Rapid risk grade analysis method for urban black and odorous water body

Abstract

The embodiment of the invention discloses a rapid risk level analysis method for urban black and odorous water, relates to the technical field of smart cities, and can improve the monitoring efficiency and accuracy of urban black and odorous water based on an unmanned aerial vehicle and reduce the long-term operation and maintenance cost. According to the method, analysis and screening are carried out by adopting the hyperspectral data instead of depending on a deep learning algorithm, so that dependence on high-computing-power resources is remarkably reduced, and the method is a solution with high cost effectiveness. The monitoring efficiency and accuracy are improved, and the long-term operation and maintenance cost is reduced. And moreover, the cost and time consumption depending on large-scale data annotation and complex model training is avoided, and the cost efficiency of the project is remarkably improved.

Description

Technical Field

[0001] The present invention relates to the field of smart city technology, and in particular to a method for quickly analyzing the risk level of black and odorous water bodies in cities and towns. Background Art

[0002] At present, river water environment governance PPP projects have developed to a certain scale. As of the end of 2019, the number of national ecological environment PPP projects in the database reached 3,196, with a total investment scale of nearly 2 trillion yuan. Among them, the comprehensive governance project has the largest investment scale, up to 1,028.869 billion yuan, accounting for 52.22% of the total investment scale of the projects in the database. Comprehensive governance projects usually package multiple types of construction content such as river regulation, new and rebuilt sewage treatment, and ecological restoration. With the continuous deepening of the battle against pollution and the central ecological and environmental protection supervision work, and the precision and refinement of the ecological environment PPP policy, it is expected that the national ecological environment PPP projects in the database will accelerate the implementation progress.

[0003] Among these, saving project costs has become a major direction that needs to be continuously studied and optimized, and the use of drones and real-time information and communication technology to monitor black and odorous water bodies is a relatively effective optimization direction. At present, dynamic remote sensing monitoring algorithms for black and odorous water treatment include but are not limited to remote sensing technology combined with machine learning algorithms, rule-based classification algorithms, etc. This type of data analysis technology based on deep learning algorithms has high requirements for computing resources and requires a large amount of sample data for manual labeling, which is not conducive to wide-area technology promotion. And this type of scheme that obtains and analyzes data through high-resolution satellites also has defects. Satellite data is difficult to meet the high spatial resolution required for dynamic monitoring, and is especially unsuitable for complex urban water systems. In complex urban water systems, especially in areas with many human buildings, when using satellite remote sensing data to extract water body information, the shadows of buildings will also cause obvious interference to the extraction of water body information.

[0004] Since drones can fly at lower altitudes, by deploying drones to obstructed areas above water bodies (such as under bridges, shadowed areas blocked by trees, etc.) to collect data, the accuracy and level of detail of black and odorous water identification can be greatly improved, making it possible to identify small-scale or early-stage black and odorous water bodies.

[0005] Therefore, how to improve the efficiency and accuracy of drone-based monitoring of urban black and odorous water bodies and reduce long-term operation and maintenance costs has become a topic that needs further research. Summary of the invention

[0006] The embodiments of the present invention provide a method for rapid risk level analysis of urban black and odorous water bodies, which can improve the efficiency and accuracy of monitoring urban black and odorous water bodies based on drones and reduce long-term operation and maintenance costs.

[0007] To achieve the above object, the embodiments of the present invention adopt the following technical solutions:

[0008] A method for rapid risk level analysis of urban black and odorous water bodies, characterized by comprising:

[0009] S1. Photograph the target water body in the target area by using a drone to obtain spectral data of the target water body;

[0010] S2, extracting spectral features of key bands from the spectral data;

[0011] S3, generating the spatial distribution of black and odorous water bodies according to the extracted spectral features;

[0012] S4. Analyze the risk level of black and odorous water bodies;

[0013] Among them, the landscape unit index factor of the town where the target area is located is obtained; the landscape unit corresponding to the target area is determined according to the landscape unit index factor, and the resistance / power value of each landscape unit is calculated through the minimum cumulative resistance model; the total resistance / power face value is obtained by using the resistance / power value of each landscape unit; the risk level and risk category of the black and odorous water body are determined according to the total resistance / power face value. If the positive driving effect of the landscape unit on the process of black and odorous water pollutants is greater than the negative blocking effect, the landscape unit is classified as a "source" landscape unit, otherwise it is a "sink" landscape unit.

[0014] S5. Upload the analysis results.

