Multi-source remote sensing data-based method for monitoring concentration of suspended matter in ocean dumping area

Abstract

A multi-source remote sensing data-based method for monitoring the concentration of suspended matter in an ocean dumping area, belonging to the technical field of remote sensing and monitoring of the concentration of suspended matter in ocean dredging and dumping areas. The method comprises: 1) selecting a certain sea area comprising a dumping area as a research area to acquire actually measured suspended matter concentration data; 2) separately acquiring the actually measured suspended matter data, and preprocessing remote sensing images; 3) matching the acquired data to form a data set, constructing a suspended matter concentration inversion model, and obtaining an image of the spatiotemporal distribution of the concentration of suspended matter in a body of water of the research area; and 4) constructing a multi-source remote sensing spatiotemporal data fusion model to obtain a high-spatiotemporal-resolution diagram of the spatiotemporal distribution of the concentration of suspended matter in the dumping area, and monitoring the conditions of a change in the concentration of suspended matter in the ocean dumping area. The present invention has advantages such as a short monitoring period, high spatiotemporal resolution, and a wide observation range, and can synchronously monitor a sea area having a large range. The present invention can also provide data support for operations such as supervision and management of ocean waste dumping, monitoring and management of an ocean waste dumping area, selection and division of a waste dumping area, supervision of use, and monitoring for illegal dumping.

Description

A method for monitoring suspended matter concentration in marine dumping areas based on multi-source remote sensing data Technical Field

[0001] The present invention belongs to the technical field of remote sensing monitoring of suspended matter concentration in marine dredging and dumping areas, and involves remote sensing satellite image processing and measured data of suspended matter concentration at monitoring sites, remote sensing reflectivity data, a suspended matter concentration inversion method, and a spatiotemporal fusion algorithm for remote sensing images. Background Art

[0002] Monitoring water quality changes in marine dumping areas is an important basis for understanding the use of dumping areas and subsequent evaluations. Among them, suspended matter concentration monitoring is an important part of water quality monitoring. The traditional monitoring method is to extract water samples from specific monitoring sites and bring them back to the laboratory to measure the suspended matter concentration value in order to control the intensity of marine dumping activities. However, the monitoring method is inefficient and costly, and the limited data cannot fully summarize the marine conditions [Yu Chuanyang, Wang Lin, Zhao Sufang, et al. Remote sensing monitoring of suspended matter concentration in the adjacent sea areas of marine engineering projects based on multi-source satellite images [J]. Environmental Impact Assessment, 2023, 45(5): 17-21+52]. At present, the use of satellite remote sensing technology for suspended matter inversion monitoring has the advantages of short monitoring cycle, high temporal and spatial resolution, and wide observation range, which improves the time-consuming and labor-intensive situation of manual monitoring. The current suspended matter concentration inversion methods mainly include analytical method, semi-analytical method, empirical method, machine learning method, etc. Among them, the empirical method has a simple construction process, is easy to use, and has a wide application. It has been well applied in many sea areas.

[0003] However, for small areas like dumping sites, where water quality changes are complex over a short period, higher spatiotemporal resolution is required for remote sensing monitoring. Conventional satellite sensors struggle to achieve both high temporal and high spatial resolution. Currently, the method of spatiotemporal fusion of multi-source remote sensing data is used to address this issue, addressing the conflict between the temporal and spatial resolutions of single remote sensing data [WU M, WU C, HUANG W, et al. An improved high spatial and temporal data fusion approach for combining Landsat and MODIS data to generate daily synthetic Landsat imagery [J]. Information Fusion, 2016, 31: 14-25.]. This method primarily fuses high temporal, low spatial resolution remote sensing data with low temporal, high spatial resolution remote sensing data using appropriate algorithms to generate high temporal, high spatial resolution remote sensing images for better observation. This method has been widely used in numerical estimation, change monitoring, and land feature classification. However, research on spatiotemporal fusion monitoring of suspended matter concentration in marine dumping sites has yet to be found. The present invention proposes a technical solution for suspended matter concentration monitoring in this area.

[0004] The present invention utilizes remote sensing inversion monitoring technology and performs spatiotemporal data fusion based on multi-source satellite images, which to a certain extent addresses the shortcomings of traditional monitoring methods and provides auxiliary assistance for monitoring and management of marine dumping areas, selection and demarcation of dumping areas, use, and supervision of illegal dumping.

