A monitoring and early warning method for the distribution of wild plants based on image recognition

Abstract

The present invention discloses a method for monitoring and early warning of the distribution of wild plants based on image recognition, and relates to the technical field of monitoring and early warning of the distribution of wild plants. The method promotes the scientific evaluation of the ecological protection status of each sub-region by calculating the coverage index fgl and the population complexity index fz. When the coverage index or the population complexity index of a certain sub-region does not reach the preset threshold, it is marked as unqualified. By calculating the environmental pressure index HJyl, when the environmental pressure index HJyl is higher than the preset first pressure threshold, a first early warning instruction will be automatically sent to indicate that the environmental pollution status is unqualified. A timely early warning mechanism helps to take countermeasures to prevent environmental deterioration. By setting a second monitoring point and constructing an invasive species interference index Rqgr, it promotes effective monitoring and evaluation of the impact of alien invasive species on the ecosystem, and sends a second early warning instruction in a timely manner to promote rapid response and control of invasive species, and protect local endangered plants and the stability of the ecosystem.

Description

Technical Field

[0001] The present invention relates to the technical field of monitoring and early warning of the distribution status of wild plants, and in particular to a method for monitoring and early warning of the distribution status of wild plants based on image recognition. Background Art

[0002] With the continuous intensification of global climate change and human activities, the living environment of wild plants has deteriorated day by day, and the survival status of endangered plants has become more worrying. Endangered plants are not only an important part of the ecosystem, but also have high scientific research and protection value. Traditional endangered plant monitoring methods mainly rely on manual surveys and fixed-point observations. These methods are inefficient and easy to miss important information when facing the vast natural environment and diverse plant populations.

[0003] The distribution and coverage of endangered plants are important indicators for assessing the health of ecosystems. Areas with high coverage usually have higher conservation value. Classifying and maintaining these areas can effectively protect the living environment of endangered plants and prevent them from further reduction or extinction. However, existing technologies have great limitations in identifying and analyzing the distribution of endangered plants, and cannot provide accurate coverage data in a timely manner, thus affecting the implementation of conservation measures. Summary of the invention

[0004] 1. Technical issues to be resolved

[0005] In view of the deficiencies of the prior art, the present invention provides a method for monitoring and early warning of the distribution status of wild plants based on image recognition to solve the problems mentioned in the background technology.

[0006] (II) Technical solution

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions: A method for monitoring and early warning of the distribution of wild plants based on image recognition comprises the following steps:

[0008] Using GIS technology and remote sensing data, a three-dimensional ecological model is established based on geographical features and ecological environment information, and the monitoring area is divided into several sub-areas. Corresponding data collection points are set in each sub-area. Each sub-area is divided based on ecosystem types, including forests, grasslands and wetlands;

[0009] A multi-rotor drone carrying multispectral satellite imaging equipment was used to collect plant distribution images in each sub-region and establish an image dataset;

[0010] Using the convolutional neural network (CNN) model or deep learning model technology, a plant recognition model is established to analyze the image data set to identify endangered plant species and coverage rates. The distribution and coverage rate of endangered plants in each sub-area are calculated and analyzed to obtain the coverage rate index fgl and population complexity index fz:

[0011] When the coverage index fgl exceeds the preset protection threshold, it means that the sub-area meets the ecological protection scope and is marked as the first qualified area in the three-dimensional ecological model;

[0012] When the population complexity index fz exceeds the preset second ecological stability threshold, it means that the diversified population structure in the sub-area is qualified, and the second qualified mark is made in the three-dimensional ecological model;

[0013] An environmental monitoring point is set for the sub-area that obtains both the first qualified mark and the second qualified mark, and the environmental information and pollution information of the sub-area are collected and acquired to establish a first ecological environment data set; and an environmental pressure index HJyl is constructed based on the first ecological environment data set. If the environmental pressure index HJyl is higher than the first pressure threshold, a first early warning instruction is sent to the outside;

[0014] A second monitoring point is set for the sub-area that obtains both the first qualified mark and the second qualified mark, and the harmful plant information and quantity data in the sub-area are collected to establish a second alien invasion data set; and an invasive species interference index Rqgr is constructed based on the second alien invasion data set. If the invasive species interference index Rqgr exceeds a safety threshold, a second early warning instruction is sent to the outside, and a corresponding repair strategy is generated based on the first early warning instruction and the second early warning instruction.

[0015] Preferably, the geographic information system (GIS) technology and remote sensing data are used to obtain the geographical features and ecological environment information of the monitoring area through high-resolution satellite images and drone aerial photography data, including topography, vegetation cover and hydrological characteristics; based on the acquired geographical and ecological data, a three-dimensional ecological model of the monitoring area is established; the three-dimensional ecological model includes the following contents: terrain elevation data DEM, vegetation type layer and hydrological layer; the vegetation type layer shows the spatial distribution of different vegetation types, and the hydrological layer shows the location and characteristics of rivers, lakes and wetlands;

[0016] According to the three-dimensional ecological model, the monitoring area is divided into several sub-areas, and multiple data collection points are set in each sub-area.

[0017] Preferably, the image data set acquired by the drone is spliced ​​into an image that completely covers the sub-area. After geometric correction and spectral correction are performed on the image data set, a plant recognition model is established. The image recognition task is processed by a convolutional neural network (CNN). Endangered plant species, non-endangered plant species, hazardous plant species, plant health conditions, vegetation type characteristics, and coverage area characteristics in the sub-area are extracted, and the image data set is marked to mark the category and location information of each plant. The image data set is used as the first input into the plant recognition model to calculate the coverage index fgl and the population complexity index fz:

[0018] The coverage index fgl is generated by the following formula:

[0019]

[0020] In the formula, CR i The population proportion of the i-th endangered plant species, A total Expressed as the total area of ​​the sub-region, A i is the coverage area of ​​the i-th endangered plant species, w i is the weight of the i-th endangered plant species, based on the endangerment level and ecological importance, n is the number of endangered plant species in the sub-region, H is the uniformity index of plant cover, calculated by the Shannon index Shannonindex, ln is expressed as a logarithmic operation;

[0021] When the coverage index fgl exceeds the preset protection threshold, it means that the sub-area meets the ecological protection scope and is marked as the first qualified one in the three-dimensional ecological model; when the coverage index fgl does not exceed the preset protection threshold, it means that the coverage of the sub-area is unqualified and is marked as the first unqualified one.