[0015] The research and development of the risk level rapid analysis scheme for urban black and odorous water bodies provided by the embodiment of the present invention is based on the actual needs of urban lifeline safety engineering construction in urban environments. Hyperspectral data is used for analysis and screening instead of relying on deep learning algorithms. This strategy significantly reduces the dependence on high computing resources and is a cost-effective solution. It not only improves the efficiency and accuracy of monitoring, but also reduces long-term operation and maintenance costs, making the technical support for environmental protection and water quality management more economical and sustainable. Compared with satellite data, the time arrangement of drone data collection is more flexible, and the route can be adjusted according to actual needs and weather conditions to adapt to specific survey areas and time requirements. This flexibility allows drones to more accurately capture short-term changes in black and odorous events in water bodies, including the formation and disappearance process of black and odorous phenomena, so as to more comprehensively understand their dynamic characteristics. In this scheme, the spectral monitoring method is used to avoid the cost and time consumption of relying on large-scale data annotation and complex model training, and significantly improves the cost efficiency of the project. This cost control strategy enables the project to maintain low operating costs while achieving high-precision monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments 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.

[0017] Figure 1a A schematic diagram of a method flow chart provided by an embodiment of the present invention;

[0018] Figure 1b , Figure 2 to Figure 5 A schematic diagram of a specific example provided for an embodiment of the present invention;

[0019] Figure 6 A schematic diagram of a verification test provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0020] In order to enable those skilled in the art to better understand the technical solution of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. The embodiments of the present invention will be described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements with the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be interpreted as limiting the present invention. It can be understood by those skilled in the art that, unless specifically stated, the singular forms "one", "one", "said" and "the" used herein may also include plural forms. It should be further understood that the term "including" used in the specification of the present invention refers to the presence of the features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or their groups. It should be understood that when we say that an element is "connected" or "coupled" to another element, it can be directly connected or coupled to other elements, or there may also be intermediate elements. In addition, the "connection" or "coupling" used here may include wireless connection or coupling. The term "and / or" used herein includes any unit and all combinations of one or more associated listed items. It will be understood by those skilled in the art that, unless otherwise defined, all terms (including technical terms and scientific terms) used herein have the same meaning as generally understood by those of ordinary skill in the art to which the present invention belongs. It should also be understood that terms such as those defined in general dictionaries should be understood to have meanings consistent with the meanings in the context of the prior art, and will not be interpreted with idealized or overly formal meanings unless defined as herein.

[0021] The embodiment of the present invention provides a method for quickly analyzing the risk level of urban black and odorous water bodies. Figure 1a As shown, including:

[0022] S1. Use a drone to photograph the target water body in the target area and obtain spectral data of the target water body.

[0023] Among them, the flight route of the drone is planned and the judgment points are marked. The planned flight route corresponds to at least two drones, and each flight collects spectral data of the water body at different zenith angles and azimuth angles. The flight route is designed according to the monitoring range to ensure that the entire water body range can be photographed by the spectral imager carried by the aircraft. In order to eliminate the uncertainty caused by single-angle data, multiple drones are used to collect spectral image data in batches at different zenith angles and azimuth angles. And the judgment points are set according to a certain radius according to the width of the river channel. The radius can be customized according to the need to obtain more accurate distribution of black and odorous water bodies. The zenith angles of each drone sampling water body include: 0° and 40°, and the azimuth angle includes: 135°. Specifically, the spectral data collected by the hyperspectral imager carried by the drone have multiple different angles. By adjusting the angle between the spectrometer cantilever and the water surface, the current scheme adopts two zenith angles of 0° (vertical positive), 40° (inclination angle) and azimuth angle 135°. This angle avoids the influence of direct sunlight, reduces the flare effect, and ensures stable data collection. Drones can be deployed at high frequencies to achieve dynamic monitoring of black and smelly water bodies.

[0024] S2. Extracting spectral features of key bands from the spectral data.

[0025] S3. Generate the spatial distribution of black and odorous water bodies according to the extracted spectral features.

[0026] The ground station marks the target areas that are judged as black and odorous water bodies on the map. Specifically, the ground station uses an automated process to determine whether each point is a black and odorous water body and automatically marks it on the map. The spatial interpolation algorithm is used to expand the classification results of these sampling points into a continuous spatial distribution map. For example Figure 1b shown.

[0027] S4. Analyze the risk level of black and smelly water bodies.

[0028] S5. Upload the analysis results.

