Multi-sensor fusion multi-modal hydrological data synchronous acquisition method

Abstract

The invention discloses a multi-sensor fusion multi-modal hydrological data synchronous acquisition method, and relates to the technical field of data acquisition, and the method comprises the steps: connecting a plurality of sensors deployed in a hydrological region to an intelligent PGA module to form a synchronous acquisition topology, issuing an acquisition instruction packet through a central main controller, hardware synchronization under an absolute timestamp of the intelligent PGA module is executed, a sampling edge mark is generated, mark alignment and channel autonomous sampling switching of multiple sensing channels are carried out, and sensing channel relatively independent acquisition of multiple sensors is executed based on self-adaptive gain regulation of the intelligent PGA module; the method and the device are used for solving the technical problem of how to consider channel adaptive regulation and data fusion quality while ensuring high-precision synchronization of multiple sensors in the prior art, can effectively improve complete synchronization of data acquisition and data quality, and improve reliability of long-term monitoring and data accuracy.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data acquisition, and particularly relates to a multi-sensor fusion multi-modal hydrological data synchronous acquisition method. BACKGROUND

[0002] In the field of hydrological monitoring, the deployment of multi-sensor and data acquisition has become an important means to obtain hydrological environmental information. The existing technology usually adopts distributed sensor nodes for data acquisition, and each node completes data acquisition through an independent acquisition circuit or a synchronous trigger signal.

[0003] However, when deploying a large number of sensors, such methods often cause time clock drift, bus delay and trigger response difference, resulting in a deviation in the sampling time between different sensor channels, and it is difficult to ensure the global synchronization consistency of the data. In addition, due to the large dynamic range of signal amplitude in the hydrological environment, the output voltage of different sensors fluctuates significantly, and the problems of signal saturation or low signal-to-noise ratio are easy to occur, thereby affecting the accuracy and reliability of subsequent data.

[0004] Some existing solutions attempt to align the data in time and compensate for the gain through software, but due to hardware delay and sampling accuracy, the synchronization and real-time performance are still difficult to meet the needs of high-precision hydrological monitoring, especially in the multi-modal data fusion scene, the accumulation of timing errors and signal distortion will further reduce the effectiveness of fusion analysis.

[0005] In summary, how to ensure high-precision synchronization of multi-sensor while considering channel adaptive regulation and data fusion quality is still a problem to be solved in the existing technology. SUMMARY

[0006] The present application provides a multi-sensor fusion multi-modal hydrological data synchronous acquisition method, which is used to solve the technical problem of how to ensure high-precision synchronization of multi-sensor while considering channel adaptive regulation and data fusion quality in the existing technology.

[0007] In view of the above problems, the present application provides a multi-sensor fusion multi-modal hydrological data synchronous acquisition method.

[0008] The application provides a multi-sensor fusion multi-modal hydrological data synchronous acquisition method, which comprises the following steps: connecting a plurality of sensors arranged in a hydrological region to an intelligent PGA module to form a synchronous acquisition topology, wherein the sensors are connected to the intelligent PGA module in a one-to-one correspondence, and the synchronous acquisition topology is a star-bus hybrid topology; according to the synchronous acquisition topology, a central main controller is used to issue an acquisition instruction package to perform hardware synchronization of the absolute timestamp of the intelligent PGA module to generate a sampling edge marker; according to the sampling edge marker, the multi-sensor channel is marked, aligned and switched for autonomous sampling, and the relative independent acquisition of the sensing channel of the multi-sensor is performed under the adaptive gain control based on the intelligent PGA module to obtain multi-channel sensing flow data.

[0009] The one or more technical solutions provided in the application have at least the following technical effects or advantages: The multi-sensor fusion multi-modal hydrological data synchronous acquisition method provided by the application is used for solving the technical problem of how to ensure the high-precision synchronization of the multi-sensor while taking into account the channel adaptive control and the data fusion quality in the prior art. BRIEF DESCRIPTION OF DRAWINGS

[0010] Figure 1 A multi-sensor fusion multi-modal hydrological data synchronous acquisition method flowchart is provided for the application. Figure 2 A structure diagram of a synchronous acquisition topology in the multi-sensor fusion multi-modal hydrological data synchronous acquisition method is provided for the application. DETAILED DESCRIPTION

[0011] The application provides a multi-sensor fusion multi-modal hydrological data synchronous acquisition method, which is used to solve the technical problem of how to ensure the high-precision synchronization of the multi-sensor while taking into account the channel adaptive control and the data fusion quality in the prior art.

[0012] AsFigure 1 As shown, the present application provides a multi-sensor fusion multi-modal hydrological data synchronous acquisition method, which comprises: S1: For the multi-sensor deployed in the hydrological region, connect to the intelligent PGA module to form a synchronous acquisition topology, wherein the sensor and the intelligent PGA module are one-to-one connection, and the synchronous acquisition topology is a star-bus hybrid topology.

[0013] In the embodiment of the present application, multi-sensors are deployed in the hydrological region, wherein the hydrological region is a target environment with representative hydrological characteristics, for example, river section, lake water area or groundwater observation point. To achieve all-round dynamic monitoring of the region, various types of sensors are arranged at different spatial positions, including but not limited to water level sensor, flow rate sensor, water quality sensor and rainfall sensor, etc., to obtain original hydrological signals with wide coverage and various data types.

