CN115358103A - Water environment detection method based on Sentinel-2 data - Google Patents
Water environment detection method based on Sentinel-2 data Download PDFInfo
- Publication number
- CN115358103A CN115358103A CN202211298732.4A CN202211298732A CN115358103A CN 115358103 A CN115358103 A CN 115358103A CN 202211298732 A CN202211298732 A CN 202211298732A CN 115358103 A CN115358103 A CN 115358103A
- Authority
- CN
- China
- Prior art keywords
- water quality
- data
- water
- model
- sentinel
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N33/00—Investigating or analysing materials by specific methods not covered by groups G01N1/00 - G01N31/00
- G01N33/18—Water
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10032—Satellite or aerial image; Remote sensing
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A20/00—Water conservation; Efficient water supply; Efficient water use
- Y02A20/152—Water filtration
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Data Mining & Analysis (AREA)
- Health & Medical Sciences (AREA)
- Chemical & Material Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Computational Mathematics (AREA)
- Pathology (AREA)
- General Health & Medical Sciences (AREA)
- Mathematical Physics (AREA)
- Pure & Applied Mathematics (AREA)
- Biochemistry (AREA)
- Immunology (AREA)
- Mathematical Optimization (AREA)
- Mathematical Analysis (AREA)
- Analytical Chemistry (AREA)
- Computer Hardware Design (AREA)
- Operations Research (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Food Science & Technology (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Evolutionary Biology (AREA)
- Geometry (AREA)
- Evolutionary Computation (AREA)
- Medicinal Chemistry (AREA)
- Probability & Statistics with Applications (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Algebra (AREA)
- Quality & Reliability (AREA)
- Databases & Information Systems (AREA)
- Software Systems (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The invention discloses a water environment detection method based on Sentinel-2 data, which relates to the field of remote sensing and comprises the following steps: acquiring a Sentinel-2 remote sensing image of an area to be monitored, preprocessing the image, and screening time phase data with high quality; collecting information of each water quality factor on the spot, judging the reliability of water quality and cleaning data; arranging the longitude and latitude of the data, and converting the two-dimensional data into space point location data with a space coordinate system; carrying out correlation analysis on the spectral reflectivity of each waveband of the remote sensing data and each water quality factor, carrying out correlation analysis between different waveband combinations and the actually measured water quality factor, and determining positive and negative correlation and correlation coefficient between the waveband and the actually measured water quality factor; establishing a semi-empirical model to invert water quality, and optimizing a water quality model through model precision verification; and (5) grading the water quality. The invention increases the diversity of the water environment detection method, and the obtained result is accurate and reliable.
Description
Technical Field
The invention relates to the technical field of remote sensing, in particular to a water environment detection method based on Sentinel-2 data.
Background
The water environment is a place on which the human society lives and develops, and is also the most serious area interfered and destroyed by human beings, the water environment detection is one of the main works in the environment monitoring work, can accurately, timely and comprehensively reflect the current situation and development trend of the water quality, and plays a vital role in the whole water environment protection, water pollution control and water environment health maintenance.
The traditional single monitoring mode for sampling the water quality on the spot consumes time and labor, is very easy to be limited by climatic and hydrological conditions, and is difficult to complete large-area and periodic continuous monitoring.
Disclosure of Invention
In order to overcome the problems in the prior art, the invention provides a water environment detection method based on Sentinel-2 data.
The technical scheme adopted by the invention for solving the technical problems is as follows: a water environment detection method based on Sentinel-2 data comprises the following steps:
step 3, sorting the longitude and latitude of the data, and converting the two-dimensional data into space point location data with a space coordinate system;
step 4, performing correlation analysis on the spectral reflectivity of each wave band of the remote sensing image data obtained in the step 1 and the water quality factor obtained in the step 2, performing correlation analysis on different wave band combinations of the remote sensing image data obtained in the step 1 and the water quality factor obtained in the step 2, and determining positive and negative correlation and correlation coefficient between the wave bands and the actually measured water quality factor;
step 5, judging the wave bands or wave band combination conditions participating in the inversion construction of various water quality factors, establishing a semi-empirical model for inverting the water quality, and optimizing a water quality model through model precision verification;
and 6, establishing a decision tree model, and grading the water quality so as to determine the water environment condition.
