Algae monitoring method, device, equipment and medium based on image analysis

Abstract

The invention relates to an algae monitoring method, device, equipment and medium based on image analysis, and belongs to the field of Beidou and artificial intelligence technology. The algae monitoring method based on image analysis comprises acquiring float monitoring data and remote sensing image data, performing spatiotemporal fusion on the float monitoring data and the remote sensing image data to construct a water quality image spatiotemporal data set, wherein the water quality image spatiotemporal data set includes an algae coverage index, training an LSTM network constructed based on the water quality image spatiotemporal data set to obtain an algae spread model, predicting float monitoring data to be predicted based on the algae spread model to obtain a predicted value of the algae coverage index; presetting an early warning threshold, and giving an early warning to algae reproduction based on the predicted value of the algae coverage index and the early warning threshold, thereby improving the accuracy of model prediction.

Description

Technical Field

[0001] The present invention relates to the fields of Beidou and artificial intelligence, and in particular to an algae monitoring method, device, equipment and medium based on image analysis. Background Art

[0002] With the continuous development of human society and the accelerated advancement of industrialization, water pollution has become one of the major environmental challenges facing the world today. Among the many forms of water pollution, the outbreak of algae blooms is particularly noteworthy. Due to its frequency and widespread occurrence, it has become a common environmental problem worldwide. Algae blooms are mainly composed of microscopic algae such as cyanobacteria, green algae, and diatoms. Under suitable environmental conditions, these algae will multiply rapidly and form large-scale blooms. This phenomenon is not only common in freshwater bodies, but also occurs frequently in some marine areas. Large-scale algae blooms will consume oxygen in the water, causing hypoxia or even oxygen enrichment in the water, seriously affecting the survival and reproduction of aquatic organisms. Some algae will release toxins during their growth, posing a direct threat to aquatic organisms and human health, causing water source pollution incidents and even food chain poisoning incidents. In addition, algae blooms will also affect water quality safety, reduce the availability of water bodies, and bring inconvenience and losses to human production and life.

[0003] At present, the prediction method for algae is mainly to combine environmental factors and water quality factors for analysis and prediction.

[0004] However, existing algae prediction methods only analyze water quality and environmental factors without verifying them with actual image data, and their accuracy needs to be further improved. Summary of the invention

[0005] In view of this, it is necessary to provide an algae monitoring method, device, equipment and medium based on image analysis to solve the technical problem that algae prediction is not combined with actual image data, resulting in low accuracy.

[0006] In order to solve the above problems, the present invention provides an algae monitoring method based on image analysis, comprising:

[0007] Acquire floating monitoring data and remote sensing image data, and perform spatiotemporal fusion on the floating monitoring data and the remote sensing image data to construct a spatiotemporal data set of water quality images, wherein the spatiotemporal data set of water quality images includes an algae coverage index;

[0008] Based on the water quality image spatiotemporal data set, the constructed LSTM network is trained to obtain an algae spread model, and based on the algae spread model, the floating monitoring data to be predicted is predicted to obtain a predicted value of the algae coverage index;

[0009] An early warning is given to algae reproduction based on the predicted value of the algae coverage index and the early warning threshold.

[0010] In a possible implementation, the spatiotemporal fusion of the floating monitoring data and the remote sensing image data to construct a spatiotemporal data set of water quality images includes:

[0011] Acquire float monitoring data in real time through float equipment, and upload the float monitoring data to the data center through Beidou short message, wherein the float monitoring data includes float positioning data and sensor data, and the sensor data includes water temperature, nitrogen content, phosphorus content, oxygen content, pH value and water depth;

[0012] The float monitoring data is sorted in time series to obtain time-ordered data, the time-ordered data is parsed to obtain the longitude and latitude information of the float and the corresponding water quality parameters, and then the longitude and latitude information and water quality parameters are sorted in time series to obtain a time-position-water quality parameter sequence;

[0013] Correcting the remote sensing image data to obtain a corrected image sequence;

[0014] The corrected image sequence and the time-position-water quality parameter sequence are subjected to spatiotemporal fusion, and the algae coverage index is determined based on the remote sensing image data to construct a spatiotemporal data set of water quality images.