[0014] Among them, the above-mentioned multi-sensor needs to be connected to the intelligent PGA module, and the intelligent PGA module is built by the intelligent programmable gain amplifier. In a specific embodiment provided by the present application, the intelligent PGA module integrates adjustable gain amplification circuit and special hardware device of local logic processing unit. Its characteristic is that it can automatically adjust the gain according to the input signal voltage range to ensure that the output signal is in the optimal dynamic range, avoiding oversaturation or too low signal-to-noise ratio. Each sensor is equipped with an independent intelligent PGA module, that is, the special connection mode in the present application, which ensures the one-to-one correspondence of the signal processing link, thereby avoiding signal interference between multiple sensors, and enhancing the stability and independence of the data acquisition path.

[0015] On this basis, multiple intelligent PGA modules form a synchronous acquisition topology according to a specific connection mode, realizing the unified clock reference and sampling boundary consistency of multi-channel signal acquisition. In the present application, the form of the synchronous acquisition topology is in the form of star-bus hybrid.

[0016] Among them, the one-to-one relationship between the sensor and the intelligent PGA module, that is, each sensor corresponds to only a single PGA module, which is simple in structure and easy to expand; while the bus structure is reflected in the connection between multiple PGA modules and the central main controller through a unified data bus, that is, the central controller as the center node, and the radial connection relationship of multiple PGA modules as multiple parallel nodes. Exemplarily, the data bus can use the commonly used CAN bus or RS-485 bus as the implementation mode, thereby ensuring the efficiency of data communication and synchronous instruction issuing between multiple modules.

[0017] In summary, through the star-bus hybrid topology, not only the parallel access and centralized control of multi-sensor are realized, but also the expansion flexibility and control centralization are considered on the physical structure.

[0018] Further, as shown in Figure 2 The step S1 of the application comprises: a plurality of intelligent PGA modules connected with the plurality of sensors, the plurality of intelligent PGA modules being connected to a central main controller through a data bus, wherein the intelligent PGA module is built by a programmable gain amplifier; a connection between the central main controller and a reference clock source is established; and wherein the connection architecture of the plurality of sensors, the plurality of intelligent PGA modules, the central main controller and the reference clock source constitutes the synchronous acquisition topology.

[0019] In the embodiment of the application, a plurality of intelligent PGA modules connected with the plurality of sensors are acquired, and in the implementation manner provided by the application, the intelligent PGA module is an intelligent programmable gain amplifier module, and the core function thereof is to perform real-time gain adjustment on the electrical signal output by the sensor through an internal programmable amplification circuit, so as to ensure that the signal amplitude is within the dynamic range that can be recognized by the sampling device.

[0020] In the specific implementation process, because the output signal amplitudes of different types of hydrological sensors are different, for example, a water level sensor can output a weak electrical signal of millivolt level, and a flow rate sensor can output a voltage signal of volt level, therefore, an independent intelligent PGA module is provided for each sensor, which can avoid the unified amplification failure problem caused by inconsistent signal strength, thereby realizing differentiated processing and high-fidelity acquisition.

[0021] Further, the plurality of intelligent PGA modules are connected to the central main controller through a data bus. The data bus here refers to a hardware link for realizing shared communication channels of multiple devices, for example, an I2C bus, an SPI bus or an industrial CAN bus. Through the data bus, the plurality of intelligent PGA modules can simultaneously access the central main controller, and the central main controller can issue instructions to each module in a broadcast or addressing manner and receive acquisition data.

[0022] For example, in a typical application, the central main controller can issue a synchronous start instruction to all intelligent PGA modules at one time, to ensure the time consistency of multi-channel sampling actions.

[0023] The intelligent PGA module is built by a programmable gain amplifier, that is, the flexibility and fine control of gain adjustment are realized through a programmable gain amplifier chip at the hardware level. Specifically, the programmable gain amplifier refers to a gain multiplier that is no longer fixed, but can be switched between a plurality of gain gears according to the instruction input by the control port, for example, programmable adjustment between 1 times, 10 times and 100 times.

[0024] In combination with the application scenario of the present application, the intelligent PGA module further adds a local logic unit for dynamically determining whether gain adjustment is needed according to the voltage state during the acquisition process, thereby not only improving adaptability but also ensuring that data acquisition remains within a high-precision range when a hydrological environment signal fluctuates greatly.

[0025] Subsequently, a connection of the central main controller and a reference clock source is established, aiming to provide a unified time reference for the entire synchronous acquisition. In a feasible implementation manner, the reference clock source can be a high-precision crystal oscillator, a GPS time service module or an atomic clock module, which functions to provide an absolute time stamp and a stable clock pulse. The central main controller can attach a unified absolute time identifier when issuing an instruction, thereby ensuring time consistency when different PGA modules execute the instruction, by connecting with the reference clock source.

[0026] For example, in a multi-channel sampling task, the central main controller generates a unified trigger time stamp according to the reference clock and broadcasts it to all PGA modules, and each module will strictly execute gain switching and sampling starting according to the time stamp after local clock calibration, thereby avoiding error accumulation caused by local crystal oscillator drift.