[0022] Preferably, the population complexity index fz is generated by the following formula:

[0023]

[0024] Where N total is the total number of plant species in the sub-area, N i is the number of individuals of the i-th endangered plant species, w i is the weight of the i-th endangered plant species, based on its endangerment level and ecological importance; n is the number of endangered plant species in the sub-region, P i is the relative proportion of the i-th endangered plant species, that is, the ratio of the number of individuals to the total number. q is an adjustment parameter, ranging from 0 to 2. When q = 1, the M value is the weighted sum of the proportions of different populations. When q ≠ 1, the M value reflects the distribution difference of the population proportions.

[0025] When the population complexity index fz exceeds the preset second ecological stability threshold, it means that the diverse population structure in the sub-area is qualified, and a second qualified mark is made in the three-dimensional ecological model; when the population complexity index fz does not exceed the preset second ecological stability threshold, it means that the population structure in the sub-area is unqualified, and a second unqualified mark is made.

[0026] Preferably, an environmental monitoring point is set in the sub-area that obtains the first qualified mark and the second qualified mark at the same time, and the environmental information and pollution information in the sub-area are collected to establish a first ecological environment data set;

[0027] The environmental information includes but is not limited to the following data: 7-15 days of average daily air temperature wd, average daily rainfall jyl and average daily sunshine time pjgz obtained through meteorological sensors;

[0028] And obtain the latitude and longitude coordinates of the industrial processing plant and the sub-area through the GPS receiver, and obtain the distance value jlz between the industrial processing plants near the sub-area;

[0029] The pollution information includes but is not limited to the following data: soil water content hsl, soil pH 1, soil heavy metal and organic matter total content hlz1, water body pH 2, water body heavy metal and organic matter total content hlz2, atmospheric pm2.5 concentration value pm1, air sulfur dioxide concentration value SO2 and air nitrogen monoxide concentration value CO;

[0030] The soil moisture content hsl is obtained through humidity sensor;

[0031] The soil pH value ph1 is measured and obtained by a soil pH sensor;

[0032] The total content of soil heavy metals and organic matter hlz1 was obtained by chemical analysis of soil samples;

[0033] The pH value of water is measured by a water quality pH sensor.

[0034] The total content of heavy metals and organic matter in water bodies, hlz2, is obtained through chemical analysis of water samples;

[0035] The atmospheric pm2.5 concentration value pm1 is obtained by measuring the atmospheric particle sensor;

[0036] The concentration of sulfur dioxide in the air (SO2) and the concentration of nitrogen monoxide in the air (CO) are measured and obtained through air quality sensors.

[0037] Preferably, after the pollution information is dimensionless processed, the soil pollution index trwr, the water pollution index swr and the air pollution index kqwr are calculated and obtained:

[0038]

[0039]

[0040] In the formula, represents the preset threshold value of soil water content, represents the preset threshold value of soil pH, represents the preset threshold value of the total content of heavy metals and organic matter in the soil that can be maximally tolerated, represents the preset threshold value of the total content of heavy metals and organic matter in the water body that can be maximally tolerated, represents the preset threshold value of water body pH, represents the preset threshold value of atmospheric pm2.5 concentration; represents the preset threshold value of the maximum tolerable sulfur dioxide concentration in the air; represents the preset threshold value of the maximum tolerable nitrogen monoxide concentration in the air; w1, w2, w3, w4, w5, w6, w7 and w8 represent weight values, and their specific values are adjusted and set by the user, and 0 < w1 < 1, 0 < w2 < 1, 0 < w3 < 1, 0 < w4 < 1, 0 < w5 < 1, 0 < w6 < 1, 0 < w7 < 1, 0 < w8 < 1, w1 + w2 + w3 = 1; w4 + w5 = 1; w6 + w7 + w8 = 1; A1 represents the first constant correction coefficient, A2 represents the second constant correction coefficient, and A3 represents the third constant correction coefficient;

[0041] Then, the soil pollution index trwr, the water pollution index swr and the air pollution index kqwr are used to generate the pollution pressure index wryl through the following associated formula;

[0042] wryl = γ * trwr + δ * swr + α * kqwr

[0043] In the formula, 0 ≤ γ ≤ 1, 0 ≤ δ ≤ 1, and γ + δ + α = 1, and γ, δ and α are the weight values of the soil pollution index trwr, the water pollution index swr and the air pollution index kqwr, which are adjusted and set by the user.

[0044] Preferably, the daily average air temperature wd, the daily average rainfall jyl, the daily average sunshine duration pjgz, the distance value jlz between industrial processing plants near the sub-region and the pollution pressure index wryl in the first ecological environment dataset are extracted; after dimensionless processing, the environmental pressure index HJyl is calculated through the following formula,

[0045]

[0046] represents the preset average air temperature threshold value, represents the preset average rainfall threshold value, It represents the preset average daily light irradiation time threshold, ln2 represents the logarithmic operation with the natural number 2 as the base, d1, d2, d3, and d4 represent weight values, and 0 < d1 < 1, 0 < d2 < 1, 0 < d3 < 1, 0 < d4 < 1, and d1 + d2 + d3 + r4 = 1.0; A4 represents the fourth constant correction coefficient;

[0047] If the environmental pressure index HJyl is higher than the first pressure threshold, it indicates that the environmental pollution is in an unqualified state, and then a first warning instruction is sent to the outside;

[0048] When the environmental pressure index HJyl is less than or equal to the first pressure threshold, it indicates that the environmental pressure is qualified.

[0049] Preferably, according to the first warning instruction, a first repair strategy is generated: vegetation restoration, adding 30% soil repair additives, dredging water bodies, adding water body repair and purification additives, establishing air purification dust collectors and desulfurization and denitrification devices;

[0050] And supervise according to the industrial processing plants in the sub-region to limit and reduce industrial emissions,

[0051] For endangered plants, delimit protected areas, conduct special protection and management, and establish habitats.