In the above method for detecting a water environment based on Sentinel-2 data, the pretreatment process in step 1 is as follows: and adopting 4 wave bands with the spatial resolution of 10 meters and 6 wave bands with the spatial resolution of 20 meters in the Sentinel-2 data, resampling the wave bands with the spatial resolution of 20 meters to 10 meters by adopting a bilinear interpolation method, carrying out wave band synthesis, inlaying and cutting by adopting a vector water area boundary to obtain the preprocessed Sentinel-2 data.
In the above method for detecting a water environment based on Sentinel-2 data, the method for determining the reliability of water quality in step 2 is as follows:
step 2.1, respectively calculating a mean value M of the n observed values of the water quality parameters and a mean value M of n-1 observed values of the removed suspicious values;
and 2.2, artificially assigning an influence coefficient k, wherein when the M/M is less than or equal to k +1, the suspicious value is not an abnormal value, and otherwise, the suspicious value is an abnormal value.
The method for detecting the water environment based on the Sentinel-2 data specifically comprises the following steps of 3: adding a unique code field for matching identification to each actual measuring point in the step 2 for subsequent matching of water quality monitoring result data and space point positions; and converting the longitude and latitude coordinate information into space point location data with a space coordinate system.
In the method for detecting the water environment based on the Sentinel-2 data, the correlation analysis in the step 4 can be performed by adopting an original spectral correlation analysis method, a normalized spectral correlation analysis method, an envelope removal spectral correlation analysis method and a first derivative micro-spectral correlation analysis method.
The method for detecting the water environment based on the Sentinel-2 data comprises the following steps of 5:
step 5.1, based on the correlation analysis result in the step 4, randomly extracting 80% of sample data and obtaining a wave band with strong correlation to carry out wave band combination, establishing a single/double/multiband water quality factor inversion model, selecting the remaining 20% of samples as verification, and finally selecting an optimal model of five water quality parameters;
step 5.2, calculating the accuracy of the inversion model obtained in the step 5.1, wherein the accuracy of the inversion model is measured by a decision coefficient and an average relative error, and the calculation formula is as follows:
determining the coefficient R 2 ：
Wherein the content of the first and second substances,
the measured value of the water quality parameter at the sampling point i is obtained;
the average value of the measured values of the water quality parameters is obtained;
the water quality parameter inversion value at the sampling point i is obtained;
average relative error MRE:
wherein the content of the first and second substances,
inverting values of the water quality parameters at the sampling point i; n is the number of modeling or verifying sampling points;
is the measured value of the water quality parameter at the sampling point i.
In the above method for detecting a water environment based on Sentinel-2 data, the optimizing the model in step 5 specifically includes: screening and matching the monitored remote sensing image data and the measured data: screening monitoring point positions with the time close to that of the monitoring images, and controlling the time of the images and the monitoring data within 1 month in the front and back; in consideration of factors such as distance areas and the like, screening point location data close to the distance of the monitored point location to participate in model optimization; and supplementing the measured data in real time and continuously optimizing the model parameters.
The invention has the beneficial effects that: (1) According to the method, a Sentinel-2 multispectral data product with 10 m spatial resolution is selected as basic data, and multispectral data of a high-resolution satellite with meter-level resolution is adopted to assist visual interpretation and proofreading of detection results, so that the diversity of water quality detection of a remote sensing satellite is increased, and support is provided for water environment control and prevention and control;
(2) According to the technical scheme, a semi-empirical model is adopted to invert water quality factors, the optimal wave band or wave band combination of the water environment index is selected and estimated according to the water environment index spectral characteristics measured by a non-imaging spectrometer or an airborne imaging spectrometer, and then a proper mathematical method is selected to establish a quantitative empirical algorithm between remote sensing data and the water environment index, so that the fitting degree of the model and the water quality condition of the region is higher;
(3) According to the invention, multiple water quality factors are inverted, and the water environment condition is graded by combining a decision tree, so that the water quality condition is clearer, the water environment condition of the area to be detected can provide data support for government departments, the water body of the detection area can be conveniently and timely treated, and the phenomenon that polluted water bodies are discharged in a messy manner is found.
