CN118115824A - Water quality variable concentration prediction method and system - Google Patents
Water quality variable concentration prediction method and system Download PDFInfo
- Publication number
- CN118115824A CN118115824A CN202410174923.2A CN202410174923A CN118115824A CN 118115824 A CN118115824 A CN 118115824A CN 202410174923 A CN202410174923 A CN 202410174923A CN 118115824 A CN118115824 A CN 118115824A
- Authority
- CN
- China
- Prior art keywords
- water quality
- band
- denotes
- reflectivity
- quality variable
- 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.)
- Pending
Links
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 title claims abstract description 85
- 238000000034 method Methods 0.000 title claims abstract description 30
- 238000002310 reflectometry Methods 0.000 claims abstract description 43
- 238000013135 deep learning Methods 0.000 claims abstract description 10
- 238000007781 pre-processing Methods 0.000 claims abstract description 6
- 239000011159 matrix material Substances 0.000 claims description 26
- 238000012937 correction Methods 0.000 claims description 23
- 238000003860 storage Methods 0.000 claims description 18
- 230000004913 activation Effects 0.000 claims description 16
- 230000005855 radiation Effects 0.000 claims description 14
- 230000015654 memory Effects 0.000 claims description 8
- 230000006870 function Effects 0.000 claims description 5
- 238000004364 calculation method Methods 0.000 claims description 4
- 238000001228 spectrum Methods 0.000 claims description 3
- 230000003595 spectral effect Effects 0.000 abstract description 11
- 230000008859 change Effects 0.000 abstract description 6
- 238000012545 processing Methods 0.000 abstract description 4
- 230000007613 environmental effect Effects 0.000 description 5
- 238000012544 monitoring process Methods 0.000 description 5
- 230000003287 optical effect Effects 0.000 description 4
- QVGXLLKOCUKJST-UHFFFAOYSA-N atomic oxygen Chemical compound [O] QVGXLLKOCUKJST-UHFFFAOYSA-N 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 3
- 238000013527 convolutional neural network Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 239000001301 oxygen Substances 0.000 description 3
- 229910052760 oxygen Inorganic materials 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 238000001303 quality assessment method Methods 0.000 description 3
- QGZKDVFQNNGYKY-UHFFFAOYSA-N Ammonia Chemical compound N QGZKDVFQNNGYKY-UHFFFAOYSA-N 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 2
- 239000000126 substance Substances 0.000 description 2
- IJGRMHOSHXDMSA-UHFFFAOYSA-N Atomic nitrogen Chemical compound N#N IJGRMHOSHXDMSA-UHFFFAOYSA-N 0.000 description 1
- OAICVXFJPJFONN-UHFFFAOYSA-N Phosphorus Chemical compound [P] OAICVXFJPJFONN-UHFFFAOYSA-N 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 229910021529 ammonia Inorganic materials 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000007789 gas Substances 0.000 description 1
- 238000011065 in-situ storage Methods 0.000 description 1
- 230000007787 long-term memory Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 239000013618 particulate matter Substances 0.000 description 1
- 229910052698 phosphorus Inorganic materials 0.000 description 1
- 239000011574 phosphorus Substances 0.000 description 1
- 238000011176 pooling Methods 0.000 description 1
- 230000005180 public health Effects 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
- 239000000758 substrate Substances 0.000 description 1
- 238000012731 temporal analysis Methods 0.000 description 1
- 230000002123 temporal effect Effects 0.000 description 1
- 238000000700 time series analysis Methods 0.000 description 1
- 238000012876 topography Methods 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/766—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using regression, e.g. by projecting features on hyperplanes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
- G06N3/0442—Recurrent networks, e.g. Hopfield networks characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Business, Economics & Management (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Multimedia (AREA)
- Artificial Intelligence (AREA)
- Software Systems (AREA)
- Computing Systems (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Biomedical Technology (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biophysics (AREA)
- Medical Informatics (AREA)
- Computational Linguistics (AREA)
- Mathematical Physics (AREA)
- Marketing (AREA)
- Databases & Information Systems (AREA)
- Molecular Biology (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Water Supply & Treatment (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Game Theory and Decision Science (AREA)
- Development Economics (AREA)
- Primary Health Care (AREA)
- Public Health (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The invention provides a water quality variable concentration prediction method and a system, which relate to the technical field of image data processing, wherein the method comprises the following steps: acquiring a remote sensing image; preprocessing the remote sensing image; extracting image features of the preprocessed remote sensing image; and predicting the water quality variable concentration according to the image characteristics through a ConvLSTM regression model based on deep learning. According to the invention, a lake or river is covered by a remote sensing image, a complex nonlinear relation between image characteristics such as spectral reflectivity and water quality parameters is accurately captured by a ConvLSTM regression model based on deep learning, the water quality variable concentration is accurately estimated by the image characteristics such as spectral reflectivity, the dynamic change of the water quality of the lake or river is monitored, and reliable water quality variable parameters are provided for estimating the water quality of the lake or river.