[0052] Preferably, for the sub-regions that simultaneously obtain the first qualified mark and the second qualified mark, set up second monitoring points, collect and obtain the information and quantity data of harmful plants in the sub-regions, and establish a second alien invasion dataset; and construct an invasive species interference index Rqgr based on the second alien invasion dataset: The invasive species interference index Rqgr is generated by the following formula:

[0053]

[0054] In the formula, O i is the individual quantity of the i-th harmful plant population, R i is the distribution density of the i-th harmful plant population, m is the number of types of harmful plants in the sub-region, B i is the area covering the distribution range of the i-th plant population, Dist i is the distance value between the i-th harmful plant population and the endangered plants;

[0055] When the invasive species interference index Rqgr exceeds the safety threshold, it indicates that there is a threat risk from invasive species. If the threat risk exceeds the safety threshold, a second warning instruction is sent to the outside;

[0056] If the invasive species interference index Rqgr does not exceed the safety threshold, it indicates that there is no threat risk from invasive species, but it is necessary to strengthen the monitoring and control of potential invasive species.

[0057] Preferably, a second restoration strategy is generated based on the second early warning instruction, including: taking removal measures for discovered invasive species, including manual removal and biological control; establishing isolation zones around areas affected by invasive species and taking isolation measures to prevent the spread of invasive species; strengthening patrol monitoring to prevent illegal destruction and human interference.

[0058] (III) Beneficial effects

[0059] The present invention provides a method for monitoring and early warning of the distribution of wild plants based on image recognition. It has the following beneficial effects:

[0060] (1) This method for monitoring and early warning of the distribution of wild plants based on image recognition, through the convolutional neural network CNN model or deep learning model technology, establishes a plant recognition model, analyzes the image data set, can efficiently and accurately identify endangered plant species and their coverage, and calculate the distribution and coverage of endangered plants in each sub-area to obtain the coverage index fgl and the population complexity index fz. This method improves the accuracy and efficiency of monitoring and realizes real-time monitoring and management of endangered plants. By calculating the coverage index fgl and the population complexity index fz, the ecological status of each sub-area can be comprehensively evaluated, and marked in the three-dimensional ecological model to identify the protected area and the area with qualified diversified population structure. When the coverage index fgl exceeds the preset protection threshold, it means that the sub-area meets the ecological protection range; when the population complexity index fz exceeds the preset second ecological stability threshold, it means that the diversified population structure in the sub-area is qualified. This marking method can timely discover the threats of environmental pressure and invasive species to the ecosystem, provide a scientific basis, and formulate corresponding maintenance strategies.

[0061] (2) This method of monitoring and early warning of the distribution of wild plants based on image recognition can scientifically evaluate the ecological protection status of each sub-area by calculating the coverage index fgl and the population complexity index fz. Qualified and unqualified markings can be made in the three-dimensional ecological model to provide a scientific basis for ecological protection and restoration decisions. The calculation and analysis of the coverage index fgl and the population complexity index fz help to timely discover changes in the ecological environment. When the coverage index or population complexity index of a sub-area does not reach the preset threshold, it can be quickly marked as unqualified, providing early warning information to ensure the timely implementation of ecological protection measures and prevent further deterioration of the ecological environment. By accurately identifying endangered plants and harmful plants and marking them accordingly, it is helpful to strengthen ecological protection and management. In particular, targeted protection measures can be taken to prevent their further reduction or extinction, thereby promoting the stability and healthy development of the ecosystem.

[0062] (3) This method of monitoring and early warning of the distribution of wild plants based on image recognition can scientifically evaluate the environmental pressure conditions in the sub-region by calculating the environmental pressure index HJyl. The environmental pressure index HJyl takes into account multiple key environmental factors, such as air temperature, rainfall, light time, distance from industrial processing plants, and pollution pressure index, and provides a comprehensive environmental pressure assessment indicator. When the environmental pressure index HJyl is higher than the preset first pressure threshold, the system will automatically send the first early warning instruction, indicating that the environmental pollution condition is unqualified. This timely early warning mechanism helps to take countermeasures quickly to prevent environmental deterioration and ensure the timeliness and effectiveness of ecological protection. The calculation method of the pollution index and the environmental pressure index can be adjusted according to the environmental characteristics of the specific area and is suitable for different types of ecosystems and environmental conditions.

[0063] (4) This method of monitoring and early warning of the distribution of wild plants based on image recognition can effectively monitor and evaluate the impact of alien invasive species on the ecosystem by setting a second monitoring point and constructing the invasive species interference index Rqgr. Timely sending of the second early warning instruction helps to quickly respond to and control invasive species and protect the stability of local endangered plants and ecosystems. Strengthening the monitoring and control of potential invasive species will help prevent and mitigate ecological risks that may arise in the future. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 The figure is a schematic flow chart of the steps of the method for monitoring and early warning of the distribution of wild plants based on image recognition according to the present invention. DETAILED DESCRIPTION

[0065] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not 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.

[0066] Example 1

[0067] See also Figure 1 The present invention provides a method for monitoring and early warning of the distribution of wild plants based on image recognition, comprising the following steps:

[0068] Using GIS technology and remote sensing data, a three-dimensional ecological model is established based on geographical features and ecological environment information, and the monitoring area is divided into several sub-areas. Corresponding data collection points are set in each sub-area. Each sub-area is divided based on ecosystem types, including forests, grasslands and wetlands;

[0069] A multi-rotor drone carrying multispectral satellite imaging equipment was used to collect plant distribution images in each sub-region and establish an image dataset;

[0070] Using the convolutional neural network (CNN) model or deep learning model technology, a plant recognition model is established to analyze the image data set to identify endangered plant species and coverage rates. The distribution and coverage rate of endangered plants in each sub-area are calculated and analyzed to obtain the coverage rate index fgl and population complexity index fz:

[0071] When the coverage index fgl exceeds the preset protection threshold, it means that the sub-area meets the ecological protection scope and is marked as the first qualified area in the three-dimensional ecological model;

[0072] When the population complexity index fz exceeds the preset second ecological stability threshold, it means that the diversified population structure in the sub-area is qualified, and the second qualified mark is made in the three-dimensional ecological model;

[0073] An environmental monitoring point is set for the sub-area that obtains both the first qualified mark and the second qualified mark, and the environmental information and pollution information of the sub-area are collected and acquired to establish a first ecological environment data set; and an environmental pressure index HJyl is constructed based on the first ecological environment data set. If the environmental pressure index HJyl is higher than the first pressure threshold, a first early warning instruction is sent to the outside;

[0074] A second monitoring point is set for the sub-area that obtains both the first qualified mark and the second qualified mark, and the harmful plant information and quantity data in the sub-area are collected to establish a second alien invasion data set; and an invasive species interference index Rqgr is constructed based on the second alien invasion data set. If the invasive species interference index Rqgr exceeds a safety threshold, a second early warning instruction is sent to the outside, and a corresponding repair strategy is generated based on the first early warning instruction and the second early warning instruction.

