Flood disaster emergency early warning method and system based on integrated radar sensor measurement
By using dynamic gain control and a pre-trained model of an integrated radar sensor, the problem of high false detection rate of water level in flood disaster emergency early warning was solved, and effective identification and accurate early warning of weak water level signals were achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- 北京信安恒久科技有限公司
- Filing Date
- 2025-12-31
- Publication Date
- 2026-05-29
AI Technical Summary
In existing flood disaster emergency early warning systems, the signal-to-noise ratio of weak water level signals is insufficient, resulting in a high false detection rate. Existing radar monitoring methods cannot dynamically adjust the amplification factor, leading to oversaturation distortion of strong signals or the drowning out of weak signals by noise.
The original echo signal is acquired by an integrated radar sensor in the first scan. Through PGA standard gain amplification and analog-to-digital conversion, rapid time-domain analysis is performed to identify suspected water level units and oversaturated interference units. Range gate gain control vector is generated, and the PGA gain curve is dynamically configured for a second scan. After adaptive amplification, analog-to-digital conversion is performed again, and the water level value is calculated by inputting it into a pre-trained water level inversion model.
It improved the signal-to-noise ratio of weak water level signals, reduced the false detection rate of water levels, and improved the accuracy and reliability of emergency early warning for flood disasters.
Smart Images
Figure CN122110035A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of disaster early warning, and in particular to a flood disaster emergency early warning method and system based on integrated radar sensing measurement. Background Technology
[0002] Real-time monitoring and early warning of floods are core tasks in water conservancy projects and emergency management, directly impacting the safety of people's lives and property and the stable development of regional economies and societies. Currently, most mainstream flood level monitoring technologies employ radar sensing, using fixed or mobile radar equipment to transmit electromagnetic waves to the monitoring area, receive reflected echo signals, and extract characteristic parameters to retrieve water level data. Existing radar monitoring methods generally use a fixed-gain signal amplification mode, which cannot dynamically adjust the amplification factor according to the actual strength of the echo signal. When there is strong near-range reflection interference or weak far-range water level signals in the monitoring area, the fixed gain can lead to oversaturation distortion of strong signals and noise-overwhelmed weak signals, thus increasing the false detection rate of water level identification.
[0003] Currently, the relevant technologies for flood disaster emergency early warning suffer from the technical problem of insufficient signal-to-noise ratio of weak water level signals, resulting in a high false detection rate of water levels. Summary of the Invention
[0004] This application provides a flood disaster emergency early warning method and system based on integrated radar sensing measurement. It employs an integrated radar sensor to acquire the original echo signal during the initial scan, amplifies it using PGA standard gain, and converts it to digital to obtain the first echo data. Through rapid time-domain analysis, it identifies suspected water level cells and oversaturated interference cells from the first echo data, generating a range gate gain control vector. Based on this vector, it dynamically configures the PGA gain curve and controls the radar for a second scan. The newly acquired original echo signal is adaptively amplified and converted to digital to obtain the second echo data. This second echo data is then input into a pre-trained water level inversion model to calculate the water level value. Based on this, flood disaster emergency early warning and other technical means are implemented. This solves the technical problem of high false detection rate of water level in existing flood disaster emergency early warning systems, achieving the technical effects of improving the signal-to-noise ratio of weak water level signals, reducing the false detection rate of water level, and improving the accuracy and reliability of flood disaster emergency early warning.
[0005] This application provides a flood disaster emergency early warning method based on integrated radar sensing measurement, comprising: controlling an integrated radar sensor to scan a monitoring area once to acquire raw echo signals; inputting the raw echo signals into a PGA for standard gain amplification; and obtaining first echo data after analog-to-digital conversion; performing rapid time-domain analysis on the first echo data to identify suspected water level units and oversaturated interference units; generating corresponding range gate gain control vectors based on the distribution of the suspected water level units and the oversaturated interference units; dynamically configuring the gain curve of the PGA based on the range gate gain control vectors; controlling the integrated radar sensor to perform a second scan of the monitoring area; adaptively amplifying the second acquired raw echo signals based on the gain curve; and obtaining second echo data after analog-to-digital conversion; inputting the second echo data into a pre-trained water level inversion model to calculate water level values; and issuing flood disaster emergency early warnings based on the water level values.
[0006] In a possible implementation, the following process is performed: the intermediate frequency analog signal output terminal of the integrated radar sensor is electrically connected to the input terminal of the PGA, the output terminal of the PGA is electrically connected to the input terminal of the analog-to-digital converter, and the processor is communicatively connected to the gain control terminal of the PGA to configure the gain curve.