[0015] In a possible implementation manner, determining the algae coverage index based on the remote sensing image data includes:

[0016] Determine a sensitive band combination of algae according to the growth characteristics of algae, wherein the sensitive band combination includes at least one or more of a green band, a red band and a near infrared band;

[0017] The remote sensing image data is analyzed based on the sensitive band combination to extract a spectral index, and the algae coverage index is determined based on the spectral index, wherein the spectral index includes a normalized difference water index, an enhanced water index and a normalized water volume fraction index.

[0018] In a possible implementation, the calculation formula of the algae coverage index is:

[0019] ,

[0020] ,

[0021] ,

[0022] ,

[0023] in, is the algae cover index, is the normalized difference water index, is the enhanced water index. is the normalized water volume fraction index, , , is the diversity index, For the green band, is the red band, The near-infrared band, , , , , , is the adjustment coefficient.

[0024] In a possible implementation, the LSTM network constructed based on the water quality image spatiotemporal dataset is trained to obtain the algae spread model, including:

[0025] The water quality image spatiotemporal dataset is divided into a training set, a validation set and a test set, wherein based on the training set, the sensor data and the algae coverage index are combined to obtain combined data, the combined data is used as an input factor and the algae coverage index is used as an output factor, and an LSTM network is constructed;

[0026] Training the LSTM network through the training set and adjusting the parameters of the LSTM network, wherein the LSTM network includes a forget gate, an input gate, a candidate memory, a unit state, an output gate, and a hidden state;

[0027] The LSTM network is evaluated using the validation set to obtain an algae spread model.

[0028] In a possible implementation, the calculation formula of the algae spread model is:

[0029] ,

[0030] in, is the algae cover index, For sensor data, is the weight parameter, is the bias term, is the activation function;

[0031] The calculation formula of the forget gate is:

[0032] ,

[0033] in, is the output of the forget gate at the current moment, is the sigmoid function, is the weight matrix of the forget gate, is the bias vector of the forget gate, is the hidden state at the previous moment, is the input at the current moment, for the current moment;

[0034] The calculation formula of the input gate is:

[0035] ,

[0036] in, is the output of the input gate at the current moment, is the weight matrix of the input gate, is the bias vector of the input gate;

[0037] The calculation formula of the candidate memory is:

[0038] ,

[0039] in, is the time step Candidate memory cells is the weight matrix, is the bias vector, is the activation function;

[0040] The calculation formula of the unit state is:

[0041] ,

[0042] in, is the unit state at the current moment, is the unit state at the previous moment;

[0043] The calculation formula of the output gate is:

[0044] ,

[0045] in, is the output of the output gate at the current moment, is the weight matrix of the output gate, is the bias vector of the output gate;

[0046] The calculation formula of the hidden state is:

[0047] ,

[0048] in, is the hidden state at the current moment.

[0049] In a possible implementation, predicting the floating monitoring data to be predicted based on the algae spread model to obtain a predicted value of the algae coverage index includes:

[0050] The float monitoring data is acquired in real time through Beidou messages, the float monitoring data is input into the algae spread model, the float monitoring data is predicted based on the algae spread model to obtain the predicted value of the algae coverage index, and the algae spread distribution image is obtained based on the predicted value of the algae coverage index.

[0051] On the other hand, the present invention also provides an algae monitoring device based on image analysis, comprising:

[0052] A data spatiotemporal fusion module is used to obtain floating monitoring data and remote sensing image data, and to perform spatiotemporal fusion on the floating monitoring data and remote sensing image data to construct a spatiotemporal data set of water quality images, wherein the spatiotemporal data set of water quality images includes an algae coverage index;

[0053] A prediction module is used to train the constructed LSTM network based on the water quality image spatiotemporal data set to obtain an algae spread model, and to predict the floating monitoring data to be predicted based on the algae spread model to obtain a predicted value of the algae coverage index;

[0054] The early warning module is used to warn of algae reproduction based on the predicted value of the algae coverage index and the early warning threshold.