[0075] In this embodiment, the distribution and coverage of endangered plants are important indicators for assessing the health of the ecosystem. Areas with high coverage usually have higher conservation value. Classifying and maintaining these areas can effectively protect the living environment of endangered plants and prevent them from further reduction or extinction. However, the existing technology has great limitations in identifying and analyzing the distribution of endangered plants, and cannot provide accurate coverage data in a timely manner, thereby affecting the implementation effect of protection measures.

[0076] This technical solution proposes an efficient and accurate monitoring method through a monitoring and early warning method for the distribution of wild plants based on image recognition. GIS technology and remote sensing data are used, combined with multi-rotor drones carrying multispectral satellite imaging equipment, to collect plant distribution images of each sub-area and establish an image data set. Through the convolutional neural network CNN model or deep learning model technology, a plant recognition model is established, and the image data set is analyzed, which can efficiently and accurately identify endangered plant species and their coverage, and calculate the distribution and coverage of endangered plants in each sub-area to obtain the coverage index fgl and the population complexity index fz. This method improves the accuracy and efficiency of monitoring and realizes real-time monitoring and management of endangered plants. By calculating the coverage index fgl and the population complexity index fz, the ecological status of each sub-area can be comprehensively evaluated, and marked in the three-dimensional ecological model to clarify the protected area and the area with qualified diversified population structure. When the coverage index fgl exceeds the preset protection threshold, it means that the sub-area meets the ecological protection range; when the population complexity index fz exceeds the preset second ecological stability threshold, it means that the diversified population structure in the sub-area is qualified. This labeling method can timely detect threats to ecosystems posed by environmental pressures and invasive species, provide a scientific basis, and formulate corresponding maintenance strategies.

[0077] This technical solution can solve the problems of low monitoring efficiency, low information utilization, and difficulty in real-time monitoring and management in existing technologies, providing strong technical support for ecological protection. Combined with the analysis of environmental information and pollution data, this solution can construct an environmental pressure index and an invasive species interference index. When these indexes exceed the safety threshold, early warning instructions are issued in a timely manner to ensure the health and stability of the ecosystem. Through the application of this system, endangered plants in the wild can be better protected and ecological balance can be maintained.

[0078] Example 2. This example is an explanation of Example 1. Specifically, the geographic information system GIS technology and remote sensing data are used to obtain the geographical features and ecological environment information of the monitoring area through high-resolution satellite images and drone aerial photography data; including terrain, vegetation coverage and hydrological characteristics; based on the acquired geographical and ecological data, a three-dimensional ecological model of the monitoring area is established; the three-dimensional ecological model includes the following contents: terrain elevation data DEM, vegetation type layer and hydrological layer; the vegetation type layer shows the spatial distribution of different vegetation types, and the hydrological layer shows the location and characteristics of rivers, lakes and wetlands;

[0079] According to the three-dimensional ecological model, the monitoring area is divided into several sub-areas, and multiple data collection points are set in each sub-area.

[0080] In this embodiment, through high-resolution satellite images and drone aerial data, the geographical features and ecological environment information of the monitoring area can be accurately obtained. Using these high-precision data, a detailed three-dimensional ecological model can be established to make the monitoring of the ecological conditions in the region more accurate and comprehensive. The three-dimensional ecological model contains terrain elevation data, vegetation type layers and hydrological layers, which can fully display the geographical and ecological characteristics of the monitoring area. The integration and display of these data can help ecologists and environmental protection workers to more comprehensively assess the health of the ecosystem and discover potential ecological problems. Multiple data collection points are set in each sub-area, and various types of environmental data can be efficiently collected. This systematic data collection method can improve the efficiency of data acquisition and ensure the comprehensiveness and timeliness of data. According to the three-dimensional ecological model, the monitoring area is divided into several sub-areas, and data collection points are set to make the monitoring of different ecosystems in the region more refined. This division method can be targeted at different ecosystem types (such as forests, grasslands, wetlands) for monitoring and management.

[0081] Example 3. This example is an explanation of Example 1. Specifically, the image data set acquired by the drone is spliced ​​into an image that completely covers the sub-area. After geometric correction and spectral correction are performed on the image data set, a plant recognition model is established. The image recognition task is processed by a convolutional neural network CNN. Endangered plant species, non-endangered plant species, hazardous plant species, plant health conditions, vegetation type characteristics, and coverage area characteristics in the sub-area are extracted, and the image data set is marked to mark the category and location information of each plant; the image data set is used as the first input into the plant recognition model to calculate the coverage index fgl and the population complexity index fz:

[0082] The coverage index fgl is generated by the following formula:

[0083]

[0084] In the formula, CR i The population proportion of the i-th endangered plant species, A total Expressed as the total area of ​​the sub-region, A i is the coverage area of ​​the i-th endangered plant species, w i is the weight of the i-th endangered plant species, based on the endangerment level and ecological importance, n is the number of endangered plant species in the sub-region, H is the uniformity index of plant cover, calculated by the Shannon index Shannonindex, ln is expressed as a logarithmic operation;

[0085] When the coverage index fgl exceeds the preset protection threshold, it means that the sub-area meets the ecological protection scope and is marked as the first qualified one in the three-dimensional ecological model; when the coverage index fgl does not exceed the preset protection threshold, it means that the coverage of the sub-area is unqualified and is marked as the first unqualified one.

[0086] Specifically, the population complexity index fz is generated by the following formula:

[0087]

[0088] Where N total is the total number of plant species in the sub-area, N i is the number of individuals of the i-th endangered plant species, w i is the weight of the i-th endangered plant species, based on its endangerment level and ecological importance; n is the number of endangered plant species in the sub-region, P i is the relative proportion of the i-th endangered plant species, that is, the ratio of the number of individuals to the total number. q is an adjustment parameter, ranging from 0 to 2. When q = 1, the M value is the weighted sum of the proportions of different populations. When q ≠ 1, the M value reflects the distribution difference of the population proportions.

