A method, system and device for estimating the suspended sediment concentration in a river

Abstract

The present invention discloses a method, system and device for estimating the suspended sediment concentration of a river, which relates to the field of retrieving the suspended sediment concentration of a river in an ultra-large value range, and includes: obtaining a measured data set of historical river suspended sediment concentration; obtaining the water body reflectance of images of multiple satellites; generating a matching sample pair data set and dividing it into training samples and validation samples; constructing a band combination mainly based on the required bands as the bands reflecting the characteristics of the suspended sediment concentration in high-concentration water bodies, and obtaining the sensitivity index RAN; constructing an inversion model using the training samples and the sensitivity index RAN; obtaining the river range; using a cloud computing platform to analyze the image sets of multiple satellites using the river range, the sensitivity index RAN and the inversion model to obtain the spatial distribution information of the river suspended sediment concentration; obtaining the spatial distribution information of the river suspended sediment concentration. The present invention can accurately evaluate the suspended sediment concentration of a river.

Description

Technical Field

[0001] The present invention relates to the field of inversion of river suspended sediment concentration in an ultra-large value range, and particularly to a method, system and device for estimating river suspended sediment concentration. Background Art

[0002] Suspended sediment concentration is an important hydrological parameter, which is of great significance for water conservancy project construction, soil and water loss control, water resource development and utilization, and hydrological forecasting. The river system is an important part of the earth's material and energy cycle, playing an important role in aspects such as water supply, shipping, and ecological value creation. Conducting river suspended sediment monitoring helps to evaluate river channel changes, river health status, and potential environmental hazards. Therefore, quickly and accurately evaluating river suspended sediment concentration and obtaining spatial distribution information play an important role in effectively strengthening river management, water resource utilization, and basin management.

[0003] In the prior art, remote sensing technology has become an effective means for quickly evaluating the suspended sediment concentration of water bodies, which can solve the problems of inconvenient field observation and difficult data acquisition for water body monitoring.

[0004] However, when studying river regions, the suspended sediment concentration in different regions of the river varies greatly, and remote sensing technology cannot accurately evaluate the river suspended sediment concentration. Summary of the Invention

[0005] Embodiments of the present invention provide a method, system and device for estimating river suspended sediment concentration, which can solve the problem in the prior art that when the suspended sediment concentration in different regions of the river varies greatly, the river suspended sediment concentration cannot be accurately evaluated.

[0006] An embodiment of the present invention provides a method for estimating the suspended sediment concentration of a river, including the following steps: obtaining a historical measured data set of the suspended sediment concentration of the river containing the collection date and spatial coordinates; screening out the river images corresponding to multiple spectral bands from the river images taken by multiple satellite remote sensors and constructing an image set; obtaining the water body reflectance spectral map information at the corresponding positions of all the images in the image set according to the spatial coordinates, and solving the multi-band water body reflectance according to the water body reflectance spectral map information; matching the measured data in the measured data set with the multi-band water body reflectance respectively according to the collection and shooting dates to obtain a data set of matching sample pairs; obtaining the water body reflectance corresponding to the selected spectral bands as the water body reflectance of the characteristic bands; obtaining the sensitivity index RAN according to the ratio of the water body reflectance of the characteristic bands to the water body reflectance corresponding to all spectral bands; constructing an inversion model by using the sensitivity index RAN and the data set of matching sample pairs through regression analysis; using the river range to determine the estimation area, screening the taken river images and constructing an image set, obtaining the water body reflectance spectral map information corresponding to all the images in the image set, obtaining the multi-band water body reflectance through the water body reflectance spectral map information, substituting the sensitivity index RAN obtained from the multi-band water body reflectance into the inversion model, obtaining the natural logarithm value of the river suspended sediment concentration, completing the estimation of the river suspended sediment concentration and obtaining the spatial distribution information of the river suspended sediment concentration.

