Figure 4 It is a structural diagram of a forest fire monitoring and alarm device shown in one embodiment of the present application. DETAILED DESCRIPTION
[0057] The following describes the embodiments of the present invention by specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict.
[0058] It should be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present invention, and thus the drawings only show the layers related to the present invention rather than being drawn according to the number, shape and size of the layers in actual implementation. In actual implementation, the type, quantity and proportion of each layer may be changed arbitrarily, and the layer layout may also be more complicated.
[0059] In the following description, numerous details are discussed to provide a more thorough explanation of embodiments of the present invention; however, it is apparent to one skilled in the art that embodiments of the present invention may be practiced without these specific details.
[0060] Figure 1 FIG. 1 is a flow chart of a forest fire detection and alarm method shown in an embodiment of the present application. Figure 1 As shown: A forest fire detection and alarm method of this embodiment may include steps S110 to S140:
[0061] S110, obtaining current environmental data of a target block at a current time point, wherein the target forest area includes a plurality of blocks, the target block is one of the plurality of blocks, and the current environmental data includes values of a plurality of environmental parameters;
[0062] This application divides the target forest area into multiple blocks in advance, which can be divided according to the area or according to the area where the original monitoring equipment is located.
[0063] One or more fire monitoring stations are set up in each block to detect the environmental parameters of one or more locations in the block, including temperature, relative humidity, wind speed, etc. The fire monitoring station is powered by solar energy or nearby power lines, and monitors environmental data by collecting data at regular intervals.
[0064] The fire detection station transmits data back to the monitoring center via a wireless network for data analysis and monitoring.
[0065] S120, determining the fire risk value of the target block at the current time point based on the current environmental data and a pre-built fire risk prediction model, wherein the fire risk prediction model is built based on historical fire data;
[0066] This application builds a fire risk prediction model based on historical data of forest fires, including:
[0067] (1) Obtain multiple data samples from the target forest area or other forest areas , wherein the data samples are the values of various environmental parameters at multiple time points within the target time period before the fire occurs, The fire occurrence sequence number;
[0068] Considering the lack of fire data samples in the target forest, this application also collects data on forest fires in other places as the basic data for building a prediction model. The sample data mainly includes various environmental parameters related to fires, such as temperature, relative humidity, and wind speed. In order to avoid data distortion and interference at a single time point before the fire, this application collects environmental data for the target time period before the fire, and the target time period can be 2-5 hours.
[0069] (2) Calculate the average value of each environmental parameter in each data sample , and calculate multiple data samples Medium environmental parameters The average Variance and average ;
[0070] This application calculates the average value of each environmental parameter in each data sample To remove the data interference problem caused by a single time point. It can better reflect the value of environmental data before the fire. In addition, for any environmental parameter, this application reflects the correlation between the parameter and the fire by calculating the variance.
[0071] Since variance is mainly used to reflect data stability, if the variance of one of the environmental parameters is small before the fire occurs and is always distributed in a small range, it can be said that the value of the environmental parameter is strongly correlated with the fire. If the variance of one of the environmental parameters is large before the fire occurs and the value distribution is less regular, it means that the value of the environmental parameter is weakly correlated with the fire.
[0072] In this application, the temperature, relative humidity, wind speed and other parameters contained in the collected data are all related to fire. Therefore, this application needs to determine the strength of the correlation based on the above parameters, that is, determine the weight. The details are as follows:
[0073] (3) Variance of environmental parameters Normalization is performed and based on the variance of multiple normalized environmental parameters Calculation environment parameters Risk weight , where the risk weight The mathematical expression is:
[0074] In the formula, is the number of environmental parameters;
[0075] The normalization process is to unify the range of variances of various environmental parameters so that the variance is always between [0,1], which can be used to reflect the relative stability of various environmental parameters.
[0076] Risk Weight Based on the normalized variance Construction, Variance The smaller it is, the higher the weight is, which conforms to the correlation mentioned above.
[0077] (4) Based on environmental parameters The average Variance and average Build environment parameters The risk reference range is as follows: The mathematical expression of the risk reference range is: ,in, is the range adjustment factor;
[0078] In addition, this application also constructs environmental parameters Risk reference range, if the environmental parameters The value of falls within the risk reference range , it means that from the environmental parameters From the perspective of fire risk.
[0079] (5) Construct a fire risk warning model based on the risk reference range and weight of each environmental parameter.