[0029] Furthermore, it also includes:

[0030] In addition to the hyperspectral imager, the drone is also equipped with a high-performance airborne computing unit NavidaManifold2, which is used for real-time data noise reduction, radiation correction, geometric correction, and spectral data normalization, and the processed data is sent back to the city lifeline data center to reduce the workload of the ground station and improve monitoring efficiency. After obtaining the spectral data of the water body, the drone performs preprocessing through the airborne computing platform, which includes:

[0031] The collected spectral data is subjected to noise reduction processing by a median filter model, and the median filter model is: Indicates the wavelength λ i , reflectivity value after denoising, Rrs(λ i-k )、Rrs(λ i )、Rrs(λ i+k ) are the wavelength λ i-k , i , i+k The reflectance value at λ represents the wavelength, which is the value corresponding to a specific position (or frequency) in the spectrum, usually in nanometers (nm). i represents the index position of the wavelength. For example, if there are multiple discrete wavelength positions (such as 400nm, 450nm, 500nm, etc.), i can be used to represent the position number of these wavelengths, such as i = 1, 2, 3, ... corresponding to each specific wavelength λi, k represents the offset of the wavelength index position, or the distance of the neighborhood range. Its value follows the following principles:

[0032] When the noise condition is high, such as under complex lighting or bad weather conditions, the image noise is large, and the k value is appropriately increased (for example, k=2 or k=3) to enhance the smoothing effect.

[0033] When the noise condition is low, such as flying under clear light, the data clarity is high, and a smaller k value (such as k=1) can be selected.

[0034] Then the atmospheric correction is performed through the radiation transfer model, which is: represents the water reflectance after atmospheric correction, L path (λ) represents path radiation, which is used to reflect the additional reflection caused by atmospheric scattering, and T(λ) represents transmittance, which is used to reflect the transmittance of sunlight by the atmosphere;

[0035] Then, the UAV aerial images are geometrically corrected. The geometric correction methods include: Here x ′represents the pixel coordinates after geometric correction, x represents the pixel coordinates in the original image, h represents the flight altitude of the drone, h0 represents the reference ground altitude, and f represents the focal length of the camera.

[0036] Finally, the spectral data collected at different angles are normalized, and the normalization processing model is: x and y are the pixel positions in the original image, which are the horizontal and vertical coordinates respectively. ′ and ′ is the pixel position after normalization, a and b are the offset and scaling parameters of the normalization, indicating the pixel position in the image, c is the scaling factor of the ordinate y, used to normalize the y coordinate to the applicable range of the model. In the preferred scheme, a=450, b=100, c=0.4. xy corresponds to the reflectance spectrum values ​​of different bands (such as blue, green, red, and near-infrared bands), 450: represents the wavelength of a reference benchmark (unit: nm), which is usually used for calibration of the blue band, 100: This value is used to scale the spectral wavelength range and adjust the difference between different bands to a comparable scale, 0.4: This is the coefficient used to standardize the reflectance value in the normalization process, which is a commonly used standardization coefficient. The pre-processed data is sent back to the urban lifeline data center to reduce the workload of the ground station and improve monitoring efficiency.

[0037] In this embodiment, S2 includes: determining the inter-band ratio model of the spectral data of the determination point of the target water body Rrs(λ1) and Rrs(λ2) are the water reflectance at wavelengths λ1 and λ2, respectively. A higher ratio, usually set to be greater than 1.2, indicates that there is a significant spectral difference between the two bands, which may be a reflection peak. A lower ratio, usually set to be less than 0.8, indicates that there is an absorption valley at that location.

[0038] The change of the spectral curve is captured by using the angle feature. The change model of the spectral curve is: Specifically, it can be determined by calculating the slope. For example, when the slope is greater than 0, it is regarded as a reflection peak, and when the slope is less than 0, it is regarded as an absorption valley.

[0039] By combining the above ratios and angle characteristics, the reflection peaks and absorption valleys of the water body can be effectively determined, and it can be further determined whether the water body at the determination point is black and smelly.

[0040] Specifically, the inter-band ratio model is used to calculate the reflectance ratio of water bodies at different wavelengths. According to the set threshold, when the ratio exceeds 1.2, it indicates that there is a significant reflection peak in the band, which may be related to the presence of phytoplankton or other pollutants in the water body; on the contrary, when the ratio is lower than 0.8, it indicates that there is an absorption valley, which usually indicates that the concentration of organic matter or pollutants in the water body is high. Then the spectral curve change model is applied to capture the changes in the spectral curve. By calculating the slope, if the slope is greater than 0, it indicates that there is a reflection peak; if the slope is less than 0, it indicates that there is an absorption valley. Combining the outputs of these two models, the characteristics of the judgment point can be effectively identified. If both are indicated as reflection peaks at the judgment point, it can be confirmed that the point may be a polluted water body. If both are absorption valleys, it is further confirmed that the water body may be rich in organic matter, showing a higher risk of pollution. Normal water bodies should have the following characteristics. First, in the inter-band ratio model, the reflectance ratio is usually between 1.0 and 1.2, showing a balance between reflection peaks and absorption valleys. This indicates that the concentration of phytoplankton and suspended solids in the water is moderate, and there is no obvious sign of pollution. In the spectral curve change model, the spectral curve of normal water shows a smooth and stable change, with a slope close to 0, reflecting no significant spectral fluctuations. This means that the environment of the water body is relatively stable and has not been polluted or disturbed by the outside world.