[0007] Further, the steps of screening out the river images corresponding to multiple spectral bands from the river images taken by multiple satellite remote sensors and constructing an image set; obtaining the water body reflectance spectral map information at the corresponding positions of all the images in the image set according to the spatial coordinates, and solving the multi-band water body reflectance are specifically as follows: constructing an image set of Landsat-7 and Landsat-5 within the time range required for the research through a cloud computing platform, screening out a new image set with a cloud cover lower than 20%, and screening out bands B1, B2, B3, B4, and B5; obtaining the multi-band water body reflectance at the corresponding positions of all the images containing the shooting date in the new Landsat-7 and new Landsat-5 image sets based on the longitude and latitude positions of the spatial coordinates of the measured data set; constructing an image set with a cloud cover lower than 20% of Landsat-8 and Landsat-9 within the time range required for the research through a cloud computing platform, and screening out bands B2, B3, B4, B8, and B11 for Landsat-8 and Landsat-9; obtaining the multi-band water body reflectance at the corresponding positions of all the images containing the shooting date in the new Landsat-8 and new Landsat-9 image sets based on the longitude and latitude positions of the spatial coordinates of the measured data set.

[0008] Further, obtain the water reflectance corresponding to the selected spectral bands as the water reflectance of the characteristic bands; calculate the sensitivity index RAN according to the ratio of the water reflectance of the characteristic bands to the water reflectance of all spectral bands. The specific steps are as follows: randomly sample the matched sample pair dataset, which is divided into a 60% training sample pair dataset and a 40% validation sample pair dataset; screen the water reflectance of the red and near-infrared bands in the training sample pair dataset as the characteristic bands reflecting the suspended sediment concentration in highly turbid water bodies, and sum them to obtain a first result; sum the water reflectance of blue, green, red, and near-infrared light to obtain a second result; divide the first result by the second result to obtain the sensitivity index RAN for the suspended sediment concentration in rivers with an extremely large value range.

[0009] Further, the inversion model is:

[0010] Ln(SSC) = a + b × RAN

[0011] Where, a is the intercept constant, b is the slope constant, and Ln(SSC) is the natural logarithm value of the river suspended sediment concentration.

[0012] Further, the specific steps to determine the river range include: obtain the image brightness values of the 1st, 2nd, 3rd, 4th, and 5th bands of Landsat-5 and Landsat-7 image bands, as well as the image brightness values of the 2nd, 3rd, 4th, 8th, and 11th bands of Landsat-8 and Landsat-9 image bands, corresponding to the BLUE, GREEN, RED, NIR, and SWIR1 parameters respectively; substitute the image brightness values into the normalized difference water index mNDWI, normalized difference vegetation index NDVI, and enhanced vegetation index EVI. The formulas are as follows:

[0013] mNDWI = (GREEN – SWIR1) / (NIR + SWIR1)

[0014] NDVI = (NIR - RED) / (NIR + RED)

[0015] EVI = 2.5*(NIR – RED) / ((NIR + 6×RED – 7.5×BLUE) + 1)

[0016] Use the logical relationship ((mNDWI > NDVI or mNDWI > EVI) and (EVI < 0.1)) to extract the river range of the detection area; when the logical relationship is true, confirm the existence of the river range, when the logical relationship is false, the river range does not exist, and select the river boundary according to the true or false result to form the river range; and use the manual visual interpretation method to delete non-river trivial patches and connect river breaks.

[0017] Further, the method for determining the estimation area using the river range, screening the captured river images and constructing an image set, obtaining the water body reflection spectral map information corresponding to all the images in the image set, calculating the multi-band water body reflectance through the water body reflection spectral map information, substituting the sensitive index RAN obtained from the multi-band water body reflectance into the inversion model, obtaining the natural logarithm value of the river suspended sediment concentration, completing the estimation of the river suspended sediment concentration and obtaining the spatial distribution information of the river suspended sediment concentration, the specific steps are as follows: screening the image set within the required research time range; within the cloud computing platform, using the river range, the sensitive index RAN and the inversion model to respectively delimit the research area range, construct the sensitive index RAN, construct the inversion model according to the sensitive index, and download the river suspended sediment concentration data for the image sets of multiple satellites; according to the classification results in the spatial distribution information of the river suspended sediment concentration, different colors represent different suspended sediment concentrations, completing the estimation of the river suspended sediment concentration and obtaining the spatial distribution information of the river suspended sediment concentration.

