A method, system, and device for adaptive high-precision measurement of seawater depth and temperature based on CTD system.

Abstract

This invention discloses a method, system, and device for adaptive high-precision measurement of seawater depth and temperature based on a CTD system, belonging to the field of marine exploration technology. The method includes the following steps: S1: Equipment initialization and algorithm parameter calibration upon water entry; S2: Synchronous real-time acquisition of multiple parameters; S3: Layered adaptive variable frequency sampling; S4: Multi-level joint error correction; S5: Time-series fusion and output of temperature and depth data. Relying on existing mature CTD hardware architecture, through process design and algorithm optimization, it achieves full-depth adaptive sampling and multi-type error layered dynamic compensation, comprehensively improving the measurement accuracy, data stability, and consistency of seawater temperature and depth without any hardware modifications.

Description

Technical Field

[0001] This invention relates to a method, system, and device for adaptive high-precision measurement of seawater depth and temperature based on a CTD system, belonging to the field of marine exploration technology. Background Technology

[0002] Ocean temperature and depth are core basic observation parameters in the fields of mariculture, marine hydrology, marine meteorology, marine engineering, and marine resource exploration. CTD (Conductivity-Temperature-Depth) is currently the mainstream equipment for ocean profiling. It calculates seawater depth using pressure sensors and collects water temperature using temperature sensors. Due to its reliable structure and convenient operation, it is widely used in marine scientific research and engineering surveys.

[0003] Existing traditional CTD measurement methods have the following inherent drawbacks: (1) Fixed sampling mode: Fixed sampling frequency is generally adopted. Hydrological parameters in shallow sea areas change rapidly and turbulence is active. Fixed low-frequency sampling is prone to losing key data; the deep sea environment is stable. Fixed high-frequency sampling will cause data redundancy, increased equipment power consumption and shortened battery life.

[0004] (2) The problem of temperature thermal hysteresis is prominent: the temperature sensor has an inherent thermal conduction delay. During the continuous sinking of the CTD, the depth sampling time is ahead of the temperature sampling time, causing the temperature and depth data to be misaligned in time, the profile curve is distorted, and the temperature and depth time misalignment error can reach up to 1.8s.

[0005] (3) Significant pressure and temperature drift error: Depth is calculated from pressure. Temperature and pressure change drastically throughout the entire ocean depth range, and pressure sensors are prone to zero drift and sensitivity drift. Traditional methods only use factory-fixed compensation parameters, which cannot dynamically adapt to environmental changes. Depth measurement has a continuous systematic error, and the comprehensive error of depth measurement in the entire ocean depth environment can reach ±0.85m.

[0006] (4) Insufficient resistance to disturbances: Turbulence occurs frequently in nearshore and mid-water areas, and the equipment is prone to tilting and shaking during the descent. Conventional static filtering algorithms cannot effectively filter out such random interferences, resulting in large data fluctuations and low efficiency.

[0007] The existing patent application, CN121594945A, discloses a temperature and salinity sensor and its real-time error compensation method. It uses a fixed sampling frequency of 25Hz and calculates the device's descent speed based on pressure changes to achieve time alignment of temperature and salinity data. However, this scheme does not set an adaptive sampling frequency according to sea depth layers, uses only first-order response time matching for error compensation, does not incorporate a temperature-pressure coupling model for dynamic temperature drift correction, and does not introduce attitude and turbulence parameters for joint filtering.

[0008] Patent application publication number CN121876923A discloses a method for temperature time-domain alignment and error compensation of a temperature, salinity, and depth (TDM) instrument. This method determines the time delay through root mean square error and combines it with a random forest model to compensate for temperature residuals. However, this scheme does not correlate depth parameters for partitioned sampling and does not design a joint correction strategy for pressure sensor temperature drift and equipment attitude sway. The algorithm is highly complex and difficult to embed.

[0009] Current improvement solutions for CTDs mainly focus on hardware structure, sensor selection, wireless transmission, and housing pressure resistance design. There are few integrated measurement methods specifically addressing sampling strategies, multi-dimensional dynamic error compensation, and data timing matching, failing to systematically solve the core pain points at the methodological level. Developing a versatile, high-precision, and highly adaptive measurement method without modifying existing CTD hardware has significant practical engineering value. Summary of the Invention

[0010] The purpose of this invention is to overcome the shortcomings of the prior art and provide a method, system and device for adaptive high-precision measurement of seawater depth and temperature based on a CTD system. Relying on the existing mature CTD hardware architecture, through process design and algorithm optimization, it realizes full-ocean-depth adaptive sampling and multi-type error hierarchical dynamic compensation, and comprehensively improves the measurement accuracy, data stability and consistency of seawater temperature and depth without any hardware modification.