[0055] In another aspect, the present invention further provides an electronic device, comprising: a processor and a memory;

[0056] The memory stores a computer-readable program executable by the processor;

[0057] When the processor executes the computer-readable program, the steps in the algae monitoring method based on image analysis as described above are implemented.

[0058] On the other hand, the present invention also provides a computer-readable storage medium, which stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps in the algae monitoring method based on image analysis as described above.

[0059] The beneficial effects of the present invention are as follows: acquiring float monitoring data and remote sensing image data, collecting monitoring data by floats, reducing costs, uploading monitoring data to a data center through Beidou short messages, ensuring the reliability and stability of data transmission, performing spatiotemporal fusion of float monitoring data and remote sensing image data to construct a spatiotemporal data set of water quality images, training the constructed LSTM network based on the spatiotemporal data set of water quality images, obtaining an algae spread model, predicting the float monitoring data to be predicted based on the algae spread model, obtaining a predicted value of the algae coverage index, and avoiding water pollution incidents caused by the spread of algae by predicting the spread of algae in real time and issuing early warnings, and deeply mining the fused data through the spatiotemporal fusion of float monitoring data and remote sensing image data, and using image analysis technology and machine learning algorithms to identify and characterize the patterns and laws of algae reproduction and spread, thereby improving the accuracy of model prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 A flow chart of an embodiment of an algae monitoring method based on image analysis provided by the present invention;

[0061] Figure 2 A schematic diagram of the LSTM network hierarchy of the algae monitoring method based on image analysis provided by the present invention;

[0062] Figure 3 A schematic diagram of input layer data elements of the algae monitoring method based on image analysis provided by the present invention;

[0063] Figure 4 A schematic structural diagram of an embodiment of an algae monitoring device based on image analysis provided by the present invention;

[0064] Figure 5 A schematic structural diagram of an embodiment of an electronic device provided by the present invention. DETAILED DESCRIPTION

[0065] The preferred embodiments of the present invention are described in detail below in conjunction with the accompanying drawings, wherein the accompanying drawings constitute a part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, but are not used to limit the scope of the present invention.

[0066] The present invention discloses an algae monitoring method, device, equipment and medium based on image analysis, which can be used in a computer. The method, device or computer-readable storage medium involved in the present invention can be integrated with the above-mentioned device or can be relatively independent.

[0067] A specific embodiment of the present invention discloses an algae monitoring method based on image analysis, which can be executed by a computer, specifically by one or more processors of the computer. Figure 1As shown, the algae monitoring methods based on image analysis include:

[0068] S101, acquiring float monitoring data and remote sensing image data, and performing spatiotemporal fusion on the float monitoring data and the remote sensing image data to construct a water quality image spatiotemporal data set, wherein the water quality image spatiotemporal data set includes an algae coverage index;

[0069] S102, training the constructed LSTM network based on the water quality image spatiotemporal data set to obtain an algae spread model, predicting the floating monitoring data to be predicted based on the algae spread model to obtain a predicted value of the algae coverage index;

[0070] S103. Issue an early warning for algae growth based on the predicted value of the algae coverage index and the early warning threshold.

[0071] Among them, the LSTM network can capture the long-term dependencies in the time series. By training the LSTM network to determine the relationship between the monitoring data and the algae coverage index, it can effectively deal with the temporal changes of algae, such as the effects of water temperature, light, and nutrient factors on algae growth, and better capture the dynamic characteristics of algae growth and spread.

[0072] Compared with the prior art, the algae monitoring method based on image analysis provided in this embodiment obtains float monitoring data and remote sensing image data, collects monitoring data through floats, reduces costs, uploads monitoring data to the data center through Beidou short messages, ensures the reliability and stability of data transmission, and performs spatiotemporal fusion of float monitoring data and remote sensing image data to construct a water quality image spatiotemporal data set, wherein the water quality image spatiotemporal data set includes an algae coverage index, realizes spatial alignment of float monitoring data and remote sensing image data, trains the constructed LSTM network based on the water quality image spatiotemporal data set to obtain an algae spread model, predicts the float monitoring data to be predicted based on the algae spread model, and obtains the predicted value of the algae coverage index; trains the model with the fused data to improve the accuracy of model prediction, presets an early warning threshold, and warns of algae reproduction based on the predicted value of the algae coverage index and the early warning threshold, and predicts the spread of algae in real time and issues an early warning to avoid water pollution caused by the spread of algae.