Description
Technical Field
The invention relates to the technical field of image data processing, in particular to a water quality variable concentration prediction method and system.
Background
Different regions of the world face water environmental problems. Extracting different spatiotemporal features using conventional in situ observation techniques is traditionally considered to be too costly for local environmental use, which inevitably poses challenges to solving aquatic and public health problems. To date, many studies have explored the potential and utility of remote sensing technology in better understanding environmental features at different temporal and spatial scales. Satellites equipped with various optical and thermal sensors offer advantages over field measurements, can provide an ever-increasing geospatial data stream, cover large areas, have high resolution, and are more economical. For many years, researchers have been used for water quality assessment, and studies conducted in large rivers and lakes, estuaries and coastal areas or in regional areas have demonstrated the applicability of satellite-based water quality assessment.
However, the current method is not suitable for monitoring small lakes or rivers due to limited spatial information of the small lakes or rivers. In urban water quality assessment, non-photosensitive parameters such as Chemical Oxygen Demand (COD), biological Oxygen Demand (BOD), total Nitrogen (TN), permanganate of chemical oxygen demand (CODMn), ammonia (NH 3-N) and Total Phosphorus (TP) are mainly used as important reference indicators. For small lakes or rivers, it is difficult to accurately evaluate the water quality variable concentration by the surface reflectivity due to the complicated nonlinear relationship between the observed value of the water quality parameter and the surface reflectivity, resulting in the lack of water quality variable parameters that can be used for evaluating the lakes or rivers.
Disclosure of Invention
In order to solve the technical problem that for small lakes or rivers, due to the complex nonlinear relation between the observed value of the water quality parameter and the surface reflectivity, the water quality variable concentration is difficult to accurately evaluate through the surface reflectivity, and the water quality variable parameter which can be used for evaluating the lakes or rivers is not yet available, the invention provides a water quality variable concentration prediction method and a water quality variable concentration prediction system.
The technical scheme provided by the invention is as follows:
First aspect
The invention provides a water quality variable concentration prediction method, which comprises the following steps:
S1: acquiring a remote sensing image;
S2: preprocessing the remote sensing image;
s3: extracting image features of the preprocessed remote sensing image;
S4: and predicting the water quality variable concentration according to the image characteristics through a ConvLSTM regression model based on deep learning.
Second aspect
The invention provides a water quality variable concentration prediction system, which comprises: a processor and a memory for storing processor-executable instructions; the processor is configured to invoke the instructions stored in the memory to perform the water quality variable concentration prediction method of the first aspect.