[0041] Furthermore, S2 also includes: using the output of the inter-band ratio model and the output of the spectral curve change model as conditions for determining the reflection peak and absorption valley, when the output values ​​of the two models meet the conditions of the reflection peak or absorption valley; according to the reflection peak and absorption valley information of the determination point of the target water body, determine whether the water body at the determination point is a black and smelly water body. For example: if the ratio and angle characteristics in a certain band meet the conditions of the reflection peak, the point is determined to be a reflection peak; if both meet the conditions of the absorption valley, it is determined to be an absorption valley. Through the obtained reflection peak and absorption valley information, it is further determined whether the water body at the determination point is a black and smelly water body.

[0042] In this embodiment, S4 includes: obtaining the landscape unit index factor of the town where the target area is located; determining the landscape unit corresponding to the target area according to the landscape unit index factor, and calculating the resistance / power value of each type of landscape unit through the minimum cumulative resistance model; wherein, for the judgment point detected as a black and odorous water body, the inverse distance weighted method is interpolated to draw the spatial distribution of the black and odorous water body in ArcGIS pro. The identified black and odorous water bodies are graded and classified by the minimum cumulative resistance model. The total resistance / power face value is obtained by using the resistance / power value of each type of landscape unit; according to the total resistance / power face value, the risk level and risk category of the black and odorous water body are determined. For example: based on multispectral data and the "source-sink" landscape theory, a minimum cumulative resistance model is constructed to evaluate the diffusion resistance and pollutant impact range of black and odorous water bodies. First, different landscape units in the urban scene are divided into "source" or "sink" units according to categories (such as farmland, residential land, forest land, etc.), and quantitative calculations are performed based on landscape resistance and power factors. Key indicators include vegetation coverage, surface roughness, rainfall, slope, etc. The model calculates the resistance and dynamic values, and finally classifies the regions according to pollution risks, and uses GIS tools to calculate the distribution ratio of each landscape unit in different risk areas. This method can realize dynamic monitoring of the risk of black and odorous water body diffusion and support more accurate pollution control measures.

[0043] In this example, if the positive driving effect of a landscape unit on the process of black and odorous water pollutants is greater than the negative blocking effect, the landscape unit is classified as a "source" landscape unit, otherwise it is a "sink" landscape unit. For example:

[0044] Based on multispectral data, a black and odorous water body diffusion resistance evaluation mechanism is developed. It can not only identify the impact of various infrastructure within the flow range of black and odorous water bodies on black and odorous water bodies, but also consider the dynamic impact of distance changes on the attenuation of pollutants in black and odorous water bodies, and monitor the change process of black and odorous water bodies in real time. At the same time, the impact range of pollutants is managed in a graded and classified manner to enhance the pertinence, effectiveness and timeliness of management measures.

[0045] Based on the design concept of "source-sink" landscape, in the process of non-point source pollution, if the positive driving effect of the landscape unit on the process of black and odorous water pollutants is greater than the negative blocking effect, then the landscape unit is classified as a "source" landscape unit, otherwise it is classified as a "sink" landscape unit.

[0046] It is generally believed that farmland, settlements, etc. are "source" landscape types, because these landscape types have more nutrient loss and pollutant output, which leads to the aggravation of black and odorous water pollution and eutrophication of lake water in the area; woodlands, grasslands, wetlands, etc. belong to "sink" landscape types, because these landscapes can effectively intercept and absorb soil erosion materials and pollutants, thereby reducing the risk of black and odorous water pollution.

[0047] First, Figure 2 The main landscapes that affect black and odorous water pollution are classified within the urban area shown. According to the survey results, they are roughly divided into several types of landscape units, including residential land, agricultural land, forest land, urban and construction land, wetlands and water bodies.