[0011] To achieve the above objectives, the present invention provides the following technical solution: The adaptive high-precision measurement method for seawater depth and temperature based on a CTD system described in this invention includes the following steps: S1: Equipment water immersion initialization and algorithm parameter calibration: After the CTD equipment is immersed in water, it completes hardware self-test, loads the preset temperature reference curve, temperature-pressure coupling error model parameters and filter initial parameters, and performs zero-point calibration on the micro inertial attitude detection unit. S2: Multi-parameter synchronous real-time acquisition: Synchronously acquire data on device motion attitude, vertical falling speed, seawater pressure, seawater temperature and turbulence intensity at a sampling rate not lower than the set frame rate, and calculate the detection depth in real time from the seawater pressure; S3: Layered adaptive variable frequency sampling: Based on the dual criteria of detection depth and vertical falling speed, the sea area is divided into multiple intervals, including at least the shallow sea high dynamic zone, the middle layer transition zone and the deep sea steady state zone. Different sampling frequencies are used in each interval, and the sampling frequency is switched instantly and seamlessly when crossing intervals. S4: Multi-level joint error correction: This involves sequentially performing three levels of correction: temperature thermal hysteresis compensation, pressure temperature drift dynamic correction, and attitude and turbulence disturbance dynamic filtering. The temperature thermal hysteresis compensation adopts a time-series differential interpolation algorithm based on the reference temperature curve. The dynamic hysteresis time is calculated based on the inherent response delay of the sensor and the real-time falling speed. The original temperature sequence is corrected point by point to eliminate the time misalignment between temperature sampling and depth sampling. The pressure-temperature drift dynamic correction is achieved by substituting the real-time seawater temperature into a pre-constructed temperature-pressure coupling model, iteratively updating the zero-point offset and sensitivity coefficient of the pressure sensor in real time, correcting the pressure value, and converting it into the corrected seawater depth value. The attitude and turbulence disturbance dynamic filtering dynamically adjusts the sliding filter window length and filter threshold according to the real-time pitch angle, roll angle and turbulence intensity to filter out random noise generated by equipment tilt and seawater turbulence. S5: Time-series fusion and output of temperature and depth data: Align the corrected temperature and depth data one by one along the time axis, identify and remove gross outliers based on statistical criteria, generate standard hydrological profile data, and complete storage and transmission.

[0012] Further, step S1 specifically includes: after the CTD device is immersed in water, it completes a self-test of the temperature sensor, pressure sensor, micro-inertial attitude detection unit, data storage module, and communication module. If the hardware is abnormal, an alarm is triggered and the detection is terminated; if the hardware is normal, all algorithm parameters, including the low temperature reference curve, the normal temperature reference curve, the temperature-pressure coupling model coefficients, the initial filtering parameters, and the sampling frequency threshold, are loaded from the built-in parameter library; the pitch angle, roll angle, and vertical velocity of the micro-inertial attitude detection unit are zero-point calibrated; after the self-test and parameter loading are completed, the device enters the detection standby state.

[0013] Furthermore, the motion attitude in step S2 includes pitch angle and roll angle; the seawater pressure is converted into the detection depth in real time using a hydrostatic formula based on seawater density and gravitational acceleration; all parameters are collected synchronously, the collection frame rate is not lower than the set reference value, and the synchronization error between parameters is controlled within the preset tolerance.

[0014] Furthermore, the method for determining the layered adaptive variable frequency sampling in step S3 is as follows: the detection depth and vertical falling velocity are compared with preset sea area partition thresholds, and the sampling frequency is dynamically switched according to the interval. The sea area partitions include at least: a first sampling frequency corresponding to the shallow high dynamic zone, a second sampling frequency corresponding to the middle transition zone, and a third sampling frequency corresponding to the deep sea steady state zone, wherein the first sampling frequency is higher than the second sampling frequency, and the second sampling frequency is higher than the third sampling frequency. When the depth and velocity conditions switch across intervals, the system instantly and seamlessly switches the sampling frequency with a response time lower than the set delay threshold.

[0015] Furthermore, the specific process of temperature thermal hysteresis compensation in step S4 is as follows: The inherent response delay of the sensor is fitted based on the dual reference temperature curves. The response delay is determined by the sensor's response time at both high and low reference temperatures in combination with the sensor's thermal conductivity. The dynamic lag time is calculated by combining real-time depth detection and descent velocity. The original temperature sequence is corrected using a time-series differential interpolation formula. The time-series differential interpolation uses at least first-order and second-order difference terms to compensate for the temperature changes within the lag time point by point, so that the compensated temperature data is aligned with the depth data on the time axis.

[0016] Furthermore, the specific process of dynamic correction of pressure temperature drift in step S4 is as follows: A pressure sensor drift model with temperature as the independent variable is constructed. The model includes at least a zero-point drift function and a sensitivity drift function. Both the zero-point drift function and the sensitivity drift function take the real-time temperature after thermal hysteresis compensation as input. The nonlinear effect of temperature on the zero point and sensitivity of the pressure sensor is fitted in polynomial form. The accurate pressure value after temperature drift correction is obtained by subtracting the zero-point offset from the original pressure value and then multiplying it by the sensitivity coefficient. The corrected seawater depth value is obtained by converting the corrected pressure value using hydrostatic formulas.

[0017] Furthermore, the specific process of dynamic filtering of attitude and turbulence disturbances in step S4 is as follows: Using the basic filter window length and basic filter threshold as initial values, the current filter window length and filter threshold are dynamically calculated by weighted summation based on the sum of the real-time pitch angle absolute value and roll angle absolute value and the turbulence intensity. The weighting coefficients of attitude angle and turbulence intensity are preset constants. Moving average filtering is used to process the temperature and depth data after the first two correction stages to filter out random fluctuations and sudden noises caused by equipment tilt and seawater turbulence.

[0018] Furthermore, the specific method for identifying and removing gross outliers based on statistical criteria in step S5 is as follows: calculate the mean and standard deviation of the fused complete data sequence; when the deviation of a data point from the mean exceeds a preset multiple of the standard deviation, it is determined to be a gross outlier and removed; align the corrected temperature and depth data point by point along a unified time axis to form a one-to-one corresponding temperature and depth data group, and finally generate continuous seawater temperature and depth profile curve data.