Compared with the prior art, the technical scheme has at least the following beneficial effects:
According to the invention, a lake or river is covered by a remote sensing image, a complex nonlinear relation between image characteristics such as spectral reflectivity and water quality parameters is accurately captured by a ConvLSTM regression model based on deep learning, the water quality variable concentration is accurately estimated by the image characteristics such as spectral reflectivity, the dynamic change of the water quality of the lake or river is monitored, and reliable water quality variable parameters are provided for estimating the water quality of the lake or river.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly described below, and it is apparent that the drawings in the following description are only some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a schematic flow chart of a water quality variable concentration prediction method provided by the invention;
fig. 2 is a schematic structural diagram of a water quality variable concentration prediction system provided by the invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention more clear, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. It will be apparent that the described embodiments are some, but not all, embodiments of the invention. All other embodiments, which can be made by a person skilled in the art without creative efforts, based on the described embodiments of the present invention fall within the protection scope of the present invention.
Unless defined otherwise, technical or scientific terms used herein should be given the ordinary meaning as understood by one of ordinary skill in the art to which this invention belongs. The terms "first," "second," and the like, as used herein, do not denote any order, quantity, or importance, but rather are used to distinguish one element from another. Likewise, the terms "a," "an," or "the" and similar terms do not denote a limitation of quantity, but rather denote the presence of at least one. The word "comprising" or "comprises", and the like, means that elements or items preceding the word are included in the element or item listed after the word and equivalents thereof, but does not exclude other elements or items. The terms "connected" or "connected," and the like, are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect.
It should be noted that "upper", "lower", "left", "right", "front", "rear", and the like are used in the present invention only to indicate a relative positional relationship, and when the absolute position of the object to be described is changed, the relative positional relationship may be changed accordingly.
Referring to the accompanying figure 1 of the specification, a flow diagram of a water quality variable concentration prediction method provided by the invention is shown.
The embodiment of the invention provides a water quality variable concentration prediction method, which comprises the following steps:
s1: and acquiring a remote sensing image.
Specifically, a large-range remote sensing image can be acquired through a satellite. Landsat 8OLI, emitted by the national aerospace agency (NASA) at 11/2/2013, is one of the most commonly used satellites for water quality monitoring.
S2: and preprocessing the remote sensing image.
In one possible embodiment, the preprocessing includes geometric correction, radiometric correction, atmospheric correction. S2 specifically includes substeps S201 to S203:
S201: and performing geometric correction on the remote sensing image.
Wherein the geometric correction comprises: orthographic correction, geographic registration, and image registration.
Further, orthographic correction (Orthorectification) is a form of geometric correction that solves for terrain-induced parallax distortions by further correcting the image in consideration of terrain elevation information. This is particularly important in mountainous areas and areas where the topography is variable.
The geographic registration process enables the image to be accurately positioned on the earth surface by establishing a relation between the pixel coordinates of the remote sensing image and the geographic coordinates of the earth surface, and geographic position information can be directly obtained through the pixel coordinates of the image.
Image registration is used to align the remote sensing image with a geographic coordinate system so that the image can be accurately positioned on the earth's surface.
In the invention, the geometric correction can correct the image deformation caused by the geometric difference between the earth surface and the satellite orbit, thereby improving the accuracy of the remote sensing image on the map.
S202: and carrying out radiation calibration on the remote sensing image subjected to geometric correction.
In one possible implementation, S202 is specifically: radiation calibration was performed according to the following formula:
L＝Gain×DN+Bias
Wherein L represents the atmospheric top layer radiation brightness of a sensor spectrum channel, DN represents the gray value of a remote sensing image, gain represents the Gain of sensor calibration, and Bias represents the offset of the sensor calibration.
It should be noted that the radiation calibration converts the digital value (DN) of the remote sensing image into the actual radiation quantity, i.e. the atmospheric top radiation brightness (L) of the sensor spectrum channel. This makes the digital values in the image physically interpretable, reflecting the radiation characteristics of the earth's surface.
In the invention, the radiometric calibration ensures that remote sensing data acquired by different sensors at different times have consistent radiation units, so that comparison and analysis of the data of different sensors at different time points can be more easily performed, and the consistency and comparability of the data are improved.
S203: and carrying out atmospheric correction on the remote sensing image after radiation calibration.