[0073] In some embodiments, in step S101, float monitoring data is obtained, and the float monitoring data is obtained in real time through the float equipment, and the float monitoring data is uploaded to the data center through the Beidou short message, wherein the float monitoring data includes float positioning data and sensor data, and the sensor data includes water temperature, nitrogen content, phosphorus content, oxygen content, pH value and water depth, and remote sensing image data is obtained by regularly scheduling drones to perform tasks over the target water body to obtain remote sensing image data; the float monitoring data and the remote sensing image data are spatiotemporally fused to construct a water quality image spatiotemporal data set. The specific process is: first, the float monitoring data is time-series sorted to obtain time-ordered data, the time-ordered data is parsed, and after obtaining the longitude and latitude information of the float and the corresponding water quality parameters, the longitude and latitude information and water quality parameters are time-series sorted to obtain a time-position-water quality parameter sequence, and its time-position-water quality parameter sequence The calculation formula is:

[0074] ,

[0075] in, is the time-location-water quality parameter sequence, i.e., the water quality vector parameter, is the nitrogen content, is the phosphorus content, is the oxygen content, is the pH value, is the water temperature, For water depth, To monitor the time point, is the longitude, is latitude;

[0076] Secondly, the remote sensing image data is corrected to obtain the corrected image sequence The correction usually adopts radiation correction, geometric correction and atmospheric correction to eliminate the influence of atmospheric scattering and terrain shadow on the image, improve the clarity and accuracy of the image, perform spatiotemporal fusion of the corrected image sequence and the time-location-water quality parameter sequence, and determine the algae coverage index based on remote sensing image data to construct a spatiotemporal data set of water quality images. , where the water quality image spatiotemporal dataset includes the algae coverage index, that is, based on the longitude and latitude coordinates of the float, the position of the float is marked on the remote sensing image to achieve spatial alignment of the float monitoring data and the remote sensing image data. After spatial alignment, the algae coverage index is determined based on the remote sensing image data, and the algae coverage index of the float point can be determined. The calculation formula of the water quality image spatiotemporal dataset is:

[0077] ,

[0078] in, is a spatiotemporal dataset of water quality images. is the spatial range of the target water body, The monitoring period is It is a universal quantifier.

[0079] The algae coverage index is determined based on remote sensing image data, and the sensitive band combination of algae is determined according to the growth characteristics of algae, that is, the most sensitive band combination is selected, wherein the sensitive band combination includes at least one or more of the green band, the red band and the near-infrared band. The remote sensing image data is analyzed based on the sensitive band combination, and a specific band combination is used to identify and analyze the algae coverage in the water body to extract the spectral index. The algae coverage index is determined based on the spectral index, wherein the spectral index includes the normalized difference water index, the enhanced water index and the normalized water volume fraction index. The calculation formula of the normalized difference water index is:

[0080] ,

[0081] in, is the normalized difference water index, For the green band, is the red band, The near-infrared band, , is the adjustment coefficient;

[0082] The calculation formula of the enhanced water index is:

[0083] ,

[0084] in, is the enhanced water index. , is the adjustment coefficient;

[0085] The calculation formula of the normalized water volume fraction index is:

[0086] ,

[0087] in, is the normalized water volume fraction index, , is the adjustment coefficient;

[0088] The calculation formula of the algae coverage index is:

[0089] ,

[0090] in, is the algae cover index, , , is the diversity index, is the reflectivity or intensity in the near-infrared band, The reflectivity or intensity of the red band, Used to adjust the proportion of near-infrared band and red band.