[0019] The adaptive high-precision measurement system for seawater depth and temperature based on a CTD system described in this invention, applied to the adaptive high-precision measurement method for seawater depth and temperature based on a CTD system, includes: The data acquisition module is used to synchronously acquire data on the motion attitude, vertical falling speed, seawater pressure, seawater temperature and turbulence intensity of the CTD detection device at a sampling rate of no less than the set frame rate, and to calculate the detection depth in real time from the seawater pressure. The adaptive frequency conversion sampling module is used to divide the sea area into multiple zones, including at least the shallow high dynamic zone, the middle layer transition zone and the deep sea steady state zone, based on the dual criteria of detection depth and vertical falling speed. Different sampling frequencies are used in each zone, and the sampling frequency is switched instantly and seamlessly when crossing zones. A multi-level joint error correction module is used to sequentially perform three-level corrections: temperature thermal hysteresis compensation, pressure temperature drift dynamic correction, and attitude and turbulence disturbance dynamic filtering. The data fusion and output module is used to align the corrected temperature and depth data one by one along the time axis, identify and remove gross outliers based on statistical criteria, generate standard hydrological profile data, and complete storage and transmission.

[0020] The adaptive high-precision measurement device for seawater depth and temperature based on a CTD system described in this invention includes: The main body of the CTD detection equipment is equipped with a temperature sensor, a pressure sensor, and a micro-inertial attitude detection unit; An embedded main control unit is built into the main body of the CTD detection device, which contains a computer program. When the computer program is executed by the embedded main control unit, it implements the steps of the adaptive high-precision measurement method for seawater depth and temperature based on the CTD system.

[0021] Compared with existing technologies, the adaptive high-precision measurement method, system, and device for seawater depth and temperature based on CTD system of the present invention exhibit the following beneficial effects in terms of technical performance and practical application: 1. This invention employs both detection depth and vertical descent velocity as criteria to divide the sea area into a shallow high-dynamic zone, a mid-level transition zone, and a deep-sea steady-state zone, using different sampling frequencies in each zone. High-frequency sampling (e.g., 20Hz) is used in the shallow high-dynamic zone to fully capture the detailed characteristics of rapidly changing hydrological parameters; low-frequency sampling (e.g., 5Hz) is used in the deep-sea steady-state zone to avoid data redundancy and significantly extend the equipment's endurance.

[0022] 2. This invention employs a time-series differential interpolation algorithm based on dual-reference temperature curves. By calculating the inherent response delay and dynamic lag of the sensor, it uses first-order and second-order differences to correct the original temperature sequence point by point. This solves the problem of time-series misalignment of temperature and depth data caused by the thermal conduction delay of the temperature sensor during the CTD equipment's descent. The overall time-series misalignment error is reduced by more than 90%, and the output temperature-depth profile curve is continuous and accurate, meeting the needs of high-precision marine scientific research.

[0023] 3. This invention constructs a pressure sensor drift model with temperature as the independent variable, and uses a polynomial form to fit the nonlinear influence of temperature on the zero point and sensitivity of the pressure sensor. Pressure values ​​are dynamically corrected through real-time iterative updates. This overcomes the shortcomings of traditional fixed-parameter compensation methods that cannot adapt to drastic temperature fluctuations across the entire ocean depth. The overall error of full-ocean-depth depth measurement is reduced by an average of approximately 42%, significantly improving the accuracy of depth measurement.

[0024] 4. This invention dynamically adjusts the sliding filter window length and filter threshold based on real-time pitch angle, roll angle, and turbulence intensity, and calculates the current filter parameters using a weighted summation method, allowing the filter intensity to adaptively change with the degree of environmental disturbance. This effectively suppresses random noise generated by equipment tilting and seawater turbulence under complex sea conditions.

[0025] 5. This invention aligns the three-level corrected temperature and depth data point by point on a unified time axis, identifying and removing coarse outliers. This technical feature ensures strict correspondence between temperature and depth data in the time dimension, with an outlier removal rate of over 98% and no valid data being mistakenly deleted. The output standard hydrological profile data is of high quality and highly reliable.

[0026] 6. The entire method of this invention is implemented purely as an algorithm, requiring no modification to the mechanical structure, sensors, or circuitry of the CTD equipment. It can be directly burned into the main control chip of existing CTD equipment as firmware, resulting in extremely low retrofitting costs for existing equipment. Furthermore, the algorithm program runs automatically throughout the entire process after being embedded in the main control unit, requiring no manual intervention. It is compatible with three mainstream operational modes: shipborne deployment, underwater fixed-point deployment, and towed detection, making it applicable to a wide range of scenarios and possessing significant engineering promotion value. Attached Figure Description

[0027] Figure 1 This is a flowchart of the overall method of the present invention; Figure 2 This is a comparison curve of the adaptive variable frequency sampling effect of the present invention; Figure 3 This is a comparison curve of time sequence alignment before and after temperature thermal hysteresis compensation in this invention; Figure 4 This is a curve comparing the depth error of the dynamic pressure-temperature drift correction of this invention with that of traditional fixed compensation. Figure 5 This is a graph comparing the perturbation data processing effects of the dynamic filtering of this invention and the traditional static filtering; Figure 6 This is a bar chart comparing the comprehensive performance indicators of the present invention with those of traditional measurement methods. Detailed Implementation

[0028] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0029] Example 1: This method embodiment uses a conventional marine scientific research CTD device as a carrier. This device is equipped with temperature sensors, pressure sensors, and a micro-inertial attitude detection unit. The entire method is automatically executed by the device's built-in main control program. Figure 1 As shown, Figure 1 The overall method flow of the present invention is shown, which includes five major steps from S1 to S5, and the overall process has a closed-loop cyclic structure.