[0018] An embodiment of the present invention provides an estimation system for river suspended sediment concentration, including:

[0019] A data module, configured to obtain a historical measured data set of river suspended sediment concentration including the collection date and spatial coordinates; a data sample module, configured to screen out the river images corresponding to multiple spectral bands according to the river images captured by multiple satellite remote sensing and construct an image set; obtain the water body reflection spectral map information corresponding to all the images in the image set according to the spatial coordinates, and solve the multi-band water body reflectance according to the water body reflection spectral map information; match the measured data in the measured data set and the multi-band water body reflectance according to the collection and shooting dates respectively to obtain a matching sample pair data set; an inversion model module, configured to obtain the water body reflectance corresponding to the screened spectral bands as the water body reflectance of the characteristic bands; calculate the sensitive index RAN according to the ratio of the water body reflectance of the characteristic bands to the water body reflectance corresponding to all the spectral bands; construct an inversion model by using the sensitive index RAN and the matching sample pair data set through regression analysis; an estimation module, configured to use the river range to determine the estimation area, screen the captured river images and construct an image set, obtain the water body reflection spectral map information corresponding to all the images in the image set, calculate the multi-band water body reflectance through the water body reflection spectral map information, substitute the sensitive index RAN obtained from the multi-band water body reflectance into the inversion model, obtain the natural logarithm value of the river suspended sediment concentration, complete the estimation of the river suspended sediment concentration and obtain the spatial distribution information of the river suspended sediment concentration.

[0020] An embodiment of the present invention provides an estimation device for river suspended sediment concentration, including: a memory and a processor; the memory stores a computer program, and when the processor executes the computer program, the above-mentioned method for estimating river suspended sediment concentration is implemented.

[0021] The embodiments of the present invention provide a method, system, and device for estimating the suspended sediment concentration in a river. Compared with the prior art, the beneficial effects are as follows:

[0022] Screen and construct an image set based on river images taken by multiple satellite remote sensing; obtain the water body reflectance spectral map information at the corresponding positions of all images in the image set according to the spatial coordinates, and solve the multi-band water body reflectance based on the water body reflectance spectral map information; match the measured data in the measured data set with the multi-band water body reflectance respectively according to the collection and shooting dates to obtain a data set of matching sample pairs; obtain the water body reflectance corresponding to the selected spectral band as the water body reflectance of the characteristic band; calculate the sensitive index RAN according to the ratio of the water body reflectance of the characteristic band to the water body reflectance of all bands; use the regression analysis method to construct an inversion model by using the sensitive index RAN and the data set of randomly sampled and screened matching sample pairs; finally, determine the estimation area using the river range, substitute the image sets of multiple satellites into the inversion model containing the sensitive index RAN, complete the estimation of the river suspended sediment concentration, and obtain the spatial distribution information of the river suspended sediment concentration.

[0023] Among them, the sensitive index RAN is obtained by analyzing the multi-band water body reflectance at the corresponding position of the spatial coordinates. The data set of matching sample pairs is obtained by matching the measured data and the multi-band water body reflectance respectively according to the collection and shooting dates. Through the regression analysis method, an inversion model is constructed by using RAN and the data set of randomly sampled and screened matching sample pairs. Finally, the result estimated by the inversion model is affected by the multi-band water body reflectance. Further, the multi-band water body reflectance is affected by the river suspended sediment concentration; the river suspended sediment concentrations at different positions in the river are different, resulting in different multi-band water body reflectances, and the situation of the river suspended sediment concentration can be accurately estimated. Description of the Drawings

[0024] Figure 1 It is a flowchart of a method, system, and device for estimating the suspended sediment concentration in a river provided by the embodiments of the present invention;

[0025] Figure 2 It is a high-quality image information diagram of a method, system, and device for estimating the suspended sediment concentration in a river provided by the embodiments of the present invention;

[0026] Figure 3 It is a model scatter diagram of a method, system, and device for estimating the suspended sediment concentration in a river provided by the embodiments of the present invention;

[0027] Figure 4 It is a river suspended sediment concentration inversion result diagram of the inversion model of a method, system, and device for estimating the suspended sediment concentration in a river provided by the embodiments of the present invention;

[0028] Figure 5The corresponding image display result map of water body extraction parameters for a method, system, and device for estimating river suspended sediment concentration provided by an embodiment of the present invention;

[0029] Figure 6 The river range map for a method, system, and device for estimating river suspended sediment concentration provided by an embodiment of the present invention;

[0030] Figure 7 The classification result map of different suspended sediment concentrations for a method, system, and device for estimating river suspended sediment concentration provided by an embodiment of the present invention. Detailed implementation manners

[0031] To make the above objects, features, and advantages of the present invention more obvious and understandable, the following detailed description of the specific implementation manners of the present invention is provided in conjunction with the accompanying drawings. Many specific details are set forth in the following description to fully understand the present invention. However, the present invention can be implemented in many other ways different from those described herein, and those skilled in the art can make similar improvements without departing from the connotation of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed below.