[0007] In a possible implementation, the first echo data is subjected to fast time-domain analysis to identify suspected water level units and oversaturated interference units, and the following processing is performed: the first echo data is divided into multiple range gates according to the radar wave round-trip time; the average signal power and signal fluctuation variance of the first echo data at the multiple range gates are calculated; range gates whose average signal power is less than a first power threshold and whose signal fluctuation variance is less than a first variance threshold are identified as suspected water level units; range gates whose average signal power is greater than or equal to a second power threshold are identified as oversaturated interference units.
[0008] In a possible implementation, based on the distribution of the suspected water level units and the oversaturated interference units, a corresponding distance gate gain control vector is generated, and the following processing is performed: a first gain value is assigned to the distance gate determined to be a suspected water level unit, a second gain value is assigned to the distance gate determined to be an oversaturated interference unit, and a standard gain value is assigned to the other distance gates among the plurality of distance gates; the gain values corresponding to the plurality of distance gates are arranged in order to form the distance gate gain control vector; wherein, the first gain value is greater than the standard gain value, and the second gain value is less than the standard gain value.
[0009] In a possible implementation, the gain curve of the PGA is dynamically configured according to the range gate gain control vector, and the following processing is performed: the mapping relationship between the range gate and the gain value is extracted according to the range gate gain control vector, and a gain curve in which the gain value changes with the range segments is generated according to the mapping relationship; the PGA is dynamically configured according to the gain curve.
[0010] In a possible implementation, the following processing is performed: The training method for the pre-trained water level inversion model includes: acquiring radar echo data samples obtained under various hydrological and interference environments, after adaptive amplification via PGA according to the range gate gain control vector, as training input; acquiring real water level values collected synchronously with the radar echo data samples, as training labels; and training the neural network model using the training input and the training labels to obtain the water level inversion model.
[0011] In a possible implementation, flood disaster emergency early warning is conducted based on the water level value, and the following processing is performed: multiple water level early warning thresholds are preset, and flood disaster graded early warning is conducted based on the relationship between the water level value and the multiple water level early warning thresholds; and / or, the water level rise rate is calculated based on the water level value, and flood disaster graded early warning is conducted based on the relationship between the water level rise rate and multiple preset water level rise rate thresholds.
[0012] This application also provides a flood disaster emergency early warning system based on integrated radar sensing measurement, comprising: a primary scan module, used to control the integrated radar sensor to perform a primary scan of the monitoring area, acquire the raw echo signal, input the raw echo signal into a PGA for standard gain amplification, and obtain first echo data after analog-to-digital conversion; a range gate gain control vector generation module, used to perform rapid time-domain analysis on the first echo data, identify suspected water level units and oversaturated interference units, and generate corresponding range gate gain control vectors according to the distribution of the suspected water level units and the oversaturated interference units; a secondary scan module, used to dynamically configure the gain curve of the PGA according to the range gate gain control vector, control the integrated radar sensor to perform a secondary scan of the monitoring area, adaptively amplify the raw echo signal acquired in the secondary scan according to the gain curve, and obtain second echo data after analog-to-digital conversion; and a flood disaster emergency early warning module, used to input the second echo data into a pre-trained water level inversion model to calculate the water level value, and conduct flood disaster emergency early warning based on the water level value.
[0013] The proposed flood disaster emergency early warning method and system based on integrated radar sensing measurement, as described in this application, firstly controls the integrated radar sensor to scan the monitoring area once, acquiring the raw echo signal. This raw echo signal is then input into a PGA (Programmable Gate Array) for standard gain amplification, and after analog-to-digital conversion, the first echo data is obtained. Next, rapid time-domain analysis is performed on the first echo data to identify suspected water level units and oversaturated interference units. Based on the distribution of these units, corresponding range gate gain control vectors are generated. Then, based on these range gate gain control vectors, the gain curve of the PGA is dynamically configured, and the integrated radar sensor is controlled to perform a second scan of the monitoring area. According to the gain curve, the second acquired raw echo signal is adaptively amplified, and after analog-to-digital conversion, the second echo data is obtained. Finally, the second echo data is input into a pre-trained water level inversion model to calculate the water level value, and flood disaster emergency early warning is performed based on this water level value. Through the above process, the proposed method and system achieve the technical effects of improving the signal-to-noise ratio of weak water level signals, reducing the false detection rate of water level, and improving the accuracy and reliability of flood disaster emergency early warning. Attached Figure Description
[0014] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings of the embodiments of the present invention will be briefly described below. Flowcharts are used in this application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed precisely in sequence. Instead, various steps can be processed in reverse order or simultaneously as needed. Furthermore, other operations can be added to these processes, or one or more steps can be removed from these processes.
[0015] Figure 1 This is a flowchart illustrating the flood disaster emergency early warning method based on integrated radar sensing measurement provided in an embodiment of this application.
[0016] Figure 2 This is a schematic diagram of the structure of an emergency early warning system for flood disasters based on integrated radar sensing measurement, provided in an embodiment of this application.
