Method for monitoring ground water level based on multispectral satellite

Abstract

The application discloses a kind of ground water level monitoring methods based on multispectral satellite, it is characterized in that, propose a kind of special water gauge, special water gauge is by red, white grid interval distribution composition, combined satellite remote sensing technology and projection amplification method, using near infrared, green band calculation and classification normalized water index, or combined near infrared, infrared and red band to identify water body.Under the satellite image resolution D2×D2, grid D1×D1 on water gauge can be identified and read, water gauge reading is amplified (H / L) times, water level in water tank is obtained by conversion, and then water area water level is obtained.The application can intuitively identify water body, read water gauge reading, convert the identification of satellite image to the identification of vertical size, and make the water level accuracy meet the use requirements, finally achieve the purpose of satellite intuitive monitoring water level.The application method is simple and effective, and has strong adaptability, and it is very easy to use satellite image to intuitively obtain water level data meeting accuracy requirements.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of hydrological measurement and water conservancy monitoring, and particularly relates to a ground water level monitoring method based on multispectral satellite. BACKGROUND

[0002] Water level refers to the elevation of the free water surface of a water body above a specified datum, and is the most basic observation item of hydrological observation, and has a very important position. On one hand, water level serves as a basic basis for engineering construction planning and design, and can provide water regime information for hydrological information forecasting and other work; on the other hand, water level can provide indirect use data for deriving other hydrological data. Traditional water level observation includes manual observation, self-recording water level meter recording, water level data coding storage, and water level automatic measurement and reporting system. In recent years, with the development of satellite remote sensing technology, the method of using satellite remote sensing images to infer and derive water level has gradually become a popular water level measurement means.

[0003] Traditional lake and river water level monitoring mainly uses hydrological stations to obtain water level information, which needs to consume a large amount of manpower, material resources and financial resources, and many lakes and rivers are located in remote areas, so it is difficult to set up observation points; in addition, the current hydrological station data sharing degree is low, and it is difficult to obtain. Compared with the traditional monitoring method, satellite altimetry has the characteristics of rapid and all-weather observation, and can monitor the changes of river and lake water level in the global range, especially the river and lake water level in the remote area without data. However, the existing satellite remote sensing method of inferring and deriving water level is not mature, and there is a lack of ground control data to verify the reliability of the inversion results.

[0004] On one hand, satellite images are easy to interpret the planar size, but it is difficult to distinguish the vertical size; on the other hand, the water level observation accuracy needs to meet the use requirements, and in general water level observation, the water level value should be accurate to 0.01m; in the observation of large floods, the water level value should be accurate to 0.1m, but the commonly used high-resolution satellite image is about 1m. Therefore, it is difficult to obtain the water level by directly reading the water gauge reading from the satellite image. In order to save the cost of manpower and material resources for building water level observation station, the satellite is used to obtain more intuitive and accurate water level measurement results, therefore, the present application provides a ground water level monitoring method based on multispectral satellite, which has important engineering significance in the field. SUMMARY

[0005] In view of the deficiencies of the prior art, the present application provides a simple, effective and intuitive and accurate ground water level monitoring method based on multispectral satellite.

[0006] In order to achieve the above object, the application provides a ground water level monitoring method based on a multi-spectral satellite, characterized in that: first, the characteristics of the multi-spectral satellite near-infrared band are used to identify water bodies, the normalized water index NDWI is calculated through the near-infrared (NIR) and green (G) band pixel brightness values, and classification is performed to identify water bodies, or a method of combining the near-infrared, infrared and red bands is used to identify water bodies. Then, a water tank is excavated near the water area where the water level needs to be measured, the water tank is connected to the water area through a channel, and a specially designed water gauge is arranged in the water tank and can be placed obliquely, or the water gauge is directly built in the river channel, the specially designed water gauge is composed of red and white grid intervals, the grids on the water gauge can be identified and read by using satellite images, the water level of the water tank is obtained through conversion, and then the water level of the water area is obtained, and the specific method is as follows:

