Dark background self-adaptive correction method for irradiance sensor

Abstract

The present application relates to a kind of irradiance sensor dark background adaptive correction method, belong to photometric technical field.The irradiance sensor dark background adaptive correction method includes double-layer field calibration step, constructs double-point linear model step, real-time dark background prediction step, aging drift adaptive compensation step and irradiance correction step.The irradiance sensor dark background adaptive correction method directly in the deep sea field dark background data of deep and shallow layer is collected, and the calibration environment is completely consistent with the actual working environment of sensor, fundamentally eliminates the system deviation caused by the difference of heat dissipation condition;Using the natural calibration opportunity of each deployment process, automatically collect dark background data and compare with historical data, real-time detection aging drift and update model parameters, realize the adaptive long-term stability maintenance, and greatly reduce maintenance cost.

Description

Technical Field

[0001] This invention belongs to the field of photometric measurement technology, specifically relating to a dark background adaptive correction method for an irradiance sensor. Background Technology

[0002] Irradiance sensors are core instruments in the field of marine optical radiation measurement, widely used in long-term ocean observations by unmanned platforms such as buoys, underwater gliders, and deep-sea landers to acquire underwater light field distribution data, thereby supporting scientific research such as photosynthesis estimation, photochemical process research, and analysis of the optical characteristics of marine organisms. However, irradiance sensors still exhibit a non-zero output signal (i.e., a dark background) under no-light conditions. This signal, mainly generated by the detector's dark current, is a key error source affecting the accuracy of low-light measurements. Studies have shown that the dark background of deep-sea irradiance sensors has a significant temperature dependence, and the measurement deviation caused by its temperature drift can be thousands of times greater than the sensor's nominal uncertainty. Therefore, accurately subtracting the dark background is a prerequisite for obtaining true irradiance data.

[0003] Currently, a typical existing solution for dark background correction of deep-sea irradiance sensors employs the "nighttime profiling method": full-depth profiling measurements are performed at night (without sunlight), and the sensor output at all depths along the entire profiling is treated as dark background values ​​for correction. However, this method has significant technical drawbacks: First, extensive observational data show that over 28% of daytime profilings still record significant downflow irradiance at a depth of 200 meters, meaning that even at night, full-depth profiling may not guarantee that measurements at all depths are dark backgrounds. Second, this method is limited to nighttime operations, requiring several hours to complete a single full-depth profiling, resulting in high energy consumption, low efficiency, and a time blind zone due to the inability to perform dark background measurements during daytime operations, making it difficult to meet the needs of continuous all-weather observation. Furthermore, this method does not consider the aging drift of the sensor over time, failing to adaptively compensate for the slow changes in dark background characteristics.

[0004] Therefore, how to solve the technical defects of the existing night profiling method, which is "limited to nighttime operation, takes several hours, has time blind spots, cannot achieve all-weather observation, and cannot compensate for sensor aging and drift", is a technical problem that urgently needs to be solved. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a dark background adaptive correction method for irradiance sensors. It utilizes the natural, light-free, and temperature-controlled nighttime environments at depths >1000 meters and 450-550 meters in the deep sea as calibration fields. The calibration conditions are completely consistent with the actual working environment of the sensor (deep sea, water-cooled heat dissipation), eliminating systematic biases caused by differences between laboratory calibration and the field environment. A linear model is constructed using measured data from two temperature points. When extrapolated to the high-temperature surface region, the model uncertainty is significantly reduced, greatly improving the accuracy of dark background correction. Furthermore, deep-layer calibration is not limited by day or night, and combined with shallow-layer nighttime calibration, it only takes a few minutes. Nighttime operation time is reduced by more than 90% compared to traditional nighttime profiling methods, effectively eliminating time blind spots and enabling continuous all-weather observation.