In one possible implementation, S203 is specifically: atmospheric correction was performed according to the following formula:
Where ρ tg denotes the surface reflectance of the target, L ds denotes the radiance value at the sensor, L ts denotes the dark matter radiance at the sensor, E 0 ·d denotes the solar radiance at the top of the atmosphere, and θ 0 denotes the solar zenith angle.
In the invention, the atmospheric correction helps to eliminate the influence of the atmospheric layer on the remote sensing image. Atmospheric gases and particulate matter absorb, scatter and emit light, causing the brightness values in the image to be disturbed. By atmospheric correction, the true reflectivity of the target surface can be restored more accurately. Meanwhile, the atmospheric correction enables remote sensing data of different time and different places to be more comparable. This is important for performing time series analysis, monitoring environmental changes, and comparing studies of different areas.
S3: and extracting the image characteristics of the preprocessed remote sensing image.
In one possible embodiment, the image features include a plurality of water quality variables, specifically:
Wherein p represents a water quality variable, R 1 represents the reflectivity of a first wave band, R 2 represents the reflectivity of a second wave band, alpha, beta and gamma represent regression coefficients, and the first wave band and the second wave band are one wave band or a combination of a plurality of wave bands.
In the invention, the reflectance values of a plurality of wave bands can be used, so that the optical information of different wave bands can be comprehensively utilized, and the spectral characteristics of the earth surface can be more comprehensively captured. Meanwhile, the complex relationship between the water quality variable and the reflectivity can be better captured by adopting the nonlinear relationship, and the accuracy of prediction is improved.
In one possible embodiment, the image features include: coastal band reflectivity R 1, blue band reflectivity R 2, green band reflectivity R 3, red band reflectivity R 4, NIR band reflectivity R 5, SWIR1 band reflectivity R 6, SWIR2 band reflectivity R 7, karst Kunlun mountain band reflectivity R 9, reflectance ratio between green and red bandsReflectance ratio between red and green bands/>Reflectance ratio between green and blue bands/>Reflectance ratio between blue and green bands/>Reflectance ratio between red and blue bands/>Reflectance ratio between blue and red bands/>Reflectance ratio between red band and near infrared band/>Reflectivity ratio between near infrared and infrared bands/>Reflectance ratio between green band and near infrared bandReflectivity ratio between near infrared band and green band/>Reflectance ratio between blue band and near infrared band/>Reflectivity ratio between near infrared band and blue band/>Normalized difference/>, of green and red bandsNormalized difference of near infrared band and red band/>Normalized difference/>, of near infrared band and SWIR1 bandNormalized difference/>, of green band and near infrared bandNormalized difference of SWIR1 band and SWIR2 band/>Normalized difference of green band and SWIR1 band/>
In the present invention, the reflectance information of a plurality of wavelength bands is used, and the optical characteristics of the earth surface can be comprehensively considered. The spectral characteristics of the earth can be more fully described by combining the different bands of wavelengths that have different responses to the earth's surface object. Different wavebands have different spatial and optical resolutions, can provide multi-scale earth surface observation, and are very important for monitoring earth surface changes and features at different scales. Meanwhile, indexes such as normalized difference indexes and reflectivity ratios are powerful tools for further analyzing the surface features. Through the indexes, the characteristics of different ground features such as water bodies and the like can be better captured, and the method has important significance for environmental monitoring and ecological research.
Further, the specific calculation mode of the reflectivity of each wave band is as follows: reflectance values represented by respective pixels of respective bands are averaged.
In the present invention, by taking the average value of each band, it is useful to reduce the influence due to noise or abnormal value in a certain band. The calculation of the average value can smooth the data to a certain extent, and the stability of the whole data is improved.
S4: and predicting the water quality variable concentration according to the image characteristics through ConvLSTM regression model based on deep learning.
Wherein ConvLSTM (convolutional long short-term memory network) is a neural network model that combines Convolutional Neural Networks (CNN) and long-term memory networks (LSTM). ConvLSTM are widely used to process sequence data with spatio-temporal information.