[0080] After building the fire risk warning model, the fire risk value can be calculated based on the environmental parameters at the current time, including:
[0081] S121, comparing the value of each environmental parameter in the current environmental data with the corresponding risk reference range to obtain a target environmental parameter, wherein the target environmental parameter is an environmental parameter whose value falls within the corresponding risk reference range;
[0082] S122, summing up the risk weights of all target environmental parameters to obtain a fire risk value.
[0083] In the present application, if the value of the environmental parameter falls within the corresponding risk reference range, then the corresponding weight is obtained, and finally the sum is calculated to obtain the risk value.
[0084] For example, the weight of temperature is 0.45, the weight of relative humidity is 0.32, and the weight of wind speed is 0.23. The temperature, humidity, and wind speed at the current time all fall within the corresponding risk reference range. Then the fire risk value = 0.45 + 0.32 + 0.23 = 1, and there is a fire risk at this time. In the specific implementation, there are more environmental parameters, such as dust concentration, vegetation index, etc. Therefore, the value of one or more unimportant environmental parameters will not affect the final risk judgment.
[0085] S130, when the fire risk value is higher than a preset threshold, controlling the drone to collect color images and infrared thermal images of the target block at multiple time points;
[0086] If the fire risk value at this time is higher than the preset threshold, it means that the environmental parameters at this time have the conditions for fire to occur and there is a fire risk. Therefore, in order to further verify whether a fire has occurred at this time, it is necessary to collect images of the corresponding location for verification. If it can be collected directly through the preset camera, it can be collected directly, but in most cases, there is no corresponding camera to collect images. Therefore, this application controls the drone to patrol the corresponding location and collect color images and infrared thermal images at multiple time points.
[0087] S140, performing smoke feature recognition on the target area based on the color images at multiple time points, and performing temperature trend recognition on the target area based on the infrared thermal images at multiple time points;
[0088] In this application, smoke features are identified based on deep learning. The specific process includes:
[0089] S1401, converting color images at multiple time points into grayscale images;
[0090] S1402, preprocessing the multiple grayscale images to obtain preprocessed images, wherein the preprocessing method includes cropping and high-pass filtering; after high-pass filtering, more high-frequency information of the image can be retained to facilitate subsequent extraction of contour features and texture features;
[0091] S1403, extracting contour features and texture features of the preprocessed image; and inputting the contour features and texture features of multiple preprocessed images into a pre-built recognition model to obtain multiple smoke recognition results;
[0092] When extracting contour features, existing contour extraction operators can be used, such as the Sobel operator, the Canny operator, etc. Texture features can be extracted through the local binary pattern (LBP), the gray level co-occurrence matrix (GLCM), etc.
[0093] The recognition model identifies contour features and texture features to see if there are smoke features in the image. In some cases, grayscale features can also be used for recognition, but since the grayscale of water mist in the forest is similar to that of smoke, using grayscale features may lead to misrecognition. Therefore, this application only uses contours and textures as recognition features.
[0094] S1404: When any smoke recognition result indicates that smoke exists, it is determined that smoke features exist in the target area.
[0095] The process of building a recognition model is similar to the recognition process, including:
[0096] (1) obtaining a positive sample image and a negative sample image, wherein the positive sample image contains smoke features and the negative sample image does not contain smoke features;
[0097] (2) converting the positive sample image and the negative sample image into grayscale sample images;
[0098] (3) Preprocessing multiple grayscale sample images to obtain preprocessed sample images; preprocessing includes cropping, flipping, scaling, filtering, etc.;
[0099] (4) Extracting contour features and texture features of the preprocessed sample image to obtain training data; the training data also includes the annotated labels.
[0100] (5) Based on the training data and in combination with the gradient descent method, an artificial neural network is trained to obtain a recognition model.