[0089] When the population complexity index fz exceeds the preset second ecological stability threshold, it means that the diverse population structure in the sub-area is qualified, and a second qualified mark is made in the three-dimensional ecological model; when the population complexity index fz does not exceed the preset second ecological stability threshold, it means that the population structure in the sub-area is unqualified, and a second unqualified mark is made.

[0090] In this embodiment, by using the convolutional neural network CNN model to process the image recognition task, endangered plants, non-endangered plants and harmful plant species in the sub-area can be efficiently and accurately identified. This efficient recognition method reduces manual intervention and improves the efficiency and accuracy of data processing. After geometric correction and spectral correction are performed on the image data obtained by the drone, the accuracy and consistency of the data can be ensured. By extracting the plant species, health status, vegetation type characteristics and coverage area characteristics in the sub-area, a comprehensive monitoring of the ecological environment is achieved, which helps to understand and grasp the overall health status of the ecosystem. By calculating the coverage index fgl and the population complexity index fz, the ecological protection status of each sub-area can be scientifically evaluated. Qualified and unqualified markings can be made in the three-dimensional ecological model to provide a scientific basis for ecological protection and restoration decisions. The calculation and analysis of the coverage index fgl and the population complexity index fz help to timely discover changes in the ecological environment. When the coverage index or population complexity index of a sub-area does not reach the preset threshold, it can be quickly marked as unqualified, provide early warning information, ensure the timely implementation of ecological protection measures, and prevent further deterioration of the ecological environment. By accurately identifying endangered plants and harmful plants and marking them accordingly, it helps to strengthen ecological protection and management efforts. In particular, for the protection of endangered plants, targeted protection measures can be taken to prevent their further reduction or extinction and promote the stability and healthy development of the ecosystem.

[0091] When the population complexity index fz exceeds the preset threshold, it means that the diversified population structure in the sub-region is qualified, which helps to maintain the diversity and stability of the ecosystem. The diverse population structure not only improves the ecosystem's ability to resist risks, but also enhances its ability to self-repair and recover.

[0092] Embodiment 4: This embodiment is an explanation of Embodiment 1. Specifically, an environmental monitoring point is set in a sub-area where the first qualified mark and the second qualified mark are obtained at the same time, and environmental information and pollution information in the sub-area are collected to establish a first ecological environment data set.

[0093] The environmental information includes but is not limited to the following data: 7-15 days of average daily air temperature wd, average daily rainfall jyl and average daily sunshine time pjgz obtained through meteorological sensors;

[0094] And obtain the latitude and longitude coordinates of the industrial processing plant and the sub-area through the GPS receiver, and obtain the distance value jlz between the industrial processing plants near the sub-area;

[0095] The pollution information includes but is not limited to the following data: soil water content hsl, soil pH 1, soil heavy metal and organic matter total content hlz1, water body pH 2, water body heavy metal and organic matter total content hlz2, atmospheric pm2.5 concentration value pm1, air sulfur dioxide concentration value SO2 and air nitrogen monoxide concentration value CO;

[0096] The soil moisture content hsl is obtained through humidity sensor;

[0097] The soil pH value ph1 is measured and obtained by a soil pH sensor;

[0098] The total content of soil heavy metals and organic matter hlz1 was obtained by chemical analysis of soil samples;

[0099] The pH value of water is measured by a water quality pH sensor.

[0100] The total content of heavy metals and organic matter in water bodies, hlz2, is obtained through chemical analysis of water samples;

[0101] The atmospheric pm2.5 concentration value pm1 is obtained by measuring the atmospheric particle sensor;

[0102] The concentration of sulfur dioxide in the air (SO2) and the concentration of nitrogen monoxide in the air (CO) are measured and obtained through air quality sensors.

[0103] Specifically, after the pollution information is dimensionless, the soil pollution index trwr, water pollution index swr and air pollution index kqwr are calculated:

[0104]

[0105] In the formula, Indicates the preset threshold of soil moisture content. Indicates the preset threshold of soil pH. Indicates the preset maximum tolerance threshold for total content of heavy metals and organic matter in soil. Indicates the preset maximum tolerance threshold for total content of heavy metals and organic matter in water bodies. Indicates the preset threshold value of water pH. Indicates the preset atmospheric PM2.5 concentration threshold; Indicates the preset maximum tolerance threshold of air sulfur dioxide concentration; Represents the preset maximum tolerable air nitric oxide concentration threshold; w1, w2, w3, w4, w5, w6, w7, and w8 represent weight values, and their specific values are adjusted and set by the user, and 0 < w1 < 1, 0 < w2 < 1, 0 < w3 < 1, 0 < w4 < 1, 0 < w5 < 1, 0 < w6 < 1, 0 < w7 < 1, 0 < w8 < 1, w1 + w2 + w3 = 1; w4 + w5 = 1; w6 + w7 + w8 = 1; A1 represents the first constant correction coefficient, A2 represents the second constant correction coefficient, and A3 represents the third constant correction coefficient;

[0106] Then, generate the pollution pressure index wryl from the soil pollution index trwr, the water pollution index swr, and the air pollution index kqwr through the following associated formula;

[0107] wryl = γ * trwr + δ * swr + α * kqwr

[0108] In the formula, 0 ≤ γ ≤ 1, 0 ≤ δ ≤ 1, and γ + δ + α = 1. γ, δ, and α are the weight values of the soil pollution index trwr, the water pollution index swr, and the air pollution index kqwr, which are adjusted and set by the user.

[0109] Specifically, extract the daily average air temperature wd, the daily average rainfall jyl, the daily average sunshine duration pjgz, the distance value jlz between industrial processing plants near the sub-region, and the pollution pressure index wryl from the first ecological environment dataset; after dimensionless processing, calculate the environmental pressure index HJyl through the following formula,

[0110]

[0111] Represents the preset average air temperature threshold, Represents the preset average rainfall threshold, Represents the preset daily average sunshine duration threshold, ln2 represents the logarithmic operation with the natural number 2 as the base, d1, d2, d3, and d4 represent weight values, and 0 < d1 < 1, 0 < d2 < 1, 0 < d3 < 1, 0 < d4 < 1, and d1 + d2 + d3 + r4 = 1.0; A4 represents the fourth constant correction coefficient;

[0112] If the environmental pressure index HJyl is higher than the first pressure threshold, it indicates that the environmental pollution is in an unqualified state, and then send the first warning instruction to the outside;

[0113] When the environmental pressure index HJyl is less than or equal to the first pressure threshold, it indicates that the environmental pressure is qualified.