[0030] Step S1: Equipment water inlet initialization and algorithm parameter calibration: The CTD device is deployed to the sea surface from the research vessel and automatically powers on and starts up. The main control unit sequentially completes hardware self-tests for the temperature sensor, pressure sensor, micro-inertial attitude detection unit, data storage module, and wired / wireless communication module. If a hardware malfunction is detected, an alarm is triggered and detection is terminated; if the hardware is normal, the parameter loading process begins.

[0031] The system loads the following calibration parameters from the built-in parameter library: (1) Temperature reference curves: sea surface normal temperature reference (25℃) and deep sea low temperature reference (0℃), using the second-order fitting curves calibrated by the manufacturer's laboratory.

[0032] (2) Temperature-pressure coupling model coefficients: zero-point drift coefficients a0, a1, a2 and sensitivity drift coefficients b0, b1, b2. The factory calibration values ​​of this equipment are: a0=12.2, a1=0.81, a2=0.021; b0=1.0012, b1=2.1×10 4 b2 = 3.5 × 10 6 .

[0033] (3) Initial filtering parameters: basic window length W0=10, basic threshold Th0=0.05℃.

[0034] (4) Attitude zero point: pitch angle θ p =0°, roll angle θ r =0°, falling speed V=0.

[0035] Step S2: Real-time synchronous acquisition of multiple parameters: After the equipment begins to descend, the main control unit synchronously collects five types of data at a frame rate of 50Hz (sampling interval Δt0≤0.02s): pitch angle θ p Roll angle θ rVertical falling speed V, seawater temperature T raw Seawater static pressure P raw Turbulence Intensity I t Simultaneously, the detection depth is calculated in real time using static pressure: ; ρ is the average density of seawater, taken as 1025 kg / m³. 3 g is the acceleration due to gravity. Actual measurements show that the synchronization error of multiple parameters does not exceed 0.5ms across the entire ocean depth range (0~4000m), which is far lower than the industry standard of 1ms.

[0036] Step S3: Layered adaptive variable frequency sampling: In this embodiment, the computing unit reads the detection depth H and the vertical falling velocity V in real time, compares the detection depth H and the vertical falling velocity V with preset sea area partition thresholds, and dynamically switches to the corresponding sampling frequency according to the current zone, as follows: When H < 200m and V > 1.2m / s, it is identified as a shallow sea high dynamic zone. High-density data is collected at a sampling frequency of 20Hz to capture the rapidly changing hydrological characteristics of the shallow sea. When 200≤H≤2000m and 0.3≤V≤1.2m / s, it is determined to be the middle layer transition zone, and sampling is performed in the normal mode at a sampling frequency of 10Hz; When H>2000m and V<0.3m / s, it is determined to be a deep-sea steady-state zone. The sampling frequency is reduced to 5Hz to save equipment energy.

[0037] When depth and velocity conditions switch across intervals, the system instantly and seamlessly switches the sampling frequency with a delay of less than 10ms, ensuring uninterrupted detection throughout the entire process.

[0038] In this embodiment, the actual sinking process is as follows: V=1.4m / s in the 0~200m range, sampled at 20Hz; V=0.7m / s in the 200~2000m range, sampled at 10Hz; and V=0.22m / s in the 2000~4000m range, sampled at 5Hz. The measured interval switching delay is 8ms, with no data interruption.

[0039] like Figure 2 As shown, traditional fixed high-frequency sampling (20Hz) generates a large amount of redundant data in deep-sea areas, resulting in high equipment power consumption; traditional fixed low-frequency sampling (5Hz) loses a large amount of critical hydrological data in shallow-sea areas, leading to incomplete profile curves. This invention dynamically switches the sampling frequency according to depth ranges: high-density sampling at 20Hz in shallow-sea areas ensures data integrity, while low-frequency sampling at 5Hz in deep-sea areas reduces power consumption, achieving a balance between sampling accuracy and equipment endurance.

[0040] The adaptive sampling scheme of this invention was verified through 10 sets of comparative tests at a full ocean depth of 4000m. The comparison between the traditional fixed-frequency sampling scheme and the scheme is as follows: Table 1. Comparison between the adaptive sampling scheme of the present invention and the traditional fixed-frequency sampling scheme.

[0041] As shown in Table 1, the adaptive sampling of this invention reduces power consumption in deep sea by about 35% (from 128mA to 83mA) compared to the traditional fixed 20Hz sampling, and significantly extends the battery life; compared to the traditional fixed 5Hz sampling, the data loss rate in shallow sea is reduced from 29.6% to 0.8%, and the data integrity is greatly improved.

[0042] Step S4: Multi-level joint error correction: The three-level correction is performed sequentially, prioritizing systematic errors and then applying random noise.

[0043] 1. Temperature thermal hysteresis compensation: Temperature sensors have an inherent thermal conduction delay, with the depth sampling time leading the temperature sampling time. This method uses a quadratic time-series differential interpolation algorithm based on dual reference temperature curves for compensation.

[0044] Sensor inherent response delay calculation: ; Where T 25 T0 is the reference response time at 25℃ sea surface, T0 is the reference response time at 0℃ deep sea, and k is the thermal conductivity of the sensor (calibrated value of this device k=1.6). The inherent response delay of this device is obtained by fitting τ=1.12s.