Further, LSTM is a variant of Recurrent Neural Network (RNN) specifically designed for processing sequence data. CNN is a neural network specifically designed for image processing, and spatial features in images are effectively extracted through structures such as convolution layers and pooling layers. ConvLSTM introduce convolution operations at each time step of the LSTM to allow the model to capture spatial structures in the time series.
Specifically, by stacking multiple recursive ConvLSTM layers (similar to the layers in LSTM, but with internal matrix multiplication exchanged) and convolution operations, the algorithm includes convolution structures in both the input to state and state-to-state transitions.
In one possible embodiment, S5 specifically includes:
s401: the image feature sequence X is input into a ConvLSTM regression model based on deep learning.
S402: determining hidden states at all moments through ConvLSTM regression models:
Ct'＝tanh(WXC*Xt+WHC*Ht-1+bC)
Wherein I t represents the activation output vector of the input gate at time t, σ () represents the activation function, W XI represents the weight matrix between the image feature sequence and the input gate, X t represents the image feature sequence at time t, W HI represents the weight matrix between the hidden state and the input gate, H t-1 represents the hidden state at time t-1, W CI represents the weight matrix between the cell storage unit and the input gate, C t-1 represents the activation output vector of the cell storage unit at time t-1, b I denotes a bias term of an input gate, F t denotes an activation output vector of a forgetting gate at time t, W XF denotes a weight matrix between an image feature sequence and the forgetting gate, W HF denotes a weight matrix between a hidden state and the forgetting gate, W CF denotes a weight matrix between a cell storage unit and the forgetting gate, b F denotes a bias term of the forgetting gate, C t denotes an activation output vector of a cell storage unit at time t, C t' denotes a candidate output vector of a cell storage unit at time t, tan h () represents a tanh activation function, W XC represents a weight matrix between the image feature sequence and the cell storage unit, W HC represents a weight matrix between the hidden state and the cell storage unit, b C represents a bias term of the cell storage unit, O t represents an activation output vector of the output gate at time t, W XO represents a weight matrix between the image feature sequence and the output gate, W HO represents a weight matrix between the hidden state and the output gate, W CO represents a weight matrix between the cell storage unit and the output gate, b O denotes the bias term of the output gate, H t denotes the hidden state at time t, x denotes the convolution operation, Representing the hadamard product operation.
The Hadamard Product refers to the operation of multiplying corresponding elements in a matrix or vector with two identical dimensions one by one.
It should be noted that, an input gate, a forget gate and an output gate are introduced, and through these gating mechanisms, the flow of information can be better controlled, which is helpful to selectively retain and forget the previous information, so that the model can adapt to the input of different time steps. Past information is maintained and updated by memory cells. The updating of the memory cells is achieved by the activation of the input gates and the candidate outputs and forgetting gates, enabling the model to effectively memorize and forget information in different time steps. Convolution operation is introduced to enable the model to process spatial information of the input image feature sequence. The convolution operation helps to extract spatial features and better capture structural information in the image.
S403: and determining the predicted value of the water quality variable concentration according to the hidden state at each moment.
In one possible implementation, S403 is specifically: determining a predicted value of the water quality variable concentration according to the following formula:
dt＝σ(WDHHt+bD)
Wherein d t represents a predicted value of the water quality variable concentration at time t, W DH represents a predicted weight matrix of the water quality variable concentration, and b D represents a predicted bias term of the water quality variable concentration.
According to the invention, by using the hidden states at all times, the model can fully utilize time sequence information to capture the dynamic change of the water quality variable concentration along with time, and is more applicable to water quality data with time sequence property, so that the model can better understand and predict the change trend of the variable, and the accuracy of the prediction of the water quality variable concentration is improved.
