High-altitude monitoring intelligent early warning system based on Internet

Abstract

The invention discloses a high-altitude monitoring intelligent early warning system based on the Internet, relates to the technical field of high-altitude video monitoring, and solves the problem that forest fire is monitored by acquiring forest abnormal information through acquired forest monitoring surface information in the prior art. The problem that forest fire detection is inaccurate and not timely due to the fact that earth surface information obtained by satellite remote sensing is easily influenced by cloud layers and smoke dust factors is not considered; according to the invention, zoning management is carried out on the total area of the forest farm; regularly acquiring satellite images of the forest farm through a satellite remote sensing technology; judging whether an abnormal area does not appear in the satellite image or not; setting early warning measures of each region and acquiring detailed information; judging whether corresponding multi-region early warning measure linkage needs to be carried out or not according to the abnormal region; the detailed information is processed, an information identification model is constructed, and abnormal conditions are identified by training the information identification model; judging whether a fire occurs in the recognition result or not; and forest fire disasters can be timely and accurately found.

Description

Technical Field

[0001] The invention belongs to the field of high-altitude monitoring, relates to intelligent early warning technology, and specifically is an Internet-based high-altitude monitoring intelligent early warning system. Background Art

[0002] Satellite remote sensing can quickly detect hotspots, monitor the spread of fires, provide fire information in a timely manner, use remote sensing methods to make forest fire risk forecasts, and use satellite digital data to estimate the burned area. It has a wide detection range, collects data quickly, and can obtain continuous data to reflect the dynamic changes of fire. Moreover, the collected data is not affected by terrain conditions and the images are real. Through high-definition cameras connected to the Internet and intelligent analysis technology, the system can capture images and videos of high-altitude areas in real time and perform rapid analysis and processing, which greatly improves the real-time and accuracy of monitoring and helps to promptly discover and warn of potential safety hazards or abnormal behaviors.

[0003] The prior art (invention patent application with publication number CN104157088A) discloses a method for monitoring forest fires using satellite remote sensing; the method is characterized in that forest monitoring surface information is obtained using satellite remote sensing, forest abnormality information is obtained based on the obtained forest monitoring surface information, and fire monitoring is performed based on the forest abnormality information; the prior art obtains forest abnormality information through the obtained forest monitoring surface information, thereby monitoring forest fires, but does not take into account that the surface information obtained by satellite remote sensing is easily affected by cloud and smoke factors, which may lead to inaccurate and untimely forest fire detection.

[0004] The present invention provides an Internet-based high-altitude monitoring intelligent early warning system to solve the above technical problems. Summary of the invention

[0005] The present invention aims to solve at least one of the technical problems existing in the prior art; to this end, the present invention proposes an Internet-based high-altitude monitoring intelligent early warning system, which is used to solve the technical problem that the prior art obtains forest abnormality information by obtaining forest monitoring surface information, thereby monitoring forest fires, but does not take into account that the surface information obtained by satellite remote sensing is easily affected by clouds and smoke factors, which will lead to inaccurate and untimely forest fire detection.

[0006] To achieve the above-mentioned object, the first aspect of the present invention provides an Internet-based high-altitude monitoring intelligent early warning system, comprising: a data acquisition module, a model building module, an early warning response module and a user interface module;

[0007] Data collection module: retrieve the total area of ​​the forest farm from the database, manage the total area of ​​the forest farm in a zoned manner, and regularly obtain satellite images of the forest farm through satellite remote sensing technology; determine whether there are no abnormal areas in the satellite images; if yes, continue to determine; if no, issue a first-level warning signal, set warning measures for each area and obtain detailed information; determine whether corresponding multi-area warning measures need to be linked based on the abnormal area; if yes, issue a reminder signal; if no, implement corresponding single-area warning measures; detailed information includes: real-time video stream, environmental sensor data, and forest farm pictures taken by drones;

[0008] Model building module: Process detailed information and build an information recognition model, and identify abnormal situations by training the information recognition model;

[0009] Early warning response module: Determine whether a fire occurs in the recognition results; if so, an alarm signal is automatically sent and the relevant department personnel carry out emergency plan processing; if not, continue to judge.