[0027] Among them, the connection architecture of the multi-sensor, the plurality of intelligent PGA modules, the central main controller and the reference clock source constitutes the synchronous acquisition topology. The core lies in multi-level division of labor and cooperation: the sensor provides the original physical quantity as the signal source; the intelligent PGA module serves as a front-end processing unit to execute signal adjustment and local synchronous response; the central main controller serves as a global scheduling core to be responsible for instruction issuing and data integration; and the reference clock source serves as a global time anchor point to ensure the accuracy of synchronization.

[0028] In summary, the synchronous acquisition topology constituted by the architecture not only realizes time alignment of multi-channel data, but also provides good scalability and maintainability in structure. For example, when a water quality sensor needs to be added, only an intelligent PGA module needs to be added and connected to the bus, which can be seamlessly integrated into the overall synchronous architecture.

[0029] S2: According to the synchronous acquisition topology, a central main controller issues an acquisition instruction package to execute absolute time stamp hardware synchronization of the intelligent PGA module, and generates a sampling edge marker.

[0030] In the embodiment of the present application, according to the synchronous acquisition topology, the central main controller can serve as a global scheduling core to issue an acquisition instruction package with a unified time identifier.

[0031] Specifically, the collection instruction package refers to a set of control information generated by the central main controller and broadcast to each intelligent PGA module via the data bus, which at least includes an instruction code for triggering sampling, an absolute timestamp for calibrating a uniform time, and a target parameter related to gain. Through unified issuance of the instruction package, it is ensured that all intelligent PGA modules receive the same synchronization instruction and establish consistency in the start of collection at the hardware level.

[0032] Subsequently, the absolute timestamp of the intelligent PGA module is executed. Specifically, after receiving the collection instruction package, each intelligent PGA module does not immediately trigger the collection action, but matches the local countdown timer or internal clock according to the absolute timestamp attached in the instruction package. When the local time and the absolute timestamp are consistent, the module will automatically trigger the corresponding operation, such as gain switching or sampling start.

[0033] Wherein, the absolute timestamp here is derived from the connection of the central main controller and the reference clock source, which has uniform and high-precision timing characteristics, thereby avoiding the triggering deviation caused by local clock differences and realizing true hardware-level global synchronization.

[0034] For example, in one sampling period, the triggering error of all PGA modules can be controlled within the nanosecond range, ensuring strict consistency of the multi-channel sampling boundary.

[0035] Synchronously, a sampling edge marker is generated, that is, when the absolute timestamp triggers the gain switching or sampling action, a transient characteristic voltage with an abrupt amplitude change will appear at the output end of the intelligent PGA module. This characteristic change is defined as the sampling edge marker.

[0036] Wherein, the sampling edge marker, as a kind of time reference symbol at the hardware level, can be used to identify the occurrence time of the sampling action of each channel, thereby realizing channel alignment and fusion in the subsequent data processing stage.

[0037] Further, by the central main controller issuing the collection instruction package, the absolute timestamp of the intelligent PGA module is executed, and step S2 of the present application comprises: The central main controller generates a collection instruction package, wherein the collection instruction package is composed of at least an instruction code, an absolute timestamp, and a target gain value, the instruction code is a gain synchronization sequence instruction, and the absolute timestamp is generated according to a reference clock source; the central main controller broadcasts the collection instruction package to the plurality of intelligent PGA modules through the data bus; each intelligent PGA module starts the countdown timer of the internal clock until the local countdown timer matches the absolute timestamp, triggers the gain switching, and updates the gain register in the intelligent PGA module to the target gain value.

[0038] In the embodiments of the present application, after the entire topology is established and it is confirmed that each intelligent PGA module is in a controllable state, a collection instruction package is generated according to a collection demand by the central main controller, that is, a data instruction set constructed by the central main controller at a specific moment according to the sampling demand, which at least consists of an instruction code, an absolute timestamp and a target gain value.

[0039] In the present application, the instruction code is specifically defined as a gain synchronization sequence instruction, that is, a command sequence for coordinating gain adjustment of each intelligent PGA module, which is used to ensure that multiple channels remain consistent when performing gain switching. The absolute timestamp is generated according to a reference clock source connected to the central main controller, and the timestamp has the characteristics of global uniformity and high precision, which is the core basis for realizing hardware synchronization. The target gain value is a preset and unified low gain value, for example, G = 1, that is, unit gain.

[0040] Subsequently, the central main controller broadcasts the collection instruction package to the multiple intelligent PGA modules through the data bus, that is, at the physical connection level, the broadcast mechanism of the bus is used to distribute the instruction to all mounted intelligent PGA modules at one time, so that the same instruction content can be received within the same time window, avoiding the difference in the order of instruction transmission caused by communication delay.

[0041] Subsequently, each intelligent PGA module starts a countdown timer of the internal clock, that is, after receiving the collection instruction package containing the absolute timestamp, each intelligent PGA module does not immediately perform gain adjustment, but sets a countdown process using the internal clock circuit. The role of the countdown timer is to delay the local operation to the target time specified by the absolute timestamp, so as to realize the alignment with the global time.