Further, smaller input scales and a modest number of convolution layers may improve model performance. The number of layers is between 2 and 7, which is generally the ideal choice for accurate modeling results. The input core size is set by optimization in consideration of the spatial resolution and the calculation cost of the remote sensing image. The overall structure of the designed regression model consists of three convolution layers and three full connection layers. There is a discard layer between the three convolutional layers and the full link layer as a control method to prevent network overfitting by randomly discarding information between the layers. Within each convolution layer we use batch normalization and ReLU activation functions. Through these structural layers, the spectral features are ultimately converted to an estimated water quality concentration.
Compared with the prior art, the technical scheme has at least the following beneficial effects:
According to the invention, a lake or river is covered by a remote sensing image, a complex nonlinear relation between image characteristics such as spectral reflectivity and water quality parameters is accurately captured by a ConvLSTM regression model based on deep learning, the water quality variable concentration is accurately estimated by the image characteristics such as spectral reflectivity, the dynamic change of the water quality of the lake or river is monitored, and reliable water quality variable parameters are provided for estimating the water quality of the lake or river.
Referring to fig. 2 of the specification, a schematic structural diagram of a water quality variable concentration prediction system provided by the invention is shown.
The present invention also provides a water quality variable concentration prediction system 20, comprising: a processor 201 and a memory 202 for storing instructions executable by the processor 201. The processor 201 is configured to call the instructions stored in the memory 202 to perform the water quality variable concentration prediction method described above.
The water quality variable concentration prediction system 20 provided by the invention can execute the water quality variable concentration prediction method and achieve the same or similar technical effects, and in order to avoid repetition, the invention is not repeated.
Compared with the prior art, the technical scheme has at least the following beneficial effects:
According to the invention, a lake or river is covered by a remote sensing image, a complex nonlinear relation between image characteristics such as spectral reflectivity and water quality parameters is accurately captured by a ConvLSTM regression model based on deep learning, the water quality variable concentration is accurately estimated by the image characteristics such as spectral reflectivity, the dynamic change of the water quality of the lake or river is monitored, and reliable water quality variable parameters are provided for estimating the water quality of the lake or river.
The following points need to be described:
(1) The drawings of the embodiments of the present invention relate only to the structures related to the embodiments of the present invention, and other structures may refer to the general designs.
(2) In the drawings for describing embodiments of the present invention, the thickness of layers or regions is exaggerated or reduced for clarity, i.e., the drawings are not drawn to actual scale. It will be understood that when an element such as a layer, film, region or substrate is referred to as being "on" or "under" another element, it can be "directly on" or "under" the other element or intervening elements may be present.
(3) The embodiments of the invention and the features of the embodiments can be combined with each other to give new embodiments without conflict.
The present invention is not limited to the above embodiments, but the scope of the invention is defined by the claims.
Claims (10)
1. A water quality variable concentration prediction method is characterized by comprising the following steps:
S1: acquiring a remote sensing image;
S2: preprocessing the remote sensing image;
s3: extracting image features of the preprocessed remote sensing image;
S4: and predicting the water quality variable concentration according to the image characteristics through a ConvLSTM regression model based on deep learning.
2. The method for predicting the concentration of a water quality variable according to claim 1, wherein the preprocessing comprises geometric correction, radiation calibration, atmospheric correction; the step S2 specifically comprises the following steps:
S201: performing geometric correction on the remote sensing image, wherein the geometric correction comprises: orthographic correction, geographic registration, and image registration;
S202: performing radiation calibration on the remote sensing image subjected to geometric correction;
s203: and carrying out atmospheric correction on the remote sensing image after radiation calibration.
3. The method for predicting the concentration of water quality variables according to claim 2, wherein S202 specifically comprises:
Radiation calibration was performed according to the following formula:
L＝Gain×DN+Bias
Wherein L represents the atmospheric top layer radiation brightness of a sensor spectrum channel, DN represents the gray value of a remote sensing image, gain represents the Gain of sensor calibration, and Bias represents the offset of the sensor calibration.