[0007] Suppose that the total length of the water gauge is S, the water gauge is obliquely placed at the diagonal of the water tank, the included angle formed between the water gauge and the bottom edge of the water tank is alpha, the vertical total height of the water gauge is H, the plane projection length of the water gauge is L, the minimum scale of the water gauge is D1, that is, the area of each grid is D1x D1, and the resolution of the satellite image is D2x D2. In order to ensure that the water gauge can be clearly identified in the satellite image, the ratio of the minimum scale of the water gauge to the resolution of the satellite image is set as lambda = D1 / D2, and the total number of grids of the water gauge is N. Therefore, the plane projection length of the water gauge L = NxD1, tan alpha = H / L, and lambda = H / (tan alphaxD2xN), that is, when the resolution D2 of the satellite image is determined, lambda can be adjusted by adjusting the included angle alpha between the water gauge and the horizontal plane, and then the minimum scale D1 of the water gauge H / (tan alpha x N) is adjusted. When the water surface intersects with the water gauge, suppose that the water level is h, and the reading of the intersection of the water surface and the water gauge is x. The normalized water index NDWI of the water tank range is calculated by using the near-infrared band pixel brightness value p(NIR) and the green band pixel brightness value p(G), that is, NDWI = (p(G)-p(NIR)) / (p(G)+p(NIR)), and the grid number a of the water body in the satellite image in the water tank range is obtained by using the water body normalized water index threshold t, that is, the grid number n of the water gauge submerged below the water surface is a, that is, the grid number a of the water body in the satellite image in the water tank range is obtained by using the water body normalized water index threshold t, that is, the grid number n of the water gauge submerged below the water surface is a; or the near-infrared, infrared and red bands of the satellite are combined as R, G and B, the part below the intersection of the water surface and the water gauge is black, the grid number m of the water gauge not submerged above the water surface is identified, and the remaining submerged grid number n = N-m. Under the condition that the grid area of the water gauge is D1x D1, x = nxD1 can be taken, that is, the accuracy of x is D1. According to the proportional relationship, H / L = h / x, the river water level value h = x x (H / L) can be calculated. By using the projection method, the accuracy of the water level h is (H / L) times the accuracy of the water gauge reading x, the vertical total height H of the water gauge and the plane projection length L of the water gauge are adjusted, the horizontal size is projected into the vertical size under the condition that the resolution of the satellite image is D2x D2, and the water level reaching the observation accuracy requirement can be obtained.

[0008] According to the above derivation, on the one hand, the method uses the ratio of the water gauge plane projection length L and the water gauge vertical total height H related to the multiple of the accuracy of the post-water level measurement. Taking the satellite image resolution of 1m*1m as an example, the water gauge vertical total height H is 1m, the water gauge plane projection length L is 20m, tan alpha is 0.05, alpha is 0.050, the ratio of the water gauge minimum scale and the satellite image resolution is 2, the water gauge minimum scale D1 is 2m, and the accuracy of the reading x of the water surface intersecting the water gauge is 2m. The accuracy of the water level h is (H / L) times the accuracy of the water gauge reading x, that is, the accuracy of h is 0.05 times the accuracy of x, so the accuracy of h is 0.1m.

[0009] On the other hand, the method has two methods for identifying water bodies using satellite data processing. The number of grids n of the water gauge submerged can be directly obtained by calculating the normalized water index NDWI; or the number of grids m of the water gauge not submerged can be directly identified by combining R, G and B of the satellite near-infrared, infrared and red wave bands, and then the number of grids n of the water gauge submerged can be obtained. Taking the satellite image resolution of 1m*1m as an example, the water tank is excavated to arrange the lower water gauge, the water gauge vertical total height H is 1m, the water gauge plane projection length L is 20m, tan alpha is 0.05, alpha is 0.050, the ratio of the water gauge minimum scale and the satellite image resolution is 2, the water gauge minimum scale D1 is 2m, and the total number of grids of the water gauge is 10. The normalized water index NDWI of the water tank range is calculated, and if the normalized water index threshold t of the water body in the region is 0.2, the number of grids NDWI>t is 7. The number of grids m of the water gauge not submerged can be directly identified by combining R, G and B of the satellite near-infrared, infrared and red wave bands. The number of grids a of the water body in the satellite image in the water tank range is 7, that is, the number of grids n of the water gauge submerged below the water surface is a=7; or by the method of combining R, G and B of the satellite near-infrared, infrared and red wave bands, the number of submerged grids n=N-m=7 can be obtained.