[0032] See Figures 1 to 7 , an embodiment of the present invention provides a method for estimating river suspended sediment concentration, including the following steps:

[0033] Step 1: Obtain field measurement data and historical river suspended sediment concentration data of hydrological stations: Use the suspended sediment concentration data recorded by hydrological stations in the detection area to construct a basic measured data set; record the spatial coordinates of the measured data.

[0034] Step 2: On the Google Earth Engine cloud platform, construct a Landsat-7 image set within the research time range (for example, the selected time range in Embodiment 1 is 1986 - 2022); then screen the new Landsat-7 image set with cloud cover less than 20%, and at the same time screen out the B1, B2, B3, B4, and B5 bands; obtain the multi-band water body reflectance at the corresponding positions of all images in the new Landsat-7 image set based on the longitude and latitude positions of the measured data in Step 1; continue to construct data sets with cloud cover less than 20% for Landsat-5, Landsat-8, and Landsat-9 in the above method, extract the water body reflectance of the B1, B2, B3, B4, and B5 bands for the Landsat-5 image set, and extract the water body reflectance of the B2, B3, B4, B8, and B11 bands for the Landsat-8 and Landsat-9 image sets.

[0035] Step 3: Match the measured suspended sediment concentration data in Step 1 and the multi-band water body reflectance data in Step 2 according to the date to obtain a dataset of matching sample pairs, and conduct random sampling, dividing them into 60% training samples and 40% validation samples.

[0036] Step 4: Construct characteristic bands using the water body reflectance in the red and near-infrared bands to represent the suspended sediment concentration in highly turbid water bodies. Sum them and divide by the sum of the water body reflectances in the blue, green, red, and near-infrared bands to obtain the river suspended sediment concentration sensitive index RAN with an extremely large value range; use the training samples in Step 3 and RAN to construct an inversion model for the river suspended sediment concentration with an extremely large value range.

[0037] Step 5: Use the validation samples in Step 3 to evaluate the accuracy of the model constructed in Step 4. The evaluation methods use the root mean square error (RMSE), mean absolute percentage error (MAPD), and the calculation formulas for the slope RMSE and MAPD are as follows:

[0038]

[0039]

[0040] In the formula and are the SSC inversion results and the actual SSC corresponding to the i-th point, respectively, with the unit of mg / L, n is the number of samples, and i represents the i-th sample pair.

[0041] Step 6: Use the RAN index in Step 4 for model transfer and correction: Screen cloud-free images of Landsat-7 and Landsat-5, Landsat-8 and Landsat-9 and construct a time window of ±1 day for image matching; extract the water body reflectance of each pixel of the river for the matching image pairs and calculate the RAN index; construct a linear transfer model to transfer the inversion model based on Landsat-7 to Landsat-5, Landsat-8, and Landsat-9 to obtain the long-term river suspended sediment concentration.

[0042] Step 7: Use mNDWI, NDVI, and EVI to extract the river boundary: Use the logical relationship ((mNDWI > NDVI or mNDWI > EVI) and (EVI < 0.1)) to extract the river range in the detection area, and use the manual visual interpretation method to delete non-river trivial patches and connect river breaks. The calculation formulas for mNDWI, NDVI, and EVI are as follows:

[0043] mNDWI = (GREEN – SWIR1) / (NIR + SWIR1)

[0044] NDVI = (NIR - RED) / (NIR + RED)

[0045] EVI = 2.5 * (NIR – RED) / ((NIR + 6 × RED – 7.5 × BLUE) + 1)

[0046] Wherein, BLUE, GREEN, RED, NIR, and SWIR1 respectively correspond to the image brightness values of the 1st, 2nd, 3rd, 4th, and 5th bands of the Landsat-5 and Landsat-7 image bands, and the image brightness values of the 2nd, 3rd, 4th, 8th, and 11th bands of the Landsat-8 and Landsat-9 image bands.

[0047] Step Eight: Use the river suspended sediment concentration inversion algorithm based on Landsat series sensors in Step Six and Step Seven and the river range in Step Eight to obtain the long-time series river suspended sediment concentration spatial distribution information;

[0048] For each scene of Landsat-5, Landsat-7, Landsat-8, and Landsat-9 images, operate using the above method for estimating river suspended sediment concentration based on remote sensing big data and cloud platform, and then the long-time series spatial variation distribution map of river suspended sediment concentration can be obtained, and further the spatio-temporal dynamic information of river suspended sediment can be mastered.

