Intelligent image stabilization method and system for unmanned intelligent turntable

Abstract

The invention provides an intelligent image stabilization method and system for an unmanned intelligent turntable. According to the method, an interference library is updated by acquiring electromagnetic interference data, and a signal-to-noise ratio is monitored to generate notch filtering parameters. Based on the parameters, a wave trap is adjusted, non-linear compensation and frequency domain equalization are synchronously triggered, image key frame abrupt change signals are identified at the same time, compensation intensity is adjusted, and closed-loop suppression is formed. In a closed loop, an inner loop extracts high-frequency jitter, an outer loop fuses interference prediction to generate primary parameters, and a cross-domain suppression instruction is output after motion analysis. And instruction-driven anti-interference coding: processing rotation suppression by an attitude coding scheme, processing electromagnetic suppression by a hierarchical error correction protocol, and performing closed-loop verification through link feedback. A verification result is fed back to update the interference library and the compensation intensity, cooperative control of electromagnetic interference and mechanical jitter is achieved, and a stable image is output. By constructing a dynamic closed-loop suppression mechanism, electromagnetic interference and mechanical jitter of the rotary table are cooperatively suppressed, and stable image output of the unmanned intelligent rotary table is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of image stabilization of intelligent rotating platform, and particularly relates to an intelligent image stabilization method and system for unmanned intelligent rotating platform. BACKGROUND

[0002] In complex application scenarios such as unmanned aerial vehicle reconnaissance and remote monitoring, the unmanned intelligent rotating platform needs to stably operate and transmit high-quality images in a strong electromagnetic interference environment. Such an environment not only interferes with the control signals and sensor data of the rotating platform, causing image jitter, frame loss or distortion, but also may induce abnormal mechanical vibration of the rotating platform power assembly, superimposing with inherent mechanical jitter to form complex "electromagnetic-mechanical" coupling interference. Therefore, an intelligent image stabilization method is urgently needed, which can dynamically perceive and distinguish electromagnetic interference and mechanical jitter, realize collaborative suppression of cross-domain interference, and adaptively adjust parameters under a closed-loop mechanism to ensure that stable, smooth and high-definition image streams can still be output in a harsh electromagnetic environment.

[0003] At present, a targeted scheme is an image stabilization method based on adaptive notch filtering and motion compensation. The core of this scheme is to monitor the signal-to-noise ratio features of the current or vibration signals of the rotating platform power assembly, and accordingly dynamically generate and adjust the center frequency, depth and bandwidth of the notch filter, so as to specifically filter out specific mechanical jitter components related to the rotating frequency and its harmonics of the rotating platform, which are induced or amplified by electromagnetic interference. At the same time, in combination with the gyroscope or accelerometer data, motion estimation and compensation are performed on the image sequence to try to offset the remaining jitter.

[0004] However, this scheme has significant defects: first, its core focuses on a one-way suppression link of "signal-to-noise ratio-mechanical jitter", and fails to effectively integrate direct perception and analysis of environmental electromagnetic interference data, and to model the mechanism of how electromagnetic interference dynamically affects the control system and induces complex mechanical vibration, resulting in no solution to non-periodic, burst electromagnetic pulse interference and abnormal mechanical response caused thereby. Second, its closed-loop mechanism is incomplete, mainly relying on internal sensors for parameter adjustment, lacking closed-loop verification and utilization of feedback of the image stream itself quality and electromagnetic environment, and having limited adaptive ability. Most importantly, it lacks an effective "cross-domain collaborative" mechanism, and cannot deeply integrate and jointly optimize electromagnetic interference feature analysis and mechanical jitter suppression, image stream quality evaluation, and is difficult to cope with the challenge of "electromagnetic-mechanical" strong coupling interference in complex electromagnetic environment, and the image stabilization is still prone to degradation in severe electromagnetic environment. SUMMARY

[0005] The present application provides an intelligent image stabilization method and system for unmanned intelligent rotating platform, to solve the problem of insufficient image stabilization in the prior art due to the difficulty in effectively collaboratively suppressing "electromagnetic-mechanical" coupling interference in a strong electromagnetic interference environment.

[0006] In a first aspect, the present application provides an intelligent image stabilization method for an unmanned intelligent turntable, comprising: Obtaining electromagnetic interference data in the turntable operating environment, updating the interference feature library of the turntable based on the electromagnetic interference data, monitoring the signal-to-noise ratio feature of the power assembly of the turntable to generate a trap wave filtering parameter; According to the trap wave filtering parameter, adjusting the trap wave depth and bandwidth related to the rotation jitter of the turntable, synchronously triggering the nonlinear compensation and frequency domain adaptive equalization processing of the turntable, identifying the mutation signal of the rotation key frame in the turntable image stream and adjusting the compensation strength combined with the signal-to-noise ratio feature, forming an interference suppression closed-loop mechanism; During the operation of the interference suppression closed-loop mechanism, the high-frequency jitter feature in the mutation signal is extracted through inner loop control, and the high-frequency jitter feature and the turntable interference prediction data are fused to generate a primary suppression parameter through outer loop prediction, and the primary suppression parameter and the electromagnetic interference data are analyzed to generate a cross-domain suppression instruction; Based on the cross-domain suppression instruction, the turntable anti-interference encoding is executed, the rotation suppression term in the cross-domain suppression instruction is used to configure the turntable posture related encoding scheme, the electromagnetic suppression term in the cross-domain suppression instruction is used to set the rotation scene error correction protocol, and the encoding data is verified through closed loop combined with the turntable link feedback; The turntable environment adaptation parameters output by the closed loop verification are fed back to the interference feature library update and the compensation strength adjustment process, realizing the cooperative control of electromagnetic interference and mechanical jitter, so as to output stable image stream.

[0007] Optionally, the turntable environment adaptation parameters output by the closed loop verification are fed back to the interference feature library update and the compensation strength adjustment process, realizing the cooperative control of electromagnetic interference and mechanical jitter, so as to output stable image stream, comprising: The position deviation report output by the closed loop verification is converted into turntable environment adaptation parameters including offset angle, displacement and environmental interference level; The offset angle and displacement in the environmental adaptation parameters are written into the interference feature library as new records to update the interference feature library, and the displacement in the environmental adaptation parameters is input into the compensation adjuster to output the correction value of the compensation strength through the displacement strength mapping table; Through the continuous update of the interference feature library and the dynamic adjustment of the compensation strength, a double-channel cooperative mechanism is formed, so that the electromagnetic interference suppression subsystem and the mechanical jitter suppression subsystem realize the control law linkage based on the environmental adaptation parameters, and output the stable video stream eliminating the frame mutation to the image processing terminal.

[0008] Optionally, the continuous updating of the interference feature library and the dynamic adjustment of the compensation strength form a dual-channel collaborative mechanism, so that the electromagnetic interference suppression subsystem and the mechanical jitter suppression subsystem realize control law linkage based on the environmental adaptation parameters, and output a stable video stream with no inter-frame mutations to the image processing terminal, including: The continuously updated interference signature library and the correction value of the compensation strength are input into the cooperative controller to activate the dual-channel cooperative mechanism, and the environmental adaptation parameters are simultaneously input into the electromagnetic interference suppression subsystem and the mechanical jitter suppression subsystem; In the electromagnetic interference suppression subsystem, a filter frequency lower limit value is calculated according to the environmental interference level, and the filter frequency lower limit value is converted into a harmonic monitoring range of the mechanical jitter suppression subsystem; In the mechanical jitter suppression subsystem, a compensation strength reference value is set according to the displacement, and the compensation strength reference value is mapped to an amplitude attenuation threshold of the electromagnetic interference suppression subsystem; An image frame processing instruction is generated based on the interactive output of the electromagnetic interference suppression subsystem and the mechanical jitter suppression subsystem, and when the offset of consecutive fixed frames is lower than a set threshold, a stable video stream is output to the terminal device.

[0009] Optionally, performing turntable anti-interference coding based on the cross-domain suppression instruction, configuring a turntable attitude-related coding scheme according to a rotation suppression item in the cross-domain suppression instruction, hierarchically setting a rotation scene error correction protocol according to the electromagnetic suppression item in the cross-domain suppression instruction, and performing closed-loop verification on the coded data in combination with turntable link feedback, includes: Parsing the rotation suppression item in the cross-domain suppression instruction, and generating a turntable attitude-related coding scheme including compensation displacement amounts for each tilt angle interval according to a mapping table of jitter suppression coefficients and tilt angles included in the rotation suppression item; parsing the electromagnetic suppression item in the cross-domain suppression instruction, dividing the rotation scene according to the interference intensity level, and configuring an error correction protocol including a data check bit length and an error correction threshold for each level of the scene; Integrate the turntable attitude-related coding scheme with the error correction protocol to construct an anti-interference coding data packet and feed it back to the turntable actuator through the turntable link; Receive the position coordinates fed back by the turntable actuator, calculate the inter-frame deviation value between the position coordinates and the expected position coordinates in the anti-interference coded data packet, and when the deviation value of multiple consecutive frames exceeds the set tolerance, generate a closed-loop verification signal to trigger the reconstruction of the anti-interference coded data packet and output a position deviation report.

[0010] Optionally, during the operation of the interference suppression closed-loop mechanism, high-frequency jitter features in the mutation signal are extracted through inner-loop control, the high-frequency jitter features are fused with turntable interference prediction data using outer-loop prediction to generate primary suppression parameters, and motion state analysis is performed on the primary suppression parameters and electromagnetic interference data to generate cross-domain suppression instructions, including: During the operation of the interference suppression closed-loop mechanism, a high-frequency jitter feature with a change rate exceeding a preset threshold is extracted from the mutation signal by an inner-loop control module; Performing spatiotemporal fusion processing on the high-frequency jitter feature and the turntable interference prediction data stored in the outer loop prediction module to generate primary suppression parameters including a frequency suppression coefficient and an amplitude correction coefficient; The primary suppression parameters and electromagnetic interference data are subjected to motion trajectory analysis, and a cross-domain suppression instruction including a rotation suppression item and an electromagnetic suppression item is generated by comparing the fluctuation forms of the primary suppression parameters and the electromagnetic interference data within the same time window.

[0011] Optionally, the notch depth and bandwidth related to turntable rotation jitter are adjusted according to the notch filter parameters, turntable nonlinear compensation and frequency domain adaptive equalization processing are synchronously triggered, sudden changes in rotation keyframes in the turntable image stream are identified, and the compensation strength is adjusted in combination with the signal-to-noise ratio characteristics, thereby forming an interference suppression closed-loop mechanism, including: Inputting the notch filter parameters into an adjustable filter, blocking the main frequency of the rotational jitter according to the center frequency value, attenuating the amplitude of the jitter harmonics according to the depth setting value, and covering the rotational jitter spectrum range according to the width setting value; Synchronously trigger the nonlinear compensation module to correct the image offset caused by rotation jitter, and trigger the frequency domain adaptive equalization module to suppress the interference energy in the rotation-related frequency band; Scanning and identifying the turntable image stream, extracting a mutation signal in the rotation key frame, and inputting the intensity vector of the mutation signal and the signal-to-noise ratio feature into a compensation regulator to adjust the compensation intensity; The compensation intensity is output to the nonlinear compensation module and a feedback signal is generated to the adjustable filter, forming an interference suppression closed-loop mechanism from image recognition to parameter adjustment.

[0012] Optionally, acquiring electromagnetic interference data in the turntable operating environment, updating an interference feature library of the turntable based on the electromagnetic interference data, and monitoring a signal-to-noise ratio feature of a turntable power assembly to generate notch filter parameters includes: The electromagnetic sensor on the unmanned intelligent turntable collects the original electromagnetic field strength data in the turntable's operating environment and outputs electromagnetic interference data including peak intensity and fluctuation frequency; The electromagnetic interference data is compared with interference patterns stored in an interference feature library, a new entry is created for an unrecorded interference pattern and added to the interference feature library, and an updated interference feature library is output; Meanwhile, an output signal of the turret power assembly is monitored to generate a signal-to-noise ratio feature containing an effective signal ratio, and when the signal-to-noise ratio feature is lower than a threshold value, based on the updated interference feature library, a notch filtering parameter containing a center frequency value, a depth setting value and a width setting value is generated.

[0013] In a second aspect, the present application provides an intelligent image stabilization system of an unmanned intelligent turret, comprising: An acquisition module is configured to acquire electromagnetic interference data in a turret operating environment, update an interference feature library of the turret based on the electromagnetic interference data, and monitor a signal-to-noise ratio feature of a turret power assembly to generate a notch filtering parameter; A processing module is configured to adjust a notch depth and bandwidth related to turret rotation jitter according to the notch filtering parameter, synchronously trigger nonlinear compensation and frequency domain adaptive equalization processing of the turret, identify a mutation signal of a rotation key frame in a turret image stream and adjust compensation strength in combination with the signal-to-noise ratio feature, and form an interference suppression closed-loop mechanism; A generation module is configured to extract a high-frequency jitter feature in the mutation signal through inner loop control during operation of the interference suppression closed-loop mechanism, generate a primary suppression parameter by fusing the high-frequency jitter feature and turret interference prediction data using outer loop prediction, and generate a cross-domain suppression instruction by performing motion state analysis on the primary suppression parameter and electromagnetic interference data; An execution module is configured to perform turret anti-interference encoding based on the cross-domain suppression instruction, configure a turret attitude related encoding scheme according to a rotation suppression term in the cross-domain suppression instruction, hierarchically set a rotation scene error correction protocol according to an electromagnetic suppression term in the cross-domain suppression instruction, and perform closed-loop verification on encoded data in combination with turret link feedback; An output module is configured to feed back turret environment adaptation parameters output by the closed-loop verification to the interference feature library update and the compensation strength adjustment process, realize collaborative control of electromagnetic interference and mechanical jitter, and output a stable image stream.

[0014] In a third aspect, the present application provides a computing device comprising a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement an intelligent image stabilization method of an unmanned intelligent turret as described in the first aspect above.

[0015] In a fourth aspect, the present application provides a computer storage medium storing a computer program, which, when executed by a computer, implements an intelligent image stabilization method of an unmanned intelligent turret as described in the first aspect.

[0016] In the present application, electromagnetic interference data in the running environment of the turntable is acquired, the interference feature library of the turntable is updated based on the electromagnetic interference data, the signal-to-noise ratio feature of the power assembly of the turntable is monitored to generate a notch filtering parameter; according to the notch filtering parameter, the notch depth and bandwidth related to the rotation jitter of the turntable are adjusted, the nonlinear compensation and frequency domain adaptive equalization processing of the turntable are triggered synchronously, the mutation signal of the rotation key frame in the image stream of the turntable is identified and the compensation strength is adjusted in combination with the signal-to-noise ratio feature, forming an interference suppression closed loop mechanism; during the operation of the interference suppression closed loop mechanism, the high-frequency jitter feature in the mutation signal is extracted through inner loop control, the high-frequency jitter feature and the turntable interference prediction data are fused by outer loop prediction to generate primary suppression parameters, and the motion state analysis is performed on the primary suppression parameters and the electromagnetic interference data to generate cross-domain suppression instructions; based on the cross-domain suppression instructions, the anti-interference encoding of the turntable is performed, the turntable posture related encoding scheme is configured according to the rotation suppression term in the cross-domain suppression instructions, the rotation scene error correction protocol is set hierarchically according to the electromagnetic suppression term in the cross-domain suppression instructions, and the closed loop verification is performed on the encoded data in combination with the turntable link feedback; the turntable environment adaptation parameters output by the closed loop verification are fed back to the interference feature library update and the compensation strength adjustment process, realizing the cooperative control of electromagnetic interference and mechanical jitter to output stable image stream.

[0017] The technical scheme of the present application has the following beneficial effects: The present application updates the interference feature library by acquiring electromagnetic interference data and generates a notch filtering parameter by monitoring the signal-to-noise ratio of the power assembly, dynamically adjusts the depth and bandwidth of the notch filter, synchronously triggers nonlinear compensation and frequency domain adaptive equalization processing, adjusts the compensation strength in combination with the key frame mutation signal and the signal-to-noise ratio feature of the image stream to form a closed loop suppression mechanism; under this mechanism, the inner loop extracts the high-frequency jitter feature and the outer loop fuses the interference prediction data to generate primary suppression parameters, and outputs cross-domain suppression instructions through motion state analysis; based on the instructions, anti-interference encoding is performed, and closed loop verification is realized in combination with the link feedback; finally, the environment adaptation parameters are fed back to the interference library update and the compensation adjustment process, realizing the dynamic cooperative suppression of electromagnetic interference and mechanical jitter, and ensuring the stable output of the turntable image stream in a complex electromagnetic environment.

[0018] The position deviation report is further converted into environmental adaptation parameters containing offset angle, displacement, and environmental interference level. Subsequently, the offset angle and displacement are used as new records to update the interference feature library. The displacement is also input into the compensation regulator, and a correction value for the compensation strength is generated based on the displacement strength mapping table. Finally, a dual-channel collaborative mechanism is formed through the continuously updated interference feature library and the dynamically adjusted compensation strength. This enables the electromagnetic interference suppression and mechanical jitter suppression subsystems to implement control law linkage based on the environmental adaptation parameters, outputting a stable video stream that eliminates inter-frame mutations. This solution achieves refined utilization of environmental adaptation parameters and dual-channel feedback collaboration: by precisely decomposing the position deviation into offset angle, displacement, and interference level, and using them to dynamically optimize the interference feature library and correct the compensation strength, a strong linkage mechanism is established between the electromagnetic suppression and mechanical suppression subsystems, significantly enhancing the system's dynamic adaptability to complex environments and ultimately ensuring the output of a high-quality, stable video stream without inter-frame mutations.

[0019] These and other aspects of the present application will become more readily apparent from the description of the following embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0021] Figure 1 A flow chart of an intelligent image stabilization method for an unmanned intelligent turntable provided by the present application is shown; Figure 2 A scene diagram showing an intelligent image stabilization method for an unmanned intelligent turntable provided by the present application is shown; Figure 3 A schematic structural diagram of an intelligent image stabilization system for an unmanned intelligent turntable provided by the present application is shown; Figure 4 A schematic structural diagram of a computing device provided by the present application is shown. DETAILED DESCRIPTION

[0022] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0023] In some of the flowcharts described in the description and claims of the present application and in the above description of the figures, a plurality of operations are included which occur in a particular order, but it should be clearly understood that the operations can be performed in an order other than that in which they appear or in parallel, and the serial numbers of the operations such as 101, 102, etc. are only used to distinguish different operations and the serial numbers themselves do not represent any execution order. In addition, the flowcharts can include more or fewer operations, and the operations can be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this paper are used to distinguish different messages, devices, modules, etc. and do not represent the order and do not limit the types of "first" and "second".

[0024] Research shows that in the image stabilization technology of unmanned intelligent turntable in complex electromagnetic environment, the existing adaptive notch filter and motion compensation scheme has the following key defects: excessive dependence on single-dimensional signal-to-noise ratio perception, difficulty in dealing with sudden electromagnetic interference and abnormal mechanical response induced thereby; closed loop is split and feedback dimension is single, lacking of closed loop utilization of image stream mutation signal and dynamic electromagnetic environment; most importantly, lacking of "electromagnetic-mechanical" cross-domain collaborative mechanism, leading to easy degradation of image stability in harsh electromagnetic environment. These defects of the existing technology are essentially due to the lack of single-dimensional perception, split closed loop and cross-domain collaboration, and an intelligent stabilization method capable of dynamically fusing multi-source information, realizing closed loop linkage and having cross-domain collaboration capability is urgently needed.

[0025] In view of the above problems, the present application proposes an intelligent image stabilization method for unmanned intelligent turntable, the core of which is to generate notch parameters by fusing electromagnetic interference data and signal-to-noise ratio, and to form a preliminary closed loop in combination with image key frame mutation signals; to innovatively build an "inner loop high frequency jitter extraction + outer loop interference prediction fusion" architecture, output cross-domain suppression instructions; and to realize double-domain continuous collaborative optimization through link closed loop verification and parameter feedback. The method integrates multi-source perception in a complete closed loop, solves the problem of lack of collaboration through cross-domain instructions, effectively blocks the conduction of electromagnetic interference and suppresses the jitter induced thereby, and finally outputs stable non-mutation image stream in complex electromagnetic environment.

[0026] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of the present application.

[0027] Figure 1 A flowchart of an intelligent image stabilization method for unmanned intelligent turntable is provided for the embodiments of the present application, as shown in Figure 1 The method comprises: 101. Obtain electromagnetic interference data in the running environment of the turntable, update the interference feature library of the turntable based on the electromagnetic interference data, monitor the signal-to-noise ratio feature of the power assembly of the turntable to generate a notch filtering parameter; Optionally, step 101 can specifically include the following steps: 1011. Collect original electromagnetic field intensity data in the running environment of the turntable through electromagnetic sensors on the unmanned intelligent turntable, and output electromagnetic interference data containing peak intensity and fluctuation frequency; 1012. Compare the electromagnetic interference data with the interference patterns stored in the interference feature library, create a new entry for unrecorded interference patterns and add them to the interference feature library, and output the updated interference feature library; 1013. Monitor the output signal of the power assembly of the turntable at the same time, generate a signal-to-noise ratio feature containing an effective signal ratio, and when the signal-to-noise ratio feature is lower than a threshold value, generate a notch filtering parameter containing a center frequency value, a depth setting value and a width setting value based on the updated interference feature library.

[0028] In the above scheme, the electromagnetic interference data refers to dynamic signals reflecting the electromagnetic disturbance characteristics of the running environment of the turntable, containing the transient change characteristics of the electromagnetic field intensity, including peak intensity and fluctuation frequency information. The interference feature library refers to a dynamic database that stores the historical evolution characteristics of electromagnetic interference patterns, which can be used to match new interference patterns and generate suppression decision basis. The notch filtering parameter refers to the core configuration set for regulating the rotation jitter suppression effect of the turntable, containing the response characteristics of the filter to the target frequency interference, including the center frequency value, the depth setting value, the width setting value information, which can be used to accurately filter out electromagnetic conduction type mechanical jitter of specific frequency. The signal-to-noise ratio feature refers to a dynamic index that quantifies the degradation of the signal quality of the turntable power assembly, containing the power comparison characteristics of the effective signal and the noise signal, including the effective signal proportion and the noise pollution degree information, which can be used to trigger the adaptive filtering mechanism and evaluate the harm level of electromagnetic interference to mechanical control.

[0029] In the embodiments of the present application, first, a broadband electromagnetic sensor array is deployed in the circumferential direction of the turntable through step 1011 to capture the original electromagnetic field intensity data of the environment at a sampling rate of 10 kHz; after 24-bit ADC analog-digital conversion, the fast Fourier transform algorithm is used to perform frequency domain analysis on the time domain signal to extract the spectral peak value of the signal in the 0-1 kHz frequency band, identify the main interference frequency and its harmonic components, and at the same time, the maximum field intensity instantaneous value in each time window is calculated through the sliding window peak value detection algorithm, and finally the structured electromagnetic interference data containing the peak value intensity and the fluctuation frequency set is output. For example, when working near a high-voltage transmission line, the sensor detects a field intensity of 156 V / m, the fast Fourier transform analysis shows that the spectral energy is concentrated at the fundamental frequency of 100 Hz and the second harmonic of 200 Hz, and the output electromagnetic interference data is the peak value intensity of 156 V / m and the fluctuation frequency set [100 Hz, 200 Hz].

[0030] Then, based on the electromagnetic interference data output by step 1011, the dynamic time warping algorithm is called to perform interference pattern matching through step 1022: the peak value intensity distribution histogram and the frequency energy weight vector of the new data are extracted, and similarity calculation is performed with the multi-dimensional features of all historical patterns in the interference feature library; if the highest similarity is lower than the preset threshold, a new entry creation process is started, the intensity interval is generated according to the current peak value intensity fluctuation range, the fundamental frequency and harmonic components in the frequency set are analyzed to generate the frequency feature label, and finally the new pattern entry containing the intensity interval, the frequency label and the initial suppression strategy is written into the interference feature library and the update result is output. For example, when the input peak value is 182 V / m and the frequency is [250 Hz, 500 Hz], the entry "switch arc interference" is created after the matching fails, the corresponding intensity interval is 180-185 V / m, the frequency label is [250 Hz±3 Hz, 500 Hz±5 Hz], and the autonomous evolution of the interference feature library is realized.

[0031] Finally, step 1013 collects three-phase current signals through the current feedback loop of the turntable motor driver and performs power spectral density analysis using a 200 ms sliding window: the effective signal band and the noise band are divided, the signal-to-noise ratio feature is generated by calculating the integral power ratio of the two bands, and the calculation formula is as follows: , wherein is the energy size of the current signal when the motor is working normally, is the current noise energy size caused by electromagnetic interference; when the signal-to-noise ratio feature SNR is lower than the preset threshold, the parameter generation module is activated, the current main interference frequency is extracted from the interference feature library updated by step 1012 as the center frequency value, the depth setting value is calculated according to the peak value intensity of the electromagnetic interference data by linear mapping formula, and the calculation formula is as follows: , the width value is automatically set according to the main frequency fluctuation range, and the setting time refers to For example, in the substation inspection, the motor signal-to-noise ratio feature SNR is detected to be less than the threshold 0.1, and the feature library contains the "switching arc interference" item, then the output notch filter parameters are: center frequency = 250Hz, depth = -0.25x182≈-45.5dB, width = 250Hzx10%=±25Hz.

[0032] In practical application, when a certain type of unmanned aerial vehicle intelligent turntable performs a patrol task in a strong electromagnetic environment of a substation, its electromagnetic sensor captures the original electromagnetic field signal of the environment, identifies that the spectral energy is concentrated at 250Hz and 500Hz, and the proportion of each is 70% and 30% respectively through fast Fourier transform analysis, and obtains the instantaneous maximum field strength 182V / m through peak detection, forming structured electromagnetic interference data; the system immediately compares the intensity distribution of the data with the single peak 182V / m and the frequency weight feature with the historical mode of the interference feature library, and the dynamic time warping algorithm calculation shows that the highest similarity is only 78%, which is lower than the threshold of 85%, and it is determined as unknown interference, and accordingly a new item named "switching arc interference" is created: the intensity interval is 173-191V / m, and the frequency label is 250Hz±3Hz and 500Hz±5Hz, which are written into the feature library; at the same time, the motor current signal monitoring window of the turntable detects that the effective signal band integral power is 120 units, the noise band integral power is 240 units, and the signal-to-noise ratio , which is much lower than the threshold 1.0, triggering the generation of the notch parameters, extracting the main frequency 250Hz from the new item as the center frequency, calculating the depth value according to the formula -0.25x182V / m to get -45.5dB, and taking the main frequency 10% as the width value, i.e. 250x0.1=±25Hz, and finally outputting the parameter group (250Hz, -45.5dB, ±25Hz) to the subsequent suppression module.

[0033] The overall scheme of the above 101, through the cooperation of electromagnetic perception and power signal monitoring, builds a dynamic anti-interference foundation: the electromagnetic sensor accurately captures the peak intensity and frequency characteristics of environmental interference, drives the interference feature library to evolve autonomously to identify unknown interference patterns; monitor the motor signal-to-noise ratio, and generate targeted notch parameters based on the updated feature library when the signal is degraded. Realize the closed-loop linkage of environmental interference cognition and suppression strategy, provide accurate input for subsequent cross-domain collaborative suppression, and block the conduction path from electromagnetic interference to mechanical jitter from the source.

[0034] 102、According to the notch filter parameters, adjust the notch depth and bandwidth related to the rotation jitter of the turntable, synchronously trigger the nonlinear compensation and frequency domain adaptive equalization processing of the turntable, identify the mutation signal of the rotation key frame in the image stream of the turntable and adjust the compensation strength combined with the signal-to-noise ratio feature, form an interference suppression closed-loop mechanism; Optionally, step 102 can specifically include the following steps: 1021. Input the notch filter parameters into the adjustable filter, block the main frequency of the rotation jitter according to the center frequency value, attenuate the amplitude of the jitter harmonics according to the depth setting value, and cover the rotation jitter spectrum range according to the width setting value; 1022. Synchronously trigger the nonlinear compensation module to correct image offset caused by rotation jitter, and trigger the frequency domain adaptive equalization module to suppress interference energy in the rotation-related frequency band; 1023. Scan and identify the turntable image stream, extract a mutation signal from the rotation key frame, and input the intensity vector of the mutation signal and the signal-to-noise ratio feature into a compensation regulator to adjust the compensation intensity; 1024. Output the compensation intensity to the nonlinear compensation module and generate a feedback signal to the adjustable filter, thereby forming an interference suppression closed-loop mechanism from image recognition to parameter adjustment.

[0035] In the above scheme, the tunable filter refers to a digital filter device capable of adjusting filtering characteristics. The notch filter parameters it receives include a center frequency value for blocking the turntable's rotational jitter (i.e., the fundamental frequency component of mechanical vibration), a depth setting for controlling the harmonic amplitude (i.e., the attenuation strength of integer multiples of the main frequency), and a width setting for determining the required rotational jitter spectrum bandwidth. The nonlinear compensation module is a pixel offset correction algorithm based on a kinematic model that inversely compensates for image deformation by establishing a position offset mapping relationship. The frequency domain adaptive equalization module is a frequency band energy suppression unit that utilizes fast Fourier transform technology. It suppresses interference by identifying rotation-related interference bands and dynamically reducing their gain. The intensity vector is a two-dimensional data set that describes the characteristics of the sudden change signal, encompassing two dimensions: pixel displacement and sudden change rate. The compensation regulator is a decision-making unit with a built-in intensity mapping table. It receives the intensity vector and signal-to-noise ratio feature data and outputs a corresponding compensation intensity level. The feedback signal is a set of control instructions generated by image quality analysis, including bandwidth adjustment recommendations and depth optimization instructions for the tunable filter, enabling dynamic iterative updates of the filter parameters.

[0036] In the embodiment of the present application, first, the notch filter parameters are input into an adjustable digital filter through step 1021. The filter adopts an IIR notch design algorithm for parameter configuration, and sets a zero point in the transfer function based on the center frequency value so that the frequency signal is completely blocked; the quality factor Q value of the filter is dynamically adjusted according to the depth setting value to control the attenuation intensity of the harmonic component and achieve the desired weakening of high-order vibration energy; finally, the stopband range is expanded according to the width setting value such as ±2Hz, and a suppression band centered on the main frequency is formed by adjusting the pole position to ensure that the jitter spectrum range during speed fluctuations is covered. For example, when the turntable detects that the main frequency drifts from 12Hz to 12.5Hz during the acceleration phase, the filter automatically adjusts the suppression band to 10.5Hz to 14.5Hz to continuously suppress the target spectrum.

[0037] Then, after the notch filtering parameter takes effect, the dual-module processing is triggered synchronously: the nonlinear compensation module calls the pixel displacement mapping algorithm based on kinematic modeling, constructs a quadratic surface offset model by analyzing the historical image sequence, and realizes geometric correction by reversing the displacement of each pixel coordinate in the current frame according to the model, for example, if the image is detected to move right by 3 pixels, a coordinate transformation of moving left by 3 pixels is applied to all pixels; at the same time, the frequency domain adaptive equalization module starts the block fast Fourier transform processing, divides the image into 8x8 pixel blocks and performs fast Fourier transform, identifies the characteristic frequency band related to strong rotation jitter, dynamically calculates the suppression coefficient k of the frequency band, and the calculation formula is as follows: , wherein is the signal-to-noise ratio weight factor, and are the energy of the interference frequency band and the total energy respectively, and finally the target frequency band spectrum is weighted and corrected to realize the cooperative suppression of spatial and frequency domain interference. For example, if the characteristic frequency band is identified as 10-15 Hz, when the interference ratio is 30%, k=0.76 is calculated, i.e. the amplitude of the 10-15 Hz frequency band is compressed to 76% of the original value.

[0038] Then, after receiving the image stream processed by step 1022, step 1023 is performed to scan the rotation key frame by frame through the inter-frame difference algorithm, and the adaptive threshold segmentation is used to lock the mutation area; then the ORB feature point detection is used to extract stable feature points in the key area, and the sparse optical flow method is used to track the feature point displacement vector between adjacent key frames, and the average displacement and the mutation rate are calculated to form a two-dimensional vector describing the mutation intensity; finally, the vector and the signal-to-noise ratio feature SNR are input into a three-dimensional lookup table decision maker, and the lookup table is established through experimental calibration, and the corresponding compensation intensity level value is output to drive the subsequent module. For example, in the high-speed turning scene of the unmanned aerial vehicle, if the 90° key frame is detected to have a 4.5-pixel displacement and the SNR drops to 14 dB, the intensity level 3 is output after lookup table matching, triggering the strong compensation mode to deal with severe jitter.

[0039] ​Finally, based on the compensation intensity level value output by step 1023, the level value is input into the parameter mapping unit of the nonlinear compensation module, the scaling ratio of the pixel displacement model is adjusted through the preset intensity gain relationship curve, the geometric correction intensity is enhanced, and more aggressive spatial morphological compensation is realized; at the same time, the mutation signal spectrum analysis engine is started, 512-point fast Fourier transform is performed on the mutation signal extracted by step 1023, the concentrated frequency band of high-frequency jitter energy is identified, and a bandwidth adjustment instruction is generated accordingly: if the peak frequency point deviates from the current notch center frequency by less than a threshold value, a narrow bandwidth instruction is output; if the deviation exceeds the threshold value or there are multiple peaks, a wide bandwidth instruction is output, which is fed back to the adjustable filter of step 1021 to complete parameter iteration; finally, through double-channel closed-loop verification, dynamic optimization of the cross-domain suppression strategy is ensured.

[0040] In actual application, when the intelligent optoelectronic turntable carried by a certain border patrol unmanned aerial vehicle performs a monitoring task in a strong electromagnetic interference area, the turntable servo system detects abnormal jitter caused by wind speed disturbance and electromagnetic pulse coupling, calculates the notch filter parameter to obtain a main frequency center value of 11.5 Hz, a depth of -28 dB, and a width of ±1.5 Hz, generates a stop band of 10-13 Hz through the adjustable filter, completely suppresses the 11.5 Hz fundamental frequency, and weakens the 23 Hz harmonic energy by 28 dB; then the dual modules are activated synchronously, the nonlinear compensation module fits the pitch axis offset model based on gyroscope data, performs reverse displacement compensation on the image frame, the actual measured pitch jitter is 3.2 pixels, the residual after compensation is 0.4 pixels, and at the same time the frequency domain equalization module analyzes the image spectrum and finds that the energy in the 8-16 Hz frequency band is abnormal, wherein , the full frequency band =22, and the weight factor table is looked up to obtain α=0.85 in combination with the signal-to-noise ratio SNR=16 dB, and the suppression coefficient is calculated, which compresses the energy in the frequency band by 21.7%; then the optical flow tracking is triggered by rotating the turntable horizontally by 90°, the average displacement of the feature point group is detected pixels, and the time window , the mutation rate is about 34.2 px / s, the intensity vector [4.1, 34.2] is generated, and the three-dimensional decision table is input with SNR=16 dB, the displacement amount is matched to 4-6 pixels, the rate is matched to 30-40 px / s, and the signal-to-noise ratio is matched to 15-18 dB, the compensation intensity level 3 is output, the compensation coefficient gain is mapped to 1.5 times, the main peak 11.7 Hz of the mutation signal spectrum is analyzed, the deviation from the current notch center 11.5 Hz is 0.2 Hz, which is less than the threshold value 0.5 Hz, the bandwidth narrowing instruction is generated, the original stop band 10-13 Hz is updated to 11.5-11.9 Hz, and after closed-loop adjustment, the inter-frame displacement of the image frame is reduced to 0.9 pixels, meeting the stability requirement of reconnaissance imaging.

[0041] The overall scheme of 102 precisely suppresses the rotating jitter core spectrum through an adjustable filter, and realizes spatial offset correction and frequency domain interference suppression through a double-module synchronous processing; the compensation strength is dynamically adjusted using the image stream mutation signal and the electrical signal characteristics, a closed-loop control chain from image quality feedback to filter parameter optimization is constructed, and the image stability and anti-interference robustness of the turntable in a complex motion environment are significantly improved.

[0042] 103、During the operation of the interference suppression closed-loop mechanism, the high-frequency jitter characteristics in the mutation signal are extracted through inner loop control, and the high-frequency jitter characteristics and turntable interference prediction data are fused to generate primary suppression parameters through outer loop prediction, and the primary suppression parameters and electromagnetic interference data are analyzed for motion state to generate cross-domain suppression instructions; Optionally, step 103 can specifically include the following steps: 1031、During the operation of the interference suppression closed-loop mechanism, the high-frequency jitter characteristics with a change rate exceeding a preset threshold are extracted from the mutation signal through the inner loop control module; 1032、The high-frequency jitter characteristics and the turntable interference prediction data stored in the outer loop prediction module are spatio-temporally fused to generate primary suppression parameters including frequency suppression coefficients and amplitude correction coefficients; 1033、The primary suppression parameters and the electromagnetic interference data are analyzed for motion trajectory, and the cross-domain suppression instructions including rotation suppression terms and electromagnetic suppression terms are generated by comparing the fluctuation patterns of the primary suppression parameters and the electromagnetic interference data in the same time window.

[0043] In the above scheme, the high-frequency jitter characteristics refer to the rapid abnormal fluctuations identified from the image stream mutation signal, with a change rate exceeding a preset threshold, which is usually caused by electromagnetic interference or mechanical vibration. The turntable interference prediction data is the future interference information pre-calculated by the system based on historical interference patterns, including the possible interference types, intensity and duration. The primary suppression parameters refer to the intermediate control parameters generated after fusing the high-frequency jitter characteristics and the interference prediction data, which are specifically composed of frequency suppression coefficients and amplitude correction coefficients. The cross-domain suppression instructions are the final generated cooperative control commands, including rotation suppression terms for adjusting the attitude of the turntable and hierarchical error correction protocols for responding to electromagnetic interference.

[0044] In the embodiment of the present application, firstly, step 1031 calculates the displacement of feature points in adjacent frames of the image stream by the inner loop control module, adopts the optical flow method to track the preset feature points, calculates the coordinate difference value in the continuous two frames of pictures and divides by the frame interval time to obtain the displacement rate. Then, dynamic threshold filtering is performed, and the displacement rate is compared with the preset threshold value. When it is detected that the rate of a certain feature point exceeds the threshold value for 3 frames, it is marked as a high-frequency jitter feature and the coordinates and rate value are output. For example, in a power inspection scene, the system identifies the displacement of the corner point of the insulator string of the electric tower, the t-th frame coordinate is (120, 80), the t+1-th frame coordinate is (158, 83), the frame interval is 40 ms, and the calculated displacement rate is , which is greater than the threshold value 25, and is marked as a high-frequency jitter feature F3.

[0045] Then, the high-frequency jitter feature output by step 1031 is received by step 1032, and the time stamp synchronization mechanism is used to perform space-time alignment with the interference prediction data in the outer loop prediction module. The Kalman filtering algorithm is adopted to compensate for the sensor transmission delay, so that both are in the same time reference. Subsequently, double-source fusion calculation is performed: the high-frequency jitter feature is analyzed by fast Fourier transform, the energy weight of the key frequency band is extracted, and the weight is multiplied by the frequency correlation coefficient in the predicted interference data to generate a frequency suppression coefficient; at the same time, the actual amplitude peak value of the jitter feature is compared with the maximum allowed interference amplitude in the prediction data, and the amplitude correction coefficient is obtained by the formula . Finally, the frequency suppression coefficient and the amplitude correction coefficient are packaged as structured primary suppression parameters and output. For example, when the corner point displacement rate of 42.3 pixels / ms is detected in the power inspection scene, the system aligns the prediction data “L3 level power frequency interference”, and the energy of the jitter at 50 Hz is 80% and the historical correlation is 90% after FFT analysis. The frequency coefficient is 0.72 obtained by multiplication calculation; based on the measured jitter amplitude of 60 units and the amplitude coefficient of 0.4, the primary parameters (0.72, 0.4) are output for subsequent use.

[0046] Finally, the dynamic time warping algorithm is used to match the waveform form of the primary suppression parameters and the electromagnetic interference data stream, convert the frequency coefficient in the primary suppression parameters into an analog frequency fluctuation curve, and perform point-by-point similarity calculation with the measured electromagnetic interference data in a 20 ms time window. When the similarity exceeds the preset threshold value 0.8, the system starts the cross-domain instruction generation: based on the spatial direction information of the high-frequency jitter feature, the frequency suppression coefficient is converted into a rotation suppression term, and the proportion mapping formula is Generate precise control instructions; simultaneously, trigger the corresponding electromagnetic suppression protocol based on the intensity level of the interference prediction data. Matching is performed according to pre-set rules: L1 enables forward error correction, L2 enables frequency domain spread spectrum, and L3 activates a combined "data retransmission + Hamming code" anti-interference mechanism. Ultimately, these instructions are packaged as complete cross-domain suppression instructions and output to the execution module. For example, when the primary input parameters (frequency coefficient 0.72, amplitude coefficient 0.4) match the measured 50Hz power frequency interference data, the DTW matching results in a similarity of 0.88, exceeding the threshold of 0.8. Combined with the corner right shift feature, the rotation suppression item "YAW axis torque reduction by 12%" is generated. Simultaneously, the electromagnetic suppression item "transmitting redundant data twice per frame with additional Hamming code error correction" is activated based on the L3 interference level.

[0047] In actual applications, when the intelligent turntable of a sea area inspection drone performs a target tracking task in a strong electromagnetic interference environment, the optical flow method is used to analyze the image stream and detect that the buoy feature point moves from the coordinates of the t frame (150,200) to the t+1 frame (192,205). The frame interval is 40 milliseconds, which is approximately equal to 42.5 pixels / millisecond, which is greater than the preset threshold of 25. It is determined to be a high-frequency jitter feature F4 and its direction and amplitude value of 60 units are output; then F4 is aligned with the "next 50 milliseconds L3 radar band interference" predicted by the outer loop. After FFT analysis, F4 accounts for 75% of the energy in the 80Hz frequency band. Combined with the historical database, the correlation coefficient between the interference in this frequency band and the jitter is 85%, and the high-frequency jitter feature F4 is calculated. , and calculate , generate primary parameters (0.64, 0.4); then convert the primary parameters into 80Hz reference waveform, and compare them with the measured electromagnetic interference data through DTW algorithm. The output similarity is 0.86, which is greater than the threshold value 0.8. Combined with the X-axis positive displacement characteristics of F4, the torque adjustment formula is used. A rotation suppression item, "increase pitch motor torque by 12%," is generated, and at the same time, the electromagnetic suppression item, "enable dual-channel turbo code error correction," is activated according to L3 interference level, ultimately synergistically suppressing image jitter caused by waves and radar interference.

[0048] The overall solution described above, 103, extracts high-frequency jitter features from the image stream's sudden change signals through an inner loop. These are then spatially and temporally integrated with the interference prediction data from the outer loop to generate primary suppression parameters that include frequency and amplitude adjustments. These parameters are then matched with the measured electromagnetic interference waveform for waveform analysis. When the similarity exceeds a threshold, rotation suppression terms and electromagnetic suppression terms are dynamically output. This entire process achieves a closed-loop correlation between electromagnetic interference features and mechanical jitter signals. Cross-domain commands are used to collaboratively suppress turntable rotation anomalies and data transmission distortion, significantly improving image stability and continuity in complex electromagnetic environments.

[0049] 104. execute the turret anti-interference encoding based on the cross-domain suppression instruction, configure a turret posture related encoding scheme according to a rotation suppression term in the cross-domain suppression instruction, hierarchically set a rotation scene error correction protocol according to an electromagnetic suppression term in the cross-domain suppression instruction, and perform closed loop verification on the encoded data in combination with turret link feedback; Optionally, step 104 can specifically include the following steps: 1041. parse the rotation suppression term in the cross-domain suppression instruction, and generate a turret posture related encoding scheme containing compensation displacement amounts of each tilt angle interval according to a jitter suppression coefficient and a tilt angle correlation mapping table contained in the rotation suppression term; 1042. parse the electromagnetic suppression term in the cross-domain suppression instruction, divide the rotation scene according to the interference intensity level, and configure an error correction protocol containing a data check bit length and an error correction threshold for each level of scene; 1043. fuse the turret posture related encoding scheme and the error correction protocol, construct an anti-interference encoded data packet, and feed back to the turret actuator through the turret link; 1044. receive the position coordinates fed back by the turret actuator, calculate the inter-frame deviation value of the position coordinates from the expected position coordinates in the anti-interference encoded data packet, and when the continuous multiple frame deviation values exceed the set tolerance, generate a closed loop verification signal to trigger the reconstruction of the anti-interference encoded data packet and output a position deviation report.

[0050] In the above scheme, the turret anti-interference encoding refers to a dynamic control strategy generated based on the cross-domain suppression instruction, containing a mechanical posture compensation rule and a data transmission protection mechanism, which is used to jointly offset the influence of rotation jitter and electromagnetic interference on image transmission. The rotation suppression term refers to the control parameter in the cross-domain suppression instruction for mechanical jitter, which is used to generate an anti-jitter displacement instruction. The turret posture related encoding scheme refers to a compensation instruction set dynamically calculated according to the rotation suppression term, containing pixel displacement amounts of each tilt angle interval, which is used to correct the image offset caused by the tilt of the turret. The electromagnetic suppression term refers to the protection parameter in the cross-domain suppression instruction for electromagnetic interference, which is used to trigger an error correction protocol adapted to the current electromagnetic environment. The error correction protocol refers to a data protection rule hierarchically configured according to the electromagnetic suppression term, which is used to guarantee the data transmission integrity under strong interference. The anti-interference encoded data packet refers to a joint instruction set fused with the posture encoding scheme and the error correction protocol, which is used to send a cooperative control command to the actuator through an anti-interference link. The position deviation report refers to the quantitative error data generated in the closed loop verification link, containing the inter-frame pixel difference value of the actual coordinates and the expected coordinates and the interference level, which is used to trigger the reconstruction of the anti-interference strategy and the optimization of the system parameters.

[0051] In the embodiment of the application, first, the rotation suppression term in the cross-domain suppression instruction is parsed through step 1041 to extract the jitter suppression coefficient k and the tilt angle correlation mapping table contained therein. By looking up the table, the current tilt angle of the turret is matched with the jitter suppression coefficient k to obtain the compensation displacement amount of the pixel in the corresponding tilt angle interval. Match to the corresponding interval, get the basic displacement amount ; and then combined with the jitter suppression coefficient k to calculate the actual compensation displacement amount . Finally, the turntable attitude related encoding scheme containing the actual compensation displacement amount and the attitude adjustment instruction is generated. For example, a certain surveying and mapping unmanned aerial vehicle turntable detects a tilt angle , and finds the 5°-10° interval in the mapping table, the current k=0.7, and calculates the pixel, and generates the encoding: "lateral compensation +0.35 pixels, pitch angle -2°".

[0052] Then, based on the electromagnetic suppression term in the cross-domain suppression instruction, the interference intensity level L contained therein is analyzed through step 1042. The pre-stored protocol configuration rule library is called, and the corresponding error correction protocol is matched according to the interference intensity level L value, including setting the check bit length and the error correction threshold , for example, when L=4 , . This protocol will be used for error detection and correction during data transmission. For example, L=4 level interference is detected near the high-voltage transmission line, the system automatically enables 32-bit check bits to cover the default 16-bit, and sets the correction to be performed only when the single-frame data error rate is less than 5%.

[0053] Then, the turntable attitude related encoding scheme generated in step 1041 is binary fused with the error correction protocol generated in step 1042 through step 1043: the turntable attitude related encoding is taken as the data payload, and the error correction protocol is taken as the data header encapsulation. The fused anti-interference encoding data packet is sent to the turntable actuator through spread spectrum communication technology, and a timestamp and a serial number identifier are attached. For example, the binary code 0x23 corresponding to the "lateral compensation +0.35 pixels" instruction is fused with the header identifier 0x1F corresponding to the 32-bit check protocol, forming a complete data packet 0x1F+0x23, which is sent to the motor controller through a 2.4GHz spread spectrum link.

[0054] Finally, the position coordinates fed back by the turntable actuator are received through step 1044, and the expected coordinates in the anti-interference encoding data packet are compared, and the inter-frame deviation value is calculated through the Euclidean distance formula . Then, the sliding window counting algorithm is used to monitor the deviation state of consecutive frames, and if the of a continuous preset number of frames exceeds the tolerance threshold, a closed-loop verification signal is generated to trigger an alarm; at the same time, a position deviation report is automatically generated, and the system is driven to backtrack to steps 1041-1043 to regenerate the anti-interference encoding data packet. For example, when 5 consecutive frames When the pixel is greater than the tolerance of 3 pixels, the report will be fed back to the mapping table adjustment, the 8° tilt compensation will be increased from 0.5 pixels to 0.6 pixels, and the error correction protocol will be upgraded, the check bit will be increased from 32 bits to 40 bits, and dynamic optimization reconstruction will be realized.

[0055] In practical application, when a surveying and mapping unmanned aerial vehicle gimbal encounters strong electromagnetic interference and complex attitude in mountainous area during terrain scanning task, the system first analyzes the cross-domain suppression instruction: reads the jitter suppression coefficient 0.8 of the rotation suppression item and the tilt mapping table, the basic displacement amount 0.5 pixels in the interval of 5°-10°, calculates the actual displacement compensation pixel, generates the attitude code "horizontal compensation + 0.4 pixels"; synchronously analyzes the interference level 4 of the electromagnetic suppression item, calls the pre-stored rule to enable the 32-bit check bit and sets the 5% error correction threshold; fuses the above data into the anti-interference code package containing the compensation instruction 0x2A+32-bit header identification 0x1F, and sends it to the gimbal motor through the spread spectrum link. After execution, the deviation between the actual coordinates (102, 203) and the expected values (100, 200) is detected: , wherein, represents the square root, when the deviation of three consecutive frames is greater than the tolerance of 3 pixels, a position deviation report is generated, and the average pixel, interference level 4, triggering system reconstruction: increasing the 7° tilt compensation from 0.5 pixels to 0.6 pixels, then , and the check bit is upgraded to 40 bits, and finally a stable image stream is output.

[0056] The overall scheme of the above 104, by analyzing the rotation suppression item in the cross-domain instruction, dynamically calculating the pixel compensation amount corresponding to the tilt angle of the gimbal, synchronously configuring the hierarchical error correction strategy according to the electromagnetic interference level; the mechanical anti-shake instruction and the data transmission rule are fused into an anti-interference code package, which is sent to the actuator through spread spectrum communication; receiving the gimbal feedback coordinates, calculating the inter-frame deviation from the expected position through the Euclidean distance, generating a position deviation report containing the interference level when multiple consecutive frames are out of limit, driving the system reconstruction strategy, increasing the tilt compensation amount, enhancing the check strength, and verifying in a loop until the deviation converges. The flow cooperates through the double channels of mechanical displacement compensation and electromagnetic error correction, which not only eliminates the image shift caused by body jitter, but also blocks the transmission errors caused by electromagnetic interference, finally outputs a stable video stream without sudden changes and stalls in a strong interference environment, significantly improving the robustness of the unmanned gimbal in complex scenes.

[0057] 105, feeding back the gimbal environment adaptation parameters output by the closed loop verification to the interference feature library updating and the compensation strength adjustment process, realizing the cooperative control of electromagnetic interference and mechanical jitter, to output a stable image stream.

[0058] Optionally, step 105 can specifically include the following steps: 1051、convert the position deviation report output by the closed loop verification into turntable environment adaptation parameters including offset angle, displacement amount, and environmental interference level; 1052、write the offset angle and displacement amount in the environment adaptation parameters as new records into the interference feature library to update the interference feature library, and input the displacement amount in the environment adaptation parameters into the compensation adjuster to output a correction value of compensation intensity through a displacement amount intensity mapping table; 1053、form a double-channel cooperative mechanism through continuous updating of the interference feature library and dynamic adjustment of the compensation intensity, so that the electromagnetic interference suppression subsystem and the mechanical jitter suppression subsystem realize control law linkage based on the environment adaptation parameters, and output stable video stream eliminating inter-frame mutation to the image processing terminal.

[0059] The step 1053 can specifically include the following process: input the continuously updated interference feature library and the correction value of the compensation intensity into the cooperative controller to activate the double-channel cooperative mechanism, and input the environment adaptation parameters into the electromagnetic interference suppression subsystem and the mechanical jitter suppression subsystem; in the electromagnetic interference suppression subsystem, calculate a filter frequency lower limit value according to the environmental interference level, and convert the filter frequency lower limit value into a harmonic monitoring range of the mechanical jitter suppression subsystem; in the mechanical jitter suppression subsystem, set a compensation intensity reference value according to the displacement amount, and map the compensation intensity reference value into an amplitude attenuation threshold value of the electromagnetic interference suppression subsystem; generate image frame processing instructions based on the interactive output of the electromagnetic interference suppression subsystem and the mechanical jitter suppression subsystem, and output stable video stream to the terminal device when the offset amount of consecutive fixed frames is lower than a set threshold value.

[0060] In the above scheme, the turntable environment adaptation parameter refers to a dynamic data set representing the spatial pose offset state of the turntable in an electromagnetic interference environment, used to drive the dynamic optimization of the cross-domain suppression strategy. The modified value of the compensation intensity refers to the mechanical dithering suppression intensity adjustment generated based on the displacement through a predefined mapping relationship, used to calibrate the response intensity of the anti-dithering control system. The dual-channel cooperative mechanism refers to a closed-loop architecture constructed by connecting the interference feature database update channel and the compensation intensity adjustment channel in parallel, used to realize the control law linkage of the electromagnetic suppression subsystem and the mechanical suppression subsystem. The electromagnetic interference suppression subsystem refers to an interference filtering module constructed based on electromagnetic field perception and frequency domain analysis, used to block the conduction path of electromagnetic noise to mechanical components. The mechanical dithering suppression subsystem refers to an actuator that suppresses physical vibration through motion compensation algorithms and torque control, used to eliminate image frame offset during turntable rotation. The harmonic monitoring range refers to the abnormal vibration frequency interval that needs to be monitored by the mechanical suppression subsystem from the electromagnetic suppression subsystem, used to locate the key frequency point of electromagnetic-induced mechanical resonance. The amplitude attenuation threshold refers to the interference signal filtering start threshold fed back by the mechanical suppression subsystem to the electromagnetic suppression subsystem, used to trigger the emergency noise reduction mechanism in strong interference scenarios. The image frame processing instruction refers to the video stream output control signal generated after cross-domain cooperation verification, used to release high-quality image stream to the terminal after eliminating frame mutations.

[0061] In the embodiment of the present application, first, the coordinate conversion process is started when the position deviation report sent by the image processing unit is received through step 1051. The pixel offset vector is processed by the arctangent function calculation module: the tilt angle is calculated with the X-axis offset as the base and the Y-axis offset as the opposite side; the pixel offset is combined with the preset pixel physical conversion coefficient to calculate the actual displacement by calling the physical scale converter, and the calculation formula is as follows: At the same time, the electromagnetic field strength classifier reads the electromagnetic sensor data, matches the interference level reference table, and finally outputs the structured environment adaptation parameters: offset angle, displacement, and environmental interference level. For example, an agricultural unmanned aerial vehicle measures pixels, pixels and field strength 110V / m, and after conversion, it gets , , and level 4.

[0062] Then, the environment adaptation parameters output by step 1051 are used to execute interference database update and compensation adjustment in parallel through step 1052. In the interference database update channel, the time and space correlation engine binds the offset angle and displacement with time stamp and position information as new records written into the distributed interference feature database. At the same time, in the compensation adjustment channel, the displacement intensity mapper receives the displacement data, queries the predefined gradient mapping table, and outputs the compensation intensity adjustment instruction. In the above example, the unmanned aerial vehicle writes 32° and 0.48mm into the database, and converts the displacement to +15% compensation correction value by looking up the table.

[0063] Finally, the cooperative controller activates the dual-system linkage based on the database updated in step 1052 and the compensation correction value. The electromagnetic suppression subsystem receives the environmental interference level, generates a harmonic monitoring range by the frequency domain calculation module according to the formula , and sets the frequency band that the mechanical subsystem needs to monitor. The mechanical vibration suppression subsystem receives the frequency band range and the compensation correction value, calculates the electromagnetic interference filtering threshold by the coupling coefficient converter according to the formula , and feeds back to the electromagnetic subsystem. Finally, the frame stability analyzer continuously monitors the subsequent image frame offset angles, generates an image frame processing instruction when all angles are less than the preset threshold, and triggers the video stream stable output protocol. After the foregoing unmanned aerial vehicle linkage, the mechanical system captures and suppresses abnormal vibration in the 40-100Hz frequency band, and the electromagnetic system starts 50% attenuation for >92N·m interference signals, and outputs the non-jump video stream to the terminal C after five consecutive frames of offset <1.5°.

[0064] In actual application, when a certain unmanned intelligent turntable executes a patrol task near a high-voltage transmission line, the image processing unit detects a position deviation report, pixels, pixels, the system executes a conversion process: first, the offset angle is calculated by the inverse tangent function to obtain , and then the pixel offset amount is substituted into the Euclidean distance formula pixels and converted into an actual displacement amount 0.49mm, and meanwhile, the interference level 4 is matched according to the 110V / m field strength read by the electromagnetic sensor by calling the interference grading table. Subsequently, the system processes the two types of parameters in parallel: the 24° offset angle and the 0.49mm displacement amount are bound to the GPS position "B zone east" and written into the interference feature library to form a historical record, and meanwhile, the 0.49mm displacement amount is input into the compensation adjuster to calculate the compensation torque from the original 100N·m to 115N·m by querying the preset mapping table. Then, the cooperative controller activates the dual-system linkage: the electromagnetic suppression subsystem calculates the lower limit of the harmonic monitoring 40Hz according to the interference level 4 according to the formula , instructs the mechanical subsystem to focus on monitoring the 40-100Hz frequency band; the mechanical subsystem calculates the 92N·m amplitude attenuation threshold based on the new compensation torque 115N·m and the safety factor 0.8, and feeds back to the electromagnetic subsystem, and the latter starts 50% attenuation for the interference signals exceeding the threshold. After the foregoing cooperative suppression, the system detects that the offset angles of five consecutive image frames (1.3°, 0.9°, 1.1°, 0.7°, 1.0°) are all less than the preset stable threshold 2°, generates an image frame processing instruction 0x01, and outputs the non-jump high-definition video stream of the transmission tower insulator to the monitoring terminal.

[0065] The overall scheme of 105 above generates environment adaptation parameters by accurately analyzing position deviation, and drives dynamic update and compensation strength adjustment of interference feature library by channel. Then, a deep coordination between the two subsystems is constructed: the electromagnetic system output harmonic monitoring range guides the mechanical system to accurately capture abnormal vibration, and the mechanical system feedback attenuation threshold constrains the electromagnetic filtering strength, forming a closed-loop linkage. Finally, a continuous and stable video stream is output after eliminating inter-frame mutations, significantly improving the robustness of image transmission in complex electromagnetic environments.

[0066] The following is a complete example for steps 101-105, as shown in Figure 2 As shown, the intelligent turntable of a certain geological exploration unmanned aerial vehicle executes a photogrammetry task in an iron ore area, and the system captures wideband electromagnetic interference data in the running environment through the three-axis electromagnetic sensor array integrated in the turntable base, including the power frequency harmonic 50Hz fundamental wave + 150Hz third harmonic of the high-voltage power transmission line, the intensity is 85dBμV / m, the transient pulse interference pulse width generated by the mine large drilling machine is 2ms, the peak value is 120V / m, and the geological radar scattering noise frequency point is 2.4GHz, with a fluctuation of ±15dB. These raw data are classified into three categories after FFT spectrum analysis: "periodic industrial interference", "transient pulse" and "wideband background noise", and are updated to the electromagnetic fingerprint partition of the turntable interference feature library. The current-to-noise ratio characteristics of the brushless motor driver are monitored synchronously: during the pulse interference, the SNR drops from the normal 28dB to 18dB, combined with the newly recorded pulse spectrum characteristics in the interference library, the center frequency is 12.5kHz, the bandwidth is ±3kHz, and the corresponding notch filter parameters are generated: center frequency 12.5kHz, depth -40dB, Q value 8.5, , rounding optimization.

[0067] Then, based on the notch filter parameters, the core parameters of the digital notch filter are dynamically adjusted: the notch depth related to rotational jitter is strengthened from the default -30dB to -40dB to enhance the pulse interference suppression capability, and the bandwidth is expanded from ±2kHz to ±3kHz to cover the complete frequency spectrum range 12.5kHz±3kHz of the mine drilling machine pulse. This adjustment synchronously triggers a dual-channel anti-interference process: on the mechanical control side, the nonlinear compensation module uses the Hammerstein feedforward model to correct the motor torque fluctuation, and calculates the compensation amount according to the signal-to-noise ratio characteristics and injects the motor drive current; on the image transmission side, the frequency domain adaptive equalizer performs energy redistribution on the 32 subbands, focusing on improving the 3-5 subbands affected by interference to compensate for the spectral notches caused by electromagnetic noise. At the same time, the YOLOv5 target detection model deployed in the image pipeline identifies that the geological marker point mutation signal appears in the 1024th rotation key frame, the horizontal offset increases by 35 pixels, exceeding the stable threshold of 20 pixels, triggering the compensation strength adjuster: combined with the current The degradation state of the battery is determined by a linear interpolation formula The motion compensation weighting coefficient is increased from the reference value of 0.6 to 0.8, thereby constructing a primary closed-loop suppression mechanism covering mechanical jitter suppression and image transmission optimization.

[0068] During the continuous operation of the closed-loop suppression mechanism, the inner loop control unit performs high-frequency feature extraction for the 1024th frame of the sudden signal with a 35-pixel offset: using a db4 wavelet basis for 5-layer wavelet packet decomposition, the 125-250Hz frequency band component is separated from the residual jitter signal, and its energy accounts for 62% of the total jitter energy, which is higher than the normal 40%, indicating that the electromagnetic pulse interference has induced mechanical resonance. The outer loop prediction engine synchronously fuses this high-frequency jitter feature with the turntable interference prediction data, which is based on an LSTM time series model to analyze the historical electromagnetic interference pattern. The pulse recurrence period in the iron ore area is 120±15ms, and the prediction of the recurrence probability of pulse interference within the next 200ms reaches 92%. The two types of data are associated through a coupling matrix to generate primary suppression parameters: high-frequency jitter suppression gain 2.3, pulse interference protection level 4. Then, the motion state analyzer performs cross-domain coupling modeling combined with electromagnetic interference data: in the mechanical domain, the correlation coefficient between high-frequency jitter energy and electromagnetic pulse intensity is analyzed, and in the electromagnetic domain, the total harmonic distortion rate of motor current is detected to rise to 12%, exceeding the 8% safety threshold. Finally, the structured cross-domain suppression instruction is output: the rotation suppression term requires the high-frequency gain mode to be set to the "intensified" state, the action frequency band is limited to 125-250Hz, and the dynamic compensation margin is increased by 30%; the electromagnetic suppression term starts the four-level protection strategy.

[0069] Then, based on the rotation suppression term in the cross-domain suppression instruction, the jitter suppression coefficient k=1.3 and the inclination angle displacement mapping table are obtained, the current turntable inclination angle θ=24° is analyzed, the basic displacement amount is calculated by looking up the 20-25° interval and linearly interpolating, and then the actual compensation displacement amount is calculated; then the attitude-related encoding scheme is generated: the is converted into motor torque increment , and the composite command is injected into the servo system by superimposing PID parameter adjustment. The electromagnetic suppression term is executed synchronously: LDPC(2048,1024) encoding and 1:0.7 dynamic spectrum compression are started. Closed-loop verification shows that the torque fluctuation standard deviation is reduced from 0.18N·m to 0.10N·m, with a reduction of 44%, and the image error rate BER is optimized to , and the subsequent frame average offset is reduced to 8.7 pixels, which is improved by 75% compared with the sudden frame.

[0070] Finally, the position deviation report is converted into three-dimensional environment adaptation parameters, the offset angle is calculated by geometric triangulation, and the actual displacement amount Rounding 0.10mm while maintaining the ambient interference level 4. Split-channel feedback processing: the spatiotemporal correlation engine binds the offset angle 4.7° with the displacement amount 0.10mm and writes the iron mine GPS coordinates into the interference feature library to form a new sample of the pulse interference mode; the compensation adjuster inputs the displacement amount 0.10mm into the preset mapping table, and the displacement <0.2mm interval corresponds to -5% compensation correction, which reduces the mechanical subsystem compensation torque from 115N·m to 109.25N·m. The collaborative controller immediately activates the dual-system deep linkage: the electromagnetic suppression subsystem sets the harmonic monitoring range lower limit 40-100Hz according to the interference level 4 formula "base frequency 10Hz x 4 = 40Hz" and notifies the mechanical subsystem; the mechanical subsystem calculates the amplitude attenuation threshold 87.4N·m based on the corrected compensation value 109.25N·m combined with the safety factor 0.8 and feeds back to the electromagnetic subsystem. Through this collaboration, the subsequent 200 frames of image offset angle standard deviation stabilizes within ±0.15°, the physical displacement is ±1.8 pixels, and the stable threshold of less than ±2.5 pixels is continuously reduced, and finally the stable video stream of the iron ore vein texture can be clearly identified to the geological analysis terminal, and the exploration task time is shortened by 40% compared with the traditional scheme.

[0071] Figure 3 A structure diagram of an intelligent image stabilization system of an unmanned intelligent turntable is provided for the embodiments of the present application, as shown in Figure 3 The system comprises: An acquisition module 31 is configured to acquire electromagnetic interference data in a turntable operating environment, update an interference feature library of the turntable based on the electromagnetic interference data, and monitor a signal-to-noise ratio feature of a dynamic component of the turntable to generate a trap wave filtering parameter. A processing module 32 is configured to adjust a trap wave depth and bandwidth related to rotation jitter of the turntable according to the trap wave filtering parameter, synchronously trigger nonlinear compensation and frequency domain adaptive equalization processing of the turntable, identify a mutation signal of a rotation key frame in a turntable image stream and adjust compensation strength in combination with the signal-to-noise ratio feature, and form an interference suppression closed-loop mechanism. A generation module 33 is configured to extract a high-frequency jitter feature in the mutation signal through inner loop control during operation of the interference suppression closed-loop mechanism, generate a primary suppression parameter by fusing the high-frequency jitter feature and turntable interference prediction data through outer loop prediction, and generate a cross-domain suppression instruction by performing motion state analysis on the primary suppression parameter and electromagnetic interference data. An execution module 34 is configured to perform turntable anti-interference encoding based on the cross-domain suppression instruction, configure a turntable posture related encoding scheme according to a rotation suppression term in the cross-domain suppression instruction, hierarchically set a rotation scene error correction protocol according to an electromagnetic suppression term in the cross-domain suppression instruction, and perform closed-loop verification on encoded data in combination with turntable link feedback. The output module 35 is configured to feed the turret environment adaptive parameter output by the closed-loop verification to the interference feature library updating and the compensation intensity adjustment process, so as to realize the collaborative control of electromagnetic interference and mechanical jitter, and output stable image flow.

[0072] Figure 3 The intelligent image stabilization system of the unmanned intelligent turret can perform Figure 1 The intelligent image stabilization method of the unmanned intelligent turret according to the embodiment described above has the same implementation principles and technical effects. The specific operation modes of each module and unit of the intelligent image stabilization system of the unmanned intelligent turret described above have been described in detail in the embodiment of the method, and will not be described in detail here.

[0073] In one possible design, Figure 3 The intelligent image stabilization system of the unmanned intelligent turret according to the embodiment described above can be implemented as a computing device, such as Figure 4 The computing device can include a storage component 41 and a processing component 42. The storage component 41 stores one or more computer instructions, wherein the one or more computer instructions are called and executed by the processing component 42.

[0074] The processing component 42 is configured to perform the above Figure 1 The intelligent image stabilization method of the unmanned intelligent turret according to the embodiment described above.

[0075] The processing component 42 can include one or more processors to execute computer instructions to complete all or part of the steps in the above method. Of course, the processing component can also be one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors or other electronic components, for executing the above method.

[0076] The storage component 41 is configured to store various types of data to support the operation of the terminal. The storage component can be realized by any type of volatile or non-volatile storage device or their combination, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk or optical disk.

[0077] Of course, the computing device can also include other components, such as input / output interface, display component, communication component, etc.

[0078] The input / output interface provides an interface between the processing component and peripheral interface modules, which can be output devices, input devices, and the like.

[0079] The communication component is configured to facilitate wired or wireless communication between the computing device and other devices, and the like.

[0080] The computing device can be a physical device or an elastic computing host provided by a cloud computing platform, and the processing component, the storage component, and the like can be basic server resources rented or purchased from the cloud computing platform.

[0081] The embodiments of the present application also provide a computer storage medium storing a computer program, and the computer program can implement the above-mentioned Figure 1 An intelligent image stabilization method of an unmanned intelligent turntable.

[0082] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-mentioned system, device and unit can refer to the corresponding processes in the foregoing method embodiments, which will not be described here.

[0083] The device embodiments described above are only schematic, and the units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, that is, they can be located in one place, or distributed on multiple network units. Part or all of the modules can be selected according to actual needs to achieve the purpose of the embodiments. Those skilled in the art can understand and implement without creative labor.

[0084] From the above description of the embodiments, those skilled in the art can clearly understand that the embodiments can be realized by means of software and necessary general hardware platforms, and of course, can also be realized by hardware. Based on such understanding, the above technical solutions can be embodied in the form of software products, and the computer software products can be stored in a computer readable storage medium, such as ROM / RAM, magnetic disk, optical disk, and the like, and include a plurality of instructions to make a computer device (which can be a personal computer, a server, and the like) execute the methods described in each embodiment or some parts of the embodiments.

[0085] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. An intelligent image stabilization method of an unmanned intelligent turntable, characterized in that, The method comprises the following steps: acquiring electromagnetic interference data in the running environment of the rotating platform, updating the interference feature library of the rotating platform based on the electromagnetic interference data, monitoring the signal-to-noise ratio feature of the power assembly of the rotating platform to generate a notch filtering parameter; adjusting the notch depth and bandwidth related to the rotating jitter of the rotating platform according to the notch filtering parameter, synchronously triggering the nonlinear compensation and frequency domain adaptive equalization processing of the rotating platform, identifying the mutation signal of the rotating key frame in the image stream of the rotating platform and adjusting the compensation strength in combination with the signal-to-noise ratio feature to form an interference suppression closed-loop mechanism; during the operation of the interference suppression closed-loop mechanism, the high-frequency jitter feature in the mutation signal is extracted through inner loop control, the high-frequency jitter feature and the rotating platform interference prediction data are fused through outer loop prediction to generate a primary suppression parameter, and the primary suppression parameter and the electromagnetic interference data are analyzed in the motion state to generate a cross-domain suppression instruction; based on the cross-domain suppression instruction, the anti-interference encoding of the rotating platform is performed, the rotating posture related encoding scheme is configured according to the rotating suppression term in the cross-domain suppression instruction, the rotating scene error correction protocol is set in stages according to the electromagnetic suppression term in the cross-domain suppression instruction, and the encoding data is verified in a closed loop in combination with the link feedback of the rotating platform; the rotating platform environment adaptation parameters output by the closed loop verification are fed back to the interference feature library updating and the compensation strength adjustment process, and the cooperative control of electromagnetic interference and mechanical jitter is realized to output a stable image stream.

2. The method of claim 1, wherein, The rotating platform environment adaptation parameters output by the closed loop verification are fed back to the interference feature library updating and the compensation strength adjustment process, and the cooperative control of electromagnetic interference and mechanical jitter is realized to output a stable image stream, which comprises: the position deviation report output by the closed loop verification is converted into rotating platform environment adaptation parameters including offset angle, displacement and environmental interference level; the offset angle and displacement in the environmental adaptation parameters are written into the interference feature library as new records to update the interference feature library, and the displacement in the environmental adaptation parameters is input into the compensation adjuster to output the correction value of the compensation strength through the displacement strength mapping table; through the continuous updating of the interference feature library and the dynamic adjustment of the compensation strength, a double-channel cooperative mechanism is formed, so that the electromagnetic interference suppression subsystem and the mechanical jitter suppression subsystem realize control law linkage based on the environmental adaptation parameters, and output a stable video stream eliminating frame-to-frame mutation to the image processing terminal.

3. The method of claim 2, wherein, Through the continuous updating of the interference feature library and the dynamic adjustment of the compensation strength, a double-channel cooperative mechanism is formed, so that the electromagnetic interference suppression subsystem and the mechanical jitter suppression subsystem realize control law linkage based on the environmental adaptation parameters, and output a stable video stream eliminating frame-to-frame mutation to the image processing terminal, which comprises: the continuously updated interference feature library and the correction value of the compensation strength are input into the cooperative controller to activate the double-channel cooperative mechanism, and the environmental adaptation parameters are input into the electromagnetic interference suppression subsystem and the mechanical jitter suppression subsystem at the same time; in the electromagnetic interference suppression subsystem, the filter frequency lower limit value is calculated according to the environmental interference level, and the filter frequency lower limit value is converted into the harmonic monitoring range of the mechanical jitter suppression subsystem; In the mechanical dithering suppression subsystem, a compensation intensity reference value is set according to the displacement amount, and the compensation intensity reference value is mapped to an amplitude attenuation threshold of the electromagnetic interference suppression subsystem; Based on the interaction output of the electromagnetic interference suppression subsystem and the mechanical dithering suppression subsystem, image frame processing instructions are generated, and when the displacement amount of consecutive fixed frames is lower than a set threshold, a stable video stream is output to the terminal device.

4. The method of claim 1, wherein, The anti-interference encoding of the turntable is performed based on the cross-domain suppression instruction, a turntable posture related encoding scheme is configured according to the rotation suppression term in the cross-domain suppression instruction, a rotation scene error correction protocol is hierarchically set according to the electromagnetic suppression term in the cross-domain suppression instruction, and closed loop verification is performed on the encoded data in combination with the turntable link feedback, including: The rotation suppression term in the cross-domain suppression instruction is parsed, a turntable posture related encoding scheme containing compensation displacement amounts of each pitch interval is generated according to the dithering suppression coefficient and the pitch angle correlation mapping table contained in the rotation suppression term; The electromagnetic suppression term in the cross-domain suppression instruction is parsed, the rotation scene is divided according to the interference intensity level, and an error correction protocol containing data check bit length and error correction threshold is configured for each level of scene; The turntable posture related encoding scheme and the error correction protocol are fused to construct an anti-interference encoding data packet and feedback to the turntable actuator through the turntable link; The position coordinates feedback by the turntable actuator are received, the inter-frame deviation value of the expected position coordinates in the anti-interference encoding data packet is calculated, and when the deviation values of consecutive multiple frames exceed a set tolerance, a closed loop verification signal is generated to trigger the reconstruction of the anti-interference encoding data packet and output a position deviation report.

5. The method of claim 1, wherein, During the operation of the interference suppression closed loop mechanism, the high frequency dithering feature in the mutation signal is extracted through the inner loop control, the high frequency dithering feature and the turntable interference prediction data are fused to generate primary suppression parameters by the outer loop prediction, the cross-domain suppression instruction is generated by performing motion state analysis on the primary suppression parameters and electromagnetic interference data, including: During the operation of the interference suppression closed loop mechanism, the high frequency dithering feature with a change rate exceeding a preset threshold is extracted from the mutation signal by the inner loop control module; The high frequency dithering feature and the turntable interference prediction data stored in the outer loop prediction module are spatio-temporally fused to generate primary suppression parameters containing frequency suppression coefficients and amplitude correction coefficients; The primary suppression parameters and the electromagnetic interference data are subjected to motion trajectory analysis, and the cross-domain suppression instruction containing rotation suppression terms and electromagnetic suppression terms is generated by comparing the fluctuation patterns of the primary suppression parameters and the electromagnetic interference data in the same time window.

6. The method of claim 1, wherein, The notch depth and bandwidth related to the rotation dithering of the turntable are adjusted according to the notch filtering parameters, the nonlinear compensation of the turntable and the frequency domain adaptive equalization processing are triggered synchronously, the mutation signal of the rotation key frame in the turntable image stream is identified, the compensation intensity is adjusted in combination with the signal-to-noise ratio feature, and the interference suppression closed loop mechanism is formed, including: The notch filtering parameters are input into the adjustable filter, the rotation dithering main frequency is blocked according to the center frequency value, the dithering harmonic amplitude is attenuated according to the depth setting value, and the rotation dithering frequency spectrum range is covered according to the width setting value; The synchronous trigger nonlinear compensation module corrects the image offset caused by the rotation jitter, and triggers the frequency domain adaptive equalization module to suppress the interference energy of the rotation related frequency band; The scanning and identification of the turntable image stream extracts the mutation signal in the rotation key frame, and the intensity vector of the mutation signal and the signal-to-noise ratio feature are input into the compensation adjuster to adjust the compensation intensity; The compensation intensity is output to the nonlinear compensation module, and a feedback signal is generated to the adjustable filter, forming an interference suppression closed loop mechanism from image recognition to parameter adjustment.

7. The method of claim 1, wherein, The electromagnetic interference data in the turntable running environment is acquired, the interference feature library of the turntable is updated based on the electromagnetic interference data, and the signal-to-noise ratio feature of the turntable power component is monitored to generate the notch filter parameter, comprising: The original electromagnetic field intensity data in the turntable running environment is collected by the electromagnetic sensor on the unmanned intelligent turntable, and the electromagnetic interference data containing peak intensity and fluctuation frequency is output; The electromagnetic interference data is compared with the interference mode stored in the interference feature library, a new item is created for the unrecorded interference mode and added to the interference feature library, and the updated interference feature library is output; Meanwhile, the output signal of the turntable power component is monitored to generate the signal-to-noise ratio feature containing the effective signal ratio, and when the signal-to-noise ratio feature is lower than the threshold, the notch filter parameter containing the center frequency value, the depth setting value and the width setting value is generated based on the updated interference feature library.

8. An intelligent image stabilization system for unmanned smart turntables, characterized by, Comprising: The acquisition module is used for acquiring electromagnetic interference data in the turntable running environment, updating the interference feature library of the turntable based on the electromagnetic interference data, and monitoring the signal-to-noise ratio feature of the turntable power component to generate the notch filter parameter; The processing module is used for adjusting the notch depth and bandwidth related to the rotation jitter of the turntable according to the notch filter parameter, synchronously triggering the nonlinear compensation and frequency domain adaptive equalization processing of the turntable, identifying the mutation signal of the rotation key frame in the turntable image stream and adjusting the compensation intensity combined with the signal-to-noise ratio feature, forming an interference suppression closed loop mechanism; The generation module is used for extracting the high frequency jitter feature in the mutation signal through inner loop control during the operation of the interference suppression closed loop mechanism, generating the primary suppression parameter by fusing the high frequency jitter feature and the turntable interference prediction data through outer loop prediction, and generating the cross-domain suppression instruction by motion state analysis on the primary suppression parameter and the electromagnetic interference data; The execution module is used for executing the anti-interference encoding of the turntable based on the cross-domain suppression instruction, configuring the turntable attitude related encoding scheme according to the rotation suppression term in the cross-domain suppression instruction, and hierarchically setting the rotation scene error correction protocol according to the electromagnetic suppression term in the cross-domain suppression instruction, and combining the turntable link feedback to perform closed loop verification on the encoded data; The output module is used for feeding back the turntable environment adaptive parameters output by the closed loop verification to the interference feature library update and the compensation intensity adjustment process, realizing the cooperative control of electromagnetic interference and mechanical jitter, and outputting stable image stream.

9. A computing device, comprising: The application relates to an unmanned intelligent rotating table and an intelligent image stabilizing method thereof.

10. A computer storage medium, characterized in that, The application relates to an unmanned intelligent rotating table and an intelligent image stabilizing method thereof.

Images