Gimbal automatic tracking control method based on spatial coordinates of leakage points

Abstract

The present application relates to the technical field of gimbal control, and particularly relates to a gimbal automatic tracking control method based on a leakage point space coordinate, comprising: synchronously triggering a signal to unify a multi-modal data acquisition time reference, standardizing to form standardized time sequence data, and respectively constructing a gimbal self-bias field and a multi-modal bias field. Relying on multi-modal dispersion to accurately distinguish the bias source, the gimbal bias calibration mode and the mobile tracking mode are adaptively switched, and three types of gimbal bias types, namely, fixed, cumulative and high-frequency oscillation, are simultaneously identified. According to the bias type and the bias amplitude, the present application divides multiple compensation levels, matches the differential treatment strategies of static compensation, dynamic parameter compensation, global calibration compensation and fault alarm. The present application effectively improves the tracking accuracy and long-term operation stability of the gimbal.

Description

Technical Field

[0001] This invention relates to the field of gimbal control technology, and in particular to an automatic tracking control method for gimbals based on the spatial coordinates of a leakage point. Background Technology

[0002] In the field of industrial leak monitoring, PTZ visual tracking is a core technology for leak source localization and continuous monitoring. Currently, most traditional PTZ tracking control methods rely on a single sensing mode to complete leak point coordinate detection and follow-up tracking, which has significant technical shortcomings. Single detection modes have weak resistance to environmental interference, and the detected coordinates are prone to noise errors. Furthermore, the inconsistent timing of various sensor acquisition modules leads to poor spatiotemporal matching of tracking data. Simultaneously, existing technologies cannot effectively distinguish between deviations caused by the PTZ's own mechanical errors and control drift, and positional deviations caused by the actual movement of the leak source, easily leading to problems such as miscalibration, tracking lag, and trajectory deviation. In addition, traditional solutions use a fixed-frequency acquisition mode, which cannot adapt to different leak conditions, such as stable and dynamic leaks. High-frequency acquisition in stable leak scenarios easily leads to computational redundancy and noise superposition, while in dynamic leak scenarios, it is difficult to capture instantaneous position changes. Moreover, existing deviation compensation methods are singular and rigid, failing to classify and grade fixed deviations, cumulative deviations, and high-frequency oscillation deviations, making them unable to adapt to different degrees and types of PTZ fault deviations. After long-term operation, tracking accuracy degrades significantly, making it difficult to meet the practical application requirements of high-precision, high-stability automatic tracking of leak sources in complex industrial scenarios. Summary of the Invention

[0003] To address this issue, the present invention provides an automatic gimbal tracking control method based on the spatial coordinates of the leak point, which overcomes the problem in the prior art that fails to distinguish between the gimbal's own deviation and the actual movement of the leak source, leading to incorrect calibration or tracking failure.

[0004] To achieve the above objectives, the present invention provides an automatic gimbal tracking control method based on the spatial coordinates of a leak point, comprising:

[0005] Acquire the acoustic spatial coordinates of the leak source, the spatial coordinates of the hydrogen-sensitive color-changing region and the spatial coordinates of the temperature anomaly region, as well as the theoretical pointing spatial coordinates formed by the gimbal based on the current tracking command, and convert them to the same coordinate system;

[0006] Within a preset time window, acoustic coordinate timing, visible light coordinate timing, infrared coordinate timing, and gimbal theoretical pointing coordinate timing are generated respectively;

[0007] The multimodal fusion leakage coordinate timing is obtained based on the acoustic coordinate timing, visible light coordinate timing, and infrared coordinate timing, and the gimbal's own deviation field is constructed based on the coordinate difference between the multimodal fusion leakage coordinate timing and the gimbal's theoretical pointing coordinate timing.

[0008] The intermodal coordinate differences between acoustic coordinate time series, visible light coordinate time series and infrared coordinate time series are calculated respectively, and the intermodal deviation field is constructed based on the intermodal coordinate differences.

[0009] The multimodal dispersion is determined based on the intermodal deviation field, it is determined whether the current gimbal itself has a deviation, and the working mode of the gimbal is determined based on the determination result, including deviation calibration mode and motion tracking mode.

[0010] In response to the gimbal being in the deviation calibration mode, the gimbal deviation type is identified based on the gimbal's own deviation field. The gimbal deviation type includes fixed deviation, cumulative deviation, and high-frequency oscillation deviation.

[0011] The compensation level is determined based on the deviation amplitude of the gimbal's own deviation field and the type of gimbal deviation, and the corresponding compensation strategy is executed according to the compensation level. The compensation strategy includes static compensation, dynamic parameter compensation and global calibration compensation.

[0012] As a preferred technical solution for the automatic tracking control method of the PTZ based on the spatial coordinates of the leak point, a unified synchronous trigger signal is used to control data acquisition when acquiring each spatial coordinate, so that the time reference of the acquired acoustic spatial coordinates of the leak source, the spatial coordinates of the hydrogen-sensitive color-changing area, and the spatial coordinates of the temperature anomaly area is the same.