[0048] Secondly, according to the pollutant influencing factors, the indicators are divided into landscape resistance coefficient indicators, namely vegetation coverage (VFC), surface roughness (M), Manning roughness coefficient (N) and water system connectivity index (G), and landscape dynamic coefficient indicators, namely average annual rainfall (P), slope (S), height (h), soil erodibility (k) and runoff curve number (CN). Based on the minimum cumulative resistance model, a landscape resistance / dynamic cost model is constructed to calculate the landscape resistance / dynamic cost of the above-mentioned landscape units. If the landscape dynamic cost of the landscape unit is greater than the landscape resistance cost, it belongs to the "source" landscape unit, otherwise, it is a "sink" landscape unit. Specific judgment process for identifying "source-sink" landscape units based on landscape resistance / dynamics Figure 3 As shown, Figure 3 The LRC in it refers to the landscape resistance cost, and the LDC refers to the landscape dynamic cost.

[0049] 1) Step 1: Calculate Figure 4 , 5 The index factor values ​​of various landscape units in the urban scene shown

[0050] ① Vegetation coverage (VFC): an important indicator used to reflect basic surface conditions.

[0051] According to the formula

[0052] in

[0053] Where: NIR is the near infrared band, R is the red band

[0054]

[0055] Where NDVIsoil is the NDVI value of the pixel covered with pure soil; NDVIveg is the NDVI value of the pixel covered with pure vegetation.

[0056] The annual maximum NDVI cumulative frequency is taken as 5%, and the values ​​corresponding to 95% of the NDVI are used to calculate NDVIsoil and NDVIveg in the area.

[0057] ②Surface roughness (M) reflects the barrier effect of volatile surfaces on the formation of non-point source pollution.

[0058] Where a is the slope

[0059] ③ Manning roughness coefficient (N). The Manning roughness coefficient defines the degree of containment of each landscape unit against pollutants flowing into the unit, as shown in Table 1.

[0060] Table 1

[0061]

[0062] ④ The water system connectivity index (G) refers to the degree of connectivity between water system units (main stream, tributaries, etc.) in each basin. If the water system connectivity index is larger, the pollutant diffusion speed will increase, but the degree of pollutant concentration attenuation will increase.

[0063]

[0064] w r is the weight of each river connectivity index, x r It is the dimensionless river connectivity index value.

[0065] ⑤Surface elevation (H) reflects the influence of gravity on the movement of non-point source pollutants.

[0066] The greater the surface elevation, the greater the gravitational potential energy, the faster the movement of pollutants, and the higher the possibility of risks.

[0067] ⑥The average annual rainfall (P) reflects the potential impact of precipitation on the formation of non-point source pollution.

[0068] The daily precipitation used in this study comes from the China National Meteorological Data Sharing Center. January, February and December of each year are divided into winter, March to May are divided into spring, and June to August are divided into summer.

[0069] September to November is divided into autumn.

[0070] ⑦ Slope (S) is an accelerating factor driving surface runoff and soil erosion.

[0071] The calculation formula is S = 10.8sin(α) + 0.03 when α < 5°,

[0072] When 5°≤α<10°, S=16.8sin(α)–0.5,

[0073] When α≥10°, S=21.9sin(α)–0.96,

[0074] Where S is the slope factor and α is the slope value.

[0075] ⑧Soil erodibility (K) reflects the sensitivity of soil erosion.

[0076] Affected by the subtropical monsoon climate conditions and geographical environment, the soil type in urban scenes is mainly yellow-brown soil.

[0077] The town scene is located in the subtropical monsoon climate zone, and the vegetation belongs to the deciduous and evergreen broad-leaved mixed forest zone.

[0078] The main vegetation in the area where the town scene is located is artificial and natural vegetation, and the ground cover plants are relatively rich.

[0079] The original landform forest and grass coverage rate of the urban scene is about 15%, and the vegetation is mainly weeds.

[0080] This time, the modified geometric mean particle size model is used to estimate the K value of Yangzhou soil erodibility. The K value estimation formula in the geometric mean particle size is as follows:

[0081] K = 7.594 {0.0017 + 0.0494 ·

[0082]

[0083] D g =exp[0.01∑f i lnm i ] (2)

[0084] Where: K value unit is short ton ac h / (100 ft short ton ac in) in the US system; D g is the geometric mean particle size, mm; f i is the mass fraction of the i-th particle size grade, %; m i is the arithmetic mean of the two ends of the i-th particle size class, mm. When i = 1, d0 = 0. Substitute the above soil erodibility K value into the corrected K m expression.

[0085] K m =-0.00911+0.55066×0.1317K(3)

[0086] This modified formula K m The international unit is t·hm 2 ·h / (MJ·mm·lm 2 ), so in the calculation process, the K value in formula (1) should be multiplied by 0.1317 to convert it into international units and then substituted into formula (3).