[0049] The embodiment of the present invention provides a system for estimating river suspended sediment concentration, including:

[0050] A data module for obtaining a historical measured dataset of river suspended sediment concentration containing the collection date and spatial coordinates; a data sample module for screening river images corresponding to multiple spectral bands from multiple satellite remote sensing images of the river and constructing an image set; obtaining the water body reflectance spectral map information at the corresponding positions of all the images in the image set according to the spatial coordinates, and solving the multi-band water body reflectance according to the water body reflectance spectral map information; matching the measured data in the measured dataset with the multi-band water body reflectance respectively according to the collection and shooting dates to obtain a matching sample pair dataset; an inversion model module for obtaining the water body reflectance corresponding to the selected spectral bands as the water body reflectance of the characteristic bands; calculating the sensitivity index RAN according to the ratio of the water body reflectance of the characteristic bands to the water body reflectance corresponding to all spectral bands; constructing an inversion model by using the sensitivity index RAN and the matching sample pair dataset through regression analysis; an estimation module for determining the estimation area using the river range, screening the river images taken and constructing an image set, obtaining the water body reflectance spectral map information corresponding to all the images in the image set, calculating the multi-band water body reflectance through the water body reflectance spectral map information, substituting the sensitivity index RAN obtained from the multi-band water body reflectance into the inversion model to obtain the natural logarithm value of the river suspended sediment concentration, completing the estimation of the river suspended sediment concentration and obtaining the spatial distribution information of the river suspended sediment concentration.

[0051] An embodiment of the present invention provides an estimation device for river suspended sediment concentration, including: a memory and a processor; the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0052] Obtaining a historical measured dataset of river suspended sediment concentration containing the collection date and spatial coordinates; screening river images corresponding to multiple spectral bands from multiple satellite remote sensing images of the river and constructing an image set; obtaining the water body reflectance spectral map information at the corresponding positions of all the images in the image set according to the spatial coordinates, and solving the multi-band water body reflectance according to the water body reflectance spectral map information; matching the measured data in the measured dataset with the multi-band water body reflectance respectively according to the collection and shooting dates to obtain a matching sample pair dataset; obtaining the water body reflectance corresponding to the selected spectral bands as the water body reflectance of the characteristic bands; calculating the sensitivity index RAN according to the ratio of the water body reflectance of the characteristic bands to the water body reflectance corresponding to all spectral bands; constructing an inversion model by using the sensitivity index RAN and the matching sample pair dataset through regression analysis; determining the estimation area using the river range, screening the river images taken and constructing an image set, obtaining the water body reflectance spectral map information corresponding to all the images in the image set, calculating the multi-band water body reflectance through the water body reflectance spectral map information, substituting the sensitivity index RAN obtained from the multi-band water body reflectance into the inversion model to obtain the natural logarithm value of the river suspended sediment concentration, completing the estimation of the river suspended sediment concentration and obtaining the spatial distribution information of the river suspended sediment concentration.

[0053] A specific embodiment is as follows:

[0054] Estimate the spatial distribution of river suspended sediment concentration for the main stream of the Yellow River in a rectangular area of the Loess Plateau in 2022. Specifically, it is described in detail in combination with Figure 1 the flowchart.

[0055] Step 1: Use the measured river suspended sediment concentration of the hydrological stations on the main stream of the Yellow River in the exploration area of the Loess Plateau to construct a basic measured data set, and record the corresponding dates and spatial coordinates of the measured data.

[0056] Step 2: On the Google Earth Engine cloud platform, construct a Landsat-7 image set screened by 20% cloud cover from 1986 to 2022, with a total of 307 scenes. Secondly, select bands B1, B2, B3, B4, and B5, and obtain the band water reflectance of the corresponding coordinates of the Landsat-7 image set based on the spatial location of the measured data obtained in Step 1. Then, construct the band water reflectance of the Landsat-5, Landsat-8, and Landsat-9 image sets according to the above method. For Landsat-8 and Landsat-9 images, select bands B2, B3, B4, B8, and B11, and the number of images is 265 scenes, 146 scenes, and 27 scenes respectively. Among them, the image information of the Landsat series sensors is as Figure 2 shown in the figure. In the figure, L5-Collection is the TM / Landsat-5 image, L7-Collection is the ETM+ / Landsat-7 image, L8-Collection is the OLI / Landsat-8 image, and L9-Collection is the OLI / Landsat-9 image.