[0013] As a preferred technical solution for the automatic tracking control method of gimbal based on the spatial coordinates of the leakage point, the acoustic coordinate timing, visible light coordinate timing, infrared coordinate timing, and gimbal theoretical pointing coordinate timing within the preset time window include:

[0014] Obtain the collection timestamp corresponding to each spatial coordinate;

[0015] Arrange multiple spatial coordinates of the same coordinate type in chronological order to form the original temporal sequence of that coordinate type;

[0016] Using the current time as the endpoint, extract an interval with a time length equal to the preset time window, extract the coordinate points that fall within the interval from each original time series, and form window time series of each coordinate type;

[0017] The time intervals between adjacent coordinate points within the same window time series are uniformly interpolated to ensure that the window time series of each coordinate type have the same time node sequence, thus obtaining the acoustic coordinate time series, visible light coordinate time series, infrared coordinate time series, and gimbal theoretical pointing coordinate time series.

[0018] As a preferred technical solution for the automatic tracking control method of gimbal based on the spatial coordinates of the leakage point, the step of constructing the gimbal's own deviation field includes:

[0019] The acoustic coordinates, visible light coordinates, and infrared coordinates corresponding to the acoustic coordinate time series, visible light coordinate time series, and infrared coordinate time series are weighted and fused to obtain the multimodal fusion leakage coordinates at that time node. The multimodal fusion leakage coordinate time series is formed by traversing each time node.

[0020] The instantaneous deviation vector at that time node is obtained by subtracting two coordinates at the same time node in the multimodal fusion leakage coordinate time sequence and the gimbal theoretical pointing coordinate time sequence. The instantaneous deviation vector time sequence is formed by traversing each time node.

[0021] The instantaneous deviation vector sequence is used as the gimbal's own deviation field.

[0022] The instantaneous deviation vector includes three components: horizontal deviation angle, vertical deviation angle, and radial distance deviation.

[0023] As a preferred technical solution for the gimbal automatic tracking control method based on the spatial coordinates of the leakage point, the steps for constructing the inter-modal deviation field based on the coordinate differences between each mode include:

[0024] The acoustic and visible light coordinates at the same time point are subtracted to obtain the instantaneous deviation vector of the acoustic and visible light modes at that time point. The time sequence of the instantaneous deviation vector of the acoustic and visible light modes is formed by traversing each time point.

[0025] The acoustic and infrared coordinates at the same time point are subtracted to obtain the instantaneous deviation vector of the acoustic and infrared modes at that time point. The time sequence of the instantaneous deviation vector of the acoustic and infrared modes is formed by traversing each time point.

[0026] The difference between the visible light coordinates and the infrared coordinates at the same time point is used to obtain the instantaneous deviation vector of the visible light and infrared modes at that time point. The time sequence of the instantaneous deviation vector of the visible light and infrared modes is formed by traversing each time point.

[0027] The instantaneous deviation vector time series of acoustic and visible light modes, acoustic and infrared modes, and visible light and infrared modes are combined to construct an intermodal deviation field;

[0028] The instantaneous deviation vectors of acoustic and visible light modes, acoustic and infrared modes, and visible and infrared modes all include three components: horizontal deviation angle, vertical deviation angle, and radial distance deviation.

[0029] As a preferred technical solution for the gimbal automatic tracking control method based on the spatial coordinates of the leakage point, determining the multimodal discreteness corresponding to the preset time window includes:

[0030] Based on the instantaneous deviation vectors of acoustic and visible light modes, acoustic and infrared modes, and visible light and infrared modes at each time node in the intermodal deviation field, the instantaneous modal dispersion at each time node is calculated, whereby the instantaneous modal dispersion is the mean of the magnitudes of the three instantaneous deviation vectors.

[0031] Within the preset time window, the relative average deviation of the instantaneous modal dispersion at each time node is calculated, and the relative average deviation is used as the multimodal dispersion.

[0032] As a preferred technical solution for the automatic tracking control method of gimbal based on the spatial coordinates of the leakage point, the method determines whether the current gimbal itself has a deviation based on the multimodal dispersion and determines the working mode of the gimbal, including:

[0033] If the multimodal dispersion is less than the preset dispersion threshold, it is determined that the current gimbal has its own deviation, and the working mode is determined to be deviation calibration mode.

[0034] Conversely, if the deviation is not found, it is determined that the current deviation originates from the actual movement of the leakage source, and the working mode is set to movement tracking mode.

[0035] As a preferred technical solution for the automatic tracking control method of gimbal based on the spatial coordinates of the leakage point, the gimbal deviation type is identified according to the gimbal's own deviation field, including:

[0036] Calculate the magnitude of the instantaneous deviation vector at each time node in the deviation field of the gimbal itself to form a deviation amplitude time series;

[0037] Analyze the statistical characteristics and trends of the deviation amplitude time series within a preset time window;

[0038] If the standard deviation of the deviation amplitude time series is less than a preset standard deviation threshold, it is identified as a fixed deviation.

[0039] If the deviation amplitude time series shows a monotonic change trend with time, it is identified as a cumulative deviation;

[0040] If the deviation amplitude time series exhibits periodic fluctuation characteristics, it is identified as a high-frequency oscillation deviation.