[0091] In some embodiments, in step S102, the LSTM network is trained based on the water quality image spatiotemporal dataset to obtain an algae spread model. First, the water quality image spatiotemporal dataset is divided into a training set, a validation set and a test set, that is, the water quality image spatiotemporal dataset is divided into a training set a1, a validation set a2 and a test set a3 according to the ratio of 75%, 15%, and 15%. In the training set, the collected sensor data including water temperature, phosphorus content, nitrogen content, oxygen content, pH value, water depth, longitude and latitude, timestamp and algae coverage index of the current point are used as the basic neuron unit of the training set. The sensor data and the algae coverage index in the water quality image spatiotemporal dataset are combined to obtain combined data. The combined data is used as an input factor and the algae coverage index is used as an output factor to construct an LSTM network. Please refer to the schematic diagram of the LSTM network hierarchy. Figure 2 ,like Figure 2 As shown in the figure, the LSTM network includes an input layer, an LSTM layer, a hidden layer, and an output layer. The input layer receives a sample sequence, and its sample sequence is composed of data elements of the input layer. For a schematic diagram of the input layer data elements, please refer to Figure 3 ,like Figure 3 As shown in the figure, the input layer data elements are composed of sensor data from multiple floating points collected on different dates and the algae coverage index calculated by remote sensing images; secondly, the LSTM network is trained through the training set, and the parameters of the LSTM network are adjusted. The LSTM network includes an input gate, a forget gate, a candidate memory, a unit state, an output gate, and a hidden state. The calculation formula of the input gate is:

[0092] ,

[0093] in, is the output of the input gate at the current moment, is the sigmoid function, is the input at the current moment, is the hidden state at the previous moment, is the weight matrix of the input gate, is the bias vector of the input gate, is the current moment, and the calculation process of its input gate is: the input at the current moment and the hidden state at the previous moment Passed into the input gate, through the weight matrix Perform linear changes and add a bias vector , use the sigmoid function to linearly change the result and get the output at the current moment ;

[0094] The calculation formula of the forget gate is:

[0095] ,

[0096] in, is the output of the forget gate at the current moment, is the weight matrix of the forget gate, is the bias vector of the forget gate;

[0097] The calculation formula of the candidate memory is:

[0098] ,

[0099] in, is the time step Candidate memory cells is the weight matrix, is the bias vector, is the activation function;

[0100] The calculation formula of the unit state is:

[0101] ,

[0102] in, is the unit state at the current moment, is the unit state at the previous moment;

[0103] The calculation formula of the output gate is:

[0104] ,

[0105] in, is the output of the output gate at the current moment, is the weight matrix of the output gate, is the bias vector of the output gate;

[0106] The calculation formula of the hidden state is:

[0107] ,

[0108] in, is the hidden state at the current moment;

[0109] In the process of training the LSTM network with the training set, at the current time step, after the LSTM layer receives the input of the input layer and the unit state of the previous time step, it calculates and outputs the unit state and hidden state of the current time step through the gating mechanism (including the forget gate, input gate and output gate) and the unit state update. The hidden layer processes the input and unit state through the gating mechanism, updates the unit state and calculates the hidden state of the current time step. The output layer receives the input of the hidden layer and completes the calculated value output. At this time, after the training set completes a sample training, the LSTM network is evaluated through the validation set, and the data of the next time step is used as the result set to calibrate and optimize the LSTM network. Through iterative training, the algae spread model is obtained. The calculation formula of the algae spread model is:

[0110] ,

[0111] in, is the algae cover index, For sensor data, is the weight parameter, is the bias term, is the activation function.

[0112] Based on the algae spread model, the float monitoring data to be predicted is predicted to obtain the predicted value of the algae coverage index. The float monitoring data is acquired in real time through Beidou messages, and the float monitoring data is input into the algae spread model. Based on the algae spread model, the float monitoring data is predicted to obtain the predicted value of the algae coverage index. Based on the predicted value of the algae coverage index, an algae spread distribution image is obtained, that is, the predicted value of the algae coverage index at each float positioning point is displayed from low to high according to different time periods, and the algae spread distribution image is obtained. The predicted value of the algae coverage index is in the range of 0-1, with 0.1 being low and 0.9 being high.