Drawings
The invention is further illustrated by the following examples in conjunction with the drawings.
FIG. 1 is a flow chart of the present invention;
FIG. 2 is a diagram illustrating accuracy evaluation of a chemical oxygen demand model according to an embodiment of the present invention;
FIG. 3 is a diagram illustrating an evaluation of the ammonia nitrogen concentration model according to an embodiment of the present invention;
FIG. 4 is a diagram illustrating the accuracy of a total phosphorus concentration model according to an embodiment of the present invention;
FIG. 5 shows the result of remote sensing monitoring of water quality grade according to the embodiment of the present invention.
Detailed Description
In order to make the technical solutions of the present invention better understood, the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.
The invention provides a water environment detection method based on Sentinel-2 data and a semi-empirical multi-factor decision tree, which is shown in figure 1; selecting a Laoshan area in Qingdao city in the area to be detected, collecting water quality of partial Laoshan river reservoirs and along a coastline by sample data, inverting the water quality through a water quality model to obtain an inversion result and a global water quality result of a corresponding point, and fitting, wherein the method comprises the following steps of:
the method comprises the steps of firstly, acquiring a Sentinel-2 remote sensing image of a region to be monitored, preprocessing the image, and screening data of a time phase with good quality;
(1) Downloading Sentinel-2 data of a required time phase from a website of the European space Bureau;
(2) Adopting 4 wave bands with the spatial resolution of 10 meters and 6 wave bands with the spatial resolution of 20 meters in the Sentinel-2 data, utilizing SNAP software, using an S2 sampling Processor tool and selecting a bilinear interpolation method to resample the wave bands with the resolution of 20 meters to 10 meters, then exporting the result to an img storage format supported by ENVI, finally utilizing ENVI 5.3 to carry out wave band synthesis and mosaic, and adopting a vector water boundary to cut to obtain preprocessed Sentinel-2 data;
(3) Visually interpreting the remote sensing image, removing interference factors such as cloud and fog solar flare and screening the remote sensing image with qualified quality.
Step two, collecting information of each water quality factor on the spot, judging the reliability of water quality and cleaning data;
(1) Arranging the distribution of the required survey point locations, wherein the point location selection is mainly distributed in rivers in the Laoshan urban area, and due to the interference of various factors, the point location selection of parts of the urban area is more, and a few point locations are distributed in coastal and reservoir areas, so that the time before and after the satellite passes through the border and the weather condition of the day are predicted, the real-time survey is combined with the satellite passing time, and the weather condition needs to be good, cloud and free of wind;
(2) Collecting water quality indexes including total phosphorus concentration (TP), ammonia nitrogen concentration (NH 3-N) and Chemical Oxygen Demand (COD); in surface water quality monitoring, transient water samples are usually collected. The required water sample amount is shown in (the preservation and management technology of the water quality sampling sample of HJ 493-2009) and the sampling amount takes the requirements of repeated analysis and quality control into consideration and leaves room; immediately adding a preservative according to the requirements (specified by the preservation and management technology of the HJ 493-2009 water quality sampling sample) after the water sample is sampled or filled into a container; oil sampling: before sampling, the oil film possibly existing is destroyed, the glass material container is arranged in a bracket of the water sampler by a vertical water sampler, the glass material container is placed to the depth of 300 mm, the glass material container is lifted upwards while water is sampled, and a proper space is left when the glass material container reaches the water surface
(3) Judge water quality monitoring data reliability, because various external factors influence, like poultry farming, sewage discharge, instrument error etc. may cause water quality monitoring data unusual, unusual data can lead to the model precision to reduce by a wide margin, consequently need reject the abnormal value, arrange in order the washing data promptly, reject the abnormal value step as follows:
the statistics of each index can show that the value of each index fluctuates in a relatively fixed range, the occurrence frequency is high, the probability of exceeding the range is very small, the range of the value range can be used as a reference for judging whether the monitored value belongs to an abnormal value, and when the monitored value exceeds the normal range, whether the monitored value is the abnormal value needs to be judged;
