CN-116682019-B - Water bloom region prediction method, device and equipment based on remote sensing image

CN116682019BCN 116682019 BCN116682019 BCN 116682019BCN-116682019-B

Abstract

The invention relates to the field of remote sensing data analysis, in particular to a water bloom region prediction method based on a remote sensing image, which is used for carrying out water quality parameter inversion on the remote sensing image to obtain water quality prediction parameters, extracting the water bloom area of the remote sensing image based on the water quality prediction parameters, and carrying out water bloom position prediction by combining with hydrological data, so that the local effects of time and space objects are considered, the influence of factors such as water quality, hydrology and water temperature on water bloom is considered, the water bloom prediction precision is improved, and the fine and efficient water bloom prediction is realized.

Inventors

ZHANG HANBO
JING WENLONG
DENG YINGBIN
YANG JI
ZHANG MINGLIANG
HU YIQIANG
Shu Sijing
WANG LIUTAO

Assignees

广东省科学院广州地理研究所

Dates

Publication Date: 20260508
Application Date: 20230421

Claims (7)

1. The water bloom region prediction method based on the remote sensing image is characterized by comprising the following steps of: Obtaining a plurality of sample remote sensing images, and wave band reflection data and hydrological data of each sample remote sensing image in a target time period; according to the wave band reflection data and the hydrological data of each sample remote sensing image in a target time period, a water bloom prediction model is constructed, wherein the water bloom prediction model comprises a water quality prediction module, a water bloom area prediction module and a water bloom position prediction module, and according to the wave band reflection data and the hydrological data of each sample remote sensing image in the target time period, the water bloom prediction model is constructed, and the method comprises the following steps: Obtaining water quality actual measurement data of each sample remote sensing image in a target time period, wherein the water quality actual measurement data comprises a plurality of water quality actual measurement parameters of a plurality of pixels; According to a plurality of preset sampling points and window sizes, sampling windows corresponding to the sampling points of each sample remote sensing image are constructed, and according to the water quality actual measurement data and the wave band reflection data, wave band reflection data and water quality actual measurement data of each sampling window in a target time period are obtained; the wave band reflection data and the water quality actual measurement data of each sampling window in the target time period are input into the water quality prediction module for training, and the target water quality prediction module is obtained; According to the wave band reflection data of each sampling window in the target time period, extracting water bloom pixels of each sampling window to obtain the number of water bloom pixels corresponding to each moment of each sampling window; Obtaining the spatial resolution of each sampling window, and obtaining the water bloom area parameters corresponding to each moment of each sampling window according to the number of the water bloom pixels, the spatial resolution and a preset water bloom area parameter calculation algorithm, wherein the water bloom area parameter calculation algorithm is as follows: in the formula, Is the area parameter of the water bloom, Is the number of the water bloom pixels, Is the spatial resolution; The wave band reflection data of each sampling window in a target time period is input to the target water quality prediction module, and water quality prediction data corresponding to each moment of each sampling window is obtained; Acquiring longitude and latitude data of each sampling window and water temperature parameters corresponding to each moment, inputting the longitude and latitude data of each sampling window, the water temperature parameters corresponding to each moment, water quality prediction data and water bloom area parameters into the water bloom area prediction module, and training the water bloom area prediction module to obtain a target water bloom area prediction module; acquiring offshore distance data of each sampling window and hydrologic data of each sampling window in a target time period, wherein the hydrologic data comprise hydrologic parameters corresponding to each moment, and the hydrologic parameters comprise wind speed parameters, wind direction parameters and water flow speed parameters; Training the water bloom position prediction module by using offshore distance of each sampling window, corresponding hydrological data, water bloom area parameters and a preset initial water bloom area parameter prediction algorithm to obtain a target water bloom position prediction module, wherein the initial water bloom position parameter prediction algorithm is as follows: in the formula, And Time parameters respectively representing the current time and the next time, Is the first The abscissa parameter corresponding to the moment in time, Is the first The ordinate parameter corresponding to the moment in time, As a parameter of the wind speed, As a parameter of the wind direction, For the water flow rate parameter, l is the offshore distance data, As a result of the first speed coefficient, As a result of the second velocity coefficient, Is a third speed coefficient; Is the first The water bloom area parameter corresponding to the moment; Obtaining wave band reflection data, water temperature data and hydrological data of a remote sensing image to be detected, and inputting the wave band reflection data of the remote sensing image to be detected into the water quality prediction module to obtain water quality prediction data of the remote sensing image to be detected; inputting water quality prediction data and water temperature data of the remote sensing image to be detected into the water bloom area prediction module, and obtaining water bloom prediction area parameters of each prediction window of the remote sensing image to be detected according to a plurality of preset prediction windows; Inputting the water bloom predicted area parameters and the hydrologic data of each predicted window of the remote sensing image to be detected to the water bloom position prediction module to obtain water bloom predicted coordinate parameters of each predicted window of the remote sensing image to be detected; And obtaining a plurality of water bloom prediction areas of the remote sensing image to be detected according to the water bloom prediction area parameters and the water bloom prediction coordinate parameters of each prediction window of the remote sensing image to be detected.