[0010] Preferably, the system also includes: a user interface module: used to display satellite images of the forest farm and the acquired detailed information; and display real-time monitoring screens and information recognition results by continuously monitoring the operating status of the system.

[0011] The present invention provides users with satellite images and data analysis results of forest farms that can be displayed intuitively and clearly through a user interface module, which is helpful for managers or users to quickly understand the information of the forest farm. This intuitive display method helps to improve decision-making efficiency and accuracy.

[0012] Preferably, the zoning management of the total area of ​​the forest farm includes:

[0013] The resource assessment results are obtained from the database, and the forest farm is divided into different functional areas based on the resource assessment results. The functional areas are then graded according to the frequency of human activities. Among them, different functional areas include: protection areas, ecological restoration areas, forestry management areas, scenic areas, and community utilization areas. The grade division includes: marking protection areas and ecological restoration areas as first-level areas, marking forestry management areas and community utilization areas as second-level areas, and marking scenic areas as third-level areas.

[0014] It should be noted that forestry management areas are where forestry management activities are carried out moderately under the premise of ensuring the health of the ecosystem; community utilization areas are where local communities use resources in a limited manner in traditional ways.

[0015] The present invention helps to better monitor and manage the risk of forest fires through reasonable functional zoning, thereby reducing losses; this classification method provides a scientific basis for formulating more specific management policies, which is conducive to managers being able to take different management measures for different areas.

[0016] Preferably, the determining whether there is no abnormal area in the satellite image includes:

[0017] The color image received by the satellite is stored in a database, a normal satellite image is obtained from the database, and the color image received by the satellite is compared with the normal satellite image to determine whether there is a color difference in the satellite image; if so, an abnormal area appears in the satellite image; if not, no abnormal area appears in the satellite image.

[0018] By comparing newly received satellite images with normal images, the present invention can quickly detect color changes in satellite images and further identify possible abnormal areas. This real-time monitoring capability helps to detect fires in a timely manner, providing a valuable time window for timely response and measures. By using computer vision and image processing technology, color differences in images can be accurately compared. This method can provide more objective and accurate anomaly detection results, helping to reduce false alarms and missed alarms.

[0019] Preferably, the step of setting early warning measures for each area and obtaining detailed information includes:

[0020] In the first-level area, drones are used to conduct inspections and take pictures of the forest farm. Different types of environmental sensors are installed in the second-level area to collect environmental data. The video surveillance installation diagram is obtained from the database, and different types of cameras are installed in the third-level area according to the video surveillance installation diagram to capture real-time video streams. Among them, environmental sensors include: temperature, humidity, smoke, and gas concentration sensors; different types of cameras include: high-definition cameras and thermal imaging cameras.

[0021] The present invention uses drones for inspections, which can quickly and efficiently cover a large area of ​​forest farms in the first-level area; environmental sensors are installed in the second-level area to monitor environmental changes in the forest farm in real time and accurately; cameras installed in the third-level area can capture real-time video streams and provide clear visual information; the entire system integrates multiple technical means such as drones, environmental sensors and cameras to achieve intelligent management of forest farms. This management method not only improves management efficiency, but also reduces labor costs, which helps to promote the modernization and intelligent development of forestry management.

[0022] Preferably, judging whether corresponding multi-region early warning measures need to be linked according to the abnormal area includes:

[0023] Retrieve the satellite image of the abnormal area to determine whether the abnormal area in the satellite image appears only in a single area; if so, there is no need to implement corresponding multi-regional early warning measures; if not, there is a need to implement corresponding multi-regional early warning measures.