[0087] According to the contents of the urban scene soil database and the requirements of relevant literature, a k value recommendation can be obtained.

[0088] ⑨The runoff curve number (CN) reflects the maximum possible surface runoff capacity of the landscape unit.

[0089] The CN value is a dimensionless parameter determined by factors such as antecedent soil moisture, soil type, vegetation cover, land use and slope. This study uses the SCS-CN model, which is based on the water balance equation and two basic assumptions.

[0090] The first assumption is that the ratio of the actual net surface flow (Q) of the catchment to the maximum possible flow in the basin is equal to the actual infiltration (F) divided by the potential storage capacity (S).

[0091] Another assumption is that the initial loss (I a ) is part of the potential storage capacity (S).

[0092] The formulas can be expressed as follows:

[0093] P=I a +F+Q,(1)

[0094]

[0095] I a =λ·S,(2)

[0096] Where: P is rainfall (mm); Q is runoff depth (mm); actual infiltration (F); I a is the initial loss, S is the potential water storage capacity, and λ is the initial loss coefficient.

[0097] From (1) and (2), we get

[0098]

[0099] The U.S. Soil Conservation Service obtained I a The empirical formula with S is

[0100] λ=0.2,

[0101] Therefore, there is

[0102]

[0103] The empirical conversion relationship between S and CN values ​​is as follows:

[0104] S=25400 / CN-25 (5)

[0105] Where:

[0106] The following formulas can be derived from (4) and (5):

[0107] S=5[P+2Q-(4Q 2 +5PQ) 1 / 2 )](6)

[0108]

[0109] Using known rainfall runoff measured data, the CN value of the urban scene can be calculated according to (6)(7).

[0110] 2) Step 2: Confirm the “source-sink” category of each landscape unit based on the indicator factor value of each landscape unit.

[0111] The resistance / dynamic value of various landscape units is calculated based on the minimum cumulative resistance model. The specific identification formula of the model is as follows:

[0112] Resistance base value:

[0113] MCRR=min∑[D ij ×(VFC i +M i +G i )×N i ]

[0114] Dij represents the distance from the i-th landscape unit to the corresponding exit j; VFCi, Mi, Ni, Gi represent the plant coverage, surface roughness, Manning roughness coefficient and water system connectivity of the i-th landscape unit.

[0115] Power base value:

[0116] MCRE=min∑[D ij ×(P i +S i +K i +H i )×CN i ]

[0117] Pi, Si, Ki, and Hi represent the average annual rainfall, slope, soil erodibility, and elevation of the i-th landscape unit.

[0118] In addition, min takes the minimum of the cumulative resistance and motivation costs from each unit to all watershed outlets.

[0119] Minimum landscape resistance and power were generated using the cost distance tool in ArcGIS 10.1.

[0120] ΔMCR=MCRE-MCRR

[0121] If the MCRE of a landscape unit is greater than MCRR, it is a “source” landscape, otherwise it is a “sink” landscape.

[0122] 3) The third step: Calculate the total resistance / dynamic face value of the “source-sink” landscape unit in the urban scene according to the dynamic / resistance model.

[0123]

[0124] Where MCR represents the minimum cumulative resistance value, that is, the total resistance / power value, dimensionless; f represents the negative correlation between the minimum cumulative resistance and the ecological process; min represents the minimum cumulative resistance value for different "sources"; D ij represents the spatial distance from source j to landscape unit i, m; R i It represents the resistance base of landscape unit i to the movement process, dimensionless; m is the number of resistance surface grids; n is the number of source areas.

[0125] 4) Step 4: Classify the total resistance value of the “source-sink” landscape unit in the urban scene

[0126] ① Calculate the standard deviation of the minimum cumulative resistance (MCR) of each landscape unit.

[0127] The total resistance surface value of each landscape unit is classified at intervals of half the variance, and the number of grids in each category and its proportion of the total number of grids, as well as the difference in proportion between the previous category and the next category, are counted.

[0128] For example, the number of A1 grids is 30,000, accounting for 39% of the total number of grids; the number of A2 grids is 7,693, accounting for 10.9% of the total number of grids. From A1 to A2, the change is 28.08%. Therefore, A1 and A2 are selected as mutation points. Subsequent results are similar.

[0129] ②The above mutation points were selected as the critical thresholds of the “source-sink” risk zones, and the “source-sink” risk zones were divided into four levels: major risk zone, larger risk zone, general risk zone, and low risk zone. The statistical function of ArcGIS software was used to calculate the area of ​​each risk zone and the area proportion of each risk zone for each land use type.