[0113] In some embodiments, in step S103, an early warning is issued for algae reproduction based on the predicted value of the algae coverage index and the early warning threshold. The algae coverage index in the next three days, one week, and one month is predicted in real time through the algae spread model. When the algae coverage index exceeds the early warning threshold, an early warning is issued. The early warning threshold is flexibly set according to different regions. For example, in area a, 0-0.1 is the normal range, 0.1-0.3 is the attention level, 0.3-0.6 is the orange warning, and more than 0.6 is the red warning. In area b, 0-0.2 is the normal range, 0.2-0.4 is the attention level, 0.4-0.6 is the orange warning, and more than 0.6 is the red warning. In area a, when the predicted value of the algae coverage index is between 0.3-0.6, an orange warning is issued. In area b, when the predicted value of the algae coverage index is between 0.4-0.6, an orange warning is issued.

[0114] In order to better implement the algae monitoring method based on image analysis in the embodiment of the present invention, based on the algae monitoring method based on image analysis, correspondingly, Figure 4 As shown, the embodiment of the present invention further provides an algae monitoring device based on image analysis. The algae monitoring device based on image analysis 400 includes:

[0115] The data spatiotemporal fusion module 401 is used to obtain the floating monitoring data and the remote sensing image data, and to perform spatiotemporal fusion on the floating monitoring data and the remote sensing image data to construct a spatiotemporal data set of water quality images, wherein the spatiotemporal data set of water quality images includes an algae coverage index;

[0116] The prediction module 402 is used to train the constructed LSTM network based on the spatiotemporal data set of water quality images to obtain an algae spread model, and to predict the floating monitoring data to be predicted based on the algae spread model to obtain a predicted value of the algae coverage index;

[0117] The early warning module 403 is used to warn of algae reproduction based on the predicted value of the algae coverage index and the early warning threshold.

[0118] like Figure 5 As shown, the present invention also provides an electronic device 500 , which can be a computing device such as a mobile terminal, a desktop computer, a notebook, a palmtop computer, a server, etc. The electronic device 500 includes a processor 501 , a memory 502 , and a display 503 . Figure 5 Only some components of the electronic device 500 are shown, but it should be understood that it is not required to implement all of the components shown, and more or fewer components may be implemented instead.

[0119] In some embodiments, the memory 502 may be an internal storage unit of the electronic device 500, such as a hard disk or memory of the electronic device 500. In other embodiments, the memory 502 may also be an external storage device of the electronic device 500, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the electronic device 500. Further, the memory 502 may also include both an internal storage unit of the electronic device 500 and an external storage device. The memory 502 is used to store application software and various data installed in the electronic device 500, such as program codes installed in the electronic device 500. The memory 502 may also be used to temporarily store data that has been output or is to be output. In one embodiment, an algae monitoring program based on image analysis is stored on the memory 502, and the algae monitoring program based on image analysis can be executed by the processor 501, thereby realizing the algae monitoring method based on image analysis of each embodiment of the present invention.

[0120] In some embodiments, the processor 501 may be a central processing unit (CPU), a microprocessor or other data processing chip, which is used to run program codes or process data stored in the memory 502, such as an algae monitoring method based on image analysis.

[0121] In some embodiments, the display 503 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode) touch device, etc. The display 503 is used to display identification information of the algae monitoring program based on image analysis and to display a visual user interface. The components 501-503 of the electronic device 500 communicate with each other via a system bus.

[0122] In some embodiments, when the processor 501 executes the image analysis-based algae monitoring program in the memory 502, the various steps of the image analysis-based algae monitoring method described in the above embodiments are implemented. Since the image analysis-based algae monitoring method has been described in detail above, it will not be repeated here.

[0123] Accordingly, the present invention also provides a computer-readable storage medium, which is used to store computer-readable programs or instructions. When the program or instructions are executed by a processor, the steps or functions of the algae monitoring method based on image analysis provided in the above-mentioned method embodiments can be implemented.