[0042] Further, until the local countdown timer matches the absolute timestamp, triggering gain switching, that is, when the internal countdown ends and is completely consistent with the absolute timestamp, each intelligent PGA module will immediately perform gain switching operation. That is, the programmable gain amplifier inside the module changes its amplification factor, and the specific operation is to update the gain register in the intelligent PGA module to the target gain value.

[0043] It is explained that at the moment when the triggering action occurs, the target gain value is written into the gain register, and the gain register is a key storage unit for controlling the amplification factor, which immediately drives the amplification circuit to complete the physical gain adjustment. Synchronous gain adjustment of multiple channels under the global time is realized, and the time consistency of multi-modal hydrological data collection is ensured.

[0044] Further, after triggering gain switching, step S2 of the present application comprises: A sampling edge mark is generated with the triggering of the gain switching, wherein the output voltage of each intelligent PGA module has an abrupt change in amplitude with the triggering of the gain switching, forming a sampling edge mark on a time axis; wherein the sampling edge mark is a hardware synchronization action mark corresponding to multiple sensing channels of the multi-sensor.

[0045] In the embodiments of the present application, generating a sampling edge mark with the triggering of the gain switching refers to that at the moment when the acquisition instruction package issued by the central main controller is executed and the gain register update is completed by each intelligent PGA module, the output signal will have a specific physical change, which is defined as a sampling edge mark. The edge here is not a virtual identifier in the logical sense, but an actual electrical signal feature, which is attached to the mutation process of the voltage waveform and is used to identify the sampling starting point.

[0046] Specifically, with the triggering of the gain switching, the output voltage of each intelligent PGA module has an abrupt change in amplitude, that is, when the internal amplification factor is switched under the target timestamp trigger, the input signal at the output end of the amplification channel will show a sharp change in voltage amplitude.

[0047] For example, when the output voltage of a certain water level sensor is 0.1V and the target gain value is switched from 1 times to 10 times, the module output voltage will instantaneously jump to 1.0V, and the voltage mutation presents a clear steep edge on the time axis. In the present application, since the present switching is a response based on absolute timestamp, it can be used as a synchronization trigger point for hardware sampling.

[0048] In the present application, the above voltage mutation event is abstracted as a unified time reference symbol as a sampling edge mark on the time axis. Each intelligent PGA module will generate a similar voltage edge when performing gain switching, thereby forming a series of consistent mark points on the time axis. These mark points have global synchronization properties because they are triggered by the same absolute timestamp and implemented at the hardware circuit level, so they can achieve time alignment across multiple channels. That is, as a shared synchronization reference for the entire multi-sensor system.

[0049] In other words, the data collected by any sensor channel can be aligned with the data of other channels through the corresponding sampling edge mark.

[0050] For example, in a river monitoring scenario, the flow rate sensor and the water quality sensor trigger gain switching at different positions at the same time, and both will generate a sampling edge mark in their output waveforms. These marks, as a manifestation of global hardware synchronization actions, enable the data from different channels to be strictly time-unified in the subsequent fusion processing stage, thereby ensuring high-precision synchronous collection of multi-modal hydrological data.

[0051] S3: According to the sampling edge mark, the mark alignment of multiple sensing channels is performed, and the relative independent collection of the sensing channels of the multiple sensors is performed under the adaptive gain regulation based on the intelligent PGA module, and the multi-channel sensing stream data is obtained.

[0052] In this application, after all the intelligent PGA modules perform gain switching and generate a unified voltage mutation at the output end, the mutation point is used as a time reference to calibrate the data stream from each sensor in time sequence by using the synchronization of the mutation point on different channels.

[0053] Specifically, the mark alignment is performed by identifying the sampling edge mark appearing in each channel waveform and taking it as the sampling starting point, so as to eliminate the sampling misalignment caused by the differences in sensor characteristics, wiring delay or local clock deviation.

[0054] For example, in river section monitoring, the flow rate sensor and the water level sensor may be located at different positions, but by aligning the edge marks, the data streams can be accurately compared and fused at the same time.

[0055] Correspondingly, after completing the global alignment based on the sampling edge mark, the central main controller switches each intelligent PGA module from the unified synchronization mode to the independent autonomous sampling mode.

[0056] Specifically, each intelligent PGA module no longer depends on the instructions of the central controller, but completes the rhythm scheduling of data collection according to the sampling state and the judgment of the local logic unit. The sampling rate or collection strategy can be flexibly adjusted according to the real-time changes of the channel signals, so as to avoid the problems of low resource utilization or high data redundancy in the global synchronization mode.

[0057] Furthermore, during the independent operation of each intelligent PGA module, the output voltage is monitored in real time, and whether the amplification factor needs to be adjusted is automatically determined in combination with the preset standard condition range, so as to ensure that the data is in the effective sampling interval.

[0058] The adaptive gain regulation proposed in this application means that when the sampling signal is too strong and close to saturation, the module automatically reduces the gain; when the signal is too weak and affects the signal-to-noise ratio, the module automatically increases the gain; and when the signal is in the normal range, the existing gain value remains unchanged. In order to maintain the quality stability of the data of each channel.