[0045] Calculation of dynamic lag time: ; Δt represents the dynamic lag time (s) of temperature relative to depth, which varies dynamically with depth and velocity.

[0046] Secondary time-series differential interpolation correction: Let the original temperature sequence be... , First-order difference: ; Second-order difference: ; Temperature after compensation: ; The temperature data after thermal hysteresis compensation has been aligned with the depth data on the time axis.

[0047] like Figure 3 As shown in the comparison, the traditional method refers to directly processing the original temperature data T without performing any thermal hysteresis compensation. raw The conventional approach is to align the temperature and depth data (H) with the depth data using timestamps. However, in traditional measurement methods, the temperature sampling time lags significantly behind the depth sampling time, resulting in a significant time shift in the temperature-depth profile and distortion. After processing with the compensation algorithm of this invention, the temperature and depth curves are precisely aligned on the time axis, the temporal misalignment is completely eliminated, and the temperature-depth profile accurately reflects the distribution characteristics of seawater temperature with depth.

[0048] The measured compensation effect is shown in Table 2 below: Table 2 Measured compensation effect

[0049] Table 2 shows that the timing misalignment in shallow sea conditions (V=1.5m / s) decreased from 1.76s to 0.15s (a reduction of 91.5%); in mid-water conditions (V=0.8m / s) it decreased from 1.21s to 0.10s (a reduction of 91.7%); and in deep sea conditions (V=0.2m / s) it decreased from 0.68s to 0.06s (a reduction of 91.2%). The timing misalignment error across the entire sea area was reduced by more than 90%.

[0050] 2. Dynamic correction of pressure and temperature drift: The zero-point offset and sensitivity coefficient of the pressure sensor exhibit nonlinear drift with temperature. This method constructs a temperature-pressure coupled polynomial model for real-time correction.

[0051] Drift model: Zero-point drift function: ; Sensitivity drift function: ; Where Z(T) is the zero-point offset (Pa), and K(T) is the sensitivity coefficient (dimensionless). cor1 This is the real-time temperature after thermal hysteresis compensation.

[0052] Dynamic pressure correction: ; Depth conversion: ; Among them, P cor This is the precise pressure value (Pa) after temperature drift correction. H cor The final seawater depth (m) after multi-level correction.

[0053] like Figure 4As shown in the figure, traditional fixed compensation refers to the use of a single fixed compensation parameter calibrated at the factory of the pressure sensor, employing a constant zero-point offset and sensitivity coefficient for pressure correction across the entire ocean depth range. Traditional factory-set fixed compensation parameters cannot adapt to the alternating temperature and pressure environment across the entire ocean depth range, resulting in large fluctuations in depth error, especially in the thermocline region. This invention employs a dynamic iterative correction method, correcting zero-point drift and sensitivity drift in real time, maintaining a low-amplitude and stable depth error curve across the entire ocean depth range.

[0054] Table 3 Measured Depth Error Data

[0055] As shown in Table 3, the average comprehensive error of full ocean depth is reduced by about 42%.

[0056] 3. Dynamic filtering of attitude and turbulence disturbances: The filter parameters are dynamically adjusted based on the equipment attitude and turbulence parameters, and a moving average filter is used to suppress random noise.

[0057] Dynamic filtering window length: ; Where W0=10 is the base window length, α=0.8 is the attitude weighting coefficient, β=6 is the turbulence weighting coefficient, and θ p θ r These are pitch angle and roll angle, respectively. t The turbulence intensity.

[0058] Dynamic filtering threshold: ; Where Th0=0.05℃ is the basic threshold, and γ=0.01 and δ=0.08 are the perturbation weighting coefficients.

[0059] Moving average filtering: ; Temperature after the first two correction stages or depth data; This is the filtered data.

[0060] like Figure 5As shown in the comparison, the original filtering refers to the practice of using a fixed window length (usually a constant value between W=10 and 15) and a fixed filtering threshold for moving average filtering. In nearshore areas with strong turbulence, equipment experiences severe shaking, resulting in significant random fluctuations in the original temperature data. While traditional static filtering can partially smooth noise, its fixed filtering parameters lead to insufficient filtering during severe disturbances and over-smoothing during mild disturbances. The dynamic threshold moving average filtering of this invention adaptively adjusts the filtering window and threshold based on real-time attitude angles and turbulence intensity, effectively suppressing random interference. The filtered temperature curve is smooth and stable while preserving the true hydrological characteristics.

[0061] In this embodiment, the average filter window W=18, the filter threshold Th=0.11℃, and the temperature fluctuation is only 0.06℃.

[0062] Step S5: Time-series fusion of temperature and depth data, outlier removal and output: Temporal fusion: merging the temperature series after three-level correction With depth sequence Alignment is performed point by point on a unified 50Hz time axis to form a one-to-one corresponding standard data set of (time-depth-temperature), with a timing alignment error of no more than 0.01s.

[0063] Outlier removal: The 3σ criterion is used to calculate the mean μ and standard deviation σ of the fused data sequence Y(i). ; ; when Outliers were identified as gross anomalies and removed. Verification with 10 sets of full-ocean-depth data showed an outlier removal rate of 98.7%, with no valid data mistakenly deleted.

[0064] Data Output: The system integrates valid data to generate continuous seawater temperature-depth profile curves, stores them in local non-volatile memory, and transmits them externally via a wired interface or wireless module. The equipment continues to descend, cyclically executing processes S2 to S5 until the exploration mission concludes.