[0017] Explanation of reference numerals in the attached diagram: 10 for primary scanning module, 20 for distance gate gain control vector generation module, 30 for secondary scanning module, and 40 for flood disaster emergency early warning module. Detailed Implementation
[0018] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.
[0019] This application provides an emergency early warning method for flood disasters based on integrated radar sensing measurements, such as... Figure 1 As shown, the method includes: Step S100: Control the integrated radar sensor to scan the monitoring area once to acquire the raw echo signal. Input the raw echo signal into the PGA for standard gain amplification, and obtain the first echo data after analog-to-digital conversion. The intermediate frequency analog signal output terminal of the integrated radar sensor is electrically connected to the input terminal of the PGA, the output terminal of the PGA is electrically connected to the input terminal of the analog-to-digital converter, and the processor is communicatively connected to the gain control terminal of the PGA to configure the gain curve.
[0020] Specifically, an integrated radar sensor is an integrated device that combines transmitting, receiving, and signal processing modules, eliminating the need for separate transmitters and receivers. It transmits radar waves to the monitored area and receives the reflected raw echo signals. A PGA (Programmable Gain Amplifier) is an electronic component whose amplification factor can be adjusted via circuitry or software, dynamically adjusting the amplification gain based on signal strength. An analog-to-digital converter (ADC) converts analog signals into digital signals. The processor is the core unit controlling the entire signal processing flow, responsible for configuring the PGA's gain curve, i.e., how the amplification factor changes with signal parameters.
[0021] First, the integrated radar sensor is activated and scans the monitoring area according to preset parameters, including scanning angle, transmission frequency, and pulse width, to receive the reflected raw echo signal. This echo signal is an analog signal, which is transmitted to the PGA through hardware circuitry. The processor sends control commands to the PGA's gain control terminal via I2C or SPI communication protocol to configure the initial gain curve. The PGA amplifies the raw echo signal according to the curve using standard gain, i.e., the amplification factor is fixed, such as 10 times. The amplified analog signal is input to the analog-to-digital converter, which converts it into a digital signal according to the set sampling rate, and finally outputs the first echo data. The first echo data is a two-dimensional array containing timestamps and signal amplitude.
[0022] Step S200: Perform fast time-domain analysis on the first echo data to identify suspected water level units and oversaturated interference units. Based on the distribution of the suspected water level units and the oversaturated interference units, generate corresponding range gate gain control vectors.
[0023] Specifically, fast time-domain analysis refers to the calculation of statistical characteristics of signal data over time. Range gating divides the round-trip time of radar waves into several equally spaced time intervals. Each time interval corresponds to the signal at a specific distance between the radar and the monitored area; that is, each range gate represents signal data within a specific range. The range gate gain control vector is an array of gain values arranged in range gate order, with each element corresponding to the target magnification factor for a given range gate.
[0024] First, the first echo data is segmented according to the radar wave round-trip time to obtain multiple range gates. The average power and fluctuation variance of the signal in each range gate are calculated. Suspected water level units and oversaturated interference units are screened out by preset power thresholds and variance thresholds. Then, corresponding gain values are assigned to different types of range gates, and the gain control vector is formed by combining them in the order of range gates.
[0025] In one possible implementation, rapid time-domain analysis is performed on the first echo data to identify suspected water level units and oversaturated interference units. Step S200 further includes step S210, which divides the first echo data into multiple range gates according to the radar wave round-trip time. Specifically, based on the radar sensor's transmission parameters and the maximum distance of the monitoring area, the maximum radar wave round-trip time is calculated, and then this maximum time is divided into N equally spaced time periods, each time period being a range gate. The first echo data is allocated to each range gate according to the correspondence between timestamps and range gates. For example, if the maximum distance of the monitoring area is set to 50 meters, the maximum radar wave round-trip time is (50 meters × 2) ÷ 300,000 kilometers per second = 333.33 nanoseconds. This 333.33 nanoseconds is divided into 50 range gates, with each range gate having a time interval of 6.67 nanoseconds, corresponding to a 1-meter range interval. In the first echo data, signals with timestamps between 0 and 6.67 nanoseconds are assigned to distance gate 1, corresponding to a distance of 0 to 1 meter; signals with timestamps between 6.67 and 13.33 nanoseconds are assigned to distance gate 2, corresponding to a distance of 1 to 2 meters, and so on, until all data are assigned to 50 distance gates.
[0026] Step S220: Calculate the average signal power and signal fluctuation variance of the first echo data across the plurality of distance gates. Specifically, for the signal data within each distance gate, the average signal power is calculated using the arithmetic mean method, i.e., first convert each signal amplitude into a power value, and then calculate the average value. The signal fluctuation variance is calculated using the variance formula, i.e., first calculate the squared difference between each signal power value and the average power, and then calculate the average value.