[0010] The advantages and beneficial effects of the present application are as follows:

[0011] Traditional water level monitoring of lakes, rivers and the like mainly uses hydrological stations to obtain water level information, which requires a large amount of manpower, material resources and financial resources, and many lakes, rivers and the like are located in remote areas, so it is difficult to set up observation points; in addition, the current hydrological station data sharing degree is low, and it is difficult to obtain. In recent years, with the development of satellite remote sensing technology, the resolution of satellite images has gradually improved from the original hundred-meter level to 1 meter or even 0.5 meter, which provides technical support for using high-resolution satellite remote sensing images to retrieve water level, which can greatly reduce the cost of manual observation. However, satellite images are easy to interpret the plane size, but it is difficult to distinguish the vertical size. Therefore, the ground water level monitoring method based on multispectral satellite in the application designs a special water gauge, and combines satellite remote sensing technology and projection magnification method, uses the characteristics that the near-infrared band of multispectral satellite is easy to identify water body, calculates the normalized water index NDWI through the near-infrared (NIR) and green (G) band pixel brightness value and classifies to identify water body, or uses the method of combining near-infrared, infrared and red band to identify water body, to achieve the purpose of directly identifying water body and reading water gauge reading; the application uses the projection magnification method, under the condition that the resolution of satellite image is D2* D2, can adjust the water gauge and the horizontal angle α to adjust λ, and then adjust the minimum scale D1 of the water gauge = H / (tan α*N), which is magnified (H / L) times, not only converts the identification of satellite image to the plane size into the identification of the vertical size, but also makes the water level accuracy meet the use requirement, and finally achieves the purpose of directly monitoring the water level by satellite. The method is simple and effective, and has strong adaptability, which not only realizes the identification of satellite image to the vertical size, but also can use satellite image for remote observation, and it is very easy to directly obtain water level data meeting the accuracy requirement by satellite image, which greatly reduces the cost of manpower, material resources and financial resources. BRIEF DESCRIPTION OF DRAWINGS

[0012] Figure 1 It is a top view of the special water gauge in the application.

[0013] Figure 2 It is a front view of the special water gauge in the application.

[0014] Figure 3 It is a satellite image band combination schematic diagram in the application.

[0015] Figure 4 It is a schematic diagram of embodiment 1 in the application.

[0016] Figure 5 It is a water gauge site selection diagram of embodiment 2 in the application.

[0017] Figure 6 It is a front view of the water gauge design of embodiment 2 in the application.

[0018] Figure 7 It is a top view of the water gauge design of embodiment 2 in the application.

[0019] Figure 8 Figure 2 is a side view of the water gauge designed in Example 2 of the present application.

[0020] Figure 9 Figure 3 is a schematic diagram of satellite images in Example 2 of the present application.

[0021] In the figure: ① is a river channel, ② is a canal, ③ is a special water gauge, ④ is a water surface line, H is the total vertical height of the water gauge, S is the total length of the water gauge, a is the included angle between the water gauge and the bottom edge of the water tank, L is the length of the planar projection of the water gauge, D1 is the smallest scale of the water gauge, N is the total number of grids of the water gauge, h is the water level, x is the reading of the intersection of the water surface and the water gauge, m is the number of grids of the water gauge that are not submerged, and n is the number of grids of the water gauge that are submerged. DETAILED DESCRIPTION

[0022] The present application is further described below in conjunction with the accompanying drawings and specific examples.

[0023] Example 1

[0024] In general water level observation, the water level value should be accurate to 0.01 m; in large flood observation, the water level value should be accurate to 0.1 m. Taking water level observation in general cases as an example, in the case of satellite image resolution of 1 m x 1 m, the production scheme of the special water gauge is as follows: the total vertical height H of the water gauge is 1 m, the planar projection length L of the water gauge is 200 m, tan a = 0.005, a = 0.005, the total number of grids N of the water gauge is 100, then the ratio of the smallest scale of the water gauge to the satellite image resolution λ = 2, the smallest scale D1 of the water gauge is 2 m, and the accuracy of the reading x of the intersection of the water surface and the water gauge is 2 m. The accuracy of the water level h is (H / L) times the accuracy of the reading x of the water gauge, i.e. the accuracy of h is 0.005 times the accuracy of x, so the accuracy of h is 0.01 m, which meets the accuracy requirement of water level observation.

[0025] The normalized water index NDWI of the water tank range is calculated using the near-infrared (NIR) and green (G) band pixel brightness values, NDWI = (p(G) - p(NIR)) / (p(G) + p(NIR)), if the normalized water index threshold t of the water body in the region is 0.2, and the number of grids NDWI > t is 47, then the number of grids a of the water body in the satellite image in the water tank range can be obtained, a = 47. Or the water body can be identified by combining R, G, and B bands of satellite near-infrared, infrared, and red bands, and the number of grids a of the water body in the satellite image in the water tank range can be obtained according to the intersection of the water surface and the water gauge in Example 1. Figure 4 The submerged part of the water gauge is black, as shown in Figure 4The black part of the water gauge shows that the number of grids m of the water gauge not submerged is 53, and the number of grids n of the water gauge submerged is a = N-m = 47. According to the formula: x = n x D1, h = x x (H / L), the water level h = 0.47 m can be calculated.