[0024] By judging that the anomaly only occurs in a single area, the present invention can avoid unnecessary linkage of early warning measures in multiple areas, thereby saving resources and time; conversely, if the abnormal area involves multiple locations, the linkage of early warning measures in multiple areas can be initiated in time to ensure that relevant areas can receive timely and effective early warning and response; according to the analysis results of satellite images, human, material and financial resources can be allocated more reasonably; for key abnormal areas, monitoring and early warning efforts can be strengthened; for non-key areas, the investment of resources can be appropriately reduced, thereby achieving optimal allocation of resources.

[0025] Preferably, the processing of the detailed information and constructing the information recognition model includes:

[0026] Retrieve detailed information, clean the detailed information, analyze the collected detailed information through computer vision technology and build an abnormal information data set; mark the abnormal information data set as an abnormal recognition sequence; retrieve the information recognition model, input the abnormal recognition sequence into the information recognition model to obtain the abnormal label; match the corresponding abnormal type according to the abnormal label; wherein the information recognition model is built based on the artificial intelligence model; the abnormal label is set to a positive integer.

[0027] The present invention helps to remove invalid, duplicate or erroneous data by cleaning information, ensuring that the data input into the computer vision analysis system is accurate, and further improving the accuracy and reliability of subsequent analysis; marking the abnormal information data set as an abnormal identification sequence and setting it to a positive integer as an abnormal label, so as to facilitate subsequent rapid matching to the corresponding abnormal type according to the abnormal label, which helps to achieve rapid classification and positioning of abnormal information.

[0028] Preferably, the information recognition model is constructed based on an artificial intelligence model, including:

[0029] Acquire standard training data; wherein the standard training data includes standard input data consistent with the content attributes of the abnormal recognition sequence, and standard output data consistent with the content attributes of the abnormal label;

[0030] The artificial intelligence model is trained using standard training data, and the trained artificial intelligence model is marked as an abnormal prediction model; wherein the artificial intelligence model includes a convolutional neural network model or a long short-term memory neural network model.

[0031] Preferably, identifying abnormal situations by training the information recognition model includes:

[0032] Retrieve the abnormal information data set, and divide the abnormal information data set into a training set and a validation set; use the abnormal information as standard input data, and use the type of abnormal information, target detection label, and annotation information as standard output data;

[0033] The information recognition model is trained using the training set; the trained information recognition model is verified using the verification set to obtain verification parameters; and abnormal situations are identified based on the verification parameters.

[0034] Compared with the prior art, the present invention has the following beneficial effects:

[0035] 1. Satellite images provide objective and comprehensive data support, which helps decision makers formulate response strategies and measures more scientifically. By comparing satellite images at different time points, it is also possible to analyze the changing trends and patterns of abnormal areas, providing a reference for long-term planning and decision-making. When multi-regional early warning measures need to be linked, the sharing and analysis of satellite images can promote collaboration and cooperation between different regions. By jointly analyzing satellite image data, different regions can understand each other's situation and needs more accurately, thereby formulating more coordinated response plans. According to the analysis results of satellite images, human, material and financial resources can be allocated more reasonably. For key abnormal areas, monitoring and early warning efforts can be strengthened. For non-key areas, resource input can be appropriately reduced, thereby achieving optimal resource allocation.

[0036] 2. Through satellite remote sensing technology and real-time video streaming, the system can monitor the forest farm conditions in real time or near real time to ensure that abnormal situations can be discovered quickly; combined with environmental sensor data and forest farm pictures taken by drones, the system can evaluate the forest farm status in multiple dimensions and all directions to improve the accuracy of abnormality detection; zoning management of the total area of ​​the forest farm helps to accurately locate abnormal areas and reduce false alarms and missed alarms; for situations that require multi-regional early warning measures to be linked, the system can automatically issue reminder signals to ensure that relevant departments can work together to jointly respond to crises; by training information recognition models, the system can automatically identify abnormal situations, provide intelligent support for decision-making, and help reduce the subjectivity and uncertainty of human judgment. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0038] Figure 1 A schematic diagram of the relationship between the modules included in the present invention;

[0039] Figure 2 It is a schematic diagram of the high-altitude monitoring intelligent early warning process of the present invention;