under the condition of monitoring beyond a normal range, judging whether the water quality parameter is an abnormal value or not, considering that the water quality parameter concentration has certain connection gradual change property in a time space, adopting an abnormal value identification and processing method for the project, and mainly adopting an influence coefficient method: respectively calculating the mean value M of the n observed values of the water quality parameters and the mean value M of the n-1 observed values after the suspicious values are removed; artificially assigning an influence coefficient k, wherein when M/M is less than or equal to k +1, the suspicious value is not an abnormal value, otherwise, the suspicious value is an abnormal value;
step three, sorting the longitude and latitude of the data, and converting the two-dimensional data into space point location data with a space coordinate system;
(1) Arranging longitude and latitude coordinate information and other information of real-site detection point positions of Laoshan mountain river reservoirs and coastal areas, adding a unique code field for matching identification for each real-site, matching water quality monitoring result data with space point positions, and storing the data in a standard format after the data are arranged;
(2) Converting the format arranged in the step (1) into space point location data with a space coordinate system by utilizing longitude and latitude coordinate information.
Performing correlation analysis on the spectral reflectivity of each waveband of the remote sensing data and each water quality factor, performing correlation analysis between different waveband combinations and the actually measured water quality factor, and determining positive and negative correlation and correlation coefficient between the waveband and the actually measured water quality factor;
(1) Spectral data and sampled water quality parameters obtained by Laoshan area actual measurement are utilized, correlation analysis is carried out by combining with current Laoshan area Sentinel-2 remote sensing image data, the correlation analysis is used as a water quality parameter inversion model construction basis, original spectrum correlation analysis can be adopted, single-band and band combination analysis is carried out on an unprocessed original spectrum, wherein the single-band inversion accuracy is lower than 50%, correlation is obviously improved after different bands are combined, and the water quality inversion accuracy can reach more than 70%; the correlation analysis of the normalized spectrum, after the spectrum is normalized, the single-band and band combination analysis is carried out, and the effect is not different from the original correlation analysis of the spectrum; spectral characteristics can be highlighted by removing an envelope line and performing first derivative on a spectrum, numerical comparison with a spectral curve is facilitated, single-band and band combination analysis is performed by envelope line removal spectral correlation analysis and first derivative differential spectral correlation analysis methods, different band combination effects are different, and an optimal band combination result is not different from an original spectral analysis result;
(2) According to the analysis result of characteristic wave bands of water quality parameters of a water area spectral curve of an actually measured Laoshan area, a regression method is selected to carry out correlation analysis on the spectral reflectivity of each wave band of remote sensing data and each water quality parameter and the correlation analysis between different wave band combinations and the actually measured water quality parameters, a single-wave band model is a wave band combination model for establishing the correlation between the remote sensing reflectivity on a single-wave band image and each water quality index concentration, a correlation model between the numerical value on a corresponding image position and the water quality index concentration is established after the wave band combination is carried out on the single-wave band image, a difference model, a ratio model and some water body index models are mainly combined, the positive and negative correlation and the correlation coefficient between the wave bands and the actually measured water quality parameters are determined, and the wave band or wave band combination condition which can participate in water quality parameter inversion construction is judged. For example, the water quality of Laoshan waters is used, the correlation coefficient between the COD concentration estimation value and the actual measurement can reach 0.72 in the constructed unary linear regression model, the correlation coefficient between the TP concentration and the NH3-N concentration can reach 0.61 in the constructed multiple linear regression model, and the correlation coefficient between the TP concentration and the NH3-N concentration can reach 0.72.