[0101] The temperature trend is determined through infrared thermal images at multiple time points. The specific process is as follows:
[0102] S1411, converting pixel values in a plurality of infrared thermal images into temperature values based on calibration parameters of the infrared camera;
[0103] In an infrared image, the grayscale value or color depth of each pixel represents the radiation intensity of that point. In order to convert the radiation intensity into a temperature value, it is necessary to use the calibration data of the infrared thermal imager (such as emissivity, atmospheric transmittance, etc.) and related physical laws (such as Stefan-Boltzmann law, Planck radiation law, etc.) for calculation. In a specific application, the calibration parameters of the infrared camera can be used for calculation to obtain the correspondence between the grayscale value and the temperature value, and then the pixel values in multiple infrared thermal images are converted into temperature values. S1412, filter out the target pixel points in the infrared thermal image whose temperature values are higher than the temperature threshold, and obtain the high temperature area and the area of the high temperature area, wherein the area of the high temperature area is the number of corresponding target pixel points;
[0104] Forest fires are generally divided into surface fires, canopy fires, and underground fires. When a fire occurs, the local temperature can reach over 400°C, and the temperature at the unburned junction can also reach over 100°C. Therefore, this application sets the threshold to 100°C, which is a temperature that does not exist under normal circumstances. Therefore, once this temperature is reached, there is a greater risk of fire or a fire has already started, and the corresponding pixel can be marked as an abnormal target pixel.
[0105] S1413, calculating the average temperature values corresponding to all target pixels in the high temperature area;
[0106] S1414, mapping the areas and average temperature values of the high temperature areas of the multiple infrared thermal images into a two-dimensional coordinate system in chronological order, extracting the high temperature area spreading trend and the average temperature changing trend in the two-dimensional coordinate system, and determining the temperature trend of the target block based on the high temperature area spreading trend and the average temperature changing trend, wherein the horizontal axis of the two-dimensional coordinate system is time, and the vertical axis is area.
[0107] In order to avoid interference caused by some human activities in the forest, this application uses the average temperature value and whether there is a spreading trend to determine whether it is an out-of-control fire or there is a trend of fire getting out of control. Therefore, two conditions need to be met: (1) the average temperature value of the high-temperature area remains or continues to rise; (2) the area of the high-temperature area continues to increase; the average temperature value reflects the overall temperature level of the high-temperature area. This application makes trend judgments based on the overall temperature level of the high-temperature area and whether there is a spreading trend, which can filter out some fixed-point high-temperature interference caused by human activities.
[0108] This application maps the area and average temperature value of the high temperature area of multiple infrared thermal images to a two-dimensional coordinate system in chronological order, and uses the two-dimensional coordinate system to quickly determine the trend, specifically including:
[0109] S14141, use a pre-built sliding window to slide along the two-dimensional coordinate system, and calculate the The average area of the high temperature area of the infrared thermal image and average temperature ,in, is the sliding step number;
[0110] The step size of the sliding window is fixed, and each time it slides, the average area within the window is calculated and average temperature , thereby avoiding the interference caused by single mutation data and obtaining the area change trend and temperature change trend as a whole.
[0111] S14142, in the average area When the average temperature increases along the time axis, When the temperature is not decreasing in one direction, it is determined that the target block has a temperature increasing trend.
[0112] One-way increase means that for the average area of any two adjacent steps, the average area of the next step is Greater than the previous step , that is, satisfy , in some cases; the average temperature value Not decreasing in one direction means that there is at least one set of average temperature values of two adjacent steps that do not satisfy .
[0113] S150, when there is a smoke feature in the target block and there is a temperature rising trend in the target block, the location information of the target block and the fire alarm information are sent to the target object based on a pre-set signal station.
[0114] In this application, only when the target block has smoke characteristics and the target block has a temperature rise trend, is it determined that there is a fire in the current target block. This is to avoid smoke interference caused by human activities or fixed-point high temperature interference. The above processing process has high fault tolerance and can help relevant staff understand various abnormal high temperature scenes in the forest.
[0115] In one embodiment of the present application, the location information of the target block and the fire alarm information are sent to the target object based on the pre-set signal station, including:
[0116] The longitude and latitude information of the drone is used as the location information of the target block, and the location information of the target block and the fire alarm information are transmitted to the nearest signal station, and then forwarded to the target object by the nearest signal station.
[0117] Since the network signal in the forest is unstable, the present application utilizes drones to transmit the current drone positioning information and fire alarm information back to the signal station, which then transmits the information in a chain until it is transmitted back to the monitoring center.
[0118] Figure 2 This is a schematic diagram of the drone signal return scenario in this application. Figure 2 As shown in the figure, there are multiple drones working simultaneously in the forest. Taking drone s2 as an example, when drone s2 detects an abnormality, it transmits the signal back to the nearest signal station A1, where the coordinates of each signal station are stored inside the drone. The distance between the drone and each signal station can be obtained through simple calculation. Selecting the nearest one for backtransmission can improve signal quality and reduce transmission time and delay.