[0040] Figure 3It is a schematic diagram of the specific steps of data collection of the present invention;

[0041] Figure 4 Schematic diagram of the specific steps of constructing the model of the present invention. DETAILED DESCRIPTION

[0042] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0043] See also Figure 1-Figure 2 , the first aspect of the present invention provides an Internet-based high-altitude monitoring intelligent early warning system, including: including: a data acquisition module, a model building module, an early warning response module and a user interface module;

[0044] Data collection module: retrieve the total area of ​​the forest farm from the database and manage the total area of ​​the forest farm in a zoned manner; regularly obtain satellite images of the forest farm through satellite remote sensing technology; determine whether there are no abnormal areas in the satellite images; if yes, continue to determine; if no, issue a first-level warning signal, set warning measures for each area and obtain detailed information; determine whether corresponding multi-area warning measures need to be linked based on the abnormal area; if yes, issue a reminder signal; if no, implement corresponding single-area warning measures; the detailed information includes: real-time video stream, environmental sensor data, and forest farm pictures taken by drones;

[0045] Model building module: Process detailed information and build an information recognition model, and identify abnormal situations by training the information recognition model;

[0046] Early warning response module: Determine whether a fire occurs in the recognition results; if so, an alarm signal is automatically sent and the relevant department personnel carry out emergency plan processing; if not, continue to judge.

[0047] See also Figure 3, the specific process of data collection, retrieve the total area of ​​the forest farm from the database, obtain the resource assessment results from the database, divide the forest farm into different functional areas according to the resource assessment results, and divide the functional areas into levels according to the frequency of personnel activities; among them, different functional areas include: protection areas, ecological restoration areas, forestry management areas, scenic areas, and community utilization areas; the level division includes: marking the protection areas and ecological restoration areas as first-level areas, marking the forestry management areas and community utilization areas as second-level areas, and marking the scenic areas as third-level areas; regularly obtain satellite images of the forest farm through satellite remote sensing technology; store the color images received by the satellite in the database, obtain normal satellite images from the database, compare the color images received by the satellite with the normal satellite images, and determine whether there is a color difference in the satellite image; if so, there is an abnormal area in the satellite image , issue a first-level warning signal; if not, no abnormal area appears in the satellite image, and continue to judge; in the first-level area, use drones to patrol and take pictures of the forest farm; install different types of environmental sensors in the second-level area, and collect environmental data through environmental sensors; obtain the video surveillance installation diagram from the database, and install different types of cameras in the third-level area according to the video surveillance installation diagram to shoot real-time video streams; among them, environmental sensors include: temperature, humidity, smoke, and gas concentration sensors; different types of cameras include: high-definition cameras and thermal imaging cameras; retrieve satellite images of the abnormal area to determine whether the abnormal area in the satellite image appears only in a single area; if yes, there is no need to carry out corresponding multi-area early warning measures linkage; if not, issue a reminder signal, and corresponding multi-area early warning measures linkage is required.

[0048] It should be noted that the specific frequency of obtaining satellite images of forest farms through satellite remote sensing technology is set by the staff based on actual conditions.

[0049] For example, it is necessary to conduct high-altitude monitoring of a forest farm, obtain resource assessment results from the database, divide the forest farm into different functional areas based on the resource assessment results, and divide the functional areas into levels according to the frequency of human activities; mark the protection area and ecological restoration area as the first-level area, the forestry management area and community utilization area as the second-level area, and the scenic area as the third-level area; set the satellite image of the forest farm to be acquired every 10 minutes through satellite remote sensing technology; obtain a total of 3 satellite images of different time periods and mark them as A, B, and C, obtain normal satellite images from the database, and compare the color image received by the satellite with the normal satellite image, where image A No color difference appears; color difference appears in the first-level area in image B, and color difference appears in both the first-level and second-level areas in image C; a first-level warning signal is issued in the time period of images B and C; in the time period of image B, inspections are carried out by using drones and pictures of the first-level area in the forest farm are taken and marked as information B1; in the time period of image C, inspections are carried out by using drones and pictures of the first-level area in the forest farm are taken and marked as information C1, and environmental data of the second-level area are collected by using environmental sensors and marked as information C2; information B1, information C1, and information C2 are input as detailed information into the model building module.