4. The method for predicting the concentration of water quality variables according to claim 2, wherein S203 specifically comprises:
Atmospheric correction was performed according to the following formula:
Where ρ tg denotes the surface reflectance of the target, L ds denotes the radiance value at the sensor, L ts denotes the dark matter radiance at the sensor, E 0 ·d denotes the solar radiance at the top of the atmosphere, and θ 0 denotes the solar zenith angle.
5. The method for predicting the concentration of a water quality variable according to claim 1, wherein the image features comprise a plurality of water quality variables, and the water quality variables are specifically:
Wherein p represents a water quality variable, R 1 represents the reflectivity of a first wave band, R 2 represents the reflectivity of a second wave band, alpha, beta and gamma represent regression coefficients, and the first wave band and the second wave band are a wave band or a combination of a plurality of wave bands.
6. The method for predicting water quality variable concentration of claim 1, wherein the image features comprise: coastal band reflectivity R 1, blue band reflectivity R 2, green band reflectivity R 3, red band reflectivity R 4, NIR band reflectivity R 5, SWIR1 band reflectivity R 6, SWIR2 band reflectivity R 7, karst Kunlun mountain band reflectivity R 9, reflectance ratio between green and red bandsReflectance ratio between red and green bands/>Reflectance ratio between green and blue bands/>Reflectance ratio between blue and green bands/>Reflectance ratio between red and blue bands/>Reflectance ratio between blue and red bands/>Reflectance ratio between red band and near infrared band/>Reflectivity ratio between near infrared and infrared bands/>Reflectivity ratio between green band and near infrared band/>Reflectivity ratio between near infrared band and green band/>Reflectance ratio between blue band and near infrared band/>Reflectance ratio between near infrared and blue bandsNormalized difference/>, of green and red bandsNormalized difference between near infrared band and red bandNormalized difference/>, of near infrared band and SWIR1 bandNormalized difference/>, of green band and near infrared bandNormalized difference of SWIR1 band and SWIR2 band/>Normalized difference of green band and SWIR1 band/>
7. The method for predicting the concentration of water quality variables according to claim 6, wherein the specific calculation mode of the reflectivity of each wave band is as follows: reflectance values represented by respective pixels of respective bands are averaged.
8. The method for predicting the concentration of a water quality variable according to claim 1, wherein the step S4 specifically comprises:
s401: inputting an image feature sequence X into a ConvLSTM regression model based on deep learning;
S402: determining hidden states at all moments through ConvLSTM regression models:
C′t＝tanh(WXC*Xt+WHC*Ht-1+bC)
Wherein I t represents the activation output vector of the input gate at time t, σ () represents the activation function, W XI represents the weight matrix between the image feature sequence and the input gate, X t represents the image feature sequence at time t, W HI represents the weight matrix between the hidden state and the input gate, H t-1 represents the hidden state at time t-1, W CI represents the weight matrix between the cell storage unit and the input gate, C t-1 represents the activation output vector of the cell storage unit at time t-1, b I denotes a bias term of an input gate, F t denotes an activation output vector of a forgetting gate at time t, W XF denotes a weight matrix between an image feature sequence and the forgetting gate, W HF denotes a weight matrix between a hidden state and the forgetting gate, W CF denotes a weight matrix between a cell storage unit and the forgetting gate, b F denotes a bias term of the forgetting gate, C t denotes an activation output vector of a cell storage unit at time t, C' t denotes a candidate output vector of a cell storage unit at time t, tan h () represents a tanh activation function, W XC represents a weight matrix between the image feature sequence and the cell storage unit, W HC represents a weight matrix between the hidden state and the cell storage unit, b C represents a bias term of the cell storage unit, O t represents an activation output vector of the output gate at time t, W XO represents a weight matrix between the image feature sequence and the output gate, W HO represents a weight matrix between the hidden state and the output gate, W CO represents a weight matrix between the cell storage unit and the output gate, b O denotes the bias term of the output gate, H t denotes the hidden state at time t, x denotes the convolution operation, Representing Hadamard product operation;
S403: and determining the predicted value of the water quality variable concentration according to the hidden state at each moment.