[0057] Step 3: Match the measured suspended sediment concentration data in Step 1 and the band water reflectance in Step 2 according to the date to construct a matching sample pair data set, with a total of 644 pairs obtained. Randomly sample them into 451 pairs of training samples and 193 pairs of validation samples.

[0058] Step 4: Construct the characteristic bands of water reflectance in the red and near-infrared bands as the suspended sediment concentration of highly concentrated turbid water bodies. Sum them and divide by the sum of the water reflectances in the blue, green, red, and near-infrared bands to obtain the river suspended sediment concentration sensitivity index RAN with an extremely large value range. Use the training samples in Step 3 and the sensitivity index RAN to construct an inversion model for the river suspended sediment concentration with an extremely large value range. The scatter plot of the model is as Figure 3 shown in the figure; the model formula is as follows:

[0059] Ln(SSC) = a + b × RAN

[0060] Where a = -4.61025, b = 20.3024, Ln(SSC) is the natural logarithm of river suspended sediment concentration.

[0061] According to the scatter plot, it can be seen that the model fitting R2 reaches 0.86, and the fitting effect is good.

[0062] Step 5: Use the verification sample from step 3 to evaluate the accuracy of the inversion model constructed in step 4, such as Figure 4 The figure shows the inversion results of river suspended sediment concentration using the inversion model constructed in step 4; the evaluation method uses the root mean square error (RMSE), mean absolute percentage error (MAPD) ​​and slope RMSE and MAPD calculation formulas as follows:

[0063]

[0064]

[0065] In the formula and They are the SSC inversion result and actual SSC corresponding to the i-th point, respectively, in mg / L, n is the number of samples, and i represents the i-th sample pair.

[0066] According to the model accuracy evaluation results, it can be seen that the inversion model of river suspended sediment concentration in a super-large range constructed using the RAN index has a good effect, with R2=0.89, RMSE=1957.41 mg / L, MAPD=34.89% and a slope of 0.90.

[0067] Step 6: Use the RAN index of step 4 to correct the water reflectance of the Landsat-5, Landsat-8 and Landsat-9 image bands constructed in step 2: Use the Landsat-5, Landsat-8 and Landsat-9 image sets to screen cloud-free images through manual visual interpretation; obtain image pairs within a time window of ±1 day through date matching, and extract water reflectance pixel by pixel. Note that the water reflectance of each pixel is calculated using the ee.Kernel.square command to calculate the mean water reflectance within a 3×3 range; then, construct a linear migration model based on the scatter plots of water reflectance of Landsat-7 and Landsat-5, Landsat-8 and Landsat-9, and migrate the inversion model built based on Landsat-7 to Landsat-5, Landsat-8 and Landsat-9 to achieve the acquisition of long-term river suspended sediment concentration.

[0068] Step 7: Using the multi-band image set in Step 2, calculate three parameters: the modified normalized difference water index (mNDWI), the normalized difference vegetation index (NDVI), and the enhanced vegetation index (EVI). The images corresponding to each parameter are shown as Figure 5 shown.

[0069] The calculation formulas for the relevant indices are as follows:

[0070] mNDWI = (GREEN – SWIR1) / (NIR + SWIR1)

[0071] NDVI = (NIR - RED) / (NIR + RED)

[0072] EVI = 2.5 × (NIR – RED) / ((NIR + 6 × RED – 7.5 × BLUE) + 1)

[0073] In the formulas, BLUE, GREEN, RED, NIR, and SWIR1 respectively correspond to the image brightness values of the 1st, 2nd, 3rd, 4th, and 5th bands of the Landsat-5 and Landsat-7 image bands, as well as the 2nd, 3rd, 4th, 8th, and 11th bands of the Landsat-8 and Landsat-9 image bands.

[0074] Using the logical relationship ((mNDWI > NDVI or mNDWI > EVI) and (EVI < 0.1)), extract the river range of the detection area, and use the manual visual interpretation method to delete non-river trivial patches and connect river breaks. The river range is shown as Figure 6 shown.