[0130] The risk impact of black and odorous water bodies should not only consider the impact of various landscape types in the area where the black and odorous water bodies flow through on the black and odorous water bodies themselves, but also consider the trend of pollutant attenuation due to changes in spatial distance when migrating from the "source" to other landscape units. The minimum cumulative resistance model can couple the resistance factor with the transport process of pollutants in black and odorous water bodies, and effectively identify the dynamic attenuation evolution process of non-point source pollution in the form of resistance cost. At the same time, from the perspective of the balance between landscape resistance and erosion, the impact of the "source-sink" landscape on the evolution of black and odorous water bodies is quantified. At present, empirical research has used the minimum cumulative resistance model to quantify and identify the impact range of black and odorous water bodies in urban scenarios.

[0131] Furthermore, this embodiment also includes: cross-validating the analysis results.

[0132] Among them, the analysis result is used as the judgment result, and the target water body is collected by a water quality analyzer to obtain the measured result, and the measured result is used to cross-validate the judgment result. For example: In order to verify whether the above method can accurately determine the distribution range of black and odorous water bodies, fixed-point sampling will be carried out on the waters defined as black and odorous water bodies to cross-validate the accuracy of the results and ensure that the differences are within an acceptable range. The sampling points of the verification experiment will be consistent with the original sampling points to reduce variable interference, ensure the uniformity of the comparison results, and facilitate verification. A multi-parameter water quality analyzer is used on site to perform in-situ measurements on surface water bodies (depth 0.5 meters). The measured physical and chemical indicators include water temperature (WT), pH value, dissolved oxygen (DO), turbidity (Tur), chlorophyll a (Chl-a), conductivity (EC), redox potential (ORP), and a Secchi disk is used to monitor transparency. At the same time, a 5-liter water sampler is used to collect surface (0.5 meters) water samples and bring them back to the laboratory for nutrient analysis. The nutrient parameters analyzed in the laboratory include total carbon (TC), total organic carbon (TOC), total nitrogen (TN), nitrite nitrogen (NO2 - -N), nitrate nitrogen (NO3 - -N), ammonia nitrogen (NH4 + -N), total phosphorus (TP), phosphate (PO4 3- -P). Total organic carbon (TOC) and total nitrogen (TN) were measured using a Shimadzu TOC-VCPH analyzer (Shimadzu, Japan). Total phosphorus (TP) was analyzed by spectrophotometry after digestion. Ammonia nitrogen (NH4 + -N), nitrite nitrogen (NO2 - -N), nitrate nitrogen (NO3 - -N) and phosphate (PO4 3- -P) was measured using a Lachat QC8500 flow injection analyzer (Lachat Instruments, USA). According to the four physical and chemical indicators (transparency, dissolved oxygen, redox potential and ammonia nitrogen) of each sampling area, if more than 60% of the data of one indicator or more than 30% of the data of no less than two indicators reach the "severe black and smelly" level, the detection area should be identified as "severe black and smelly", otherwise it should be identified as "mild black and smelly". The classification standards for the degree of pollution of urban black and smelly water bodies are shown in Table 2.

[0133] Table 2

[0134] Classification standard for pollution degree of urban black and odorous water bodies

[0135] Table Classification standard of urban black and smelly waterpollution degree

[0136]

[0137] Note: *When the water depth is less than 25cm, the indicator is taken as 40% of the water depth.

[0138] The data collected at the sampling points are processed and analyzed, and the water sample test results are compared with the results determined by the drone spectral data at the original sampling point. The data of important points are compared to determine their errors and coincidence rates. If the error is small or the coincidence rate is high, it can be determined that the drone monitoring method is feasible in identifying black and odorous water bodies.

[0139] Further, such as Figure 6 The experimental results shown in the figure, the cross-validation step verifies the accuracy of the drone monitoring data through the traditional black and odorous water sampling method, ensures that the monitoring method is reliable, and uses modern scientific and technological means to improve the efficiency of black and odorous water identification and monitoring E.

[0140]

[0141] Qo represents the observed value, Qm represents the simulated value, represents the simulation value at time t, represents the observed value at time t, represents the total average of the observations, and T represents the maximum time.

[0142] For the monitored black and odorous water bodies, feedback information will be given to the relevant departments or governance teams in real time. The classification of each black and odorous water body will be informed so as to initiate corresponding emergency governance measures to minimize the duration and scope of pollution.