[0041] As a preferred technical solution for the automatic tracking control method of gimbal based on the spatial coordinates of the leakage point, the compensation level is determined according to the deviation amplitude of the gimbal's own deviation field and the type of gimbal deviation, and the corresponding compensation strategy is executed according to the compensation level, including:

[0042] The mean value of the time sequence of the deviation amplitude of the gimbal's own deviation field is obtained as the average deviation amplitude;

[0043] If the gimbal deviation type is high-frequency oscillation deviation, a fault alarm will be output directly without any compensation strategy.

[0044] If the gimbal deviation type is fixed deviation and the average deviation amplitude is less than the fixed deviation threshold, the compensation level is determined to be level one, and the corresponding static compensation strategy is executed.

[0045] If the gimbal deviation type is fixed deviation and the average deviation amplitude is greater than or equal to the fixed deviation threshold, the compensation level is determined to be level three, and a global calibration compensation strategy is executed accordingly.

[0046] If the gimbal deviation type is cumulative deviation and the average deviation amplitude is less than the cumulative deviation threshold, the compensation level is determined to be level two, and the corresponding dynamic parameter compensation strategy is executed.

[0047] If the gimbal deviation type is cumulative deviation and the average deviation amplitude is greater than or equal to the cumulative deviation threshold, the compensation level is determined to be level three, and a global calibration compensation strategy is executed accordingly.

[0048] As a preferred technical solution for the automatic tracking control method of gimbal based on the spatial coordinates of the leakage point, the static compensation strategy is to fit a compensation function to the gimbal's own deviation field and then superimpose the compensation function into the subsequent gimbal control commands.

[0049] The dynamic parameter compensation strategy is to reduce the upper limit of the gimbal's rotation speed and acceleration while superimposing the compensation function.

[0050] The global calibration and compensation strategy involves controlling the gimbal to sequentially point to multiple fixed reference leak points with known absolute spatial coordinates. While the gimbal is pointing to each fixed reference leak point, the acoustic spatial coordinates, hydrogen-sensitive color-changing area spatial coordinates, and temperature anomaly area spatial coordinates of each fixed reference leak point are simultaneously acquired. Based on the difference between the known absolute spatial coordinates of each fixed reference leak point and the corresponding acquired spatial coordinates, the angle compensation value of the gimbal in each direction is calculated, and the angle compensation value is applied to subsequent gimbal control commands.

[0051] Compared with existing technologies, the advantages of this invention lie in its ability to simultaneously acquire the spatial coordinates of the acoustic, visible light, and infrared three-modal leakage sources and compare them with the theoretical pointing coordinates of the gimbal, thereby constructing the gimbal's own deviation field and inter-modal deviation field. The inter-modal dispersion accurately distinguishes between the gimbal's own tracking deviation and the actual movement of the leakage source, avoiding erroneous gimbal self-calibration triggered by leakage source movement. Furthermore, it identifies three types of deviations: fixed deviation, cumulative deviation, and high-frequency oscillation deviation, and implements graded compensation strategies for different deviation types and amplitudes. This achieves accurate attribution, adaptive graded handling, and closed-loop correction of gimbal tracking deviations, significantly improving the tracking reliability, positioning accuracy, and fault identification capabilities of the gimbal in complex industrial environments. Attached Figure Description

[0052] Figure 1 This is a flowchart of the automatic tracking control method for a gimbal based on the spatial coordinates of a leak point, according to an embodiment of the present invention.

[0053] Figure 2 A flowchart for constructing the intermodal deviation field in an embodiment of the present invention;

[0054] Figure 3 This is a logic diagram for determining whether the gimbal itself has a deviation in an embodiment of the present invention. Detailed Implementation

[0055] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0056] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0057] It should be noted that in the description of this invention, the terms "upper", "lower", "left", "right", "inner", "outer", etc., which indicate directions or positional relationships, are based on the directions or positional relationships shown in the accompanying drawings. This is only for the convenience of description and is not intended to indicate or imply that the device or element must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation of this invention.

[0058] Furthermore, it should be noted that, in the description of this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0059] Please see Figure 1 As shown in the figure, this invention provides an automatic gimbal tracking control method based on the spatial coordinates of a leak point, including:

[0060] Step S1: Obtain the acoustic spatial coordinates of the leak source, the spatial coordinates of the hydrogen-sensitive color-changing region and the spatial coordinates of the temperature anomaly region, as well as the theoretical pointing spatial coordinates formed by the gimbal based on the current tracking command, and convert them to the same coordinate system;

[0061] Step S2: Within a preset time window, acoustic coordinate timing, visible light coordinate timing, infrared coordinate timing, and gimbal theoretical pointing coordinate timing are generated respectively.

[0062] Step S3: Obtain the multimodal fusion leakage coordinate time sequence based on the acoustic coordinate time sequence, visible light coordinate time sequence, and infrared coordinate time sequence; and construct the gimbal's own deviation field based on the coordinate difference between the multimodal fusion leakage coordinate time sequence and the gimbal's theoretical pointing coordinate time sequence.