9. The method for predicting water quality variable concentration according to claim 8, wherein S403 specifically comprises:
determining a predicted value of the water quality variable concentration according to the following formula:
dt＝σ(WDHHt+bD)
Wherein d t represents a predicted value of the water quality variable concentration at time t, W DH represents a predicted weight matrix of the water quality variable concentration, and b D represents a predicted bias term of the water quality variable concentration.
10. A water quality variable concentration prediction system, comprising a processor and a memory for storing instructions executable by the processor; the processor is configured to invoke the instructions stored in the memory to perform the water quality variable concentration prediction method of any one of claims 1 to 9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410174923.2A CN118115824A (en) | 2024-02-07 | 2024-02-07 | Water quality variable concentration prediction method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410174923.2A CN118115824A (en) | 2024-02-07 | 2024-02-07 | Water quality variable concentration prediction method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN118115824A true CN118115824A (en) | 2024-05-31 |
Family
ID=91208020
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410174923.2A Pending CN118115824A (en) | 2024-02-07 | 2024-02-07 | Water quality variable concentration prediction method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118115824A (en) |
-
2024
- 2024-02-07 CN CN202410174923.2A patent/CN118115824A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112836610B (en) | Land use change and carbon reserve quantitative estimation method based on remote sensing data | |
| CN113128134B (en) | Mining area ecological environment evolution driving factor weight quantitative analysis method | |
| Tran Anh et al. | Downscaling rainfall using deep learning long short‐term memory and feedforward neural network | |
| Matsunobu et al. | Cloud detection using convolutional neural networks on remote sensing images | |
| CN112949414B (en) | Intelligent surface water body drawing method for wide-vision-field high-resolution six-satellite image | |
| Imani | Particulate matter (PM2. 5 and PM10) generation map using MODIS Level-1 satellite images and deep neural network | |
| CN114297938B (en) | Inversion method of optical shallow water depth based on neural network | |
| CN113553697B (en) | Long-time-sequence multi-source data-based vegetation disturbance analysis method for coal mining | |
| CN114813651A (en) | Remote sensing water quality inversion method combining difference learning rate and spectrum geometric characteristics | |
| CN114494821A (en) | Remote sensing image cloud detection method based on feature multi-scale perception and self-adaptive aggregation | |
| CN116148188A (en) | Air-space-ground integrated lake water quality tracing method, system, equipment and storage medium | |
| CN114120036A (en) | Lightweight remote sensing image cloud detection method | |
| CN113989612A (en) | Remote sensing image target detection method based on attention and generation countermeasure network | |
| CN117035066A (en) | Ground surface temperature downscaling method coupling geographic weighting and random forest | |
| CN116977311A (en) | Flood disaster area detection method, system, electronic equipment and storage medium | |
| Liu et al. | Bi-LSTM model for time series leaf area index estimation using multiple satellite products | |
| CN116879192B (en) | Water bloom prediction method, device, equipment and medium based on satellite remote sensing data | |
| CN116721243B (en) | Deep learning atmosphere correction method and system based on spatial spectrum feature constraint | |
| CN117274831B (en) | Offshore turbid water body depth inversion method based on machine learning and hyperspectral satellite remote sensing image | |
| Mercier et al. | Vision transformer models to measure solar irradiance using sky images in temperate climates | |
| CN118115824A (en) | Water quality variable concentration prediction method and system | |
| CN116070132A (en) | Method for predicting transparency of seawater and sea surface temperature based on multi-source remote sensing data | |
| CN114296061A (en) | Cross calibration method based on multivariate variable detection and different radiation transmission models | |
| CN113657351A (en) | High-resolution remote sensing image forest and grass change detection device and method based on deep learning | |
| Gu et al. | A downscaling-calibrating framework for generating gridded daily precipitation estimates with a high spatial resolution |
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 |