[0119] Figure 3 This is a schematic diagram of the signal transmission scenario of the signal station in this application, such as Figure 3 As shown, the signal stations use the short message service of the Beidou system to transmit signals back to the monitoring center.
[0120] S160, when there is no smoke feature in the target block, or there is no temperature rising trend in the target block, marking the target block as a high-risk block, and marking a block whose fire risk value is not higher than a preset threshold as a low-risk block;
[0121] The high-risk blocks are patrolled according to a preset first patrol frequency, and the low-risk blocks are patrolled according to a preset second patrol frequency, wherein the first patrol frequency is greater than the second patrol frequency.
[0122] Finally, if there is no smoke feature in the target area, or if there is no temperature rise trend in the target area, there is no need to send back an alarm signal, and only enhanced patrols can be carried out by increasing the frequency of drone patrols. Monitoring data is sent back regularly to facilitate staff to connect to the overall situation of the forest.
[0123] A forest fire monitoring and alarm method of the present invention constructs a fire risk prediction model through historical data of a period of time before the forest fire, and then predicts the fire risk value of the target block at the current time based on the environmental data at the current time and the fire risk prediction model. If the fire risk value is large, it immediately triggers a drone patrol, and the drone hovers above the target block to take color images and infrared thermal images. If a fire has already occurred at this time, then the smoke characteristics or temperature rising trends can be extracted from the color images and infrared thermal images. At this time, the drone returns the positioning information to the preset signal station, and the signal station forwards the alarm information. The present application performs preliminary fire prediction based on environmental information, and controls the drone to perform image acquisition and feature extraction after being triggered, so that fires or fire hazards can be discovered more accurately and quickly.
[0124] like Figure 4 As shown, the present application also provides a forest fire monitoring and alarm device, comprising:
[0125] An acquisition module, used to acquire current environmental data of a target block at a current time point, wherein the target forest area includes multiple blocks, the target block is one of the multiple blocks, and the current environmental data includes values of multiple environmental parameters;
[0126] A risk determination module, used to determine the fire risk value of the target block at the current time point based on the current environmental data and a pre-built fire risk prediction model, wherein the fire risk prediction model is built based on historical fire data;
[0127] An image acquisition module, used to control the drone to acquire color images and infrared thermal images of the target block at multiple time points when the fire risk value is higher than a preset threshold;
[0128] A feature recognition module, used to perform smoke feature recognition on the target block based on color images at multiple time points, and to perform temperature trend recognition on the target block based on infrared thermal images at multiple time points;
[0129] The alarm module is used to send the location information of the target block and the fire alarm information to the target object based on a pre-set signal station when there is a smoke feature in the target block or there is a temperature rising trend in the target block.
[0130] A forest fire detection and alarm device of the present invention builds a fire risk prediction model through historical data of a period of time before the forest fire, and then predicts the fire risk value of the target block at the current time based on the environmental data at the current time and the fire risk prediction model. If the fire risk value is large, it will immediately trigger a drone patrol, and the drone will hover above the target block to take color images and infrared thermal images. If a fire has already occurred at this time, then the smoke characteristics or temperature rising trends can be extracted from the color images and infrared thermal images. At this time, the drone will return the positioning information to the preset signal station, and the signal station will forward the alarm information. The present application performs a preliminary prediction of the fire based on environmental information, and controls the drone to perform image acquisition and feature extraction after being triggered, so that fires or fire hazards can be discovered more accurately and quickly.
[0131] This embodiment further provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, any one of the methods in this embodiment is implemented, wherein the method is the execution logic of this system.
[0132] This embodiment also provides an electronic terminal, including: a processor and a memory;
[0133] The memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory, so that the terminal executes any one of the methods in this embodiment.
[0134] The computer-readable storage medium in this embodiment can be understood by ordinary technicians in this field: all or part of the steps of implementing the above-mentioned method embodiments can be completed by hardware related to the computer program. The aforementioned computer program can be stored in a computer-readable storage medium. When the program is executed, the steps of the above-mentioned method embodiments are executed; and the aforementioned storage medium includes: ROM, RAM, magnetic disk or optical disk and other media that can store program codes.
[0135] The electronic terminal provided in this embodiment includes a processor, a memory, a transceiver and a communication interface. The memory and the communication interface are connected to the processor and the transceiver and complete communication with each other. The memory is used to store computer programs, the communication interface is used to communicate, and the processor and the transceiver are used to run computer programs so that the electronic terminal executes each step of the above method.
[0136] In this embodiment, the memory may include a random access memory (RAM), and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.