[0027] Step S230: Distance gates whose average signal power is less than a first power threshold and whose signal fluctuation variance is less than a first variance threshold are identified as suspected water level units. Specifically, based on the environmental noise level and historical water level signal characteristics of the monitoring area, the first power threshold and the first variance threshold are experimentally calibrated. Noise signal power is typically low but fluctuates greatly, while water level signal power is low and fluctuates little. The average power and fluctuation variance of each distance gate are compared with the first power threshold and the first variance threshold, respectively. If both conditions are met, the gate is identified as a suspected water level unit. For example, through multiple experimental measurements, the environmental noise power range of the monitoring area is -70 to -60 dBmW, and the fluctuation variance range is 8 to 15 dB². Therefore, the first power threshold is set to -60 dBmW, and the first variance threshold is set to 5 dB². If the average power of distance gate 12 is -62 dBmW and the fluctuation variance is 3 dB², then distance gate 12 is identified as a suspected water level unit.
[0028] Step S240: Range gates with average signal power greater than or equal to a second power threshold are identified as oversaturated interference units. Specifically, the second power threshold is determined based on the maximum power handling capacity of the PGA and the transmit power of the radar sensor. When the signal power is too high, the PGA will enter a saturated state, causing signal distortion. This type of signal is oversaturated interference. The saturation power threshold of the PGA is experimentally measured and set as the second power threshold. The average power of each range gate is compared with this threshold; if it meets the condition of being greater than or equal to, it is identified as an oversaturated interference unit. For example, if an AD8331 PGA is selected, its saturation power threshold is 0 dB / mW, meaning that when the signal power reaches 0 dB / mW, the PGA output signal is distorted. Therefore, the second power threshold is set to 0 dB / mW. If the average power of range gate 8 is 2 dB / mW, then range gate 8 is identified as an oversaturated interference unit.
[0029] In one possible implementation, based on the distribution of the suspected water level units and the oversaturated interference units, a corresponding range gate gain control vector is generated. Step S200 further includes step S250, which assigns a first gain value to the range gate identified as a suspected water level unit, assigns a second gain value to the range gate identified as an oversaturated interference unit, and assigns a standard gain value to the other range gates among the plurality of range gates. Specifically, the first gain value is greater than the standard gain value, and the second gain value is less than the standard gain value. Specifically, the standard gain value is the initial gain of the PGA during the first scan, set according to the conventional signal strength. The first gain value is determined based on the difference between the signal strength of the suspected water level unit and the target signal strength, i.e., the weak signal needs to be amplified to a suitable strength. The second gain value is determined based on the difference between the signal strength of the oversaturated interference unit and the saturation power of the PGA, i.e., the strong signal needs to be attenuated to a non-saturated state. The gain value is selected through the gain adjustment range of the PGA. For example, the standard gain value is set to 10 times, the suspected water level unit selects a first gain value of 25 times, the oversaturation interference unit selects a second gain value of 2 times, and other distance gates still retain the standard gain value of 10 times.
[0030] Step S260: Arrange the gain values corresponding to the multiple distance gates in sequence to form the distance gate gain control vector. Specifically, according to the distance gate numbering order, from 1 to N, store the gain value corresponding to each distance gate, such as the first gain value, the second gain value, or the standard gain value, into an array. This array is the distance gate gain control vector, and the vector length is the same as the number of distance gates. For example, if the monitoring area is divided into 50 distance gates, where distance gates 12 and 13 are suspected water level units with a corresponding gain of 25 times, distance gate 8 is an oversaturation interference unit with a corresponding gain of 2 times, and the other 47 distance gates have a standard gain of 10 times, then the distance gate gain control vector is [10,10,10,10,10,10,10,2,10,10,10,25,25,10,...,10], a total of 50 elements, arranged in the order of distance gates 1 to 50.
[0031] Step S300: Based on the range gate gain control vector, dynamically configure the gain curve of the PGA, control the integrated radar sensor to perform a secondary scan of the monitoring area, and adaptively amplify the original echo signal acquired in the secondary scan according to the gain curve, and obtain the second echo data after analog-to-digital conversion.
[0032] Specifically, dynamically configuring the PGA gain curve refers to adjusting the PGA's amplification factor at different times, corresponding to different range gates, based on the mapping relationship between range gates and gain values, so that the gain curve exhibits a segmented variation characteristic. The mapping relationship between the sequence number of each range gate and its corresponding gain value is extracted from the range gate gain control vector. A segmented gain curve is generated based on this mapping relationship, with each range gate corresponding to a segment of the curve representing a constant gain. The processor writes the gain curve parameters into the PGA's control register via a communication protocol. After dynamic configuration is completed, a secondary scan is initiated. The new raw echo signal is amplified according to the gain value corresponding to the current range gate and then undergoes analog-to-digital conversion to obtain the second echo data.