[0026] Example 2

[0027] In general water level observation, the water level value should be accurate to 0.01 m; in large flood observation, the water level value should be accurate to 0.1 m. The Yulin store rubber dam in Linqu County of Weifang City in Shandong Province is selected as the observation object, and the water gauge is built near the dam bottom and the dam top, which is mainly used for monitoring the process of dam collapse and dam overflow. Therefore, the precision of water level monitoring near the dam top and the dam bottom should be improved as much as possible, and the water gauge site selection position is shown in Figure 5 . In the case of satellite image resolution of 1 m x 1 m, according to the monitoring needs, the ratio of the minimum scale of the water gauge to the satellite image resolution is λ = 2, three water gauges with different precision are designed by adjusting the angle α between the water gauge and the horizontal plane, and the production scheme of the special water gauge is as follows: five sections of water gauges are laid, one section is used for monitoring the dam bottom water level, which is recorded as water gauge A; one section is used for monitoring the dam lower water level, which is recorded as water gauge B; one section is used for monitoring the dam middle water level, which is recorded as water gauge C; one section is used for monitoring the dam upper water level, which is recorded as water gauge D; one section is used for monitoring the dam top water level, which is recorded as water gauge E; and the five sections of water gauges are connected. According to the actual height of the Yulin store rubber dam, the bottom end of the water gauge A is laid at-0.5 m, and the top end of the water gauge E is laid at 4.2 m, the design front view of the water gauge is shown in Figure 6 , the plan view is shown in Figure 7 , and the side view is shown in Figure 8 . The length L A of the water gauge A is 40 m, the height H A is 0.2 m, tanα A is 0.005, and α A is 0.005; the length L B of the water gauge B is 24 m, the height H B is 1.2 m, tanα B is 0.05, and α B is 0.050; the length L C of the water gauge C is 4 m, the height H C is 2 m, tanα C is 0.5, and α C is 0.464; the length L D of the water gauge D is 22 m, the height H D is 1.1 m, tanα D is 0.05, and α D is 0.050; the length L E of the water gauge E is 40 m, the height H E is 0.2 m, tanαE =0.005, α E =0.005; Number of grid cells N in water gauge A A =20, water gauge B grid number N B =12, water gauge C grid number N C =2, water gauge D grid number N D =11, water gauge E grid number N E =20, total grid number N=65, minimum scale of each water gauge D1=2m, accuracy of the reading x at the intersection of the water surface and the water gauge is 2m. Based on the fact that the accuracy of the water level h is (H / L) times the accuracy of the water gauge reading x, the measurement accuracy of water gauge A is 0.01m, water gauge B is 0.1m, water gauge C is 1m, water gauge D is 0.1m, and water gauge E is 0.01m. This meets the accuracy requirements for water level observation under general flood conditions and major flood conditions, and also conforms to actual engineering conditions, making it suitable for monitoring dam breaches and overtopping processes.

[0028] The normalized water index (NDWI) for the water tank area is calculated using near-infrared (NIR) and green (G) band pixel brightness values: NDWI = (p(G) - p(NIR)) / (p(G) + p(NIR)). If the NDWI threshold for water in this area is t = 0.3, and the number of grids with NDWI > t is 52, then the number of grids representing water in the satellite imagery within the water tank area is a = 52. Therefore, the number of grids not submerged in the water gauge is m = Na = 13. Alternatively, water bodies can be identified using a combination of R, G, and B bands from the satellite, as described in Example 2. Figure 9 At the point where the water level intersects with the water gauge, the submerged portion of the water gauge will appear to be the same color as the water. Figure 9 As shown in the submerged portion of the water gauge, the number of grid cells m in the unsubmerged section is 13. This indicates that water gauges A, B, C, and D are completely submerged, and water gauge E is partially submerged, posing a risk of overtopping. If the water only reaches water gauge A, the water level calculation formula is: h=(Nm)×D1 / λ×(H A / L A -0.5; If the water level reaches water gauge B, the water level calculation formula is: h = H A+ (NN A -m)×D1 / λ×(H B / L B -0.5; If the water level reaches water gauge C, the water level calculation formula is: h = H A +H B +(NN A -N B -m)×D1 / λ×(H C / L C -0.5; If the water level reaches the water gauge D, the water level calculation formula is: h = H A +HB +H C +(N D +N E -m)×D1 / λ×(H D / L D )-0.5; if the water is flooded to the water level E, the water level calculation formula is: h=H A +H B +H C +H D +(N E -m)×D1 / λ×(H E / L E )-0.5. Thus the water level h=4.08m can be calculated.