[0124] In summary, the algae monitoring method, device, equipment and medium based on image analysis provided by the present invention obtain float monitoring data and remote sensing image data, and perform spatiotemporal fusion of the float monitoring data and remote sensing image data to construct a water quality image spatiotemporal data set, wherein the water quality image spatiotemporal data set includes an algae coverage index, the algae coverage index is determined based on the remote sensing image data, and the constructed LSTM network is trained based on the water quality image spatiotemporal data set to obtain an algae spread model, and the float monitoring data to be predicted is predicted based on the algae spread model to obtain a predicted value of the algae coverage index; a warning threshold is preset, and algae reproduction is warned based on the predicted value of the algae coverage index and the warning threshold, thereby improving the accuracy of model prediction.

[0125] Those skilled in the art will appreciate that all or part of the processes of the above-mentioned embodiments can be implemented by instructing related hardware through a computer program, and the program can be stored in a computer-readable storage medium, wherein the computer-readable storage medium is a disk, an optical disk, a read-only storage memory, or a random access memory, etc.

[0126] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by any technician familiar with the technical field within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention.

Claims

1. An algae monitoring method based on image analysis, characterized in that: include: Acquiring floating monitoring data and remote sensing image data, performing spatiotemporal fusion on the floating monitoring data and the remote sensing image data to construct a spatiotemporal data set of water quality images, wherein the spatiotemporal data set of water quality images includes an algae coverage index; performing spatiotemporal fusion on the floating monitoring data and the remote sensing image data to construct a spatiotemporal data set of water quality images, including: Acquire float monitoring data in real time through float equipment, and upload the float monitoring data to the data center through Beidou short message, wherein the float monitoring data includes float positioning data and sensor data, and the sensor data includes water temperature, nitrogen content, phosphorus content, oxygen content, pH value and water depth; The float monitoring data is sorted in time series to obtain time-ordered data, the time-ordered data is parsed to obtain the longitude and latitude information of the float and the corresponding water quality parameters, and then the longitude and latitude information and water quality parameters are sorted in time series to obtain a time-position-water quality parameter sequence; Correcting the remote sensing image data to obtain a corrected image sequence; The corrected image sequence and the time-position-water quality parameter sequence are subjected to spatiotemporal fusion, and an algae coverage index is determined based on the remote sensing image data to construct a spatiotemporal data set of water quality images, wherein the algae coverage index is determined based on the remote sensing image data, including: Determine a sensitive band combination of algae according to the growth characteristics of algae, wherein the sensitive band combination includes at least one or more of a green band, a red band and a near infrared band; The remote sensing image data is analyzed based on the sensitive band combination to extract a spectral index, and the algae coverage index is determined based on the spectral index, wherein the spectral index includes a normalized difference water index, an enhanced water index, and a normalized water volume fraction index; the algae coverage index is calculated as follows: ， ， ， ， in, is the algae cover index, is the normalized difference water index, is the enhanced water index. is the normalized water volume fraction index, , , is the diversity index, For the green band, is the red band, The near-infrared band, , , , , , is the adjustment coefficient; Based on the water quality image spatiotemporal data set, the constructed LSTM network is trained to obtain an algae spread model, and based on the algae spread model, the floating monitoring data to be predicted is predicted to obtain a predicted value of the algae coverage index; An early warning is given to algae reproduction based on the predicted value of the algae coverage index and the early warning threshold.

2. The algae monitoring method based on image analysis according to claim 1, characterized in that: The LSTM network constructed based on the water quality image spatiotemporal dataset is trained to obtain an algae spread model, including: The water quality image spatiotemporal dataset is divided into a training set, a validation set and a test set, wherein based on the training set, the sensor data and the algae coverage index are combined to obtain combined data, the combined data is used as an input factor and the algae coverage index is used as an output factor, and an LSTM network is constructed; Training the LSTM network through the training set and adjusting the parameters of the LSTM network, wherein the LSTM network includes a forget gate, an input gate, a candidate memory, a unit state, an output gate, and a hidden state; The LSTM network is evaluated using the validation set to obtain an algae spread model.