[0075] Step 8: Using the river range in Step 7, estimate the river suspended sediment concentration in the detection area in 2022. Screen all the images in 2022 from the image set in Step 2 to construct an estimation image set, and perform the estimation of the suspended sediment concentration using the RAN index and the inversion model in Step 4 on all the images in the set on the Google Earth Engine cloud platform, then the spatial distribution information of the river suspended sediment can be obtained, that is, the estimation of the river suspended sediment concentration is completed. That is, construct an inversion model on the GEE platform, and the output result is the spatial distribution information of the river suspended sediment concentration. The classification result is shown as Figure 7 shown, where different colors in the figure represent different suspended sediment concentrations. From Figure 7 it can be seen that there are significant differences in the suspended sediment concentrations at different spatial positions of the river.

[0076] The above-described embodiments merely represent several implementation manners of the present invention. Their descriptions are relatively specific and detailed, but they should not be construed as limiting the scope of the invention patent. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present invention, several modifications and improvements can still be made, and these all fall within the protection scope of the present invention. Therefore, the protection scope of the present invention patent shall be subject to the appended claims.

Claims

1. A method for estimating suspended sediment concentration in a river, characterized in that: The following steps are involved: Obtain a historical river suspended sediment concentration measured data set containing the acquisition date and spatial coordinates; Based on river images taken by multiple satellite remote sensing, river images corresponding to multiple spectral bands are screened out and an image set is constructed; water body reflectance spectrum information at corresponding positions of all images in the image set is obtained according to spatial coordinates, and multi-band water body reflectance is solved based on the water body reflectance spectrum information; Match the measured data in the measured data set with the multi-band water body reflectance according to the collection and shooting dates, respectively, to obtain a matching sample pair data set; The water body reflectance corresponding to the screened spectral band is obtained as the water body reflectance of the characteristic band; the sensitivity index RAN is obtained according to the ratio of the water body reflectance of the characteristic band to the water body reflectance corresponding to all spectral bands; the inversion model is constructed by using the sensitivity index RAN and the matching sample data set through regression analysis; The estimation area is determined by using the river range, the captured river images are screened and an image set is constructed, the water body reflectance spectrum information corresponding to all the images in the image set is obtained, the multi-band water body reflectance is calculated through the water body reflectance spectrum information, the sensitivity index RAN obtained from the multi-band water body reflectance is substituted into the inversion model, the natural logarithm of the river suspended sediment concentration is obtained, the river suspended sediment concentration is estimated, and the spatial distribution information of the river suspended sediment concentration is obtained.

2. A method for estimating suspended sediment concentration in a river according to claim 1, characterized in that: The specific steps of selecting river images corresponding to multiple spectral bands and constructing an image set based on multiple river images taken by satellite remote sensing are as follows: obtaining water body reflection spectrum information at corresponding positions of all images in the image set based on spatial coordinates, and solving multi-band water body reflectance based on the water body reflection spectrum information. Through the cloud computing platform, we constructed Landsat-7 and Landsat-5 image sets within the time range required for the study, screened new image sets with cloud cover less than 20%, and selected the B1, B2, B3, B4, and B5 bands; Based on the spatial coordinate latitude and longitude positions of the measured data set, the multi-band water reflectance of the corresponding positions of all images with shooting dates in the new Landsat-7 and new Landsat-5 image sets is obtained; Through the cloud computing platform, we constructed a set of Landsat-8 and Landsat-9 images with cloud cover less than 20% within the time range required for the study, and screened the B2, B3, B4, B8 and B11 bands of Landsat-8 and Landsat-9; Based on the latitude and longitude of the spatial coordinates of the measured data set, the multi-band water reflectance of the corresponding positions of all images containing the shooting date in the new Landsat-8 and new Landsat-9 image sets is obtained.

3. A method for estimating suspended sediment concentration in a river according to claim 1, characterized in that: The water body reflectance corresponding to the screened spectral band is obtained as the water body reflectance of the characteristic band; the sensitivity index RAN is obtained according to the ratio of the water body reflectance of the characteristic band to the water body reflectance corresponding to all spectral bands. The specific steps are: Randomly sample the matching sample pair data set and divide it into 60% training sample pair data set and 40% validation sample pair data set; The water body reflectance in the red light and near infrared bands is selected in the training sample pair data set as the characteristic bands reflecting the suspended sediment concentration of high-concentration turbid water bodies, and they are summed to obtain the first result; The water body reflectances of blue light, green light, red light and near infrared light are summed to obtain a second result; Divide the first result by the second result to obtain the river suspended sediment concentration sensitivity index RAN in a very large range.