[0033] In one possible implementation, the gain curve of the PGA is dynamically configured according to the range gate gain control vector. Step S300 further includes step S310, which involves extracting the mapping relationship between the range gate and the gain value according to the range gate gain control vector, and generating a gain curve in which the gain value changes piecewise with the range according to the mapping relationship. Specifically, the key-value pair (i.e., the mapping relationship) between each range gate number and its corresponding gain value is extracted from the range gate gain control vector. The range gate number is converted into a corresponding range interval. The range interval is calculated based on the time interval of the range gate and the radar wave propagation speed. A piecewise constant gain curve is generated with the distance as the horizontal axis and the gain value as the vertical axis, meaning that the gain value remains unchanged within the same range interval.
[0034] Step S320: Dynamically configure the PGA according to the gain curve. Specifically, the processor establishes a connection with the PGA through communication protocols such as I2C or SPI, and writes the time range and gain value parameters corresponding to each distance interval in the gain curve into the PGA's control register. The timing control module inside the PGA automatically switches to the set gain value within the corresponding time period according to the input time range, realizing dynamic gain configuration. For example, the processor uses an STM32F4 microcontroller, which connects to the PGA through the SPI communication protocol. The time range and gain value of each segment in the gain curve are converted into binary instructions that the PGA can recognize. The microcontroller writes the instructions of all segments sequentially into the PGA's control register. After the PGA starts, when it detects that the round-trip time of the radar wave enters a certain time range, it automatically switches the gain to the corresponding multiple.
[0035] Step S400: Input the second echo data into the pre-trained water level inversion model to calculate the water level value, and conduct flood disaster emergency early warning based on the water level value.
[0036] Specifically, a neural network-based water level inversion model is first constructed and trained. This model calculates the actual water level based on radar echo data. The training data includes radar echo data samples under various environments and synchronously collected real water level values. After the model is trained, it is input as second echo data and outputs the water level value. Then, by pre-setting multiple water level warning thresholds or water level rise rate thresholds, the warning level is determined based on the calculation results, and a warning is triggered.
[0037] In one possible implementation, the training method of the pre-trained water level inversion model, step S400 further includes step S410, acquiring radar echo data samples obtained by adaptive amplification via PGA according to the range gate gain control vector under various hydrological and interference environments, as training input. Specifically, different hydrological environments and different interference environments are selected, such as calm water surface, wavy water surface, water surface containing floating objects, etc., and interference environments such as rainy days, foggy days, vegetation obstruction, building reflection, etc. In each environment, radar echo data is collected by radar sensors according to the process of steps S100-S300. Multiple sets of samples are collected for each environment. Each set of samples contains second echo data and the corresponding range gate gain control vector. All samples constitute the training input dataset. For example, 20 environment combinations are selected, including 5 hydrological environments and 4 interference environments, including calm water surface, level 1 wave water surface, level 2 wave water surface, water surface containing branches and floating objects, water surface containing plastic waste and floating objects, light rain, moderate rain, heavy fog, and shoreline tree obstruction. For each environmental combination, 100 sets of samples were collected. Each set of samples contained 1024 data points of second echo data and a range gate gain control vector of length 50. The training input dataset contained 2000 sets of samples. The input features of each set of samples were a 1074-dimensional feature vector resulting from the concatenation of the 1024-dimensional second echo data and the 50-dimensional range gate gain control vector.
[0038] Step S420: Obtain the actual water level values collected synchronously with the radar echo data samples, as training labels. Specifically, while collecting radar echo data samples, high-precision water level measurement equipment, such as submersible level gauges or ultrasonic level gauges, is used to simultaneously measure the actual water level values in the monitoring area. Each radar echo data sample corresponds to one actual water level value, and all actual water level values constitute the training label dataset, with each label corresponding one-to-one with the training input sample. For example, a submersible level gauge with an accuracy of ±0.01 meters is selected and installed at the center of the radar sensor's monitoring area, synchronizing with the radar sensor's data acquisition. When the radar sensor collects a second set of echo data, the level gauge simultaneously records the current water level value, which is the training label for the corresponding sample. 2000 sets of training input samples correspond to 2000 training labels.
[0039] Step S430: Train the neural network model using the training input and the training labels to obtain the water level inversion model. Specifically, the neural network model adopts a fully connected neural network structure, including an input layer, hidden layers, and an output layer. The activation function for the hidden layers is ReLU, and the activation function for the output layer is Linear. The training process includes data preprocessing, dividing the training set and validation set, setting training parameters, minimizing the loss function using an optimizer, and stopping the training when the validation set loss converges. The model parameters are then saved to obtain the water level inversion model.