[0005] Summary of the Invention

[0006] In order to address the problem that existing remote sensing monitoring technology is limited by the spatiotemporal resolution of satellite sensors and cannot take into account all aspects at the same time, thus failing to effectively monitor small areas such as dumping areas with complex water quality changes in a short period of time, a spatiotemporal fusion algorithm is used to generate a spatiotemporal distribution map of suspended matter concentration in temporary marine dumping areas with higher spatiotemporal resolution, so as to realize hourly dynamic monitoring of suspended matter concentration in temporary marine dumping areas and analyze the spatiotemporal variation characteristics of suspended matter generated by dredging and dumping operations.

[0007] In order to achieve the above object, the technical solution adopted in the present invention is:

[0008] A method for monitoring suspended matter concentration in marine dumping areas based on multi-source remote sensing data fusion, the monitoring method comprising the following steps: Step 1: selecting a sea area including a certain dumping area as the study area, and obtaining more than 40 sets of measured suspended matter concentration data at different locations in the study area. Step 2: obtaining a pair of high spatial resolution (less than 50 meters) satellite images and multiple pairs of high temporal resolution (less than 1 hour) satellite images that are consistent with the acquisition time of the measured suspended matter data and cover the study area, and pre-processing the remote sensing images to obtain remote sensing reflectivity. Step 3: matching the data obtained in steps 1 and 2 according to time and location to form a data set. Step 4: constructing a suspended matter concentration inversion model based on the data set obtained in step 3, inputting the remote sensing image obtained in step 2 into the suspended matter concentration inversion model, and obtaining a spatiotemporal distribution image of the suspended matter concentration in the water body of the study area, including a pair of high spatial resolution (less than 50 meters) suspended matter concentration images and multiple pairs of high temporal resolution (less than 1 hour) suspended matter concentration images. Step 5: Use the suspended matter concentration image obtained in step 4 to build a multi-source remote sensing spatiotemporal data fusion model, input the high temporal resolution (less than 1 hour) spatiotemporal distribution image of suspended matter concentration into the fusion model, and obtain a high temporal resolution spatiotemporal distribution map of suspended matter concentration in the dumping area. Step 6: Monitor the changes in suspended matter concentration in the marine dumping area based on the spatiotemporal distribution map of suspended matter concentration in the dumping area generated in step 5. The specific steps are as follows:

[0009] Step 1: Select a sea area including a dumping area as the study area, and obtain more than 40 sets of measured suspended matter concentration data at different locations in the study area.

[0010] The measured suspended matter concentration data were obtained through water quality monitoring stations in the study area.

[0011] Step 2: Obtain one high spatial resolution (less than 50 meters) satellite image and multiple high temporal resolution (less than 1 hour) satellite images that coincide with the acquisition time of the measured suspended matter data and cover the study area, and preprocess the remote sensing images to obtain remote sensing reflectance.

[0012] Furthermore, the remote sensing images are from satellites with different temporal and spatial resolutions, such as GOCI and Landsat 8, and are available free of charge on the website. These remote sensing images are divided into a single low temporal, high spatial resolution remote sensing image and multiple high temporal, low spatial resolution remote sensing images. The high temporal, low spatial resolution remote sensing images include multiple remote sensing images with a temporal resolution of one hour or less, taken on the day of dumping operations at the dumping site.

[0013] Furthermore, the preprocessing process of the satellite image data is as follows: radiometric calibration, geometric correction, atmospheric correction, image cropping, and finally obtaining remote sensing reflectance data of different bands of satellite images at the water quality monitoring station location;

[0014] Step 3: Match the data obtained in steps 1 and 2 based on time and location to form a data set. That is, match the water sampling time and longitude and latitude of the water quality station corresponding to the measured water quality station with the shooting time and longitude in the satellite image to form a suspended matter concentration and remote sensing reflectance data set.

[0015] Step 4: Based on the data set obtained in step 3, the suspended matter concentration inversion method is used to input the remote sensing image obtained in step 2 to obtain the corresponding spatiotemporal distribution image of suspended matter concentration in water bodies for each remote sensing image. The spatiotemporal distribution image of suspended matter concentration in water bodies obtained by a low temporal and high spatial resolution remote sensing image is set as F, and the spatiotemporal distribution image of suspended matter concentration in water bodies obtained by multiple high temporal and low spatial resolution remote sensing images is set as C. i , i is an integer from 1 to n, n is the number of high temporal and low spatial resolution remote sensing images, and F is resampled to the same resolution as C1 and set to G. Specifically:

[0016] Based on the suspended matter concentration inversion method, a remote sensing inversion model of suspended matter concentration in the water body of the study area was constructed, and the spatiotemporal distribution image of suspended matter concentration in each remote sensing image of the study area was obtained by inversion.