4. A method for estimating suspended sediment concentration in a river according to claim 1, characterized in that: The inversion model is: Where a is the intercept constant, b is the slope constant, and Ln(SSC) is the natural logarithm of river suspended sediment concentration.

5. A method for estimating suspended sediment concentration in a river according to claim 1, characterized in that: The specific steps to determine the extent of the river include: Get the image brightness values ​​of the 1st, 2nd, 3rd, 4th, and 5th bands of Landsat-5 and Landsat-7, and the image brightness values ​​of the 2nd, 3rd, 4th, 8th, and 11th bands of Landsat-8 and Landsat-9, corresponding to the BLUE, GREEN, RED, NIR, and SWIR1 parameters respectively; Substitute the image brightness value into the normalized water index mNDWI, normalized vegetation index NDVI and enhanced vegetation index EVI, the formula is: mNDWI=(GREEN–SWIR1) / (NIR+SWIR1) NDVI=(NIR-RED) / (NIR+RED) EVI=2.5*(NIR–RED) / ((NIR+6×RED–7.5×BLUE)+1) The logical relationship (mNDWI>NDVI or mNDWI>EVI) and (EVI<0.1) are used to extract the river range of the detection area; when the logical relationship is true, the river boundary is confirmed to exist, and when the logical relationship is false, the river boundary does not exist. The river boundary is selected and the river range is constructed according to the true and false results; and the manual visual interpretation method is used to delete non-river trivial patches and connect the river break-up points.

6. A method for estimating suspended sediment concentration in a river according to claim 1, characterized in that: The method uses the river range to determine the estimation area, screens the captured river images and constructs an image set, obtains the water body reflection spectrum information corresponding to all the images in the image set, obtains the multi-band water body reflectance through the water body reflection spectrum information, substitutes the sensitivity index RAN obtained from the multi-band water body reflectance into the inversion model, obtains the natural logarithm of the river suspended sediment concentration, completes the estimation of the river suspended sediment concentration and obtains the spatial distribution information of the river suspended sediment concentration, and the specific steps are as follows: Screening of imaging sets within the time frame of the desired study; In the cloud computing platform, river range, sensitivity index RAN and inversion model are used to define the study area, construct sensitivity index RAN, build inversion model based on sensitivity index and download river suspended sediment concentration data for multiple satellite image sets; According to the classification results in the spatial distribution information of river suspended sediment concentration, different colors represent different suspended sediment concentrations. The estimation of river suspended sediment concentration is completed and the spatial distribution information of river suspended sediment concentration is obtained.

7. A system for estimating suspended sediment concentration in a river, characterized in that: include: The data module is used to obtain the historical river suspended sediment concentration measured data set containing the acquisition date and spatial coordinates; The data sample module is used to screen river images corresponding to multiple spectral bands and construct an image set based on river images taken by multiple satellite remote sensing; obtain water body reflection spectrum information at corresponding positions of all images in the image set according to spatial coordinates, and solve multi-band water body reflectance according to the water body reflection spectrum information; match the measured data in the measured data set with the multi-band water body reflectance according to the collection and shooting dates, and obtain a matching sample pair data set; The inversion model module is used to obtain the water body reflectance corresponding to the screened spectral band as the water body reflectance of the characteristic band; the sensitivity index RAN is obtained according to the ratio of the water body reflectance of the characteristic band to the water body reflectance corresponding to all spectral bands; the inversion model is constructed by using the sensitivity index RAN and the matching sample data set through regression analysis; The estimation module is used to determine the estimation area using the river range, screen the captured river images and construct an image set, obtain the water body reflectance spectrum information corresponding to all the images in the image set, obtain the multi-band water body reflectance through the water body reflectance spectrum information, substitute the sensitivity index RAN obtained from the multi-band water body reflectance into the inversion model, obtain the natural logarithm of the river suspended sediment concentration, complete the estimation of the river suspended sediment concentration and obtain the spatial distribution information of the river suspended sediment concentration.

8. A device for estimating suspended sediment concentration in a river, comprising: Memory and processor; The memory stores a computer program, wherein the processor implements a method for estimating river suspended sediment concentration according to any one of claims 1 to 6 when executing the computer program.

Images