[0040] For example, the model structure is as follows: 1074 neurons in the input layer, receiving a 1074-dimensional feature vector → 256 neurons in the first hidden layer with ReLU activation function → 128 neurons in the second hidden layer with ReLU activation function → 1 neuron in the output layer with Linear activation function. Data preprocessing uses Min-Max normalization to map the input features and label values to the 0-1 range. The training and validation sets are divided in an 8:2 ratio. The optimizer is Adam, the learning rate is set to 0.001, the number of iterations is set to 1000, the batch size is set to 32, and the loss function is mean squared error. During training, the validation set loss is calculated every 100 iterations. Training stops when the validation set loss does not decrease for 50 consecutive iterations. The final saved water level inversion model has a mean absolute error of ±0.02 meters on the validation set.
[0041] In one possible implementation, flood disaster emergency early warning is conducted based on the water level value. Step S400 further includes step S440, which involves presetting multiple water level early warning thresholds and conducting graded early warning for flood disasters based on the relationship between the water level value and the multiple water level early warning thresholds. Specifically, based on the terrain features of the monitoring area, historical flood disaster data, and flood control standards, multiple early warning levels are defined, such as four levels. A corresponding water level early warning threshold is set for each level. The water level value output by the water level inversion model is compared with the threshold to determine the corresponding early warning level and trigger corresponding early warning measures, such as audible and visual alarms, SMS notifications, and platform push notifications. For example, if the monitoring area is an urban river with a flood control standard of 3.0 meters, four early warning thresholds are set based on historical data: blue warning 1.0 meter, yellow warning 1.5 meters, orange warning 2.0 meters, and red warning 2.5 meters. When the water level value calculated by the water level inversion model is 1.8 meters, this value is greater than the yellow warning threshold of 1.5 meters and less than the orange warning threshold of 2.0 meters, and is therefore determined to be a yellow warning. The triggered warning measures include: the local audible and visual alarm will activate a flashing yellow light and intermittent alarm sound, a yellow warning text message will be sent to the flood control command center, and the yellow warning information will be pushed to the city's flood control monitoring platform along with the current water level and monitoring location.
[0042] And / or step S450, calculate the water level rise rate based on the water level value, and conduct flood disaster classification early warning based on the relationship between the water level rise rate and multiple preset water level rise rate thresholds. Specifically, water level values are collected at fixed time intervals, such as 5 minutes, and the ratio of the difference between two adjacent water level values to the time interval is calculated to obtain the water level rise rate. Based on the water level rise risk in the monitored area, such as the risk of flooding due to a rapid rise in a short period of time, multiple rise rate warning thresholds are preset. The calculated rise rate is compared with the thresholds to determine the warning level and trigger an early warning. For example, if the water level collection time interval is set to 5 minutes, the water level rise rate calculation formula is: rise rate = (current water level value - previous water level value) ÷ 5 minutes. Four rise rate thresholds are preset: blue warning 0.05 meters per minute, yellow warning 0.1 meters per minute, orange warning 0.15 meters per minute, and red warning 0.2 meters per minute. The previous water level was 1.0 meter, and the current water level is 1.5 meters. The calculated rate of rise is (1.5 - 1.0) ÷ 5 = 0.1 meters per minute, which equals the yellow warning threshold, thus triggering a yellow warning. The triggered warning measures include: activating the local audible and visual alarm with flashing yellow lights and continuous audible alarms; sending a yellow warning text message containing the rate of rise to the flood control command center; and pushing the warning information to the monitoring platform, indicating the rate of rise and the estimated time to reach the next warning threshold.
[0043] This application embodiment uses an integrated control radar sensor to acquire the original echo signal during the first scan. After PGA standard gain amplification and analog-to-digital conversion, the first echo data is obtained. Suspected water level units and oversaturated interference units are identified from the first echo data through rapid time-domain analysis, and a range gate gain control vector is generated. Based on this vector, the PGA gain curve is dynamically configured and the radar is controlled for a second scan. The newly acquired original echo signal is adaptively amplified and then converted to digital to obtain the second echo data. The second echo data is input into a pre-trained water level inversion model to calculate the water level value. Based on this, flood disaster emergency early warning and other technical means are used to solve the technical problem of high water level false detection rate in existing flood disaster emergency early warning systems. This achieves the technical effects of improving the signal-to-noise ratio of weak water level signals, reducing the water level false detection rate, and improving the accuracy and reliability of flood disaster emergency early warning.
[0044] In the above text, refer to Figure 1 This paper describes in detail a flood disaster emergency early warning method based on integrated radar sensing measurement according to an embodiment of the present invention. Next, we will refer to... Figure 2 This invention describes an emergency early warning system for flood disasters based on integrated radar sensing measurement, according to an embodiment of the present invention.
[0045] The flood disaster emergency early warning system based on integrated radar sensing measurement according to embodiments of the present invention addresses the technical problem of high false detection rate of water level in existing flood disaster emergency early warning systems. It achieves the technical effects of improving the signal-to-noise ratio of weak water level signals, reducing the false detection rate of water level, and improving the accuracy and reliability of flood disaster emergency early warning. The flood disaster emergency early warning system based on integrated radar sensing measurement includes: a primary scanning module 10, a range gate gain control vector generation module 20, a secondary scanning module 30, and a flood disaster emergency early warning module 40.