[0017] The suspended matter concentration remote sensing inversion model is an exponential model: SPM = ae bx

[0018] Where SPM is the surface suspended matter concentration (g / m 3 ), a and b are fitting coefficients, and x is the ratio of the red band value to the green band value in the remote sensing reflectance of the dataset obtained in step 3.

[0019] Step 5: Construct a multi-source remote sensing spatiotemporal data fusion model based on the spatiotemporal distribution image of suspended matter concentration obtained in step 4. This patent adopts the FSDA spatiotemporal data fusion model. The FSDAF algorithm was proposed by Zhu et al. [ZHU X, HELMER EH, GAO F, et al. A flexible spatiotemporal method for fusing satellite images with different resolutions [J]. Remote Sensing of Environment, 2016, 172: 165-177] in 2016. The input data of the algorithm is a pair of low and high spatial resolution images obtained from different satellite images at a certain moment and the low spatial resolution image data at the prediction moment. The output data is the high-resolution remote sensing image at the prediction moment. This patent makes certain changes to the algorithm. The image pair of input data is F obtained in step 4 and G obtained by resampling F. The input low spatial resolution image data at another prediction moment is C.i , i is an integer from 1 to n, and the output data is C i The high temporal and spatial resolution image P of the suspended matter concentration temporal and spatial distribution at the corresponding moment i , i is an integer from 1 to n.

[0020] Step 6: Based on the high temporal and spatial resolution image P of suspended matter concentration generated in step 5 i , where i is an integer from 1 to n. This data is used to monitor changes in suspended solids concentration in marine dumping areas and provide auxiliary assistance for dumping supervision, thereby assisting management departments in their monitoring work. Specifically, the data is used to analyze the suspended solids concentration in the dumping area from the spatiotemporal distribution image and its concentration change trend.

[0021] The effects and benefits of the present invention are:

[0022] (1) The remote sensing technology used in the present invention has the advantages of short monitoring period, high spatiotemporal resolution, and wide observation range. It can synchronously monitor a large sea area with good instantaneity, overcoming the shortcomings of traditional monitoring methods. By constructing a multi-source remote sensing spatiotemporal data fusion model to solve the spatiotemporal contradiction problem of remote sensing images, the suspended matter concentration monitoring in temporary marine dumping areas becomes more refined.

[0023] (2) The present invention can provide auxiliary assistance for the supervision and management of marine dumping, the monitoring and management of marine dumping areas, the selection and use of dumping areas, and the supervision of illegal dumping, and can also assist management departments in conducting supervisory work. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] FIG1 is a flow chart of the method of the present invention.

[0025] Figure 2 is a schematic diagram of the study area and the location of the dumping area.

[0026] Figure 3 is a temporal and spatial distribution diagram of the suspended matter concentration in the dumping area; Figure 3(a) is a temporal and spatial distribution diagram of the suspended matter concentration in the dumping area at 8:15; Figure 3(b) is a temporal and spatial distribution diagram of the suspended matter concentration in the dumping area at 9:15; Figure 3(c) is a temporal and spatial distribution diagram of the suspended matter concentration in the dumping area at 10:15; Figure 3(d) is a temporal and spatial distribution diagram of the suspended matter concentration in the dumping area at 11:15; Figure 3(e) is a temporal and spatial distribution diagram of the suspended matter concentration in the dumping area at 12:15; Figure 3(f) is a temporal and spatial distribution diagram of the suspended matter concentration in the dumping area at 13:15. DETAILED DESCRIPTION

[0027] The present invention will be further described below with reference to the accompanying drawings.

[0028] Taking the remote sensing monitoring of suspended matter concentration in the temporary ocean dumping area of ​​dredged materials in Huanghua Port as an example, the implementation steps of this method are divided into the following four steps.

[0029] Step 1: Determine the study area and obtain the location information of the monitoring stations and the suspended matter data they monitor.

[0030] In this study, the Huanghua Port temporary ocean dumping area was selected as the study area. The study area is shown in Figure 1. V1 to V9 are nine water quality stations. Water samples were collected by ship, and surface water samples were brought back to the laboratory for suspended matter concentration measurement. A total of 41 water samples were collected and their suspended matter concentrations were measured. The dumping area is the area enclosed by the dotted line in Figure 1.

[0031] Step 2: Select appropriate satellite images according to the study area and perform data preprocessing.

[0032] The satellite remote sensing data used in this study are from the Geostationary Ocean Color Imager II (GOCIII) satellite in South Korea and Landsat 8, the eighth satellite in the U.S. Landsat program. Based on the time of field water quality sampling, this study selected 16 GOCIII remote sensing images and one Landsat 8 remote sensing image. The GOCIII image was obtained from the Korea Ocean Satellite Center (https: / / www.nosc.go.kr / ), and the Landsat 8 image was obtained from EarthExplorer (https: / / earthexplorer.usgs.gov). The images were processed using ENVI software for radiometric calibration, atmospheric correction, and geometric correction to obtain remote sensing reflectance. The remote sensing reflectance data were then matched with the corresponding water suspended matter concentration data, resulting in 41 corresponding remote sensing reflectance data sets for measured sites.