[0006] This invention provides a dark background adaptive correction method for an irradiance sensor, comprising the following steps: Dual-layer field calibration: Acquire first dark background data and corresponding first temperature data collected by the irradiance sensor at a first depth, and acquire second dark background data and corresponding second temperature data collected by the irradiance sensor at a second depth; wherein, the first depth is greater than 1000 meters, and the temperature corresponding to the first temperature data is 2℃ to 4℃; the second depth is 450 meters to 550 meters, the second dark background data is collected during periods without sunlight, and the temperature corresponding to the second temperature data is 10℃ to 12℃; Constructing a two-point linear model: For each pixel of the irradiance sensor, establish a linear model between the dark background value and the temperature; Real-time dark background prediction: During the actual measurement process, the current temperature data of the location of the irradiance sensor is acquired in real time. The current temperature data is substituted into the linear model corresponding to each pixel to calculate the dark background prediction value of each pixel at the current temperature. Aging drift adaptive compensation: During each deployment, the first and second depth dark background data collected this time are acquired and compared with the corresponding depth dark background data collected in the previous deployment. If the change in dark background exceeds the preset threshold, it is determined that the sensor has experienced dark background drift. The model coefficients of the linear model are updated using the dark background data collected this time, and the updated model coefficients are used for subsequent dark background prediction. Irradiance correction: Subtract the corresponding dark background prediction value from the original measurement data of each pixel to obtain the corrected irradiance measurement data.

[0007] This technical solution utilizes the natural, light-free, and constant-temperature environment at depths >1000 meters and 450-550 meters at night as the field calibration sites. The calibration conditions are completely consistent with the actual working environment of the sensor (deep sea, water-cooled heat dissipation), eliminating systematic biases caused by differences between laboratory calibration and the field environment. A linear model is constructed using measured data from the two temperature points. When extrapolated to the high-temperature range of the surface, the model uncertainty is significantly reduced, greatly improving the accuracy of dark background correction. Furthermore, deep-sea calibration is not limited by day or night, and combined with shallow-sea nighttime calibration, it only takes a few minutes. Compared to traditional nighttime profiling methods, nighttime operation time is reduced by more than 90%, effectively eliminating time blind spots and enabling continuous all-weather observation.

[0008] In some embodiments, the expression for the linear model in the step of constructing the two-point linear model is: (1); In equation (1), , Total number of pixels; For the first Predicted values ​​of pixels in the dark background at temperature T; and For the first The regression coefficients of each pixel.

[0009] This technical solution constructs a linear model through two-point constraints, using measured dark background data points at two different temperatures to determine model parameters. Compared with the traditional single-point lookup table method, it effectively shortens the extrapolation distance from the low-temperature calibration point (2-4℃ and 10-12℃) to the surface high-temperature range (25-30℃), reducing model extrapolation error. Furthermore, the linear model is simple in form, has low computational load, and is suitable for real-time operation on embedded platforms.

[0010] In some embodiments, the aging drift adaptive compensation step uses the first deep dark background data obtained through a two-layer field calibration step. Second-depth dark background data As the data collected in this study, the change in the first depth of dark background was calculated according to equation (2). Second depth dark background variation The expression for equation (2) is: (2); In equation (2), This is the first deep dark background data from the last deployment and storage. This is the second-depth dark background data from the previous deployment and storage; for and Assign weights separately and Calculate the overall dark background variation: (3); In equation (3), the weights and Determined based on the confidence levels of the first and second depth dark background data, and ;like If the threshold is exceeded, it is determined that the sensor has experienced dark background drift, and the data collected in this study is used to determine the cause. and The model coefficients of the linear model are updated, and the updated model coefficients are used for subsequent dark background prediction.

[0011] This technical solution achieves a comprehensive and accurate assessment of sensor aging drift by simultaneously monitoring dark background changes at both deep and shallow temperature points and assigning different weights to each. The mechanism of simultaneous monitoring at two depths effectively avoids misjudgments caused by anomalies in data from a single depth, improving the reliability of aging detection. Once drift is detected exceeding a threshold, the model coefficients are immediately updated, enabling adaptive compensation for long-term sensor aging. This eliminates the need for factory recalibration, significantly extending the calibration cycle and reducing maintenance costs.