Step five, judging wave bands or wave band combination conditions participating in inversion construction of various water quality factors, establishing a semi-empirical model for inverting water quality, and optimizing a water quality model through model precision verification, wherein the semi-empirical model established in the embodiment and the verification on the model precision are shown in figures 2-4;
(1) Based on the correlation analysis result, randomly extracting 80% of sample data and obtaining a wave band with strong correlation to carry out wave band combination, and establishing a single/double/multiband water quality factor inversion model in the form of:
wherein a is 0 ，a 1 ，…，a n Is an adjustable parameter; z is water concentration, X i The radiation value of the i wave band received by the sensor;
selecting the rest 20% of samples as verification, and finally selecting the optimal models of the five water quality parameters;
(2) And (3) calculating the accuracy of the inversion model obtained in the step (1), wherein the accuracy of the inversion model is measured by a decision coefficient and an average relative error, and the calculation formula is as follows:
determining the coefficient R 2 ：
Wherein, the first and the second end of the pipe are connected with each other,
the measured value of the water quality parameter at the sampling point i is obtained;
the average value of the measured values of the water quality parameters is obtained;
inverting values of the water quality parameters at the sampling point i;
average relative error MRE:
wherein the content of the first and second substances,
inverting values of the water quality parameters at the sampling point i; n is the number of modeling or verifying sampling points;
is the measured value of the water quality parameter at the sampling point i.
The closer the decision coefficient is to 1, the higher the correlation degree between the decision coefficient and the actual measurement result, and the strength of the correlation indicates the proximity of the inversion result and the actual measurement result. The relative error is used for quantitatively describing the difference between the inversion water quality parameter and the actually measured water quality parameter, and the smaller the value is, the higher the inversion precision is, and the closer the inversion result is to the actually measured result is.
As shown in fig. 2-4, in which the chemical oxygen demand model determines the coefficient R 2 0.72 percent, 16.90 percent of MRE and the ammonia nitrogen concentration model determination coefficient R 2 0.73, 23.14% MRE, total phosphorus concentration modeType determining coefficient R 2 It was 0.61 and the MRE was 29.35%.
(3) In order to ensure the accuracy of the monitoring model, the model established in the process of carrying out water quality factor inversion is optimized and updated, and the optimization comprises the steps of screening and matching the monitoring remote sensing image data and the measured data: (1) screening monitoring point positions with the time close to that of the monitoring images, and controlling the time of the images and the monitoring data within 1 month in the front and back; (2) in consideration of factors such as distance areas and the like, screening point location data close to the distance of the monitored point location to participate in model optimization; (3) and supplementing the measured data in real time and continuously optimizing the model parameters.
And step six, establishing a decision tree model according to the surface water environmental quality standard (GB 3838-2002), and grading the water quality so as to determine the water environment condition.
According to the ' surface water environmental quality standard ' (GB 3838-2002) ', surface water environmental quality evaluation is to select corresponding category standards according to water area function categories to be realized, a decision tree is established by a comprehensive pollution index method for evaluation, the evaluation result indicates that the water quality reaches the standard, and the overproof items and overproof multiples are indicated when the overproof water exists. The standard limit values of the basic items of the surface water environment quality standard are shown in table 1, and the surface water environment quality standard is classified according to the limits of total phosphorus, ammonia nitrogen and chemical oxygen demand.