[0050] See also Figure 4 , the specific steps of model construction are: retrieve detailed information, clean the detailed information, analyze the collected detailed information through computer vision technology and build an abnormal information data set; mark the abnormal information data set as an abnormal recognition sequence; retrieve the information recognition model, input the abnormal recognition sequence into the information recognition model, and obtain the abnormal label; match the corresponding abnormal type according to the abnormal label; wherein the information recognition model is built based on the artificial intelligence model; the abnormal label is set to a positive integer; obtain standard training data; wherein the standard training data includes standard input data consistent with the content attributes of the abnormal recognition sequence, and standard output data consistent with the content attributes of the abnormal label;

[0051] Using standard training data to train an artificial intelligence model, and marking the trained artificial intelligence model as an anomaly prediction model; wherein the artificial intelligence model includes a convolutional neural network model or a long short-term memory neural network model; retrieving an abnormal information data set, and dividing the abnormal information data set into a training set and a validation set; using the abnormal information as standard input data, and using the type of the abnormal information, target detection label, and annotation information as standard output data;

[0052] The information recognition model is trained using the training set; the trained information recognition model is verified using the verification set to obtain verification parameters; and abnormal situations are identified based on the verification parameters.

[0053] For example, information B1, information C1, and information C2 are retrieved and input into the information recognition model built based on the artificial intelligence model. The information recognition model shows that the appearance of information B1 is caused by smoke, so no fire occurred in the time period corresponding to information B1; the information recognition model shows that the appearance of information C1 is caused by clouds, and the appearance of information C2 is caused by a fire. Therefore, if a fire occurs in the time period corresponding to information C1 and information C2, an alarm signal will be automatically sent immediately, and personnel from relevant departments will carry out emergency plan processing.

[0054] Part of the data in the above formula is calculated by removing the dimension and taking its numerical value. The formula is a formula closest to the actual situation obtained by software simulation of a large amount of collected data; the preset parameters and preset thresholds in the formula are set by technical personnel in this field according to actual conditions or obtained through simulation of a large amount of data.

[0055] The working principle of the present invention is as follows: the present invention retrieves the total area of ​​the forest farm from a database and performs zoning management on the total area of ​​the forest farm; regularly obtains satellite images of the forest farm through satellite remote sensing technology; determines whether there is no abnormal area in the satellite image; if yes, continues to judge; if not, issues a first-level early warning signal, sets early warning measures for each area and obtains detailed information; determines whether corresponding multi-area early warning measures need to be linked according to the abnormal area; if yes, issues a reminder signal; if not, implements corresponding single-area early warning measures; processes the detailed information and constructs an information recognition model, identifies abnormal situations by training the information recognition model; and determines whether a fire occurs in the recognition result.

[0056] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. An Internet-based high-altitude monitoring intelligent early warning system, characterized in that: include: Data collection module, model building module, warning response module and user interface module; Data collection module: retrieve the total area of ​​the forest farm from the database and carry out zoning management of the total area of ​​the forest farm; regularly obtain satellite images of the forest farm through satellite remote sensing technology; Determine whether there are no abnormal areas in the satellite image; If yes, continue to judge; If not, a first-level warning signal is issued, warning measures for each area are set, and detailed information is obtained; based on the abnormal area, it is determined whether corresponding multi-area warning measures need to be linked; If yes, a reminder signal is issued; If not, corresponding single-region early warning measures will be implemented; detailed information includes: real-time video streams, environmental sensor data, and forest farm pictures taken by drones; Model building module: Process detailed information and build an information recognition model, and identify abnormal situations by training the information recognition model; Early warning response module: Determine whether a fire occurs in the recognition results; if so, an alarm signal is automatically sent and the relevant department personnel carry out emergency plan processing; if not, continue to judge.