[0063] Step S4: Calculate the intermodal coordinate differences between acoustic coordinate time series, visible light coordinate time series and infrared coordinate time series respectively, and construct the intermodal deviation field based on the intermodal coordinate differences;

[0064] Step S5: Determine the multimodal dispersion based on the intermodal deviation field, determine whether the current gimbal itself has a deviation, and determine the working mode of the gimbal, including deviation calibration mode and motion tracking mode, based on the determination result.

[0065] Step S6: In response to the gimbal being in the deviation calibration mode, the gimbal deviation type is identified based on the gimbal's own deviation field. The gimbal deviation type includes fixed deviation, cumulative deviation, and high-frequency oscillation deviation.

[0066] Step S7: Determine the compensation level based on the deviation amplitude of the gimbal's own deviation field and the gimbal deviation type, and execute the corresponding compensation strategy according to the compensation level. The compensation strategy includes static compensation, dynamic parameter compensation, and global calibration compensation.

[0067] Different types of sensor acquisition modules have inherent differences in hardware sampling frequency and startup response latency. The time base of multimodal coordinate data acquired in asynchronous acquisition mode is not uniform, which will directly lead to systematic errors in subsequent time sequence construction, deviation calculation and modal fusion results. A unified synchronization triggering mechanism can realize instantaneous synchronization of acquisition actions of multiple modules.

[0068] In implementation, a periodic square wave trigger signal is generated by an FPGA or microcontroller, simultaneously triggering the acoustic acquisition card, visible light camera, and infrared thermal imager. The trigger frequency is set according to the dynamic change rate of the leakage source. The rotation speed of a typical explosion-proof pan-tilt unit is usually 0.1° / s to 30° / s. Even at the fastest rotation speed of 30° / s, it only moves 0.6° per frame (0.02 seconds). The encoder resolution is generally at the 0.01° level, and the trigger frequency can be set to 5Hz to 20Hz. For rapidly changing leakage sources (such as intermittent spraying), the frequency is set to 10Hz to 20Hz, and for stable leakage sources, it is set to 5Hz to 10Hz. After synchronous triggering, all coordinate data are accompanied by the same timestamp, and subsequent processing can directly align according to the timestamp.

[0069] The original time series coordinates are messy and have inconsistent sampling densities, making it impossible to directly perform point-to-point difference calculations and modal fusion. Standardization processing, such as window truncation and sorted interpolation, is required to unify the time series rules of all time series.

[0070] The purpose of generating time sequences for each coordinate within a preset time window is to organize the original discrete coordinate data into a time-aligned, equally spaced sequence, facilitating subsequent deviation field calculations. The length of the preset time window needs to cover the dynamic response time of the gimbal and the possible changes in the leakage source, typically ranging from 5 to 30 seconds. For fast-moving dynamic scenarios, 5 to 10 seconds can be used. This window length can be determined by measuring the response time of the gimbal from receiving a command to completing its rotation under typical operating conditions, and taking twice that time as the window length to ensure that the window contains sufficient data points.

[0071] The specific steps include: first, obtaining the acquisition timestamp for each coordinate; then, sorting by time to obtain the original time series; next, extracting coordinate points within a window with the current time as the endpoint; and finally, uniformly interpolating the time intervals between adjacent points within the window using linear interpolation, i.e., calculating the coordinate value of the intermediate point based on the time weights of the two preceding and following measured points. The resulting four time series all have corresponding coordinates at the same time node, laying the foundation for subsequent fusion and subtraction.

[0072] Coordinate data from a single sensing mode contains independent detection noise and error. Multimodal weighted fusion can minimize single-point detection error. The fused coordinates are used as the true location reference of the leakage source. By comparing them with the preset theoretical pointing coordinates of the gimbal, the tracking deviation caused by the mechanical rotation of the gimbal and the execution of control commands can be accurately separated.

[0073] The weights for weighted fusion are dynamically determined based on the dispersion of each modal coordinate within a preset time window. The dispersion can be calculated using variance or mean absolute deviation: first, the mean of the coordinate values ​​of the modal coordinate time series at each time node within the window is calculated, including three components: horizontal angle, vertical angle, and radial distance. Then, the sum of squares (variance) or the sum of absolute deviations (mean absolute deviation) of the coordinates of each node from the mean is calculated.

[0074] For each mode, the dispersion of the three components—horizontal angle, vertical angle, and radial distance—is calculated separately. The average dispersion of these three components is then taken as the overall dispersion of the mode. The principle of weight allocation is: modes with smaller overall dispersion are assigned larger weights, indicating more stable coordinates and lower noise. Specifically, the reciprocal of the overall dispersion of each mode is first taken, and then these three reciprocals are normalized. That is, the weight of each mode is equal to its reciprocal dispersion divided by the sum of the three reciprocals.

[0075] For example, if the overall dispersion of acoustic coordinates is 0.01, visible light is 0.02, and infrared is 0.03, then their reciprocals are 100, 50, and 33.33 respectively. After normalization, the acoustic weights are approximately 0.546, visible light is approximately 0.273, and infrared is approximately 0.181. The dynamic weighting method can adapt to changes in the signal-to-noise ratio of each mode under different environmental conditions: under strong background noise, the acoustic dispersion increases, and its weight automatically decreases; under nighttime or low light conditions, the visible light dispersion increases, and its weight automatically decreases; when the leakage is extremely small and the infrared temperature change is weak, the infrared dispersion increases, and its weight automatically decreases. The final fused coordinates are obtained by summing the three modal coordinates according to the above dynamic weights, thereby maximally suppressing the random errors and systematic biases of a single mode.