TABLE 1 Standard limits of basic project of surface water environment quality standards
| Class I | Class II | Class III | Class IV | Class V | |
| The total phosphorus concentration is less than or equal to | 0.02 | 0.1 | 0.2 | 0.3 | 0.4 |
| Chemical oxygen demand concentration is less than or equal to | 15 | 15 | 20 | 30 | 40 |
| Ammonia nitrogen concentration is less than or equal to | 0.015 | 0.5 | 1.0 | 1.5 | 2.0 |
According to the environmental function and the protection target of the surface water area, five types are sequentially divided according to the function height:
the class I is mainly suitable for head water and national natural protection areas;
the II type is mainly suitable for primary protection areas of surface water sources of centralized drinking water areas, habitats of rare aquatic organisms, fish and shrimp production fields, cable bait fields of larvae and juvenile fishes and the like;
class III is mainly suitable for fishery water areas and swimming areas such as centralized drinking water surface water source areas, fish and shrimp overwintering fields, migration passages, aquaculture areas and the like;
the IV class is mainly suitable for general industrial water areas and recreational water areas which are not directly contacted with human bodies;
class V is mainly suitable for agricultural water areas and water areas with general landscape requirements.
And classifying the standard values of the surface water environment quality standard basic items into five classes corresponding to the five water area functions of the surface water, and executing the standard values of the corresponding classes by different functional classes respectively. The standard value for the water area function type is higher than the standard value for the water area function type. The same water area has multiple using functions and executes the standard value corresponding to the highest function category. The water area function and the function reaching category standard are the same, the water area in Laoshan region is monitored according to the method of the embodiment of the invention, and the monitoring result is as follows: the water quality in the monitoring area of Laoshan is mainly II-type water, the total phosphorus concentration is mainly 0-0.2mg/L, the ammonia nitrogen concentration is mainly 0-1mg/L, and the chemical oxygen demand concentration is mainly 10-20mg/L. As shown in fig. 5, fig. 5 shows the water quality level monitoring result of the laoshan reservoir in the laoshan region, the reservoir range is the water area range extracted by the Sentinel-2 satellite image through the water body index, the water quality level in the monitored region can be reflected visually, and as can be seen from fig. 5, the whole water body of the laoshan reservoir mainly contains type II water, and a small amount of type III water is arranged at the edge.
The above embodiments are only exemplary embodiments of the present invention, and are not intended to limit the present invention, and the scope of the present invention is defined by the claims. Various modifications and equivalents of the invention may be made by those skilled in the art within the spirit and scope of the invention, and such modifications and equivalents should also be considered as falling within the scope of the invention.
Claims (7)
1. A water environment detection method based on Sentinel-2 data is characterized in that: the method comprises the following steps:
step 1, acquiring a Sentinel-2 remote sensing image of a region to be monitored, preprocessing the image and screening out data of a required time phase;
step 2, collecting information of each water quality factor on site, judging water quality reliability and cleaning data;
step 3, sorting the longitude and latitude of the data, and converting the two-dimensional data into space point location data with a space coordinate system;
step 4, performing correlation analysis on the spectral reflectivity of each wave band of the remote sensing image data obtained in the step 1 and the water quality factor obtained in the step 2, performing correlation analysis on different wave band combinations of the remote sensing image data obtained in the step 1 and the water quality factor obtained in the step 2, and determining positive and negative correlation and correlation coefficients between the wave bands and the actually measured water quality factor;
step 5, judging the wave bands or wave band combination conditions participating in the inversion construction of various water quality factors, establishing a semi-empirical model for inverting the water quality, and optimizing a water quality model through model precision verification;
and 6, establishing a decision tree model, and grading the water quality so as to determine the water environment condition.
2. The method for detecting water environment based on Sentinel-2 data according to claim 1, wherein the preprocessing in step 1 comprises: and adopting 4 wave bands with the spatial resolution of 10 meters and 6 wave bands with the spatial resolution of 20 meters in the Sentinel-2 data, resampling the wave bands with the spatial resolution of 20 meters to 10 meters by adopting a bilinear interpolation method, carrying out wave band synthesis, inlaying and cutting by adopting a vector water area boundary to obtain the preprocessed Sentinel-2 data.
3. The method for detecting the water environment based on the Sentinel-2 data according to claim 1, wherein the method for judging the reliability of the water quality in the step 2 is as follows:
step 2.1, respectively calculating a mean value M of the n observed values of the water quality parameters and a mean value M of n-1 observed values of the removed suspicious values;
and 2.2, artificially assigning an influence coefficient k, wherein when the M/M is less than or equal to k +1, the suspicious value is not an abnormal value, and otherwise, the suspicious value is an abnormal value.