[0114] In this embodiment, by setting environmental monitoring points in the sub-area where the first qualified mark and the second qualified mark are obtained, the environmental information and pollution information in the area can be fully collected. This monitoring method covers a variety of environmental factors such as meteorological data, soil, water and air pollution, ensuring comprehensive monitoring of environmental conditions. The method of calculating the soil pollution index, water pollution index and air pollution index after dimensionless processing can effectively integrate a variety of environmental data and provide an intuitive and easy-to-understand assessment of the pollution status. These indices can be flexibly adjusted according to the specific conditions of different regions through weight adjustment and the application of correction coefficients, which improves the accuracy and applicability of data analysis.

[0115] By calculating the environmental pressure index HJyl, the environmental pressure status in the sub-region can be scientifically evaluated. The environmental pressure index takes into account multiple key environmental factors, such as air temperature, rainfall, sunshine time, distance from industrial processing plants and pollution pressure index, and provides a comprehensive environmental pressure assessment indicator. When the environmental pressure index HJyl is higher than the preset first pressure threshold, the system will automatically send the first warning instruction, indicating that the environmental pollution status is unqualified. This timely early warning mechanism helps to take countermeasures quickly to prevent environmental deterioration and ensure the timeliness and effectiveness of ecological protection. The calculation method of the pollution index and environmental pressure index can be adjusted according to the environmental characteristics of the specific area and is applicable to different types of ecosystems and environmental conditions. This method is not only applicable to natural ecosystems such as forests, grasslands, and wetlands, but also to environmental monitoring and management in urban and industrial areas. By calculating and analyzing the pollution indexes of soil, water and air respectively, targeted control measures can be taken for different pollution sources and pollution types. For example, through the use of specific soil remediation, cleaning of water pollutants and air purification devices, different types of environmental pollution can be effectively controlled and reduced.

[0116] Example 5, this example is an explanation of Example 4. Specifically, according to the first warning instruction, a first restoration strategy is generated: vegetation restoration, adding 30% soil restoration additives, desilting the water body, adding water body restoration and purification additives, and establishing an air purification dust collector and a desulfurization and denitrification device;

[0117] And regulate industrial processing plants according to sub-regions, limit and reduce industrial emissions,

[0118] Protected areas are designated for endangered plants, special protection and management are carried out, and habitats are established.

[0119] In this embodiment, the first restoration strategy generated covers comprehensive management measures for vegetation, soil, water and air. This multi-level ecological restoration method ensures the comprehensive restoration of the environment and helps to improve the overall ecological environment quality. The vegetation restoration strategy can promote the growth of plants in the sub-area, increase the green coverage area, and improve the ecological stability and biodiversity of the region. This not only helps to prevent soil erosion, but also improves air quality and enhances the beauty of the landscape. Adding 30% soil restoration additives can effectively improve soil structure and fertility, promote the reproduction of soil microorganisms, and improve soil health. The use of restoration additives can quickly and effectively reduce soil heavy metals and organic pollutants, improve soil water holding capacity and plant growth conditions. By desilting the water body and adding water restoration and purification additives, sediments and pollutants in the water body can be effectively removed, and the transparency and water quality of the water body can be improved. Purification additives help reduce harmful substances in the water body, improve the water ecological environment, and promote the healthy growth of aquatic plants and animals. The establishment of air purification dust collectors and desulfurization and denitrification devices can effectively reduce the concentration of harmful substances such as particulate matter, sulfur dioxide and nitrogen monoxide in the atmosphere, improve air quality, and reduce the impact on residents' health. The implementation of air purification measures will help improve the air environment in the sub-area and provide a fresh and healthy living environment.

[0120] Example 6: This example is an explanation of Example 1. Specifically, a second monitoring point is set for the sub-area that obtains the first qualified mark and the second qualified mark at the same time, and the harmful plant information and quantity data in the sub-area are collected to establish a second alien invasion data set; and an invasive species interference index Rqgr is constructed based on the second alien invasion data set: the invasive species interference index Rqgr is generated by the following formula:

[0121]

[0122] In the formula, O i is the number of individuals in the i-th harmful plant population, R i is the distribution density of the i-th harmful plant population, m is the number of harmful plant species in the sub-region, B i is the area of ​​the distribution range of the i-th plant population, Dist i is the distance value between the ith harmful plant population and the endangered plant;

[0123] When the invasive species interference index Rqgr exceeds the safety threshold, it means that the invasive species has a threat risk. If the threat risk exceeds the safety threshold, a second warning instruction will be sent to the outside.

[0124] If the invasive species interference index Rqgr does not exceed the safety threshold, it means that there is no threat risk from invasive species, but it is necessary to strengthen the monitoring and control of potential invasive species.

[0125] In this embodiment, by setting up a second monitoring point and constructing an invasive species interference index Rqgr, the impact of alien invasive species on the ecosystem can be effectively monitored and evaluated. Timely sending of the second early warning instruction helps to quickly respond to and control invasive species and protect the stability of local endangered plants and ecosystems. Strengthening the monitoring and control of potential invasive species helps prevent and mitigate ecological risks that may arise in the future.

[0126] Example 7. This example is an explanation of Example 6. Specifically, according to the second early warning instruction, a second restoration strategy is generated, including: taking removal measures for discovered invasive species, including manual removal and biological control; establishing isolation zones around areas affected by invasive species, and taking isolation measures to prevent the spread of invasive species; strengthening patrol monitoring to prevent illegal destruction and human interference.