[0012] In some embodiments, during the irradiance correction step, the corrected irradiance measurement data is calculated according to equation (4), which is expressed as: (4); In equation (4), Indicates wavelength. Indicates water depth. Indicates the current temperature. Indicates wavelength , water depth The corrected irradiance value, Indicates wavelength , water depth The raw measurement data output by the irradiance sensor. Indicates wavelength At the current temperature The predicted value for the dark background can be determined based on... It is calculated by combining the correspondence between wavelength and pixel.

[0013] This technical solution achieves complete removal of the dark background component from irradiance measurement data through precise subtraction on a pixel-by-pixel and wavelength-by-wavelength basis, directly obtaining the true irradiance value. The formula applies the temperature-compensated dark background prediction value to the actual measurement correction, effectively eliminating the impact of temperature changes and aging drift on measurement accuracy.

[0014] In some embodiments, the first dark background data and the second dark background data are collected during the ascent or descent of the irradiance sensor mounting platform.

[0015] This technical solution can make full use of the ascent or descent motion of unmanned platforms such as underwater gliders and buoys during normal operation, and automatically complete dark background data collection at the depths that the platform will inevitably pass through. It does not require additional hovering or special navigation, does not change the original observation operation mode, and does not increase additional energy consumption and time costs, thus achieving "zero additional cost" on-site calibration.

[0016] In some embodiments, when the first dark background data and the second dark background data are acquired, the sampling rate is not less than 1Hz, and temperature, depth and time data are recorded simultaneously.

[0017] In this technical solution, a sampling rate of over 1Hz ensures the statistical reliability of the dark background data, and averaging a large number of data points effectively suppresses random noise. Simultaneous recording of temperature, depth, and time data provides complete physical environment information for subsequent data filtering, quality control, and temperature-dark background relationship modeling, ensuring that the temperature-dark background data pairs used in model building are strictly corresponding in time and space, thus improving model accuracy. Simultaneously, the time information can be used to verify whether the shallow calibration was indeed conducted at night (without sunlight).

[0018] In some embodiments, the first dark background data is the average of multiple dark background data continuously collected at a first depth, and the second dark background data is the average of multiple dark background data continuously collected at a second depth.

[0019] This technical solution effectively suppresses the influence of random noise such as sensor readout noise and shot noise through multi-point averaging, making the obtained dark background values ​​closer to the true values. Averaging reduces the impact of outliers in a single measurement on the calibration results, improving the repeatability and robustness of the calibration data.

[0020] In some embodiments, the irradiance sensor is a spectral irradiance sensor, with each pixel corresponding to one spectral channel.

[0021] Based on the above scheme, the irradiance sensor dark background adaptive correction method in this embodiment of the invention utilizes the natural, light-free, and constant-temperature environment at night (>1000 meters and 450-550 meters in the deep sea) as the field calibration field. The calibration conditions are completely consistent with the actual working environment of the sensor (deep sea, water-cooled heat dissipation), eliminating the systematic deviation caused by the difference between laboratory calibration and the field environment. A linear model is constructed using measured data from two temperature points. When extrapolated to the high-temperature range of the surface layer, the uncertainty of the model is significantly reduced, greatly improving the accuracy of dark background correction. At the same time, deep calibration is not limited by day and night, and combined with shallow nighttime calibration, it only takes a few minutes. The nighttime operation time is shortened by more than 90% compared with the traditional nighttime profiling method, effectively eliminating the time blind spot and enabling continuous observation around the clock. Attached Figure Description

[0022] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings: Figure 1 This is a flowchart of the dark background adaptive correction method for the irradiance sensor of the present invention. Detailed Implementation

[0023] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0024] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature.