2. According to the Internet-based high-altitude monitoring intelligent early warning system of claim 1, it is characterized in that: The system also includes: a user interface module: used to display satellite images of the forest farm and the acquired detailed information; by continuously monitoring the operating status of the system, displaying real-time monitoring images and information recognition results.

3. The high-altitude monitoring intelligent early warning system based on the Internet according to claim 1 is characterized in that: The zoning management of the total area of ​​the forest farm includes: The resource assessment results are obtained from the database, and the forest farm is divided into different functional areas based on the resource assessment results. The functional areas are then graded according to the frequency of human activities. Among them, different functional areas include: protection areas, ecological restoration areas, forestry management areas, scenic areas, and community utilization areas. The grade division includes: marking protection areas and ecological restoration areas as first-level areas, marking forestry management areas and community utilization areas as second-level areas, and marking scenic areas as third-level areas.

4. The high-altitude monitoring intelligent early warning system based on the Internet according to claim 1 is characterized in that: The determining whether there is no abnormal area in the satellite image includes: The color image received by the satellite is stored in a database, a normal satellite image is obtained from the database, and the color image received by the satellite is compared with the normal satellite image to determine whether there is a color difference in the satellite image; if so, an abnormal area appears in the satellite image; if not, no abnormal area appears in the satellite image.

5. The high-altitude monitoring intelligent early warning system based on the Internet according to claim 1 is characterized in that: Setting early warning measures for each area and obtaining detailed information include: In the first-level area, drones are used to conduct inspections and take pictures of the forest farm. Different types of environmental sensors are installed in the second-level area to collect environmental data. The video surveillance installation diagram is obtained from the database, and different types of cameras are installed in the third-level area according to the video surveillance installation diagram to capture real-time video streams. Among them, environmental sensors include: temperature, humidity, smoke, and gas concentration sensors; different types of cameras include: high-definition cameras and thermal imaging cameras.

6. The high-altitude monitoring intelligent early warning system based on the Internet according to claim 1 is characterized in that: The determining whether corresponding multi-region early warning measures need to be linked according to the abnormal area includes: Retrieve the satellite image of the abnormal area to determine whether the abnormal area in the satellite image appears only in a single area; if so, there is no need to implement corresponding multi-regional early warning measures; if not, there is a need to implement corresponding multi-regional early warning measures.

7. The high-altitude monitoring intelligent early warning system based on the Internet according to claim 1 is characterized in that: The detailed information is processed and an information recognition model is constructed, including: Retrieve detailed information, clean the detailed information, analyze the collected detailed information through computer vision technology and build an abnormal information data set; mark the abnormal information data set as an abnormal recognition sequence; retrieve the information recognition model, input the abnormal recognition sequence into the information recognition model to obtain the abnormal label; match the corresponding abnormal type according to the abnormal label; wherein the information recognition model is built based on the artificial intelligence model; the abnormal label is set to a positive integer.

8. The Internet-based high-altitude monitoring intelligent early warning system according to claim 7 is characterized in that: The information recognition model is constructed based on an artificial intelligence model, including: Acquire standard training data; wherein the standard training data includes standard input data consistent with the content attributes of the abnormal recognition sequence, and standard output data consistent with the content attributes of the abnormal label; The artificial intelligence model is trained using standard training data, and the trained artificial intelligence model is marked as an abnormal prediction model; wherein the artificial intelligence model includes a convolutional neural network model or a long short-term memory neural network model.

9. The high-altitude monitoring intelligent early warning system based on the Internet according to claim 1 is characterized in that: The identifying of abnormal situations by training the information identification model includes: Retrieve the abnormal information data set, and divide the abnormal information data set into a training set and a validation set; use the abnormal information as standard input data, and use the type of abnormal information, target detection label, and annotation information as standard output data; The information recognition model is trained using the training set; the trained information recognition model is verified using the verification set to obtain verification parameters; and abnormal situations are identified based on the verification parameters.

Images