[0046] The first scanning module 10 controls the integrated radar sensor to perform a single scan of the monitoring area, acquire the raw echo signal, input the raw echo signal into the PGA for standard gain amplification, and obtain the first echo data after analog-to-digital conversion. The range gate gain control vector generation module 20 performs rapid time-domain analysis on the first echo data, identifies suspected water level units and oversaturated interference units, and generates corresponding range gate gain control vectors based on the distribution of the suspected water level units and the oversaturated interference units. The second scanning module 30 dynamically configures the gain curve of the PGA according to the range gate gain control vector, controls the integrated radar sensor to perform a second scan of the monitoring area, adaptively amplifies the raw echo signal acquired in the second scan according to the gain curve, and obtains the second echo data after analog-to-digital conversion. The flood disaster emergency early warning module 40 inputs the second echo data into a pre-trained water level inversion model to calculate the water level value, and issues a flood disaster emergency early warning based on the water level value.
[0047] The system may further include: the intermediate frequency analog signal output terminal of the integrated radar sensor is electrically connected to the input terminal of the PGA, the output terminal of the PGA is electrically connected to the input terminal of the analog-to-digital converter, and the processor is communicatively connected to the gain control terminal of the PGA to configure the gain curve.
[0048] The detailed description of the specific configuration of the range gate gain control vector generation module 20 is explained as follows: As mentioned above, to perform rapid time-domain analysis on the first echo data and identify suspected water level units and oversaturated interference units, the range gate gain control vector generation module 20 may further include: a range gate partitioning unit for dividing the first echo data into multiple range gates according to the radar wave round-trip time; a signal information calculation unit for calculating the average signal power and signal fluctuation variance of the first echo data on the multiple range gates; a suspected water level determination unit for determining range gates whose average signal power is less than a first power threshold and whose signal fluctuation variance is less than a first variance threshold as suspected water level units; and an oversaturated interference determination unit for determining range gates whose average signal power is greater than or equal to a second power threshold as oversaturated interference units.
[0049] The distance gate gain control vector generation module 20 generates corresponding distance gate gain control vectors based on the distribution of the suspected water level units and the oversaturated interference units. The distance gate gain control vector generation module 20 may further include: a gain value allocation unit for allocating a first gain value to the distance gates identified as suspected water level units, a second gain value to the distance gates identified as oversaturated interference units, and a standard gain value to the other distance gates among the plurality of distance gates; and a distance gate gain control vector composition unit for arranging the gain values corresponding to the plurality of distance gates in sequence to form the distance gate gain control vector; wherein the first gain value is greater than the standard gain value, and the second gain value is less than the standard gain value.
[0050] The detailed description of the specific configuration of the secondary scanning module 30 is explained as follows: As mentioned above, the gain curve of the PGA is dynamically configured according to the distance gate gain control vector. The secondary scanning module 30 may further include: a gain curve generation unit for extracting the mapping relationship between the distance gate and the gain value according to the distance gate gain control vector, and generating a gain curve in which the gain value changes piecewise with distance according to the mapping relationship; and a PGA configuration unit for dynamically configuring the PGA according to the gain curve.
[0051] The detailed description of the specific configuration of the flood disaster emergency early warning module 40 is explained as follows: As mentioned above, the flood disaster emergency early warning module 40 may further include: a water level inversion model pre-training unit for acquiring radar echo data samples obtained by adaptive amplification via PGA according to the range gate gain control vector under various hydrological and interference environments, as training input; acquiring real water level values synchronously collected with the radar echo data samples, as training labels; and using the training input and the training labels to train the neural network model to obtain the water level inversion model.
[0052] The flood disaster emergency early warning module 40, which performs flood disaster emergency early warning based on the water level value, may further include: a first flood disaster graded early warning unit for presetting multiple water level early warning thresholds and performing flood disaster graded early warning based on the relationship between the water level value and the multiple water level early warning thresholds; and a second flood disaster graded early warning unit for calculating the water level rise rate based on the water level value and performing flood disaster graded early warning based on the relationship between the water level rise rate and the preset multiple water level rise rate thresholds.
[0053] The flood disaster emergency early warning system based on integrated radar sensing measurement provided in the embodiments of the present invention can execute the flood disaster emergency early warning method based on integrated radar sensing measurement provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.
[0054] Although this application makes various references to certain modules in the system according to the embodiments of this application, any number of different modules can be used and run on user terminals and / or servers. The various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of each functional unit are only for easy distinction between each other and are not used to limit the scope of protection of this invention.
[0055] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.