[0059] In summary, under the cooperation of the above alignment and independent collection mechanism, the continuous data stream collected by multiple sensors in parallel is finally obtained. The data stream is not only strictly synchronized in the time dimension, but also has a high accuracy after gain optimization in the amplitude characteristics.

[0060] For example, when monitoring rainfall intensity, river water level and flow rate simultaneously, the aligned data stream of the three types of signals can be obtained at the same time point, and the amplitude of each signal is ensured to be within the optimal sampling range, thereby providing a solid foundation for subsequent multi-modal hydrological data fusion and comprehensive analysis.

[0061] Further, according to the sampling edge mark, the marking alignment and channel autonomous sampling switching of the multiple sensing channels are performed, and the step S3 includes: The multiple sensing channels are marked and aligned, and the sampling edge mark is used as the alignment basis; after the marking alignment is completed, the central main controller generates a mode switching instruction, which is broadcast to the multiple intelligent PGA modules through a data bus, each intelligent PGA module is switched from the synchronous mode to the autonomous sampling mode, and the multiple sensors perform autonomous sampling management of the sensing channels.

[0062] In the process of multi-channel data acquisition, the sampling edge mark generated in the previous step is used as a time anchor point to unify the sampling starting points or sampling sequences of different channels.

[0063] In the actual sampling process, due to the differences in spatial position distribution, line length and local response speed of the multiple sensors, the sampling signals inevitably have time deviations in the original state. By identifying the sampling edge mark formed by the voltage mutation in the sampling waveform of each channel, the unified starting point of sampling can be accurately located, thereby realizing the synchronization correction between channels.

[0064] In other words, the marking alignment process establishes a shared time reference system for different sensing channels through the hardware edge mark, so that the channel data realizes a strict one-to-one correspondence in time sequence.

[0065] The sampling edge mark is used as the alignment basis, which further limits the execution mode of the alignment process, that is, it does not rely on software post-processing or time compensation, but directly uses the edge mark generated by the hardware as the only alignment standard.

[0066] Compared with the traditional software post-processing method, the method proposed in the application has higher precision and lower delay, because the sampling edge mark is derived from the voltage mutation under the unified absolute timestamp, and can ensure consistency on the nanosecond time scale. For example, in a collection system including ten sensing channels, the edge mark will appear in the output waveform of each channel at the same time point, and by identifying these marks, the sampling streams of the ten channels can be aligned.

[0067] Subsequently, after completing the mark alignment, the central main controller generates a mode switching instruction, which means that after completing the mark alignment phase, the central main controller will generate a new control instruction to inform each intelligent PGA module to exit the current synchronization mode and enter a relatively independent running mode for the collection process.

[0068] In the specific implementation, the switching command is broadcast to the plurality of intelligent PGA modules through a data bus, received and executed in the same time window, each intelligent PGA module exits the synchronization mode and switches to the autonomous sampling mode, that is, the dependence on the global synchronization clock and the edge mark is closed, the synchronization collection alignment has been completed, and then the independent sampling logic of the intelligent PGA module is entered.

[0069] Specifically, in the autonomous sampling mode, each intelligent PGA module autonomously determines the sampling frequency, trigger condition and gain adjustment strategy through internal algorithms and circuit mechanisms, thereby realizing localized data collection management.

[0070] Each sensor independently runs and collects data under the regulation of its corresponding intelligent PGA module. The connotation of management is that the intelligent PGA module will decide whether to adjust the gain or change the sampling parameters according to the real-time monitored signal state, so as to ensure that the collected data is always in the optimal range.

[0071] Further, switching to the autonomous sampling mode, the plurality of sensors perform autonomous sampling management of the sensing channel, and the step S3 of the present application comprises: According to the first intelligent PGA module, the data sampling management of the sensing channel of the first sensor is performed, wherein the first intelligent PGA module is any intelligent PGA module, and the first sensor is the sensor associated with the first intelligent PGA module; wherein the sampling signal state of the sensing channel of the first sensor is determined by monitoring the output voltage, and gain adaptive control is performed.

[0072] In the embodiment of the present application, the data sampling management of the sensing channel of the first sensor is performed according to the first intelligent PGA module. Specifically, any connected intelligent PGA module is taken as an execution object, and the local data sampling management of the sensor associated therewith is carried out.

[0073] In the present application, the first intelligent PGA module is not limited to a specific number, but is used to illustrate that the mechanism is applicable to any module; similarly, the first sensor refers to any sensor corresponding thereto. It is ensured that the collection, amplification and regulation of each sensor signal are completed in a dedicated hardware unit, avoiding the interference or priority conflict that may be caused by multi-channel shared processing.

[0074] For example, although the water level sensor and the water quality sensor are connected to the system at the same time, their signal collection paths are completely independent in hardware and do not affect each other.

[0075] Wherein, by monitoring the output voltage, the sampling signal state of the sensing channel of the first sensor is judged, that is, the intelligent PGA module detects the voltage amplitude of the output end in the acquisition process in real time, which is used as the judgment basis of signal quality and dynamic range.