2. The method for predicting a water bloom area based on a remote sensing image as set forth in claim 1, wherein the band reflection data includes a green band reflectance, a near infrared band reflectance, and a red band reflectance; according to the band reflection data, extracting water bloom pixels from each sampling window to obtain the number of water bloom pixels corresponding to each moment of each sampling window, including the steps of: obtaining normalized vegetation indexes and normalized water indexes corresponding to each moment of each pixel according to the band reflection data, a preset normalized vegetation index calculation algorithm and a normalized water index calculation algorithm, wherein the normalized vegetation index calculation algorithm is as follows: in the formula, In order to normalize the water body index, For the reflectivity of the green band of wavelengths, Is the reflectivity of the near infrared band; the normalized water index is: in the formula, In order to normalize the vegetation index, Is the reflectivity of the red wave band; And according to the normalized vegetation index and the normalized water body index corresponding to each moment of each pixel, and a preset normalized vegetation index threshold and a preset normalized water body index threshold, respectively extracting the water bloom pixels in each sampling window to obtain the number of the water bloom pixels corresponding to each moment of each sampling window.
3. The water bloom area prediction method based on the remote sensing image as set forth in claim 2, wherein the water quality prediction data comprises a plurality of water quality prediction parameters of a plurality of pixels, the longitude and latitude data comprises longitude parameters and latitude parameters, and the water quality prediction parameters comprise total phosphorus content, total nitrogen content, pH value, dissolved oxygen content and chlorophyll content of a water area; The method comprises the steps of inputting longitude and latitude data of each sampling window, water temperature parameters, water quality prediction data and water bloom area parameters corresponding to each moment to the water bloom area prediction module, training the water bloom area prediction module, and obtaining a target water bloom area prediction module, and comprises the following steps: According to the water quality prediction data corresponding to each moment of each sampling window and a preset arithmetic average value calculation algorithm, obtaining arithmetic average values of a plurality of water quality prediction parameters corresponding to each moment of each sampling window, wherein the arithmetic average value calculation algorithm is as follows: Wherein x is the arithmetic average value of the water quality prediction parameters, D is the total number of pixels in the sampling window, D represents the D-th pixel, A value of a water quality prediction parameter for the d-th pixel; Training the water bloom area prediction module according to longitude and latitude data of each sampling window, water temperature parameters corresponding to each moment, water bloom area parameters, arithmetic average values of a plurality of water quality prediction parameters and a preset initial water bloom area parameter prediction algorithm to obtain a target water bloom area prediction module, wherein the initial water bloom area parameter prediction algorithm is as follows: in the formula, Is the water bloom area parameter corresponding to the t moment of the ith sampling window, Is the regression constant of the ith sampling window, For the longitude parameter of the i-th sampling window, For the latitude parameter of the ith sampling window, N is the total number of categories of the water quality prediction parameters, k represents the water quality prediction parameters of the kth category, Regression coefficients for the kth class of water quality prediction parameters to the ith sampling window, For the arithmetic average of the water quality prediction parameters in the ith sampling window, The residual value of the water bloom area prediction module in the ith sampling window is calculated, and m is the total number of the sampling windows; The expression of the target water bloom area prediction module is as follows: in the formula, TP is the total phosphorus content in a water area, TN is the total nitrogen content, pH is the pH value, DO is the dissolved oxygen content, Q is the water temperature parameter, chla is the chlorophyll content, As a residual value, f () is a nonlinear function.
4. The method for predicting water bloom areas based on remote sensing images as set forth in claim 1, wherein the obtaining the plurality of water bloom predicted areas of the remote sensing images to be measured according to the water bloom predicted area parameters and the water bloom predicted coordinate parameters of each prediction window of the remote sensing images to be measured includes the steps of: Calculating radius parameters of each prediction window according to the water bloom prediction area parameters of each prediction window; And constructing a circular area of each prediction window as the water bloom prediction area according to the radius parameter of each prediction window and the corresponding water bloom prediction coordinate parameter as the center, and obtaining a plurality of water bloom prediction areas of the remote sensing image to be detected.
5. The method for predicting a water bloom area based on a remote sensing image as recited in claim 1, further comprising the steps of: and responding to a display instruction, acquiring an electronic map, and displaying and labeling a plurality of water bloom prediction areas on the electronic map according to the plurality of water bloom prediction areas of the remote sensing image to be detected.