[0127] In this embodiment, human forces are organized to physically remove, cut or burn the invasive species to ensure that their roots and propagation materials are completely removed. Introduce natural enemies or competitive species of invasive species, and use natural ecological relationships to inhibit the spread and growth of invasive species. Use appropriate biological agents, such as pathogens or insects, to biologically control invasive species. Establish isolation zones around areas invaded by invasive species to prevent the spread of invasive species through physical barriers, chemical barriers or plant isolation zones. Isolation zones can include protective nets, fences or plant native plants to form a natural protective barrier. Increase the frequency and intensity of patrols to ensure that newly emerging invasive species are discovered and dealt with in a timely manner. Use technical means such as drones and monitoring equipment to comprehensively monitor the dynamics of invasive species to ensure that there are no blind spots in monitoring. Strengthen law enforcement in the region to prevent illegal logging, grazing and other behaviors from causing damage to the ecosystem. Through publicity and education, raise public awareness of the hazards of invasive species and encourage communities to participate in the monitoring and prevention of invasive species.

[0128] Contain the spread: Directly reduce the number of invasive species through artificial removal and biological control to curb their further spread and reproduction.

[0129] Ecological protection: Establish isolation zones and strengthen patrol monitoring to effectively protect local endangered plants and ecosystems and prevent invasive species from causing long-term damage to local ecosystems.

[0130] Improve response capabilities: By strengthening monitoring and preventing human interference, we can improve early warning and response capabilities to invasive species, ensure timely measures and reduce ecological risks.

[0131] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for monitoring and early warning of plant distribution in the wild based on image recognition, characterized in that: The following steps are included: Using GIS technology and remote sensing data, a three-dimensional ecological model is established based on geographical features and ecological environment information, and the monitoring area is divided into several sub-areas. Corresponding data collection points are set in each sub-area. Each sub-area is divided based on ecosystem types, including forests, grasslands and wetlands; A multi-rotor drone carrying multispectral satellite imaging equipment was used to collect plant distribution images in each sub-region and establish an image dataset; Using the convolutional neural network (CNN) model or deep learning model technology, a plant recognition model is established to analyze the image data set to identify endangered plant species and coverage rates. The distribution and coverage rate of endangered plants in each sub-area are calculated and analyzed to obtain the coverage rate index fgl and population complexity index fz: When the coverage index fgl exceeds the preset protection threshold, it means that the sub-area meets the ecological protection scope and is marked as qualified in the three-dimensional ecological model; when the population complexity index fz exceeds the preset second ecological stability threshold, it means that the diversified population structure in the sub-area is qualified and is marked as qualified in the three-dimensional ecological model; An environmental monitoring point is set for the sub-area that obtains both the first qualified mark and the second qualified mark, and the environmental information and pollution information of the sub-area are collected and acquired to establish a first ecological environment data set; and an environmental pressure index HJyl is constructed based on the first ecological environment data set. If the environmental pressure index HJyl is higher than the first pressure threshold, a first early warning instruction is sent to the outside; A second monitoring point is set for the sub-area that obtains both the first qualified mark and the second qualified mark, and the harmful plant information and quantity data in the sub-area are collected to establish a second alien invasion data set; and an invasive species interference index Rqgr is constructed based on the second alien invasion data set. If the invasive species interference index Rqgr exceeds a safety threshold, a second early warning instruction is sent to the outside, and a corresponding repair strategy is generated based on the first early warning instruction and the second early warning instruction.

2. The method for monitoring and early warning of wild plant distribution based on image recognition according to claim 1, characterized in that: Using geographic information system (GIS) technology and remote sensing data, high-resolution satellite images and drone aerial photography data are used to obtain the geographical features and ecological environment information of the monitoring area, including topography, vegetation cover and hydrological characteristics. Based on the acquired geographical and ecological data, a three-dimensional ecological model of the monitoring area is established. The three-dimensional ecological model includes the following contents: Terrain elevation data DEM, vegetation type layer and hydrological layer; the vegetation type layer shows the spatial distribution of different vegetation types, and the hydrological layer shows the location and characteristics of rivers, lakes and wetlands; According to the three-dimensional ecological model, the monitoring area is divided into several sub-areas, and multiple data collection points are set in each sub-area.

3. The method for monitoring and early warning of wild plant distribution based on image recognition according to claim 1, characterized in that: The image datasets acquired by the drone are spliced ​​into images that completely cover the sub-area. After geometric and spectral corrections are performed on the image datasets, a plant recognition model is established. The image recognition task is processed by the convolutional neural network CNN. Endangered plant species, non-endangered plant species, hazardous plant species, plant health conditions, vegetation type characteristics, and coverage area characteristics in the sub-area are extracted, and the image datasets are marked to mark the category and location information of each plant. The image dataset is used as the first input into the plant recognition model to calculate the coverage index fgl and the population complexity index fz: The coverage index fgl is generated by the following formula: In the formula, CR i The population proportion of the i-th endangered plant species, A total Expressed as the total area of ​​the sub-region, A i is the coverage area of ​​the i-th endangered plant species, w i is the weight of the i-th endangered plant species, based on the endangerment level and ecological importance, n is the number of endangered plant species in the sub-region, H is the uniformity index of plant cover, calculated by the Shannon index Shannonindex, ln is expressed as a logarithmic operation; When the coverage index fgl exceeds the preset protection threshold, it means that the sub-area meets the ecological protection scope and is marked as the first qualified one in the three-dimensional ecological model; when the coverage index fgl does not exceed the preset protection threshold, it means that the coverage of the sub-area is unqualified and is marked as the first unqualified one.

4. The method for monitoring and early warning of wild plant distribution based on image recognition according to claim 3 is characterized in that: The population complexity index fz is generated by the following formula: Where N total is the total number of plant species in the sub-area, N i is the number of individuals of the i-th endangered plant species, w i is the weight of the i-th endangered plant species, based on its endangerment level and ecological importance; n is the number of endangered plant species in the sub-region, P i is the relative proportion of the i-th endangered plant species, that is, the ratio of the number of individuals to the total number. q is an adjustment parameter, ranging from 0 to 2. When q = 1, the M value is the weighted sum of the proportions of different populations. When q ≠ 1, the M value reflects the distribution difference of the population proportions. When the population complexity index fz exceeds the preset second ecological stability threshold, it means that the diversified population structure in the sub-area is qualified, and the second qualified mark is made in the three-dimensional ecological model; When the population complexity index fz does not exceed the preset second ecological stability threshold, it means that the population structure in the sub-area is unqualified and a second unqualified mark is performed.