[0076] Wherein, if the output voltage is close to full range, it means that the signal is too strong and there is a risk of saturation; if the output voltage is too low, it means that the signal is weak and the signal-to-noise ratio is poor; if the voltage is in the normal range, it means that the current gain setting is reasonable. Through this real-time monitoring, the intelligent PGA module can quickly and accurately obtain the channel sampling state, and based on this, the gain adaptive control is carried out.

[0077] In summary, it is ensured that the channel data of the first sensor is always in the optimal sampling interval, thereby improving the stability and reliability of the overall acquisition system.

[0078] Further, the sampling signal state of the sensing channel of the first sensor is judged, and the gain adaptive control is carried out, and the step S3 of the present application comprises: If the output voltage monitored by the first intelligent PGA module is close to full range, a first sampling signal state is generated, the first intelligent PGA module is down-regulated, and a gain down-regulation event is recorded, wherein the first sampling signal state represents that the sampling signal of the channel sensing data is too strong and there is a risk of distortion; if the output voltage monitored by the first intelligent PGA module is low, a second sampling signal state is generated, the first intelligent PGA module is up-regulated, and a gain up-regulation event is recorded, wherein the second sampling signal state represents that the sampling signal of the channel sensing data is too weak and the signal-to-noise ratio is poor; if the output voltage monitored by the first intelligent PGA module is in the normal range, the gain is kept, wherein each intelligent PGA module corresponds to a normal range based on the sensing type.

[0079] In the embodiment of the present application, if the output voltage monitored by the first intelligent PGA module is close to full range, that is, the voltage amplitude has reached the upper limit region of the set preset threshold value, for example, when the full range is 5V, the voltage exceeding 4.5V enters the state close to full range, this state is judged as the first sampling signal state. It means that the signal is too strong and will be distorted, the first intelligent PGA module will automatically reduce the gain in microseconds to prevent saturation, and record the gain down-regulation event and time.

[0080] Specifically, the amplification factor of the internal programmable amplifier is immediately reduced to ensure that the output signal returns to the safe range, thereby avoiding overload distortion. At the same time, the occurrence time, down-regulation amplitude and corresponding sensing channel of the gain adjustment are recorded for subsequent data tracking and state analysis.

[0081] Synchronous, if the output voltage monitored by the first intelligent PGA module is in the low position, that is, when the output voltage value is continuously maintained in the lower limit area close to zero, specifically, for example, the voltage amplitude is lower than the set preset threshold, for example, when the full range is 5V, if the voltage is lower than 0.5V, it indicates that the amplified signal is insufficient to reflect the original data characteristics of the sensor, and the second sampling signal state is generated.

[0082] Among them, the problem corresponding to the second sampling signal state is that the input signal is too small, which is not only difficult to be accurately captured by the sampling circuit, but also may be submerged in noise, resulting in a decrease in effective data rate. Therefore, gain up is a necessary compensation means.

[0083] Specifically, after determining the second sampling signal state, the gain multiple is automatically increased to amplify the weak signal and increase the signal amplitude so that it enters a reasonable dynamic range. At the same time, this operation is also recorded as a gain up event.

[0084] Synchronous, if the output voltage monitored by the first intelligent PGA module is in the low position, that is, when the output voltage value is continuously maintained in the lower limit area close to zero, specifically, for example, the voltage amplitude is lower than the set preset threshold, for example, when the full range is 5V, if the voltage is lower than 0.5V, it indicates that the amplified signal is insufficient to reflect the original data characteristics of the sensor, and the second sampling signal state is generated.

[0085] Further, after gain self-adaptive control, step S3 of the present application comprises: According to the recorded gain down event and gain down event, a gain event sequence is generated, wherein each sensor sampling corresponds to a gain event sequence, and wherein the gain event element at least includes time and gain amount; for the gain event sequence, a gain change curve is generated; and according to the gain change curve, hydrological event verification is performed.

[0086] In the embodiment of the present application, in the foregoing adaptive control process, each gain adjustment operation will form an event record. Whether it is a down adjustment triggered by an excessively strong output voltage or an up adjustment triggered by an excessively weak voltage, it will be recorded by the system. The recorded discrete event points are arranged in time sequence, that is, the gain event sequence is formed.

[0087] Among them, the gain event sequence reflects the response of a sensor channel to signal dynamics in a specific sampling period, so its essence is the gain adjustment trajectory in the time dimension.

[0088] Each intelligent PGA module independently monitors and controls the sampling of the sensor connected thereto, and thus each channel generates an independent gain event sequence.

[0089] For example, a water level sensor may frequently trigger a down-regulation event during a flood peak, while a water quality sensor may trigger an up-regulation event when background noise is strong.

[0090] The gain event element includes, but is not limited to, time and gain amount, wherein the time is used to identify the specific time of the event occurrence, and the gain amount is used to represent the specific adjustment amplitude, such as a down-regulation of 5 dB or an up-regulation of 10 dB. The occurrence background and amplitude change of a gain regulation behavior can be accurately described.