6. The utility model provides a water bloom regional prediction device based on remote sensing image which characterized in that includes: The sample data acquisition module is used for acquiring a plurality of sample remote sensing images, and wave band reflection data and hydrological data of each sample remote sensing image in a target time period; The model construction module is used for constructing a water bloom prediction model according to wave band reflection data and hydrological data of each sample remote sensing image in a target time period, wherein the water bloom prediction model comprises a water quality prediction module, a water bloom area prediction module and a water bloom position prediction module, and the water bloom prediction model is constructed according to the wave band reflection data and the hydrological data of each sample remote sensing image in the target time period and comprises the following steps: Obtaining water quality actual measurement data of each sample remote sensing image in a target time period, wherein the water quality actual measurement data comprises a plurality of water quality actual measurement parameters of a plurality of pixels; According to a plurality of preset sampling points and window sizes, sampling windows corresponding to the sampling points of each sample remote sensing image are constructed, and according to the water quality actual measurement data and the wave band reflection data, wave band reflection data and water quality actual measurement data of each sampling window in a target time period are obtained; the wave band reflection data and the water quality actual measurement data of each sampling window in the target time period are input into the water quality prediction module for training, and the target water quality prediction module is obtained; According to the wave band reflection data of each sampling window in the target time period, extracting water bloom pixels of each sampling window to obtain the number of water bloom pixels corresponding to each moment of each sampling window; Obtaining the spatial resolution of each sampling window, and obtaining the water bloom area parameters corresponding to each moment of each sampling window according to the number of the water bloom pixels, the spatial resolution and a preset water bloom area parameter calculation algorithm, wherein the water bloom area parameter calculation algorithm is as follows: in the formula, Is the area parameter of the water bloom, Is the number of the water bloom pixels, Is the spatial resolution; The wave band reflection data of each sampling window in a target time period is input to the target water quality prediction module, and water quality prediction data corresponding to each moment of each sampling window is obtained; Acquiring longitude and latitude data of each sampling window and water temperature parameters corresponding to each moment, inputting the longitude and latitude data of each sampling window, the water temperature parameters corresponding to each moment, water quality prediction data and water bloom area parameters into the water bloom area prediction module, and training the water bloom area prediction module to obtain a target water bloom area prediction module; acquiring offshore distance data of each sampling window and hydrologic data of each sampling window in a target time period, wherein the hydrologic data comprise hydrologic parameters corresponding to each moment, and the hydrologic parameters comprise wind speed parameters, wind direction parameters and water flow speed parameters; training the water bloom position prediction module by using offshore distance of each sampling window, corresponding hydrological data, water bloom area parameters and a preset initial water bloom area parameter prediction algorithm to obtain a target water bloom position prediction module, wherein the initial water bloom area parameter prediction algorithm is as follows: in the formula, Time parameters respectively representing the current time and the next time, Is the first The abscissa parameter corresponding to the moment in time, Is the first The ordinate parameter corresponding to the moment in time, As a parameter of the wind speed, As a parameter of the wind direction, For the water flow rate parameter, l is the offshore distance data, As a result of the first speed coefficient, As a result of the second velocity coefficient, Is a third speed coefficient; Is the first The water bloom area parameter corresponding to the moment; the water quality prediction module is used for obtaining wave band reflection data, water temperature data and hydrological data of the remote sensing image to be detected, inputting the wave band reflection data of the remote sensing image to be detected into the water quality prediction module and obtaining water quality prediction data of the remote sensing image to be detected; The water bloom area prediction module is used for inputting water quality prediction data and water temperature data of the remote sensing image to be detected into the water bloom area prediction module, and obtaining water bloom prediction area parameters of each prediction window of the remote sensing image to be detected according to a plurality of preset prediction windows; the water bloom position prediction module is used for inputting water bloom prediction area parameters and hydrologic data of each prediction window of the remote sensing image to be detected to the water bloom position prediction module to obtain water bloom prediction coordinate parameters of each prediction window of the remote sensing image to be detected; And the water bloom region prediction module is used for obtaining a plurality of water bloom prediction regions of the remote sensing image to be detected according to the water bloom prediction area parameters and the water bloom prediction coordinate parameters of each prediction window of the remote sensing image to be detected.
7. A computer device comprising a processor, a memory and a computer program stored on the memory and executable on the processor, the computer program when executed by the processor implementing the steps of the remote sensing image based water bloom region prediction method as claimed in any one of claims 1 to 5.