[0025] Example 1 like Figure 1 As shown, in one embodiment of the adaptive dark background correction method for irradiance sensor of the present invention, the adaptive dark background correction method for irradiance sensor includes a two-layer field calibration step, a two-point linear model construction step, a real-time dark background prediction step, an aging drift adaptive compensation step, and an irradiance correction step.

[0026] I. Double-layer field calibration During the deployment of irradiance sensors on platforms (such as buoys, underwater gliders, and deep-sea landers), the naturally occurring lightless and constant-temperature environment in the deep sea is utilized to obtain real dark background data at two different depths.

[0027] Deep-sea calibration: When the platform descends or ascends to the first depth (h1 > 1000 meters), the seawater temperature at this depth is constant at 2°C to 4°C, and there is naturally no light. At this time, the control sensors continuously collect dark background data at a sampling rate of no less than 1Hz, simultaneously recording temperature, depth, and time data. Let the collected deep-sea temperature data be... (2℃ ≤) ≤ 4℃), No. The first dark background data of pixels is Output dataset .

[0028] Shallow calibration: During periods without sunlight (i.e., nighttime), when the platform passes the second depth (450m ≤ h2 ≤ 550m), the nighttime seawater temperature at this depth is 10℃ to 12℃. Dark background data at this depth is collected at a sampling rate of no less than 1Hz, simultaneously recording temperature, depth, and time data. Let the collected shallow temperature data be... (10℃ ≤ ≤ 12℃), No. The second dark background data for each pixel is Output dataset .

[0029] Preferably, the first dark background data and the second dark background data are respectively taken as the average value of multiple dark background data continuously collected at the corresponding depth, so as to reduce the influence of random noise.

[0030] Deep and shallow calibrations can be completed during a single dive-ascent of the platform, or they can be completed separately during multiple deployments.

[0031] II. Steps for constructing a two-point linear model Dark current exhibits an approximately linear relationship with temperature over a finite temperature range. For each pixel of the irradiance sensor... ( = 1, 2, …, , (Total number of pixels), establish linear models between dark background values ​​and temperature respectively: (1); In equation (1), For the first Predicted values ​​of pixels in the dark background at temperature T; and For the first Model coefficients (regression coefficients) for each pixel.

[0032] For spectral irradiance sensors, each pixel corresponds to a spectral channel, and each pixel is modeled independently. The solved model coefficients for each pixel are stored in the data processing unit of the sensor or the platform for subsequent dark background prediction.

[0033] III. Real-time Dark Background Prediction During the actual measurement process (i.e., with stratospheric measurement), the current temperature data at the location of the irradiance sensor is acquired in real time. For each pixel ,Will Substitute into the established linear model:

[0034] Calculate the predicted dark background value for each pixel at the current temperature.

[0035] Due to the temperature gradient between the surface and deep sea layers, the temperature varies when the sensor operates at different depths, resulting in changes in the dark background. By using real-time temperature measurement and linear model prediction, dynamic compensation for the dark background under arbitrary temperature conditions can be achieved, avoiding the correction bias caused by temperature changes in traditional methods.

[0036] IV. Adaptive Compensation for Aging Drift During long-term use, the dark background characteristics of irradiance sensors will slowly change due to factors such as detector aging, optical window contamination, and electronic component parameter drift. To compensate for this aging drift, an adaptive compensation step is performed during each deployment.

[0037] First, acquire the deep dark background data collected during the double-layer on-site calibration process in this deployment. and shallow dark background data .

[0038] Then, calculate the changes in the deep dark background and the shallow dark background separately: (2); In equation (2), This is the first deep dark background data from the last deployment and storage. This is the second-depth dark background data from the previous deployment and storage.

[0039] for and Assign weights separately and Calculate the overall dark background variation: (3); In equation (3), the weights and Determined based on the confidence levels of the first and second depth dark background data, and Because the temperature at depths (>1000m) remains constant at 2-4℃ year-round with almost no fluctuations, the confidence level of its dark background data is higher, therefore it is preferred. ,For example = 0.6-0.8, = 0.2-0.4. The specific value can be determined according to the sensor model and the actual working environment.