[0033] Step 3: Invert the suspended matter concentration based on the measured data and remote sensing reflectivity.

[0034] The 41 corresponding measured site and remote sensing reflectance data obtained above were randomly divided into 31 groups and 10 groups. The former served as the training set for model construction, and the latter was used to validate the inversion model. Pearson correlation analysis revealed the strength of the correlation between different bands or band combinations and suspended matter concentration. The correlation coefficients between remote sensing reflectance and suspended matter concentration in each band were obtained from the training set. The results showed that bands 555, 620, 660, 680, and 709 exhibited significant correlations with SPM. Therefore, these five bands can be used to establish the SPM inversion algorithm. The correlation coefficients between the five bands and the band ratios are shown in Table 1.

[0035] Table 1 Pearson correlation coefficients for each band and band combination

[0036] It can be seen that the SPM concentration is most significantly correlated with 680 / 555 and 660 / 555. The exponential model was established using the ratio bands 680 / 555 and 660 / 555 for simulation analysis, and the model equation and determination coefficient R were calculated. 2 , the root mean square error RMSE is shown in Table 2.

[0037] Table 2 Band / band ratio model expression

[0038] The 660 / 555nm band has the best correlation coefficient and the smallest absolute error. The corresponding ratio band index and polynomial model have the best results, which are verified using the test set.

[0039] This example uses the root mean square error (RMSE) and mean absolute percentage error (RE) to evaluate the inversion accuracy of the model. The calculation formula is as follows:

[0040] Where: x mod,i is the calculated value of the i-th measured site; x obs,i is the measured value of the i-th measurement station; n represents the number of measurement stations. Substituting 10 sets of validation data into the model for validation, the RMSE is 4.38 and the RE is 18.95%. The RE is less than 20%, indicating good inversion performance and can be well applied in the inversion of suspended matter concentration. Therefore, the inversion formula is: SPM = 2.7542e 3.2132x

[0041] Where SPM is the surface suspended matter concentration (g / m3), and x is the band ratio, i.e., the ratio of the red band at a wavelength of 660 to the green band at a wavelength of 555. This inversion model is used to input the preprocessed remote sensing image in step 2 to obtain the spatiotemporal distribution image of suspended matter concentration in each remote sensing image of the study area and crop the area retaining the dumping area. The spatiotemporal distribution image of suspended matter concentration obtained by Landsat 8 image inversion is set as F, which is resampled to a resolution of 250m and saved as G. The six images of the spatiotemporal distribution image of suspended matter concentration obtained by GOCI II image inversion from 8:15 to 13:15 on November 18, 2023 are denoted as Ci, where i is an integer from 1 to 8.

[0042] Step 4: Input F and G from step 3 to construct the FSDAF spatiotemporal fusion model. Input Ci obtained by inversion from GOCI II, where i is an integer from 1 to 8, to generate a high-spatiotemporal resolution spatiotemporal distribution map of suspended matter concentration in the dumping area from 8:15 to 13:15 Beijing time on November 18, 2023, as shown in Figure 3, where points A and B are the dumping points where the dredging vessel performs dumping operations in the dumping area.

[0043] Step 5: Using the high-resolution spatial and temporal distribution images of suspended sediment concentrations within the dumping area generated in Step 4, we monitored the changes in suspended sediment concentration within the marine dumping area and analyzed the dumping. Areas of high suspended sediment concentration, distinct from the surrounding background water, were observed within and near the dumping area. These clumps of water diffused and moved with time and current, consistent with the diffusion of dredged material after entry into the water. Figure 3a shows that after the New Manatee dredger began dumping at Point A within the dumping area at 6:30 AM, after 1 hour and 45 minutes of dilution, the average suspended sediment concentration within the dumping area was 36.92 mg / L at 8:15 AM, with a minimum of 22.65 mg / L, showing little difference from the surrounding background water. Figure 3b shows that after the New Manatee dredger began dumping at point B within the dumping area at 9:06 a.m., the average suspended sediment concentration in the dumping area increased significantly, reaching an average concentration of 50.27 mg / L, with a minimum value of 30.92 mg / L within the area, significantly different from the surrounding background water. After the dumping operation ceased, as shown in Figures 3c-f, the average suspended matter concentration in the area gradually decreased from 45.85 mg / L to 31.68 mg / L, with a rate of decrease of 4.72 mg / L per hour. The area with high concentrations of clumped suspended sediment spread to the southwest with the water flow, and the area gradually decreased. By 1:15 p.m., the minimum suspended sediment concentration in the dumping area dropped to 20.34 mg / L, which was not much different from the surrounding sea water. In summary, dumping operations in the dumping area can directly affect the suspended matter concentration in the area and the surrounding sea areas. The hourly suspended sediment map inverted based on FSDAF can better monitor the water quality of the sea area in a small area and further analyze the short-term changes in the sea area caused by dredging and dumping activities.