[0076] After weighted fusion, multimodal fused leakage coordinates are obtained. Then, the difference between the fused coordinates and the theoretical pointing coordinates of the gimbal is taken to obtain the instantaneous deviation vector. This vector contains three components: horizontal deviation angle, vertical deviation angle, and radial distance deviation, which correspond to the gimbal's horizontal rotation error, pitch rotation error, and distance direction error, respectively. By traversing all time points, the deviation vector time series is obtained, which is the gimbal's own deviation field.

[0077] When the PTZ is working normally and has no self-deviation, the coordinate data detected by the three types of sensor modes are highly consistent, and the deviation between modes is minimal, regardless of whether the leakage source is stationary or moving smoothly. If the coordinate deviation between the multiple modes is significantly discrete, it indicates that the coordinate fluctuation comes from the actual movement of the leakage source or environmental interference, rather than a malfunction of the PTZ itself.

[0078] In practice, the differences between acoustic and visible light coordinates, acoustic and infrared coordinates, and visible and infrared coordinates are calculated for each time point. Each difference is represented by three components: horizontal, vertical, and radial. For example, if the acoustic coordinates indicate a leak point at (0.1°, 0.2°, 5m) and the visible light coordinates indicate (0.12°, 0.19°, 5.02m), then the difference is (-0.02°, 0.01°, -0.02m). Combining these three sets of differences across all time points forms the intermodal deviation field.

[0079] Understandably, this deviation field does not rely on any external reference, but only utilizes the cross-verification between different physical quantities. If the leakage source is fixed, the positioning results of the three modes should coincide within the sensor calibration accuracy. If the leakage source is moving, due to the different response speeds of each mode (acoustic being the fastest and infrared the slowest), they will exhibit regular differences. This step provides the raw data for subsequent discreteness calculations.

[0080] The instantaneous modal deviation at a single time point cannot reflect the overall discrete characteristics within a preset time window. By using the relative average deviation of the instantaneous dispersion mean, the overall fluctuation and dispersion of multimodal detection data within a time period can be objectively and comprehensively characterized, providing a unified standard for subsequent threshold determination.

[0081] Determining the multimodal dispersion is to quantify the degree of inconsistency among the three modes. First, the instantaneous modal dispersion at each time point is calculated, which is achieved by taking the average of the magnitudes of the three instantaneous deviation vectors. The average of the three magnitudes is taken as the instantaneous dispersion at that time point. Then, within a preset time window, the relative average deviation of all instantaneous dispersions is calculated.

[0082] The relative average deviation is defined as the sum of the absolute deviations of each instantaneous dispersion from the average dispersion within the window, divided by the product of the average dispersion and the number of time nodes. For example, if there are 10 nodes within the window, the average dispersion is 0.03, and the dispersions of each node are [0.028, 0.032, ...], and the sum of the absolute deviations is 0.02, then the relative average deviation = 0.02 / (0.03 × 10) = 0.0667. This value is used as the multimodal dispersion for the entire window. The relative average deviation can eliminate the influence of overall scaling of dispersion within the window, and more sensitively reflect the relative fluctuations between modes.

[0083] When the leakage source is fixed and the gimbal is without deviation, the positioning results of the three modes will highly overlap with very small dispersion. When the gimbal itself has a deviation, the positioning results of the three modes will still overlap because the sensors are fixed relative to the gimbal, but they will have a common offset from the theoretical direction, and the dispersion will still be very small. When the leakage source actually moves, the different response speeds of the three modes will cause the positioning results to disperse, and the dispersion will increase significantly. Therefore, a preset dispersion threshold is set to distinguish between these two situations.

[0084] This threshold is determined by measuring multiple sets of dispersion data in a laboratory environment using a fixed leak source, with the pan-tilt unit in both unbiased and biased states. The boundary value of the dispersion distribution under the two states is taken. For example, the dispersion is ≤0.02 when there is no deviation, ≤0.025 when there is a deviation, and ≥0.05 when the leak moves. Therefore, the threshold can be set to 0.03.

[0085] In practice, if the calculated multimodal dispersion is less than the threshold, the three modes are considered to be pointing in the same direction, and the current deviation originates from the gimbal itself, thus entering deviation calibration mode; otherwise, the deviation is considered to originate from the movement of the leakage source, thus entering motion tracking mode. In motion tracking mode, the gimbal no longer attempts self-calibration, but instead uses the acoustic coordinates with the fastest response as the primary target for real-time tracking.

[0086] Different types of hardware and control faults in a gimbal will exhibit different temporal fluctuation characteristics. Fixed deviations show steady-state small fluctuations with no obvious trend changes, while cumulative deviations show a gradual characteristic of increasing or decreasing unidirectionally over time. High-frequency oscillation deviations show regular periodic fluctuations. These three types of deviations can be accurately distinguished through time-series statistical characteristics and trend analysis.