[0091] Subsequently, a gain change curve is generated for the gain event sequence, that is, a continuous curve trajectory is formed through the mapping relationship between the time horizontal axis and the gain vertical axis. The gain dynamic change law of a certain sensor channel in the entire sampling period is intuitively displayed, and the process of signal strength fluctuation can also be reflected. For example, when the water level rises sharply in a short time, the curve will show continuous down-regulation actions; when the signal environment tends to be stable, the curve enters a relatively flat stage.

[0092] In a further preferred implementation process, hydrological event verification is performed according to the gain change curve, that is, the dynamic characteristics of gain regulation are combined with actual hydrological events to cross-verify the occurrence and development of hydrological events.

[0093] Specifically, since hydrological events such as sudden flood rise, sudden drop or water quality mutation are usually accompanied by sharp fluctuations in signal amplitude, the gain change curve can serve as evidence for the authenticity and reasonableness of the events.

[0094] For example, when the raw data of a flow rate sensor indicates that the water flow is accelerating, and the corresponding gain curve shows continuous down-regulation in the same period, it can be verified that the flow rate data indeed comes from a real hydrological phenomenon, rather than accidental noise or sensor failure.

[0095] For another example, at a certain time point, the gain starts to continuously increase to cope with the weakening of the turbidity signal. The starting time of a high-turbidity sewage group can be clearly indicated. The amplitude and rate of gain increase can indirectly reflect the intensity and speed of the increase in turbidity of the sewage group. When the gain sequence returns to stability, it indicates the end of the sewage group passing.

[0096] In summary, not only data is collected, but also the reliability of hydrological events is verified based on the dynamic characteristics of hardware, which enhances the credibility of the overall monitoring results.

[0097] Further, after obtaining the multi-channel sensor flow data, the step of the application further includes: The time alignment of the multi-channel sensing flow data is performed through the sampling edge marks of the channels, and synchronous flow data is determined; the comprehensive hydrological state of the hydrological region is evaluated by performing multi-sensing fusion by traversing the synchronous flow data.

[0098] Further, the step also includes generating directional feedback information according to the comprehensive hydrological state, and sending the directional feedback information to a central main controller for sensing collection constraint.

[0099] In the embodiment of the application, for the multi-channel sensing flow data, after the multi-sensor autonomous sampling is completed, the data streams generated by different channels need to be integrated through a unified time reference. The sampling edge mark is a voltage mutation signal generated by the intelligent PGA module when the gain is switched, and has natural synchronization characteristics, which can be used as the only reference point for cross-channel data alignment. By identifying and matching the edge mark positions in the data streams of the channels, all sensing signals can be rearranged on a unified time axis, thereby obtaining accurate alignment effect.

[0100] In summary, a set of data stream sets completely consistent in time scale is obtained. Thus, the deviation caused by hardware delay or asynchronous sampling is eliminated.

[0101] Subsequently, the synchronous flow data is traversed, and the data of all channels are jointly processed and comprehensively analyzed. Specifically, different physical quantities, i.e., data from different sensors, such as water level, flow rate, rainfall, water quality indicators, etc., are cooperatively calculated and feature extracted, so as to make up for the deficiencies of a single sensor in spatial coverage, time response or signal characteristics.

[0102] For example, the fusion method can include any one of weighted average, comprehensive discrimination based on feature vector, etc.

[0103] For example, in flood monitoring, relying solely on water level data may not be able to accurately predict sudden water rise, but combining flow rate and rainfall data for fusion can greatly improve the accuracy of prediction.

[0104] Subsequently, through the above fused multi-modal data, the overall hydrological condition of the target region can be scientifically judged. Not only the water level, flow rate, but also the rainfall intensity, water quality change and their internal relationship are included. Through fusion analysis, a comprehensive evaluation result of the regional hydrological environment can be output.

[0105] For example, whether there is a flood risk, whether there is a sudden pollution event, or whether it is in the dry season. The advantage is that hydrological monitoring not only stays at the single-point data collection level, but also realizes the overall dynamic evaluation of the region.

[0106] Further, according to the comprehensive hydrological state, the collected focused direction is determined, such as a sudden pollution event, and the water quality is taken as the corresponding collection focused direction, that is, the generated directional feedback information. Then, the directional feedback information is sent to the central main controller, and the central controller performs cooperative coordination between multiple sensors according to the directional feedback information, performs sensor collection constraint, and further ensures the effectiveness of data collection.

[0107] The person skilled in the art can clearly understand the multi-sensor fusion multi-modal hydrological data synchronous collection method in the embodiment according to the foregoing detailed description of the multi-sensor fusion multi-modal hydrological data synchronous collection method. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the related part is referred to the method part description.

[0108] The above description of disclosed embodiments enables a person skilled in the art to implement or use the present application. Various modifications to the embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A multi-sensor fusion multi-modal hydrological data synchronous acquisition method, characterized in that, The method comprises: A plurality of sensors deployed in a hydrological region are connected to an intelligent PGA module to form a synchronous acquisition topology, wherein the sensors and the intelligent PGA module are one-to-one connected, and the synchronous acquisition topology is a star-bus hybrid topology; According to the synchronous acquisition topology, a central main controller is used to issue an acquisition instruction package to perform absolute timestamp-based hardware synchronization of the intelligent PGA module to generate a sampling edge marker; According to the sampling edge marker, the marking alignment and channel autonomous sampling switching of the plurality of sensing channels are performed, and the relative independent acquisition of the sensing channels of the plurality of sensors is performed under the adaptive gain control of the intelligent PGA module to obtain multi-channel sensing flow data.