Claims

1. A method for monitoring ground water level based on multispectral satellite, characterized in that: Comprising the following steps: S1: The characteristics of water bodies are easily identified by using multispectral satellite near-infrared bands or by using a combination of near-infrared, infrared and red bands, and a certain characteristic value is calculated based on the characteristics of the satellite image bands to identify the water level, and the water level is obtained by classifying the water gauge reading; the method for calculating the characteristic value to identify the water gauge reading to obtain the water level includes NDWI, NDVI, MNDWI, NDBI and NDMI; S2: A water tank is excavated near the water area where the water level needs to be measured, the water tank is connected to the water area through a channel, and a specially designed water gauge is placed in the water tank at an angle; the characteristics of water bodies being easily identified as black by using multispectral satellite near-infrared bands are used to distinguish water bodies from other ground objects by using a combination of near-infrared, infrared and red bands in R, G and B, and the water level is obtained by identifying the water gauge reading: The specially designed water gauge is composed of two different color grids that are sensitive to satellite bands, and the grids on the water gauge can be identified and read from the satellite image, and the water level of the water tank is obtained by conversion, and then the water level of the water area is obtained; the color of the specially designed water gauge is composed of two different colors of red, white, green, blue or yellow; the specific calculation method is as follows: Assuming that the total length of the water gauge is S, it is placed at an angle at the corner of the water tank, forming an angle of a with the bottom edge of the water tank, the vertical total height of the water gauge is H, the plane projection length of the water gauge is L, the minimum scale of the water gauge is D1, that is, the area of each grid is D1x D1, and the resolution of the satellite image is D2x D2, in order to ensure that the water gauge can be clearly identified in the satellite image, the ratio of the minimum scale of the water gauge to the resolution of the satellite image is set as λ=D1 / D2, and the total number of grids of the water gauge is N; therefore, the plane projection length of the water gauge L=NxD1, tan a=H / L, and λ=H / (tan axD2x N), that is, when the resolution D2 of the satellite image is determined, λ can be adjusted by adjusting the angle a between the water gauge and the horizontal plane, and then the minimum scale D1 of the water gauge is adjusted as H / (tan a x N); When the water surface intersects with the water gauge, assuming that the water level is h, and the reading of the intersection of the water surface and the water gauge is x; the normalized water index NDWI of the water tank range is calculated by using the near-infrared band pixel brightness value p(NIR) and the green band pixel brightness value p(G), that is, NDWI=(p(G)-p(NIR)) / (p(G)+p(NIR)), and by using the threshold value t of the water body normalized water index, the number of grids a in the satellite image within the water tank range that is water body can be obtained, that is, the number of grids n of the water gauge submerged below the water surface is a; Or by combining the satellite near-infrared, infrared and red bands in R, G and B, the part below the intersection of the water surface and the water gauge is black, and the number of grids m of the water gauge above the water surface that is not submerged can be identified, and the remaining submerged grid number is n=N-m; under the condition that the grid area of the water gauge is D1x D1, x=nxD1 can be taken, and the accuracy of x is D1; according to the proportional relationship, H / L=h / x, the river water level value h=x x(H / L) can be calculated. With this projection method, the accuracy of water level h is (H / L) times the accuracy of the staff reading x. By adjusting the total vertical height H of the staff and the length L of the staff plane projection, the horizontal size is projected into the vertical size under the satellite image resolution of D2×D2, and the water level reaching the observation accuracy requirement is obtained.

2. The multispectral satellite-based ground water level monitoring method according to claim 1, characterized in that: Combined with the water tank and the special staff, the characteristics that the water body is easy to be recognized in the near-infrared band of the multispectral satellite are utilized, the normalized water index NDWI is calculated and classified in the near-infrared and green bands to recognize the water body, and then the staff reading is recognized to obtain the water level, or other characteristic values including NDVI, MNDWI, NDBI and NDMI are calculated and classified to recognize the water body.

Images