[0087] In practice, the magnitude of the instantaneous deviation vector at each time point is first calculated to obtain the time series of deviation amplitude. Then, the statistical characteristics of the time series are analyzed within a preset time window: the standard deviation is calculated to determine the degree of dispersion, the slope of the linear regression is calculated to determine the monotonic trend, and the autocorrelation function or the number of peaks is calculated to determine the periodicity.

[0088] The preset standard deviation threshold is calculated by collecting a segment of data with no or fixed deviation under normal gimbal operating conditions, and then setting the standard deviation to 1.5 times this value as the threshold, for example, 0.01°. If the standard deviation of the amplitude time series within the window is less than this threshold, it indicates that the amplitude is basically constant and is identified as a fixed deviation. For cumulative deviation, to determine if there is a monotonic trend: perform linear fitting on the amplitude; if the absolute value of the fitting slope is greater than the preset slope threshold (e.g., 0.005° / s) and the goodness of fit is good... A value >0.8 is identified as a cumulative deviation. For high-frequency oscillation deviations, periodicity is determined by calculating the autocorrelation function of the amplitude time series. If a significant peak occurs within a lag time of less than 0.5 seconds (correlation coefficient >0.6), or the amplitude fluctuation frequency is greater than 5Hz, it is identified as a high-frequency oscillation deviation. Fixed deviations typically originate from installation alignment errors or sensor coaxiality deviations; cumulative deviations originate from encoder cumulative errors or transmission slippage; high-frequency oscillation deviations originate from mechanical wear or external vibrations.

[0089] The fault mechanisms and impact of the three types of gimbal deviations are significantly different. High-frequency oscillation deviations are mostly caused by loose hardware or component failures, which cannot be compensated by software algorithms and require direct alarm and repair. Fixed deviations and cumulative deviations are systematic deviations that can be corrected by software, and the smaller the deviation amplitude, the less severe the fault. Different compensation levels can be divided according to the amplitude, and corresponding lightweight, standardized, and full-dimensional compensation strategies can be matched.

[0090] The average deviation amplitude is obtained by taking the mean of the deviation amplitude timing. For high-frequency oscillation deviations, a fault alarm is directly output because this deviation is a hardware mechanical problem that cannot be solved by any software compensation; further compensation may only exacerbate the damage.

[0091] For fixed deviations, a minimum threshold value is set, for example, 0.05°. This threshold is determined by measuring the allowable error range during normal operation after initial calibration following gimbal installation, and then taking 80% of the upper limit as the minimum threshold value. If the average deviation amplitude is less than this threshold, it indicates a small deviation, and a static compensation strategy (superimposed compensation function) is used; if it is greater than or equal to the threshold, it indicates a large deviation, requiring global calibration.

[0092] For cumulative deviation, a cumulative deviation threshold is set, for example, 0.1°. This threshold is determined by simulating tracking errors under different levels of cumulative deviation and using the maximum error value at which the gimbal can still maintain basic tracking. If the average deviation amplitude is less than this threshold, dynamic parameter compensation is used; if it is greater than or equal to the threshold, global calibration is still required. This hierarchical strategy balances compensation effectiveness and system response efficiency.

[0093] Static compensation strategy: Based on the gimbal's own deviation field fitting compensation function, since the fixed deviation is basically constant, the average value of the deviation amplitude time sequence can be taken as a constant compensation value, which is directly superimposed on the expected angle of subsequent gimbal control commands. For example, if the theoretical horizontal pointing angle of the gimbal is 30° and the fixed deviation is +0.2°, then the actual command issued is 30°-0.2°=29.8°.

[0094] Dynamic parameter compensation strategy: Based on the superimposed compensation function, the upper limit of the gimbal's rotation speed is reduced from the default 30° / s to 15° / s, and the upper limit of acceleration is reduced from 10° / s. Reduce to 5° / Reducing dynamic parameters can decrease the impact and slippage of the transmission mechanism, thereby suppressing the further expansion of accumulated deviation.

[0095] Global calibration compensation strategy: The pan-tilt unit is controlled to sequentially point to 3-5 fixed reference leakage points. The absolute spatial coordinates of these reference points are pre-calibrated precisely using laser ranging or a total station. For each reference point, its acoustic, visible light, and infrared coordinates are simultaneously acquired. The angle difference between the measured and known coordinates of each reference point is then calculated, and the average difference across multiple reference points is taken as the global angle compensation value for the pan-tilt unit in each direction. This compensation value not only corrects fixed deviations but also compensates for systematic errors caused by encoder nonlinearity or temperature drift. The compensation value is applied to all subsequent control commands until the next global calibration is triggered.

[0096] It should be noted that the fixed reference leak point involved in this implementation does not refer to the actual leak point, but rather to a fixed reference object with known absolute spatial coordinates pre-positioned in the monitoring scenario to execute a global calibration compensation strategy. In actual engineering deployment, this can be achieved by setting up a sound wave generator (such as a miniature speaker), a controllable heating element (such as an electric heating element), and a hydrogen-sensitive color-changing label at multiple fixed locations within the pan-tilt-zoom (PTZ) field of view. These three components can be integrated onto the same reference bracket.