Description

Water bloom region prediction method, device and equipment based on remote sensing image Technical Field The invention relates to the field of remote sensing data analysis, in particular to a water bloom area prediction method, a device, equipment and a storage medium based on remote sensing images. Background Developing refined water bloom prediction has important significance for preventing disasters caused by water bloom outbreaks and formulating reasonable prevention and treatment measures. At present, the research on the prediction of the water bloom is commonly based on the water bloom area prediction of satellite remote sensing images or the indirect prediction of the water bloom spatial distribution through the water bloom explosion probability, but the water bloom explosion is rapid, the spatial resolution and the timeliness of the satellite remote sensing are difficult to meet the requirements of the water bloom prediction, and the application of the satellite remote sensing in a cloudy rain area is limited. In addition, the explosion of the water bloom is a complex process commonly influenced by multiple factors such as water environment, water temperature, climate and the like, and the refined water bloom prediction is difficult to realize by a single optical remote sensing method. Disclosure of Invention Based on the water bloom region prediction method, device, equipment and storage medium based on the remote sensing image, the water bloom region prediction method, device, equipment and storage medium based on the remote sensing image are provided, the water quality parameter inversion is carried out on the remote sensing image to obtain the water quality prediction parameter, the water bloom region extraction is carried out on the remote sensing image based on the water quality prediction parameter, the water bloom position prediction is carried out by combining with hydrological data, the local effects of time and space objects are considered, the influence of factors such as water quality, hydrology and water temperature on the water bloom is considered, the water bloom prediction precision is improved, and the fine and efficient water bloom prediction is realized. In a first aspect, an embodiment of the present application provides a water bloom area prediction method based on a remote sensing image, including the following steps: Obtaining a plurality of sample remote sensing images, and wave band reflection data and hydrological data of each sample remote sensing image in a target time period; according to wave band reflection data and hydrological data of each sample remote sensing image in a target time period, a water bloom prediction model is constructed, wherein the water bloom prediction model comprises a water quality prediction module, a water bloom area prediction module and a water bloom position prediction module; Obtaining wave band reflection data, water temperature data and hydrological data of a remote sensing image to be detected, and inputting the wave band reflection data of the remote sensing image to be detected into the water quality prediction module to obtain water quality prediction data of the remote sensing image to be detected; inputting water quality prediction data and water temperature data of the remote sensing image to be detected into the water bloom area prediction module, and obtaining water bloom prediction area parameters of each prediction window of the remote sensing image to be detected according to a plurality of preset prediction windows; Inputting the water bloom predicted area parameters and the hydrologic data of each predicted window of the remote sensing image to be detected to the water bloom position prediction module to obtain water bloom predicted coordinate parameters of each predicted window of the remote sensing image to be detected; And obtaining a plurality of water bloom prediction areas of the remote sensing image to be detected according to the water bloom prediction area parameters and the water bloom prediction coordinate parameters of each prediction window of the remote sensing image to be detected. In a second aspect, an embodiment of the present application provides a bloom area prediction apparatus based on a remote sensing image, including: The sample data acquisition module is used for acquiring a plurality of sample remote sensing images, and wave band reflection data and hydrological data of each sample remote sensing image in a target time period; the model construction module is used for constructing a water bloom prediction model according to wave band reflection data and hydrological data of each sample remote sensing image in a target time period, wherein the water bloom prediction model comprises a water quality prediction module, a water bloom area prediction module and a water bloom position prediction module; the water quality prediction module is used for obtaining wave band reflection data, water temperature