[0065] like Figure 6As shown, the five core indicators are summarized: timing matching, measurement error, data quality, and equipment power consumption. The traditional method in the figure refers to conventional CTD data acquisition and processing methods that use a fixed sampling frequency of 20Hz or 5Hz, factory-fixed pressure compensation parameters, conventional static sliding filtering, and no targeted outlier handling. This is a typical solution with defects such as sampling rigidity, thermal hysteresis, pressure and temperature drift, and insufficient disturbance immunity. Compared to traditional CTD measurement methods, this invention achieves significant optimization in all indicators: the timing misalignment error of temperature and depth data is reduced by 91.3%; the comprehensive error of full ocean depth measurement is reduced by 42.3%; the temperature measurement error is reduced by 38.1%; the data effectiveness under complex sea conditions is increased to 96.1%; and the average power consumption of equipment in deep-sea areas is reduced by 35.1%, fully verifying the overall superiority and engineering application value of the method.

[0066] Example 2: This system embodiment provides a high-precision adaptive seawater depth and temperature measurement system based on a CTD system, applicable to the measurement method described in any step of Embodiment 1. The system includes a data acquisition module, an adaptive frequency conversion sampling module, a multi-level joint error correction module, and a data fusion and output module. The algorithm programs for each module are embedded in the CTD main control unit, enabling fully automatic operation.

[0067] Data acquisition module: The data acquisition module is used to synchronously acquire the motion attitude (pitch angle) of the CTD detection device at a frame rate of not less than 50Hz. Roll angle ), vertical falling velocity V, seawater pressure Seawater temperature and turbulence intensity Data, and through formulas The detection depth H is calculated in real time.

[0068] The core function of this module is to establish a unified time reference, ensuring that all parameters are acquired on the same time axis. Field measurements show that the synchronization error of multiple parameters across the entire ocean depth range does not exceed 0.5ms, providing timing consistency assurance for subsequent adaptive frequency conversion sampling and multi-level error correction. Each sensor and the main control unit are driven by a synchronous clock signal to ensure strict alignment of the acquisition timing.

[0069] Adaptive frequency conversion sampling module: The adaptive frequency conversion sampling module is used to divide the sea area into shallow high dynamic zone, mid-level transition zone and deep steady-state zone based on the dual criteria of detection depth H and vertical falling velocity V. Different sampling frequencies are used in each zone and seamlessly switched in real time when crossing zones.

[0070] The core working logic of this module is to receive depth H and velocity V data from the data acquisition module in real time and compare them with the preset sea area zoning thresholds.

[0071] Multi-level joint error correction module: The multi-level joint error correction module is used to sequentially perform three-level corrections: temperature thermal hysteresis compensation, pressure temperature drift dynamic correction, and attitude and turbulence disturbance dynamic filtering. It is the core module of the system.

[0072] The temperature thermal hysteresis compensation unit is based on fitting the sensor's inherent response delay using dual-reference temperature curves. The dynamic lag time is calculated by combining the real-time falling speed V. The original temperature series was analyzed using a quadratic time-difference interpolation formula. Make corrections: ; in , These are first-order and second-order differences, respectively.

[0073] The pressure-temperature drift dynamic correction unit constructs a pressure sensor drift model with temperature as the independent variable: ; ; pass Corrections were made, and the depth was calculated. .

[0074] The attitude and turbulence dynamic filtering unit adjusts the dynamic filtering based on the real-time pitch angle. Roll angle and turbulence intensity Dynamically calculate the filter window length W and the threshold Th: ; ; And a moving mean filter is used. Process it.

[0075] Data fusion and output module: The data fusion and output module is used to process the temperature data after three-level correction. With depth data Alignment was performed point by point along the 50Hz time axis. The 3σ criterion was used to identify and remove coarse outliers, with an outlier removal rate of over 98% and no valid data was mistakenly deleted.

[0076] This module integrates effective data to generate continuous seawater temperature and depth profile curves, supports local non-volatile storage and wired or wireless transmission, and is compatible with three operating modes: shipborne deployment, underwater fixed-point deployment, and towed detection.

[0077] in: Shipborne deployment type: The equipment is deployed by a research vessel, sinks freely throughout the entire process, and its adaptive sampling and correction logic is effective throughout the process, making it suitable for full-ocean-depth profile exploration.

[0078] Underwater fixed-point deployment type: The equipment is deployed statically with a descent speed V≈0, automatically determining a steady-state environment. It defaults to the lowest sampling frequency of 5Hz, resulting in minimal power consumption. Under static conditions, the temperature measurement error does not exceed 0.04℃, and the depth measurement error does not exceed 0.38m.

[0079] Towed detection: The device moves by being towed, and its attitude changes frequently (θ). p θ r (Maximum ±12°), the dynamic filtering module responds automatically, the filtering window W automatically expands to 22, and the filtering threshold Th is adjusted accordingly to enhance the anti-disturbance capability.

[0080] Example: Based on Example 1, this example provides a high-precision adaptive measurement device for seawater depth and temperature based on a CTD system, including a CTD detection device body and an embedded main control unit.

[0081] CTD detection equipment main body: The main body of the CTD detection equipment is a conventional CTD device used in marine scientific research, equipped with the following sensors: (1) Temperature sensor: A high-precision platinum resistance thermometer is used to collect seawater temperature data. Measurement range: -2~30℃, accuracy: ±0.002℃.

[0082] (2) Pressure sensor: A quartz oscillating pressure sensor is used to collect seawater static pressure. , used to convert seawater depth.