[0097] When global calibration is required, the pan-tilt unit sequentially points to each reference support, simultaneously controlling the acoustic generator to emit simulated leak sound signals, the heating element to generate simulated leak temperature rises, and the hydrogen-sensitive color-changing markers to change color under heating or specific lighting conditions. This allows for the simultaneous acquisition of spatial coordinates in acoustic, infrared, and visible light modes. The characteristics of these simulated signals are highly consistent with the real leak source, effectively stimulating each sensing mode to perform localization calculations. The absolute spatial coordinates of each reference support are pre-calibrated using high-precision measuring instruments (such as total stations or laser trackers).

[0098] Therefore, the fixed reference leak point in this method should be interpreted broadly, including but not limited to the simulated leak source or a permanent calibration object with known coordinates. Its function is to provide a fixed, multimodal, and perceptible spatial reference for the gimbal, rather than relying on the occurrence of actual leak events.

[0099] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.

[0100] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A gimbal automatic tracking control method based on a leakage point spatial coordinate, characterized in that, include: Acquire the acoustic spatial coordinates of the leak source, the spatial coordinates of the hydrogen-sensitive color-changing region and the spatial coordinates of the temperature anomaly region, as well as the theoretical pointing spatial coordinates formed by the gimbal based on the current tracking command, and convert them to the same coordinate system; Within a preset time window, acoustic coordinate timing, visible light coordinate timing, infrared coordinate timing, and gimbal theoretical pointing coordinate timing are generated respectively; The multimodal fusion leakage coordinate timing is obtained based on the acoustic coordinate timing, visible light coordinate timing, and infrared coordinate timing, and the gimbal's own deviation field is constructed based on the coordinate difference between the multimodal fusion leakage coordinate timing and the gimbal's theoretical pointing coordinate timing. The intermodal coordinate differences between acoustic coordinate time series, visible light coordinate time series and infrared coordinate time series are calculated respectively, and the intermodal deviation field is constructed based on the intermodal coordinate differences. The multimodal dispersion is determined based on the intermodal deviation field, it is determined whether the current gimbal itself has a deviation, and the working mode of the gimbal is determined based on the determination result, including deviation calibration mode and motion tracking mode. In response to the gimbal being in the deviation calibration mode, the gimbal deviation type is identified based on the gimbal's own deviation field. The gimbal deviation type includes fixed deviation, cumulative deviation, and high-frequency oscillation deviation. The compensation level is determined based on the deviation amplitude of the gimbal's own deviation field and the type of gimbal deviation, and the corresponding compensation strategy is executed according to the compensation level. The compensation strategies include static compensation, dynamic parameter compensation, and global calibration compensation.

2. The gimbal automatic tracking control method based on the spatial coordinates of the leakage point according to claim 1, characterized in that, When acquiring various spatial coordinates, a unified synchronous trigger signal is used to control data acquisition, so that the time references of the acquired acoustic spatial coordinates of the leakage source, the spatial coordinates of the hydrogen-sensitive color-changing area, and the spatial coordinates of the temperature anomaly area are the same.

3. The automatic tracking control method for a gimbal based on the spatial coordinates of a leak point according to claim 2, characterized in that, Within the preset time window, the acoustic coordinate timing, visible light coordinate timing, infrared coordinate timing, and gimbal theoretical pointing coordinate timing include: Obtain the collection timestamp corresponding to each spatial coordinate; Arrange multiple spatial coordinates of the same coordinate type in chronological order to form the original temporal sequence of that coordinate type; Using the current time as the endpoint, extract an interval with a time length equal to the preset time window, extract the coordinate points that fall within the interval from each original time series, and form window time series of each coordinate type; The time intervals between adjacent coordinate points within the same window time series are uniformly interpolated to ensure that the window time series of each coordinate type have the same time node sequence, thus obtaining the acoustic coordinate time series, visible light coordinate time series, infrared coordinate time series, and gimbal theoretical pointing coordinate time series.

4. The automatic tracking control method for a gimbal based on the spatial coordinates of a leak point according to claim 3, characterized in that, The steps for constructing the gimbal's own deviation field include: The acoustic coordinates, visible light coordinates, and infrared coordinates corresponding to the acoustic coordinate time series, visible light coordinate time series, and infrared coordinate time series are weighted and fused to obtain the multimodal fusion leakage coordinates at that time node. The multimodal fusion leakage coordinate time series is formed by traversing each time node. The instantaneous deviation vector at that time node is obtained by subtracting two coordinates at the same time node in the multimodal fusion leakage coordinate time sequence and the gimbal theoretical pointing coordinate time sequence. The instantaneous deviation vector time sequence is formed by traversing each time node. The instantaneous deviation vector time sequence is used as the gimbal's own deviation field; The instantaneous deviation vector includes three components: horizontal deviation angle, vertical deviation angle, and radial distance deviation.