2. The multi-sensor fusion multi-modal hydrological data synchronous acquisition method of claim 1, wherein, The synchronous acquisition topology comprises: A plurality of intelligent PGA modules connected to the plurality of sensors are obtained, and the plurality of intelligent PGA modules are connected to a central main controller through a data bus, wherein the intelligent PGA module is built by a programmable gain amplifier; The connection between the central main controller and a reference clock source is established; The connection architecture of the plurality of sensors, the plurality of intelligent PGA modules, the central main controller, and the reference clock source forms the synchronous acquisition topology.

3. The multi-sensor fusion multi-modal hydrological data synchronous acquisition method of claim 2, wherein, The central main controller is used to issue an acquisition instruction package to perform absolute timestamp-based hardware synchronization of the intelligent PGA module, comprising: The central main controller generates an acquisition instruction package, wherein the acquisition instruction package is composed of at least an instruction code, an absolute timestamp, and a target gain value, the instruction code is a gain synchronization sequence instruction, and the absolute timestamp is generated according to a reference clock source; The central main controller broadcasts the acquisition instruction package to the plurality of intelligent PGA modules through the data bus; Each intelligent PGA module starts an internal clock countdown timer until the local countdown timer matches the absolute timestamp to trigger gain switching and update the gain register in the intelligent PGA module to the target gain value.

4. The multi-sensor fusion multi-modal hydrological data synchronous acquisition method of claim 3, wherein, After triggering gain switching, comprising: With the triggering of the gain switching, a sampling edge marker is generated, wherein with the triggering of the gain switching, the output voltage of each intelligent PGA module has an amplitude abrupt change to form a sampling edge marker on the time axis; The sampling edge marker is a hardware synchronization action marker corresponding to the plurality of sensing channels of the plurality of sensors.

5. The multi-sensor fusion multi-modal hydrological data synchronous acquisition method of claim 2, wherein, According to the sampling edge marker, the marking alignment and channel autonomous sampling switching of the plurality of sensing channels are performed, comprising: The plurality of sensing channels are marked and aligned, wherein the sampling edge marker is used as the alignment basis; After completing the marking alignment, the central main controller generates a mode switching instruction and broadcasts it to the plurality of intelligent PGA modules through the data bus, each intelligent PGA module is switched from the synchronous mode to the autonomous sampling mode, and the plurality of sensors perform sensing channel autonomous sampling management.

6. The multi-sensor fusion multi-modal hydrological data synchronous acquisition method of claim 5, wherein, Switching to the autonomous sampling mode, the plurality of sensors perform sensing channel autonomous sampling management, comprising: According to a first intelligent PGA module, a first sensor is subjected to sensing channel data sampling control, wherein the first intelligent PGA module is any intelligent PGA module, and the first sensor is a sensor associated with the first intelligent PGA module; The gain adaptive control is performed by monitoring the output voltage to determine the sampling signal state of the sensing channel of the first sensor.

7. The multi-sensor fusion multi-modal hydrological data synchronous acquisition method of claim 6, wherein, The gain adaptive control is performed by monitoring the output voltage to determine the sampling signal state of the sensing channel of the first sensor. If the output voltage monitored by the first intelligent PGA module is close to the full range, a first sampling signal state is generated, the first intelligent PGA module performs gain down-regulation, and a gain down-regulation event is recorded, wherein the first sampling signal state indicates that the sampling signal of the channel sensing data is too strong and there is a risk of distortion; If the output voltage monitored by the first intelligent PGA module is low, a second sampling signal state is generated, the first intelligent PGA module performs gain up-regulation, and a gain up-regulation event is recorded, wherein the second sampling signal state indicates that the sampling signal of the channel sensing data is too weak and the signal-to-noise ratio is poor; If the output voltage monitored by the first intelligent PGA module is in a standard condition range, the gain is kept, wherein each intelligent PGA module corresponds to a standard condition range based on the sensing type.

8. The multi-sensor fusion multi-modal hydrological data synchronous acquisition method of claim 7, wherein, After the gain adaptive control is performed, the following steps are included: According to the recorded gain down-regulation event and gain down-regulation event, a gain event sequence is generated, wherein each sensor sampling corresponds to a gain event sequence, and the gain event element at least includes time and gain amount; For the gain event sequence, a gain change curve is generated; According to the gain change curve, a hydrological event verification is performed.

9. The multi-sensor fusion multi-modal hydrological data synchronous acquisition method of claim 1, wherein, After the multi-channel sensing flow data is obtained, the following steps are included: The time alignment of the multi-channel sensing flow data is performed by the sampling edge mark of each channel to determine the synchronous flow data; The comprehensive hydrological state of the hydrological region is evaluated by performing multi-sensing fusion by traversing the synchronous flow data.

10. The method of claim 9, wherein, According to the comprehensive hydrological state, directional feedback information is generated; The directional feedback information is sent to the central main controller for sensing collection constraint.

Images