[0040] like If the threshold is exceeded, it is determined that the sensor has experienced significant dark background drift. This data is then used in conjunction with the collected data. and Resolve the model coefficients of the linear model and The updated model coefficients are then stored for subsequent dark background prediction.

[0041] like If the preset threshold is not exceeded, the existing model coefficients will remain unchanged and will continue to be used for subsequent measurements.

[0042] The preset threshold can be set in advance according to the sensor model, service life, and accuracy requirements. For example, it can be set to 5%-10% of the initial dark background value.

[0043] V. Irradiance Correction After obtaining the predicted dark background value of each pixel at the current temperature in the real-time dark background prediction step, the corresponding predicted dark background value is subtracted from the original measurement data of each pixel to obtain the corrected irradiance measurement data: (4); In equation (4), Indicates wavelength. Indicates water depth. Indicates the current temperature. Indicates wavelength , water depth The corrected irradiance value, Indicates wavelength , water depth The raw measurement data output by the irradiance sensor. Indicates wavelength At the current temperature The predicted value for the dark background can be determined based on... It is calculated by combining the correspondence between wavelength and pixel.

[0044] It should be noted that there is a one-to-one correspondence between the pixels and wavelengths of an irradiance sensor, and the manufacturer provides wavelength calibration coefficients. As an illustrative example, the wavelength calibration coefficient for a certain brand of irradiance sensor is... , , , The correspondence between wavelength and pixel is as follows: .

[0045] From this, we can deduce that: , Thus obtain The value of .

[0046] By following the steps above, the effects of temperature changes and sensor aging on the dark background can be effectively eliminated, and high-precision underwater irradiance measurement data can be obtained.

[0047] Through the description of several embodiments of the adaptive dark background correction method for irradiance sensors of the present invention, it can be seen that the embodiments of the adaptive dark background correction method for irradiance sensors of the present invention have at least one or more of the following advantages: 1. The irradiance sensor dark background adaptive correction method provided by the present invention directly collects dark background data in the deep and shallow layers of the deep sea. The calibration environment is completely consistent with the actual working environment of the sensor, which fundamentally eliminates the system deviation caused by the difference in heat dissipation conditions.

[0048] 2. The irradiance sensor dark background adaptive correction method provided by the present invention only requires collecting data for a few minutes when the platform passes through a depth of 500 meters. Combined with the deep calibration (>1000 meters) already completed during the normal descent / ascent of the platform, dual-point calibration can be completed. Nighttime operation time is shortened by more than 90%, and deep calibration is not limited by day and night, enabling continuous observation in all weather conditions.

[0049] 3. The irradiance sensor dark background adaptive correction method provided by the present invention establishes a temperature-dark background linear model independently for each pixel, and stores and applies the model parameters pixel by pixel, effectively eliminating fixed pattern noise caused by differences in the signal links of each pixel.

[0050] 4. The irradiance sensor dark background adaptive correction method provided by the present invention utilizes the natural calibration opportunity during each deployment process to automatically collect dark background data and compare it with historical data, detect aging drift in real time and update model parameters, thereby achieving adaptive long-term stability maintenance and significantly reducing maintenance costs.

[0051] 5. The irradiance sensor dark background adaptive correction method provided by the present invention does not require any moving mechanical parts, thus avoiding the problems of reduced mechanical structure reliability and increased power consumption in the deep-sea high-pressure environment, and has higher reliability and lower power consumption.

[0052] 6. The irradiance sensor dark background adaptive correction method provided by the present invention makes full use of the naturally occurring lightless and constant temperature conditions in the deep sea as the calibration environment, without the need for any additional light-shielding devices or standard light sources, thus reducing system complexity and cost.