5. The automatic tracking control method for a gimbal based on the spatial coordinates of a leak point according to claim 4, characterized in that, The steps for constructing the intermodal deviation field based on the coordinate differences between each mode include: The acoustic and visible light coordinates at the same time point are subtracted to obtain the instantaneous deviation vector of the acoustic and visible light modes at that time point. The time sequence of the instantaneous deviation vector of the acoustic and visible light modes is formed by traversing each time point. The acoustic and infrared coordinates at the same time point are subtracted to obtain the instantaneous deviation vector of the acoustic and infrared modes at that time point. The time sequence of the instantaneous deviation vector of the acoustic and infrared modes is formed by traversing each time point. The difference between the visible light coordinates and the infrared coordinates at the same time point is used to obtain the instantaneous deviation vector of the visible light and infrared modes at that time point. The time sequence of the instantaneous deviation vector of the visible light and infrared modes is formed by traversing each time point. The instantaneous deviation vector time series of acoustic and visible light modes, acoustic and infrared modes, and visible light and infrared modes are combined to construct an intermodal deviation field; The instantaneous deviation vectors of acoustic and visible light modes, acoustic and infrared modes, and visible and infrared modes all include three components: horizontal deviation angle, vertical deviation angle, and radial distance deviation.

6. The automatic tracking control method for a gimbal based on the spatial coordinates of a leak point according to claim 5, characterized in that, Determining the multimodal discreteness corresponding to the preset time window includes: Based on the instantaneous deviation vectors of acoustic and visible light modes, acoustic and infrared modes, and visible light and infrared modes at each time node in the intermodal deviation field, the instantaneous modal dispersion at each time node is calculated, whereby the instantaneous modal dispersion is the mean of the magnitudes of the three instantaneous deviation vectors. Within the preset time window, the relative average deviation of the instantaneous modal dispersion at each time node is calculated, and the relative average deviation is used as the multimodal dispersion.

7. The automatic tracking control method for a gimbal based on the spatial coordinates of a leak point according to claim 6, characterized in that, Based on the multimodal dispersion, determine whether the current gimbal itself has a deviation, and determine the working mode of the gimbal, including: If the multimodal dispersion is less than the preset dispersion threshold, it is determined that the current gimbal has its own deviation, and the working mode is determined to be the deviation calibration mode. Conversely, if the deviation is not found, it is determined that the current deviation originates from the actual movement of the leakage source, and the working mode is set to movement tracking mode.

8. The automatic tracking control method for a gimbal based on the spatial coordinates of a leak point according to claim 7, characterized in that, Identifying the gimbal deviation type based on the gimbal's own deviation field includes: Calculate the magnitude of the instantaneous deviation vector at each time node in the deviation field of the gimbal itself to form a deviation amplitude time series; Analyze the statistical characteristics and trends of the deviation amplitude time series within a preset time window; If the standard deviation of the deviation amplitude time series is less than a preset standard deviation threshold, it is identified as a fixed deviation. If the deviation amplitude time series shows a monotonic change trend with time, it is identified as a cumulative deviation; If the deviation amplitude time series exhibits periodic fluctuation characteristics, it is identified as a high-frequency oscillation deviation.

9. The automatic tracking control method for a gimbal based on the spatial coordinates of a leak point according to claim 8, characterized in that, The compensation level is determined based on the deviation amplitude of the gimbal's own deviation field and the type of gimbal deviation, and the corresponding compensation strategy is executed according to the compensation level, including: The mean value of the time series of the deviation amplitude of the gimbal's own deviation field is obtained as the average deviation amplitude; If the gimbal deviation type is high-frequency oscillation deviation, a fault alarm will be output directly without any compensation strategy. If the gimbal deviation type is fixed deviation and the average deviation amplitude is less than the fixed deviation threshold, the compensation level is determined to be level one, and the corresponding static compensation strategy is executed. If the gimbal deviation type is fixed deviation and the average deviation amplitude is greater than or equal to the fixed deviation threshold, the compensation level is determined to be level three, and a global calibration compensation strategy is executed accordingly. If the gimbal deviation type is cumulative deviation and the average deviation amplitude is less than the cumulative deviation threshold, the compensation level is determined to be level two, and the corresponding dynamic parameter compensation strategy is executed. If the gimbal deviation type is cumulative deviation and the average deviation amplitude is greater than or equal to the cumulative deviation threshold, the compensation level is determined to be level three, and a global calibration compensation strategy is executed accordingly.

10. The automatic tracking control method for a gimbal based on the spatial coordinates of a leak point according to claim 9, characterized in that, The static compensation strategy is to fit a compensation function based on the gimbal's own deviation field and then superimpose the compensation function into subsequent gimbal control commands. The dynamic parameter compensation strategy is to reduce the upper limit of the gimbal's rotation speed and acceleration while superimposing the compensation function. The global calibration and compensation strategy involves controlling the gimbal to sequentially point to multiple fixed reference leak points with known absolute spatial coordinates. While the gimbal is pointing to each fixed reference leak point, the acoustic spatial coordinates, hydrogen-sensitive color-changing area spatial coordinates, and temperature anomaly area spatial coordinates of each fixed reference leak point are simultaneously acquired. Based on the difference between the known absolute spatial coordinates of each fixed reference leak point and the corresponding acquired spatial coordinates, the angle compensation value of the gimbal in each direction is calculated, and the angle compensation value is applied to subsequent gimbal control commands.

Images