[0137] The above-mentioned processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.
[0138] In the above-mentioned embodiments, although the present invention has been described in conjunction with the specific embodiments of the present invention, many replacements, modifications and variations of these embodiments will be apparent to those of ordinary skill in the art based on the foregoing description. The embodiments of the present invention are intended to cover all such replacements, modifications and variations falling within the broad scope of the appended claims.
[0139] The above embodiments are merely illustrative of the principles and effects of the present invention, and are not intended to limit the present invention. Anyone familiar with the art may modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by a person of ordinary skill in the art without departing from the spirit and technical concept disclosed by the present invention shall still be covered by the claims of the present invention.
Claims
1. A forest fire monitoring and alarm method, It is characterized in that Includes steps: Acquire current environmental data of a target block at a current time point, wherein the target forest area includes a plurality of blocks, the target block is one of the plurality of blocks, and the current environmental data includes values of a plurality of environmental parameters; Determining the fire risk value of the target block at the current time point based on the current environmental data and a pre-built fire risk prediction model, wherein the fire risk prediction model is built based on historical fire data; When the fire risk value is higher than a preset threshold, controlling the drone to collect color images and infrared thermal images of the target block at multiple time points; Based on the color images at multiple time points, smoke features are identified for the target block, and based on the infrared thermal images at multiple time points, temperature trends are identified for the target block; based on the color images at multiple time points, smoke features are identified for the target block, including: converting the color images at multiple time points into grayscale images; preprocessing the multiple grayscale images to obtain preprocessed images, wherein the preprocessing method includes cropping and high-pass filtering; extracting contour features and texture features of the preprocessed images; and inputting the contour features and texture features of the multiple preprocessed images into a pre-built recognition model to obtain multiple smoke recognition results; when any smoke recognition result is that smoke exists, it is determined that the target block has smoke features; based on the infrared thermal images at multiple time points, temperature trends are identified for the target block, including: Based on the calibration parameters of the infrared camera, pixel values in multiple infrared thermal images are converted into temperature values; target pixel points in the infrared thermal images whose temperature values are higher than the temperature threshold are screened out to obtain the high temperature area and the area of the high temperature area, wherein the area of the high temperature area is the number of corresponding target pixel points; the average temperature value corresponding to all target pixel points in the high temperature area is calculated; the area and the average temperature value of the high temperature area of multiple infrared thermal images are mapped to a two-dimensional coordinate system in chronological order, and the high temperature area spreading trend and the average temperature change trend are extracted in the two-dimensional coordinate system, and the temperature trend of the target block is determined based on the high temperature area spreading trend and the average temperature change trend, wherein the horizontal axis of the two-dimensional coordinate system is time and the vertical axis is area; extracting the high temperature area spreading trend and the average temperature change trend in the two-dimensional coordinate system includes: using a pre-constructed sliding window to slide along the two-dimensional coordinate system, and calculating the area in the window at each sliding step. The average area of the high temperature area of the infrared thermal image and average temperature ,in, is the sliding step number; in the average area When the average temperature increases along the time axis, When the temperature is not decreasing in one direction, it is determined that the target block has a temperature rising trend; When there are smoke features in the target block and there is a temperature increase trend in the target block, the location information of the target block and the fire alarm information are sent to the target object based on a pre-set signal station.
2. A forest fire monitoring and alarm method according to claim 1, It is characterized in that The method for constructing the fire risk prediction model includes: Obtain multiple data samples of the target forest area or other forest areas , wherein the data samples are the values of various environmental parameters at multiple time points within the target time period before the fire occurs, The fire occurrence sequence number; Calculate the mean of the values of each environmental parameter in each data sample , and calculate multiple data samples Medium environmental parameters The average Variance and average ; Variance of environmental parameters Normalization is performed and based on the variance of multiple normalized environmental parameters Calculation environment parameters Risk weight , where the risk weight The mathematical expression is: In the formula, is the number of environmental parameters; Based on environmental parameters The average Variance and average Build environment parameters The risk reference range is as follows: The mathematical expression of the risk reference range is: ,in, is the range adjustment factor; A fire risk warning model is constructed based on the risk reference range and weight of each environmental parameter.