[0083] (3) Micro-inertial attitude detection unit: used to detect the motion attitude of the equipment in real time and output the pitch angle. and roll angle Simultaneously calculate the vertical falling speed. and turbulence intensity .

[0084] Embedded main control unit: The embedded main control unit is built into the main body of the CTD detection device and uses a low-power embedded processor, which contains a computer program. When the embedded main control unit executes this computer program, it implements all steps S1 to S5 in Embodiment 1.

[0085] Initialization and Calibration Program Module: After power-on, the device executes a hardware self-test program, sequentially checking the operating status of the temperature sensor, pressure sensor, micro-inertial attitude detection unit, data storage module, and communication module. Upon successful self-test, it loads the factory-calibrated temperature reference curve parameters and temperature-pressure coupling model coefficients from the built-in non-volatile memory. , ), filter initial parameters ( , The initial values ​​of pitch, roll, and vertical velocity were set to zero. The sampling frequency threshold was then determined. Subsequently, zero-point calibration was performed on the micro-inertial attitude detection unit, setting the initial values ​​of pitch, roll, and vertical velocity to zero. This process was completed within 0.3 seconds.

[0086] Data acquisition and control program module: Drives each sensor to acquire data synchronously at a 50Hz frame rate, generates multi-parameter data frames containing timestamps, and uses hydrostatic formulas... Real-time calculation of detection depth. This module ensures that the synchronization error of multiple parameters does not exceed 0.5ms across the entire ocean depth range.

[0087] The adaptive sampling control module reads depth H and velocity V in real time, compares them with preset partition thresholds, and outputs the corresponding sampling frequency control signal. The partitioning rules are embedded in the program logic, and the interval switching response time is less than 10ms.

[0088] Error correction program module: Temperature hysteresis compensation algorithm: calculation based on dual reference temperature curves , combined , calculate Perform a second-order time difference interpolation correction; Pressure temperature drift correction algorithm: Substitute into the thermo-pressure coupled polynomial model for calculation and ,implement Correct and convert the depth; Dynamic filtering algorithm: based on , , Dynamic calculation and Perform moving average filtering.

[0089] Data output module: It merges the corrected temperature and depth data on a unified time axis, performs the 3σ criterion to remove outliers, generates standard hydrological profile data, and controls the storage and transmission of data.

[0090] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A high-precision adaptive measurement method for seawater depth and temperature based on a CTD system, characterized in that, Includes the following steps: S1: Equipment water immersion initialization and algorithm parameter calibration: After the CTD equipment is immersed in water, it completes hardware self-test, loads the preset temperature reference curve, temperature-pressure coupling error model parameters and filter initial parameters, and performs zero-point calibration on the micro inertial attitude detection unit. S2: Multi-parameter synchronous real-time acquisition: Synchronously acquire data on device motion attitude, vertical falling speed, seawater pressure, seawater temperature and turbulence intensity at a sampling rate not lower than the set frame rate, and calculate the detection depth in real time from the seawater pressure; S3: Layered adaptive variable frequency sampling: Based on the dual criteria of detection depth and vertical falling speed, the sea area is divided into multiple intervals, including at least the shallow sea high dynamic zone, the middle layer transition zone and the deep sea steady state zone. Different sampling frequencies are used in each interval, and the sampling frequency is switched instantly and seamlessly when crossing intervals. S4: Multi-level joint error correction: This involves sequentially performing three levels of correction: temperature thermal hysteresis compensation, pressure temperature drift dynamic correction, and attitude and turbulence disturbance dynamic filtering. The temperature thermal hysteresis compensation adopts a time-series differential interpolation algorithm based on the reference temperature curve. The dynamic hysteresis time is calculated based on the inherent response delay of the sensor and the real-time falling speed. The original temperature sequence is corrected point by point to eliminate the time misalignment between temperature sampling and depth sampling. The pressure-temperature drift dynamic correction is achieved by substituting the real-time seawater temperature into a pre-constructed temperature-pressure coupling model, iteratively updating the zero-point offset and sensitivity coefficient of the pressure sensor in real time, correcting the pressure value, and converting it into the corrected seawater depth value. The attitude and turbulence disturbance dynamic filtering dynamically adjusts the sliding filter window length and filter threshold according to the real-time pitch angle, roll angle and turbulence intensity to filter out random noise generated by equipment tilt and seawater turbulence. S5: Time-series fusion and output of temperature and depth data: Align the corrected temperature and depth data one by one along the time axis, identify and remove gross outliers based on statistical criteria, generate standard hydrological profile data, and complete storage and transmission.

2. The adaptive high-precision measurement method for seawater depth and temperature based on a CTD system according to claim 1, characterized in that, Step S1 specifically includes: after the CTD equipment is immersed in water, it completes a self-test of the temperature sensor, pressure sensor, micro-inertial attitude detection unit, data storage module, and communication module. If the hardware is abnormal, an alarm is triggered and the detection is terminated; if the hardware is normal, all algorithm parameters, including the low temperature reference curve, the normal temperature reference curve, the temperature-pressure coupling model coefficients, the initial filtering parameters, and the sampling frequency threshold, are loaded from the built-in parameter library; the pitch angle, roll angle, and vertical velocity of the micro-inertial attitude detection unit are zero-point calibrated; after the self-test and parameter loading are completed, the device enters the detection standby state.