[0044] The above-described embodiments merely express the implementation methods of the present invention, but should not be understood as limiting the scope of the patent of the present invention. It should be pointed out that for those skilled in the art, several variations and improvements can be made without departing from the concept of the present invention, and these all fall within the scope of protection of the present invention.

Claims

1. A method for monitoring suspended matter concentration in ocean dumping areas based on multi-source remote sensing data, characterized in that: The monitoring method: Step 1: Select a sea area including a dumping site as the study area, and obtain more than 40 sets of measured suspended matter concentration data at different locations in the study area through water quality monitoring stations in the study area; Step 2: Obtain one high-spatial-resolution satellite image and multiple high-temporal-resolution satellite images that coincide with the acquisition time of the measured suspended matter data and cover the study area, and preprocess the remote sensing images to obtain remote sensing reflectance; Step 3: Match the data obtained in steps 1 and 2 based on time and location to form a data set. Specifically, match the water sampling time and longitude and latitude of the corresponding water quality station with the shooting time and longitude and latitude in the satellite image to form a suspended matter concentration and remote sensing reflectance data set; Step 4: Based on the data set obtained in step 3, a suspended matter concentration inversion model is constructed. The remote sensing image obtained in step 2 is input into the suspended matter concentration inversion model to obtain the spatiotemporal distribution image of the suspended matter concentration in the water body for each remote sensing image corresponding to the study area, including one high-spatial-resolution suspended matter concentration image and multiple high-temporal-resolution suspended matter concentration images. Step 5: Using the suspended matter concentration image obtained in step 4, a multi-source remote sensing spatiotemporal data fusion model is constructed, and the high-temporal-resolution spatiotemporal distribution image of suspended matter concentration is input into the fusion model to obtain a high-temporal-resolution spatiotemporal distribution map of suspended matter concentration in the dumping area; Step 6: Based on the spatiotemporal distribution map of suspended matter concentration in the dumping area generated in step 5, monitor the changes in suspended matter concentration in the marine dumping area.

2. A method for monitoring suspended matter concentration in an ocean dumping area based on multi-source remote sensing data according to claim 1, characterized in that: In the step 2, the preprocessing process of the satellite image data is: radiation calibration, geometric correction, atmospheric correction, image cropping, and finally obtaining the remote sensing reflectance data of different bands of the satellite image at the water quality monitoring station location.

3. A method for monitoring suspended matter concentration in an ocean dumping area based on multi-source remote sensing data according to claim 1, characterized in that: In the spatiotemporal distribution image of suspended solids concentration in water body in step 4, the spatiotemporal distribution image of suspended solids concentration in water body obtained by a low temporal and high spatial resolution remote sensing image is set as F, and the spatiotemporal distribution image of suspended solids concentration in water body obtained by multiple high temporal and low spatial resolution remote sensing images is set as C. i , i is an integer from 1 to n, n is the number of high temporal and low spatial resolution remote sensing images, and F is resampled to the same resolution as C1 and set to G; the specific process of step 4 is as follows: A remote sensing inversion model for suspended matter concentration in the study area was constructed to obtain the spatiotemporal distribution of suspended matter concentration in each remote sensing image of the study area. The suspended matter concentration remote sensing inversion model is an exponential model: SPM=yes bx Where SPM is the surface suspended matter concentration (g / m 3 ), a and b are fitting coefficients, and x is the ratio of the red band value to the green band value in the remote sensing reflectance of the dataset obtained in step 3.

4. A method for monitoring suspended matter concentration in an ocean dumping area based on multi-source remote sensing data according to claim 1, characterized in that: The step 5 is specifically as follows: adopting the FSDA spatiotemporal data fusion model, the input data image pair is F obtained in step 4 and G obtained by resampling F, and the input low spatial resolution image data at another prediction moment is C i , i is an integer from 1 to n, and the output data is C i The high temporal and spatial resolution image P of the suspended matter concentration temporal and spatial distribution at the corresponding moment i , i is an integer from 1 to n.

Images