3. A forest fire monitoring and alarm method according to claim 2, It is characterized in that Determining the fire risk value of the target block at the current time point based on the current environmental data and a pre-built fire risk prediction model includes: Comparing the value of each environmental parameter in the current environmental data with the corresponding risk reference range to obtain a target environmental parameter, wherein the target environmental parameter is an environmental parameter whose value falls within the corresponding risk reference range; The risk weights of all target environmental parameters are summed up to obtain the fire risk value.
4. A forest fire monitoring and alarm method according to claim 1, It is characterized in that The method for constructing the recognition model includes: Acquire a positive sample image and a negative sample image, wherein the positive sample image contains smoke features, and the negative sample image does not contain smoke features; Converting the positive sample image and the negative sample image into grayscale sample images; Preprocessing a plurality of grayscale sample images to obtain preprocessed sample images; Extracting contour features and texture features of the preprocessed sample image to obtain training data; The artificial neural network is trained based on the training data to obtain a recognition model.
5. A forest fire monitoring and alarm method according to claim 1, It is characterized in that The location information of the target block and the fire alarm information are sent to the target object based on the pre-set signal station, including: The longitude and latitude information of the drone is used as the location information of the target block, and the location information of the target block and the fire alarm information are transmitted to the nearest signal station, and then forwarded to the target object by the nearest signal station.
6. A forest fire monitoring and alarm method according to claim 1, It is characterized in that Also includes: When there is no smoke feature in the target block, or when there is no temperature rising trend in the target block, the target block is marked as a high-risk block, and blocks with a fire risk value not higher than a preset threshold are marked as low-risk blocks; The high-risk blocks are patrolled according to a preset first patrol frequency, and the low-risk blocks are patrolled according to a preset second patrol frequency, wherein the first patrol frequency is greater than the second patrol frequency.
7. A forest fire monitoring and alarm device, It is characterized in that A method for implementing a forest fire monitoring and alarm method as claimed in claim 1, comprising: An acquisition module, used to acquire current environmental data of a target block at a current time point, wherein the target forest area includes multiple blocks, the target block is one of the multiple blocks, and the current environmental data includes values of multiple environmental parameters; A risk determination module, used to determine the fire risk value of the target block at the current time point based on the current environmental data and a pre-built fire risk prediction model, wherein the fire risk prediction model is built based on historical fire data; An image acquisition module, used to control the drone to acquire color images and infrared thermal images of the target block at multiple time points when the fire risk value is higher than a preset threshold; A feature recognition module is used to perform smoke feature recognition on the target block based on the color images at multiple time points, and to perform temperature trend recognition on the target block based on the infrared thermal images at multiple time points; the smoke feature recognition on the target block based on the color images at multiple time points includes: converting the color images at multiple time points into grayscale images; preprocessing the multiple grayscale images to obtain preprocessed images, wherein the preprocessing method includes cropping and high-pass filtering; extracting contour features and texture features of the preprocessed images; and inputting the contour features and texture features of the multiple preprocessed images into a pre-built recognition model to obtain multiple smoke recognition results; when any smoke recognition result is that smoke exists, it is determined that the target block has smoke features; the temperature trend recognition on the target block based on the infrared thermal images at multiple time points includes: converting multiple infrared thermal images into grayscale images based on the calibration parameters of the infrared camera; The pixel values in the external thermal image are converted into temperature values; the target pixel points whose temperature values in the infrared thermal image are higher than the temperature threshold are screened out to obtain the high temperature area and the area of the high temperature area, wherein the area of the high temperature area is the number of corresponding target pixel points; the average temperature value corresponding to all target pixel points in the high temperature area is calculated; the area and the average temperature value of the high temperature area of multiple infrared thermal images are mapped to a two-dimensional coordinate system in chronological order, and the high temperature area spreading trend and the average temperature change trend are extracted in the two-dimensional coordinate system, and the temperature trend of the target block is determined based on the high temperature area spreading trend and the average temperature change trend, wherein the horizontal axis of the two-dimensional coordinate system is time and the vertical axis is area; the high temperature area spreading trend and the average temperature change trend are extracted in the two-dimensional coordinate system, including: using a pre-constructed sliding window to slide along the two-dimensional coordinate system, and calculating the window area at each sliding step. The average area of the high temperature area of the infrared thermal image and average temperature ,in, is the sliding step number; in the average area When the average temperature increases along the time axis, When the temperature is not decreasing in one direction, it is determined that the target block has a temperature rising trend; The alarm module is used to send the location information of the target block and the fire alarm information to the target object based on a pre-set signal station when there is a smoke feature in the target block and there is a temperature rising trend in the target block.