5. The method for monitoring and early warning of wild plant distribution based on image recognition according to claim 1, characterized in that: Setting environmental monitoring points in the sub-areas that have obtained the first qualified mark and the second qualified mark at the same time, collecting environmental information and pollution information in the sub-areas, and establishing a first ecological environment data set; The environmental information includes but is not limited to the following data: 7-15 days of average daily air temperature wd, average daily rainfall jyl and average daily sunshine time pjgz obtained through meteorological sensors; And obtain the latitude and longitude coordinates of the industrial processing plant and the sub-area through the GPS receiver, and obtain the distance value jlz between the industrial processing plants near the sub-area; The pollution information includes but is not limited to the following data: soil water content hsl, soil pH 1, soil heavy metal and organic matter total content hlz1, water body pH 2, water body heavy metal and organic matter total content hlz2, atmospheric pm2.5 concentration value pm1, air sulfur dioxide concentration value SO2 and air nitrogen monoxide concentration value CO; The soil moisture content hsl is obtained through humidity sensor; The soil pH value ph1 is measured and obtained by a soil pH sensor; The total content of soil heavy metals and organic matter hlz1 was obtained by chemical analysis of soil samples; The pH value of water is measured by a water quality pH sensor. The total content of heavy metals and organic matter in water bodies, hlz2, is obtained through chemical analysis of water samples; The atmospheric pm2.5 concentration value pm1 is obtained by measuring the atmospheric particle sensor; The concentration of sulfur dioxide in the air (SO2) and the concentration of nitrogen monoxide in the air (CO) are measured and obtained through air quality sensors.

6. The method for monitoring and early warning of wild plant distribution based on image recognition according to claim 5, characterized in that: After the pollution information is dimensionless, the soil pollution index trwr, water pollution index swr and air pollution index kqwr are calculated: In the formula, represents the preset threshold value of soil water content, represents the preset threshold value of soil pH, represents the preset threshold value of the total content of heavy metals and organic matters in soil that can be maximally tolerated, represents the preset threshold value of the total content of heavy metals and organic matters in water body that can be maximally tolerated, represents the preset threshold value of water body pH, represents the preset threshold value of atmospheric pm2.5 concentration; represents the preset threshold value of the maximum tolerated sulfur dioxide concentration in air; represents the preset threshold value of the maximum tolerated nitrogen monoxide concentration in air; w1, w2, w3, w4, w5, w6, w7 and w8 represent weight values, and their specific values are adjusted and set by the user, and 0 < w1 < 1, 0 < w2 < 1, 0 < w3 < 1, 0 < w4 < 1, 0 < w5 < 1, 0 < w6 < 1, 0 < w7 < 1, 0 < w8 < 1, w1 + w2 + w3 = 1; w4 + w5 = 1; w6 + w7 + w8 = 1; A1 represents the first constant correction coefficient, A2 represents the second constant correction coefficient, and A3 represents the third constant correction coefficient; Then the soil pollution index trwr, water pollution index swr and air pollution index kqwr are related by the following formula to generate the pollution pressure index wryl; wryl=γ*trwr+δ*swr+α*kqwr Wherein, 0≤γ≤1, 0≤δ≤1, and γ+δ+α=1, γ, δ and α are the weight values ​​of the soil pollution index trwr, the water pollution index swr and the air pollution index kqwr, which are adjusted and set by the user.

7. The method for monitoring and early warning of wild plant distribution based on image recognition according to claim 1, characterized in that: The average daily air temperature wd, average daily rainfall jyl, average daily sunshine time pjgz, distance value jlz between industrial processing plants near the sub-region and pollution pressure index wryl are extracted from the first ecological environment data set; after dimensionless processing, the environmental pressure index HJyl is calculated by the following formula: represents the preset average air temperature threshold value, represents the preset average rainfall threshold value, represents the preset average daily sunlight irradiation time threshold value, ln2 represents the logarithmic operation with the natural number 2 as the base, d1, d2, d3, and d4 represent weight values, and 0 < d1 < 1, 0 < d2 < 1, 0 < d3 < 1, 0 < d4 < 1, and d1 + d2 + d3 + d4 = 1.0; A4 represents the fourth constant correction coefficient; If the environmental pressure index HJyl is higher than the first pressure threshold, it means that the environmental pollution is in an unqualified state, and a first early warning instruction is sent to the outside; When the environmental pressure index HJyl is less than or equal to the first pressure threshold, it indicates that the environmental pressure is qualified.

8. The method for monitoring and early warning of wild plant distribution based on image recognition according to claim 7, characterized in that: According to the first warning instruction, the first restoration strategy was generated: vegetation restoration, adding 30% soil restoration additives, desilting water bodies, adding water restoration and purification additives, and establishing air purification dust collectors and desulfurization and denitrification devices; And regulate industrial processing plants according to sub-regions, limit and reduce industrial emissions, Protected areas are designated for endangered plants, special protection and management are carried out, and habitats are established.

9. The method for monitoring and early warning of wild plant distribution based on image recognition according to claim 1, characterized in that: A second monitoring point is set for the sub-area that obtains the first qualified mark and the second qualified mark at the same time, and the harmful plant information and quantity data in the sub-area are collected to establish a second alien invasion data set; and an invasive species interference index Rqgr is constructed based on the second alien invasion data set: the invasive species interference index Rqgr is generated by the following formula: In the formula, O i is the number of individuals in the i-th harmful plant population, R i is the distribution density of the i-th harmful plant population, m is the number of harmful plant species in the sub-region, B i is the area of ​​the distribution range of the i-th plant population, Dist i is the distance value between the ith harmful plant population and the endangered plant; When the invasive species interference index Rqgr exceeds the safety threshold, it means that the invasive species has a threat risk. If the threat risk exceeds the safety threshold, a second warning instruction will be sent to the outside. If the invasive species interference index Rqgr does not exceed the safety threshold, it means that there is no threat risk from invasive species, but it is necessary to strengthen the monitoring and control of potential invasive species.

10. The method for monitoring and early warning of wild plant distribution based on image recognition according to claim 9, characterized in that: Based on the second early warning instruction, the second restoration strategy is generated, including: taking removal measures for the discovered invasive species, including manual removal and biological control; establishing isolation zones around areas affected by invasive species and taking isolation measures to prevent the spread of invasive species; strengthening patrol monitoring to prevent illegal destruction and human interference.

Images