3. The adaptive high-precision measurement method for seawater depth and temperature based on a CTD system according to claim 1, characterized in that: In step S2, the motion attitude includes pitch angle and roll angle; the seawater pressure is converted into the detection depth in real time using a hydrostatic formula based on seawater density and gravitational acceleration; all parameters are collected synchronously, the collection frame rate is not lower than the set reference value, and the synchronization error between parameters is controlled within the preset tolerance.

4. The adaptive high-precision measurement method for seawater depth and temperature based on a CTD system according to claim 1, characterized in that: The determination step of the layered adaptive variable frequency sampling in step S3 is as follows: compare the detection depth and the vertical falling speed with the preset sea area partition threshold respectively, and dynamically switch to the corresponding sampling frequency according to the interval. The sea area partitioning thresholds include at least: a first sampling frequency corresponding to the shallow high-dynamic zone, a second sampling frequency corresponding to the mid-level transition zone, and a third sampling frequency corresponding to the deep-sea steady-state zone, wherein the first sampling frequency is higher than the second sampling frequency, and the second sampling frequency is higher than the third sampling frequency; when the depth and velocity conditions switch across zones, the system instantly and seamlessly switches the sampling frequency with a response time lower than the set delay threshold.

5. The adaptive high-precision measurement method for seawater depth and temperature based on a CTD system according to claim 1, characterized in that: The specific process of temperature thermal hysteresis compensation in step S4 is as follows: The sensor's inherent response delay is fitted based on dual reference temperature curves. The sensor's inherent response delay is determined by combining the sensor's response time at high and low reference temperatures with the sensor's thermal conductivity. The dynamic lag time is calculated by combining real-time depth detection and descent velocity. The original temperature sequence is corrected using a time-series differential interpolation formula. The time-series differential interpolation formula uses at least first-order and second-order difference terms to compensate for temperature changes within the lag time point by point, so that the compensated temperature data is aligned with the depth data on the time axis.

6. The adaptive high-precision measurement method for seawater depth and temperature based on a CTD system according to claim 1, characterized in that: The specific process of dynamic correction of pressure temperature drift in step S4 is as follows: A pressure sensor drift model with temperature as the independent variable is constructed. The pressure sensor drift model includes at least a zero-point drift function and a sensitivity drift function. Both the zero-point drift function and the sensitivity drift function take the real-time temperature after thermal hysteresis compensation as input. The nonlinear effect of temperature on the zero point and sensitivity of the pressure sensor is fitted in polynomial form. The accurate pressure value after temperature drift correction is obtained by subtracting the zero-point offset from the original pressure value and then multiplying it by the sensitivity coefficient. The corrected seawater depth value is obtained by converting the corrected pressure value using hydrostatic formulas.

7. The adaptive high-precision measurement method for seawater depth and temperature based on a CTD system according to claim 1, characterized in that: The specific process of dynamic filtering of attitude and turbulence disturbances in step S4 is as follows: Using the basic filter window length and basic filter threshold as initial values, the current filter window length and filter threshold are dynamically calculated by weighted summation based on the sum of the real-time pitch angle absolute value and roll angle absolute value and the turbulence intensity. The weighting coefficients of attitude angle and turbulence intensity are preset constants. Moving average filtering is used to process the temperature and depth data after the first two correction stages to filter out random fluctuations and sudden noises caused by equipment tilt and seawater turbulence.

8. The adaptive high-precision measurement method for seawater depth and temperature based on a CTD system according to claim 1, characterized in that: The specific process of identifying and removing gross outliers based on statistical criteria in step S5 is as follows: calculate the mean and standard deviation of the fused complete data sequence; when the deviation of a data point from the mean exceeds a preset multiple of the standard deviation, it is determined to be a gross outlier and removed; align the corrected temperature and depth data point by point along a unified time axis to form a one-to-one corresponding temperature and depth data group, and finally generate continuous seawater temperature and depth profile curve data.

9. A high-precision adaptive measurement system for seawater depth and temperature based on a CTD system, applied to the high-precision adaptive measurement method for seawater depth and temperature based on a CTD system as described in any one of claims 1-8, characterized in that, include: The data acquisition module is used to synchronously acquire data on the motion attitude, vertical falling speed, seawater pressure, seawater temperature and turbulence intensity of the CTD detection device at a sampling rate of no less than the set frame rate, and to calculate the detection depth in real time from the seawater pressure. The adaptive frequency conversion sampling module is used to divide the sea area into multiple zones, including at least the shallow high dynamic zone, the middle layer transition zone and the deep sea steady state zone, based on the dual criteria of detection depth and vertical falling speed. Different sampling frequencies are used in each zone, and the sampling frequency is switched instantly and seamlessly when crossing zones. A multi-level joint error correction module is used to sequentially perform three-level corrections: temperature thermal hysteresis compensation, pressure temperature drift dynamic correction, and attitude and turbulence disturbance dynamic filtering. The data fusion and output module is used to align the corrected temperature and depth data one by one along the time axis, identify and remove gross outliers based on statistical criteria, generate standard hydrological profile data, and complete storage and transmission.

10. A high-precision adaptive measurement device for seawater depth and temperature based on a CTD system, characterized in that, include: The main body of the CTD detection equipment is equipped with a temperature sensor, a pressure sensor, and a micro-inertial attitude detection unit; An embedded main control unit is built into the main body of the CTD detection device, wherein a computer program is embedded therein. When the computer program is executed by the embedded main control unit, it implements the steps of the adaptive high-precision measurement method for seawater depth and temperature based on the CTD system as described in any one of claims 1 to 8.

Images