[0143] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment. The above is only a specific implementation method of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or replacements that can be easily thought of by any technician familiar with the technical field within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A rapid risk level analysis method for urban black and odorous water bodies, characterized in that: include: S1. Photograph the target water body in the target area by using a drone to obtain spectral data of the target water body; S2, extracting spectral features of key bands from the spectral data; S3, generating the spatial distribution of black and odorous water bodies according to the extracted spectral features; S4. Analyze the risk level of black and odorous water bodies; S5. Upload the analysis results.

2. The method according to claim 1, characterized in that Also includes: The flight route of the UAV is planned and the decision points are marked, wherein the planned flight route corresponds to at least two flights of the UAV, and each flight collects spectral data of the water body at different zenith angles and azimuth angles.

3. The method according to claim 1, characterized in that The zenith angles of each drone when sampling water bodies include: 0° and 40°, and the azimuth angles include: 135°.

4. The method according to claim 1, characterized in that: After acquiring the spectral data of the water body, the drone performs preprocessing through the onboard computing platform, which includes: The collected spectral data is subjected to noise reduction processing by a median filter model, and the median filter model is: Indicates the wavelength λ i , reflectivity value after denoising, Rrs(λ i-k )、Rrs(λ i )、Rrs(λ i+k ) are the wavelength λ i-k , i , i+k The reflectance value at , λ represents the wavelength, is the value corresponding to a specific position (or frequency) in the spectrum, i represents the index position of the wavelength, wherein the value corresponding to each specific wavelength is λi, and k represents the offset of the wavelength index position, wherein when the weather is sunny, k=1; when the weather is not sunny, k=2 or k=3. Then the atmospheric correction is performed through the radiation transfer model, which is: represents the water reflectance after atmospheric correction, L path (λ) represents path radiation, which is used to reflect the additional reflection caused by atmospheric scattering, and T(λ) represents transmittance, which is used to reflect the transmittance of sunlight by the atmosphere; Perform geometric correction on the drone aerial image, the geometric correction includes: x′ represents the pixel coordinate after geometric correction, x represents the pixel coordinate in the original image, h represents the flight altitude of the drone, h0 represents the reference ground altitude, and f represents the focal length of the camera; The spectral data collected at different angles are normalized, and the normalization processing model is: x and y are the horizontal and vertical coordinates of the pixel position in the original image, x′ and y′ are the horizontal and vertical coordinates of the pixel position after normalization, a and b are the offset and scaling parameters of the normalization processing, indicating that the pixel position in the image is changed, and c is the scaling factor of the vertical coordinate y.

5. The method according to claim 1, characterized in that The key bands include at least green light, red light and near infrared light; the types of spectral features include inter-band ratios and spectral curves; S2 includes: determining the inter-band ratio model of the spectral data of the determination point of the target water body Rrs(λ1) and Rrs(λ2) are the water reflectance at wavelengths λ1 and λ2, respectively; The change of the spectral curve is captured by using the angle feature. The change model of the spectral curve is: Specifically, it can be determined by calculating the slope. For example, when the slope is greater than 0, it is considered as a reflection peak, and when the slope is less than 0, it is considered as an absorption valley. By combining the above ratios and angle characteristics, the reflection peaks and absorption valleys of the water body can be effectively determined, and it can be further determined whether the water body at the determination point is black and smelly.

6. The method according to claim 5, characterized in that S2 also includes: When R ratio When (λ1,λ2)>1.2, it indicates that there is a significant spectral difference between the two bands and they are determined to be reflection peaks; When the ratio R ratio When (λ1,λ2)<0.8, it is judged as an absorption valley.

7. The method according to claim 1, characterized in that S3 includes: The ground station will mark the target areas determined to be black and smelly water bodies on the map.

8. The method according to claim 1, characterized in that S4 include: Obtaining the landscape unit index factor of the town where the target area is located; Determine the landscape unit corresponding to the target area according to the landscape unit index factor, and calculate the resistance / power value of each landscape unit through the minimum cumulative resistance model; Using the resistance / power values ​​of various landscape units, the total resistance / power face value is obtained; The risk level and risk category of the black and odorous water body are determined based on the total resistance / power face value.

9. The method according to claim 8, characterized in that include: If the positive driving effect of a landscape unit on the process of black and odorous water pollutants is greater than the negative blocking effect, then the landscape unit is classified as a "source" landscape unit, otherwise it is a "sink" landscape unit.

10. The method according to claim 1, characterized in that Also includes: The analysis result is cross-validated, wherein the analysis result is used as a judgment result, and the target water body is collected by a water quality analyzer to obtain a measured result, and the measured result is used to cross-validate the judgment result.

Images