4. The method for detecting the water environment based on the Sentinel-2 data according to claim 1, wherein the step 3 specifically comprises: adding a unique code field for matching identification to each actual measuring point in the step 2 for subsequent matching of water quality monitoring result data and space point positions; and converting the longitude and latitude coordinate information into space point location data with a space coordinate system.
5. The method for detecting water environment according to claim 1, wherein the correlation analysis in step 4 can be performed by using original spectral correlation analysis, normalized spectral correlation analysis, envelope elimination spectral correlation analysis, and first derivative micro-spectral correlation analysis.
6. The method for detecting water environment based on Sentinel-2 data according to claim 1, wherein the step 5 specifically comprises:
step 5.1, based on the correlation analysis result in the step 4, randomly extracting 80% of sample data and the obtained wave band with strong correlation to carry out wave band combination, establishing a single/double/multiband water quality factor inversion model, selecting the remaining 20% of samples as verification, and finally selecting the optimal model of five water quality parameters;
step 5.2, calculating the accuracy of the inversion model obtained in the step 5.1, wherein the accuracy of the inversion model is measured by a decision coefficient and an average relative error, and the calculation formula is as follows:
determining the coefficient R 2 ：
Wherein the content of the first and second substances,
the measured value of the water quality parameter at the sampling point i is obtained;
the average value of the measured values of the water quality parameters is obtained;
the water quality parameter inversion value at the sampling point i is obtained;
average relative error MRE:
wherein the content of the first and second substances,
inverting values of the water quality parameters at the sampling point i; n is the number of modeling or verifying sampling points;
is the measured value of the water quality parameter at the sampling point i.
7. The method for detecting the water environment based on the Sentinel-2 data according to claim 1, wherein the optimizing the model in the step 5 specifically comprises: screening and matching the monitored remote sensing image data and the measured data: screening monitoring point positions with the time close to that of the monitoring images, and controlling the time of the images and the monitoring data within 1 month in the front and back; in consideration of factors such as distance areas and the like, images of the monitoring model are screened, point location data close to the monitoring point location distance are screened, and the point location data participate in optimization of the model; and supplementing measured data in real time and continuously optimizing model parameters.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211298732.4A CN115358103B (en) | 2022-10-24 | 2022-10-24 | Water environment detection method based on Sentinel-2 data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211298732.4A CN115358103B (en) | 2022-10-24 | 2022-10-24 | Water environment detection method based on Sentinel-2 data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115358103A true CN115358103A (en) | 2022-11-18 |
| CN115358103B CN115358103B (en) | 2023-04-07 |
Family
ID=84008378
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211298732.4A Active CN115358103B (en) | 2022-10-24 | 2022-10-24 | Water environment detection method based on Sentinel-2 data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115358103B (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105868533A (en) * | 2016-03-23 | 2016-08-17 | 四川理工学院 | River basin water environment integrated perception and application method based on Internet of Things and 3S technology |
| US20200408675A1 (en) * | 2019-06-25 | 2020-12-31 | Horiba, Ltd. | Analyzer, analysis method, and a program recording medium recorded with a program for analyzer |
| CN114739919A (en) * | 2022-04-14 | 2022-07-12 | 常州市新翼空间信息科技有限公司 | Water quality detection method based on spectrum inversion analysis |
-
2022