Claims
1. A flood disaster emergency early warning method based on integrated radar sensing measurement, characterized in that, The method includes: The integrated radar sensor is controlled to scan the monitoring area once to obtain the raw echo signal. The raw echo signal is then input into the PGA for standard gain amplification and converted from analog to digital to obtain the first echo data. A fast time-domain analysis is performed on the first echo data to identify suspected water level units and oversaturated interference units. Based on the distribution of the suspected water level units and the oversaturated interference units, a corresponding range gate gain control vector is generated. According to the range gate gain control vector, the gain curve of the PGA is dynamically configured to control the integrated radar sensor to perform a secondary scan of the monitoring area. According to the gain curve, the original echo signal acquired in the secondary scan is adaptively amplified and then converted from analog to digital to obtain the second echo data. The second echo data is input into a pre-trained water level inversion model to calculate the water level value, and flood disaster emergency early warning is carried out based on the water level value.
2. The flood disaster emergency early warning method based on integrated radar sensing measurement as described in claim 1, characterized in that, The intermediate frequency analog signal output terminal of the integrated radar sensor is electrically connected to the input terminal of the PGA, the output terminal of the PGA is electrically connected to the input terminal of the analog-to-digital converter, and the processor is communicatively connected to the gain control terminal of the PGA to configure the gain curve.
3. The flood disaster emergency early warning method based on integrated radar sensing measurement as described in claim 1, characterized in that, A rapid time-domain analysis was performed on the first echo data to identify suspected water level cells and oversaturation interference cells, including: The first echo data is divided into multiple range gates according to the radar wave round-trip time; Calculate the average signal power and signal fluctuation variance of the first echo data at the plurality of distance gates; Distance gates that indicate the average power of the signal is less than a first power threshold and the variance of the signal fluctuation is less than a first variance threshold are identified as suspected water level units. Distance gates whose average signal power is greater than or equal to the second power threshold are identified as oversaturated interference units.
4. The flood disaster emergency early warning method based on integrated radar sensing measurement as described in claim 3, characterized in that, Based on the distribution of the suspected water level units and the oversaturated interference units, a corresponding range gate gain control vector is generated, including: Assign a first gain value to the distance gate identified as the suspected water level unit, assign a second gain value to the distance gate identified as the oversaturated interference unit, and assign a standard gain value to the other distance gates among the plurality of distance gates; The gain values corresponding to the plurality of distance gates are arranged in order to form the distance gate gain control vector; Wherein, the first gain value is greater than the standard gain value, and the second gain value is less than the standard gain value.
5. The flood disaster emergency early warning method based on integrated radar sensing measurement as described in claim 1, characterized in that, Based on the distance gate gain control vector, the gain curve of the PGA is dynamically configured, including: Based on the distance gate gain control vector, the mapping relationship between the distance gate and the gain value is extracted, and a gain curve in which the gain value changes with the distance segment is generated based on the mapping relationship. The PGA is dynamically configured based on the gain curve.
6. The flood disaster emergency early warning method based on integrated radar sensing measurement as described in claim 1, characterized in that, Training methods for pre-trained water level inversion models include: Under various hydrological and interference environments, radar echo data samples obtained by adaptive amplification via PGA based on the range gate gain control vector are used as training inputs. The actual water level values collected synchronously with the radar echo data samples are used as training labels; The neural network model is trained using the training input and the training labels to obtain the water level inversion model.
7. The flood disaster emergency early warning method based on integrated radar sensing measurement as described in claim 1, characterized in that, Flood disaster emergency warnings are issued based on the aforementioned water level values, including: Multiple water level warning thresholds are preset, and flood disaster classification warnings are carried out based on the relationship between the water level value and the multiple water level warning thresholds; And / or, calculate the water level rise rate based on the water level value, and conduct flood disaster classification and early warning based on the relationship between the water level rise rate and a plurality of preset water level rise rate thresholds.
8. A flood disaster emergency early warning system based on integrated radar sensing measurement, characterized in that, The system is used to implement the flood disaster emergency early warning method based on integrated radar sensing measurement as described in any one of claims 1-7, and the system includes: The single-scan module is used to control the integrated radar sensor to perform a single scan of the monitoring area, acquire the raw echo signal, input the raw echo signal into the PGA for standard gain amplification, and obtain the first echo data after analog-to-digital conversion; The distance gate gain control vector generation module is used to perform fast time-domain analysis on the first echo data, identify suspected water level units and oversaturated interference units, and generate corresponding distance gate gain control vectors based on the distribution of the suspected water level units and the oversaturated interference units. The secondary scanning module is used to dynamically configure the gain curve of the PGA according to the range gate gain control vector, control the integrated radar sensor to perform a secondary scan of the monitoring area, and adaptively amplify the original echo signal acquired in the secondary scan according to the gain curve, and obtain the second echo data after analog-to-digital conversion. The flood disaster emergency early warning module is used to input the second echo data into a pre-trained water level inversion model to calculate the water level value, and to issue an emergency early warning for flood disasters based on the water level value.