3. The algae monitoring method based on image analysis according to claim 2, characterized in that: The calculation formula of the algae spread model is: ， in, is the algae cover index, For sensor data, is the weight parameter, is the bias term, is the activation function; The calculation formula of the forget gate is: ， in, is the output of the forget gate at the current moment, is the sigmoid function, is the weight matrix of the forget gate, is the bias vector of the forget gate, is the hidden state at the previous moment, is the input at the current moment, for the current moment; The calculation formula of the input gate is: ， in, is the output of the input gate at the current moment, is the weight matrix of the input gate, is the bias vector of the input gate; The calculation formula of the candidate memory is: ， in, is the time step Candidate memory cells is the weight matrix, is the bias vector, is the activation function; The calculation formula of the unit state is: ， in, is the unit state at the current moment, is the unit state at the previous moment; The calculation formula of the output gate is: ， in, is the output of the output gate at the current moment, is the weight matrix of the output gate, is the bias vector of the output gate; The calculation formula of the hidden state is: ， in, is the hidden state at the current moment.

4. The algae monitoring method based on image analysis according to claim 2, characterized in that: The method of predicting the floating monitoring data to be predicted based on the algae spread model to obtain a predicted value of the algae coverage index includes: The float monitoring data is acquired in real time through Beidou messages, the float monitoring data is input into the algae spread model, the float monitoring data is predicted based on the algae spread model to obtain the predicted value of the algae coverage index, and the algae spread distribution image is obtained based on the predicted value of the algae coverage index.

5. An algae monitoring device based on image analysis, characterized in that: include: The data spatiotemporal fusion module is used to obtain the floating monitoring data and the remote sensing image data, and to perform spatiotemporal fusion on the floating monitoring data and the remote sensing image data to construct a water quality image spatiotemporal data set, wherein the water quality image spatiotemporal data set includes an algae coverage index; the floating monitoring data and the remote sensing image data are spatiotemporally fused to construct a water quality image spatiotemporal data set, including: Acquire float monitoring data in real time through float equipment, and upload the float monitoring data to the data center through Beidou short message, wherein the float monitoring data includes float positioning data and sensor data, and the sensor data includes water temperature, nitrogen content, phosphorus content, oxygen content, pH value and water depth; The float monitoring data is sorted in time series to obtain time-ordered data, the time-ordered data is parsed to obtain the longitude and latitude information of the float and the corresponding water quality parameters, and then the longitude and latitude information and water quality parameters are sorted in time series to obtain a time-position-water quality parameter sequence; Correcting the remote sensing image data to obtain a corrected image sequence; The corrected image sequence and the time-position-water quality parameter sequence are subjected to spatiotemporal fusion, and an algae coverage index is determined based on the remote sensing image data to construct a spatiotemporal data set of water quality images, wherein the algae coverage index is determined based on the remote sensing image data, including: Determine a sensitive band combination of algae according to the growth characteristics of algae, wherein the sensitive band combination includes at least one or more of a green band, a red band and a near infrared band; The remote sensing image data is analyzed based on the sensitive band combination to extract a spectral index, and the algae coverage index is determined based on the spectral index, wherein the spectral index includes a normalized difference water index, an enhanced water index, and a normalized water volume fraction index; the algae coverage index is calculated as follows: ， ， ， ， in, is the algae cover index, is the normalized difference water index, is the enhanced water index. is the normalized water volume fraction index, , , is the diversity index, For the green band, is the red band, The near-infrared band, , , , , , is the adjustment coefficient; A prediction module is used to train the constructed LSTM network based on the water quality image spatiotemporal data set to obtain an algae spread model, and to predict the floating monitoring data to be predicted based on the algae spread model to obtain a predicted value of the algae coverage index; The early warning module is used to warn of algae reproduction based on the predicted value of the algae coverage index and the early warning threshold.

6. An electronic device, characterized in that: including memory and processor; The memory stores a computer-readable program executable by the processor; When the processor executes the computer-readable program, the steps of the algae monitoring method based on image analysis as described in any one of claims 1 to 4 are implemented.

7. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps in the algae monitoring method based on image analysis as described in any one of claims 1-4.

Images