[0053] Finally, it should be noted that the various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0054] The above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them; although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications can still be made to the specific implementation of the present invention or equivalent substitutions can be made to some technical features without departing from the spirit of the technical solutions of the present invention, and all such modifications and substitutions should be covered within the scope of the technical solutions claimed in the present invention.

Claims

1. A dark background adaptive correction method for an irradiance sensor, characterized in that, Includes the following steps: Dual-layer field calibration: Acquire first dark background data and corresponding first temperature data collected by the irradiance sensor at a first depth, and acquire second dark background data and corresponding second temperature data collected by the irradiance sensor at a second depth; wherein, the first depth is greater than 1000 meters, and the temperature corresponding to the first temperature data is 2℃ to 4℃; the second depth is 450 meters to 550 meters, the second dark background data is collected during periods without sunlight, and the temperature corresponding to the second temperature data is 10℃ to 12℃; Constructing a two-point linear model: For each pixel of the irradiance sensor, establish a linear model between the dark background value and the temperature; Real-time dark background prediction: During the actual measurement process, the current temperature data of the location of the irradiance sensor is acquired in real time. The current temperature data is substituted into the linear model corresponding to each pixel to calculate the dark background prediction value of each pixel at the current temperature. Aging drift adaptive compensation: During each deployment, the first and second depth dark background data collected this time are acquired and compared with the corresponding depth dark background data collected in the previous deployment. If the change in dark background exceeds the preset threshold, it is determined that the sensor has experienced dark background drift. The model coefficients of the linear model are updated using the dark background data collected this time, and the updated model coefficients are used for subsequent dark background prediction. Irradiance correction: Subtract the corresponding dark background prediction value from the original measurement data of each pixel to obtain the corrected irradiance measurement data.

2. The irradiance sensor dark background adaptive correction method according to claim 1, characterized in that, In the steps of constructing a two-point linear model, the expression for the linear model is: （1）； In equation (1), , Total number of pixels; For the first Predicted values ​​of pixels in the dark background at temperature T; and For the first The regression coefficients of each pixel.

3. The irradiance sensor dark background adaptive correction method according to claim 1, characterized in that, In the aging drift adaptive compensation step, the first deep dark background data is obtained through a two-layer field calibration step. Second-depth dark background data As the data collected in this study, the change in the first depth of dark background was calculated according to equation (2). Second depth dark background variation The expression for equation (2) is: （2）； In equation (2), This is the first deep dark background data from the last deployment and storage. This is the second-depth dark background data from the previous deployment and storage; for and Assign weights separately and Calculate the overall dark background variation: （3）； In equation (3), the weights and Determined based on the confidence levels of the first and second depth dark background data, and ;like If the threshold is exceeded, it is determined that the sensor has experienced dark background drift, and the data collected in this study is used to determine the cause. and The model coefficients of the linear model are updated, and the updated model coefficients are used for subsequent dark background prediction.

4. The irradiance sensor dark background adaptive correction method according to claim 1, characterized in that, In the irradiance correction step, the corrected irradiance measurement data is calculated according to equation (4), which is expressed as: （4）； In equation (4), Indicates wavelength. Indicates water depth. Indicates the current temperature. Indicates wavelength , water depth The corrected irradiance value, Indicates wavelength , water depth The raw measurement data output by the irradiance sensor. Indicates wavelength At the current temperature The predicted value for the dark background can be determined based on... It is calculated by combining the correspondence between wavelength and pixel.

5. The irradiance sensor dark background adaptive correction method according to claim 1, characterized in that, The first and second dark background data were collected during the ascent or descent of the irradiance sensor mounting platform.

6. The irradiance sensor dark background adaptive correction method according to claim 1, characterized in that, When acquiring the first and second dark background data, the sampling rate is no less than 1Hz, and temperature, depth, and time data are recorded simultaneously.

7. The irradiance sensor dark background adaptive correction method according to claim 1, characterized in that, The first dark background data is the average of multiple dark background data continuously collected at the first depth, and the second dark background data is the average of multiple dark background data continuously collected at the second depth.

Images