- 2022-10-24 CN CN202211298732.4A patent/CN115358103B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105868533A (en) * | 2016-03-23 | 2016-08-17 | 四川理工学院 | River basin water environment integrated perception and application method based on Internet of Things and 3S technology |
| US20200408675A1 (en) * | 2019-06-25 | 2020-12-31 | Horiba, Ltd. | Analyzer, analysis method, and a program recording medium recorded with a program for analyzer |
| CN114739919A (en) * | 2022-04-14 | 2022-07-12 | 常州市新翼空间信息科技有限公司 | Water quality detection method based on spectrum inversion analysis |
Non-Patent Citations (3)
| Title |
|---|
| GEHAD HASSAN: "Hybrid Predictive Model for Water Quality Monitoring Based on Sentinel-2A L1C Data", 《IEEE》 * |
| 姜腾龙等: "基于Landsat系列卫星分析1985―2015年东平湖叶绿素a浓度变化趋势", 《中国环境监测》 * |
| 江辉: "基于多源遥感的鄱阳湖水质参数反演与分析", 《工程科技Ⅰ辑》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115358103B (en) | 2023-04-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109190538B (en) | Sediment-laden river delta coastal zone evolution analysis method based on remote sensing technology | |
| Novoa et al. | Water quality assessment using satellite-derived chlorophyll-a within the European directives, in the southeastern Bay of Biscay | |
| B. Knoll et al. | Predicting eutrophication status in reservoirs at large spatial scales using landscape and morphometric variables | |
| Greenwood et al. | Utilizing eutrophication assessment directives from transitional to marine systems in the Thames Estuary and Liverpool Bay, UK | |
| Kratzer et al. | Using bio-optics to investigate the extent of coastal waters: A Swedish case study | |
| Veytia et al. | Overwinter sea-ice characteristics important for Antarctic krill recruitment in the southwest Atlantic | |
| Ollevier et al. | Noctiluca scintillans: Dynamics, size measurements and relationships with small soft-bodied plankton in the Belgian Part of the North Sea | |
| Javari | Geostatistical and spatial statistical modelling of precipitation variations in Iran | |
| Wang et al. | Light regulation of phytoplankton growth in San Francisco Bay studied using a 3D sediment transport model | |
| Kathiravan et al. | Spatio-temporal variability of hydro-chemical characteristics of coastal waters of Gulf of Mannar Marine Biosphere Reserve (GoMMBR), South India | |
| CN114241331A (en) | Wetland reed aboveground biomass remote sensing modeling method taking UAV as ground and Sentinel-2 intermediary | |
| CN115358103B (en) | Water environment detection method based on Sentinel-2 data | |
| Beentjes et al. | Blue cod relative abundance, size and age structure, and habitat surveys of Marlborough Sounds in 2013 | |
| Kazemzadeh et al. | Remote sensing of temporal and spatial variations of suspended sediment concentration in Bahmanshir Estuary, Iran | |
| El-Mezayen et al. | Assessment of water quality and modeling trophic level of Lake Manzala, Egypt using remotely sensed observations after recent Enhancement Project | |
| Dörnhöfer et al. | Water colour analysis of Lake Kummerow using time series of remote sensing and in situ data | |
| Torres et al. | Preliminary investigation on chlorophyll-a and total suspended matter concentration in Manila Bay using Himawari-8 AHI and Sentinel-3 OLCI C2RCC | |
| CN110988286B (en) | Intelligent water resource long-term detection system | |
| CN115078261A (en) | Inland lake total phosphorus concentration remote sensing monitoring method based on random forest model | |
| Litts et al. | Mapping potential spawning substrate for Shortnose and Atlantic sturgeon in coastal plain rivers of Georgia using low-cost side-scan sonar | |
| Amran et al. | Estimation of coastal waters turbidity using Sentinel-2 imagery | |
| Coffin et al. | Dissolved oxygen as a marine environmental quality (MEQ) measure in upper estuaries of the southern Gulf of St. Lawrence: Implications for nutrient management and eelgrass (Zostera marina) coverage | |
| Koponen et al. | Blue Carbon Habitats:–a comprehensive mapping of Nordic salt marshes for estimating Blue Carbon storage potential | |
| Dewi et al. | Study of sea surface temperature and chlorophyll-a influence on the quantity of fish caught in the waters of Sadeng, Yogyakarta, Indonesia. | |
| Chatterjee et al. | Assessing the Water Quality of Indian Sundarban Estuaries using Remote Sensing Techniques |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |