An FPV unmanned aerial vehicle identification method and device based on analog image transmission signals

Abstract

This invention discloses an FPV unmanned aerial vehicle (UAV) identification method and apparatus based on analog image transmission signals. The method includes: S1, setting acquisition parameter information, including acquisition center frequency and acquisition bandwidth; S2, acquiring spatial electromagnetic wave signals based on the acquisition parameter information to obtain an acquisition signal; S3, performing identification processing on the acquisition signal to obtain an identification signal.

Description

Technical Field

[0001] This invention relates to the fields of intelligent signal processing, strategy optimization, and signal recognition technology, and specifically to an FPV unmanned aerial vehicle (UAV) identification method and apparatus based on analog image transmission signals. Background Technology

[0002] The FPV (Field-Proof Video) drone system utilizes real-time video transmission and remote control technology to enable operators to control the drone with low latency and precision from a first-person perspective. Its core operating principle relies on the collaborative work of multiple modules, including three main subsystems: an image transmission system, a remote control system, and a power system. The FPV drone primarily uses a camera to capture images and convert them into analog image transmission signals. These signals are then transmitted to an image transmission transmitter, which transmits them via radio waves. The receiving end demodulates and reconstructs the analog image transmission signals, ultimately displaying them as images on the remote control terminal.

[0003] In terms of data interaction, the UAV and the ground control station transmit data via a wireless communication link. The wireless communication link system is a crucial component of the UAV system, its primary task being to establish a two-way air-to-ground data transmission channel for long-distance remote control, telemetry, and mission information transmission from the ground control station. Remote control enables remote operation of the UAV and mission equipment, while telemetry monitors the UAV's status, including not only image data but also flight status data and sensor data. This real-time transmission of data allows operators to promptly understand the UAV's operational status and make adjustments as needed.

[0004] The signal transmission link of an FPV drone's data link is divided into a downlink and an uplink. The link where the drone's control console sends remote control commands to the drone is called the uplink. The link where the drone sends relevant information and data to the control console is called the downlink. FPV drones primarily use the downlink for transmission because it employs a unidirectional transmission method, allowing for real-time image transmission without establishing a connection. This method results in lower latency and higher real-time performance for FPV image transmission. During signal data transmission, wireless signals are affected by terrain, objects, and atmospheric conditions, causing reflection, scattering, and diffraction of radio waves, leading to multipath propagation. Furthermore, the channel is susceptible to various noise interferences, causing a degrade in data transmission quality. Due to its high flight speed, long transmission distance, and low-latency, real-time transmission capabilities, FPV drones typically use the 5.8 GHz frequency band for signal transmission.

[0005] The simulated image transmission signal from an FPV drone essentially employs CVBS modulation with a carrier frequency of 5.8 GHz. The wide bandwidth of the carrier and the complexity of the modulation present numerous challenges for traditional signal detection and identification methods. From the perspective of the signal's own spectral characteristics, the use of mixed modulation such as AM (VSB) and FM results in significant nonlinearity and complex spectral features, making it difficult to identify using common digital signal detection and modulation methods. From the perspective of the signal monitoring environment, this simulated image transmission signal is generally modulated onto a 5.8 GHz carrier, potentially within the range of approximately 5650 MHz to 5950 MHz. Due to the wide spectral distribution, real-time rapid detection and identification are difficult. From an electromagnetic environment perspective, the monitoring environment is subject to significant electromagnetic interference, such as various communication signals in urban environments and electromagnetic noise generated by industrial equipment. This interference severely affects the acquisition and analysis of FPV drone image transmission signals. Furthermore, different geographical environments can also affect signal propagation to varying degrees, further increasing the difficulty of identifying the image transmission signal.

[0006] Traditional signal processing methods are inefficient when dealing with complex FPV drone image transmission signals, failing to meet the demands of real-time monitoring. Furthermore, establishing accurate signal recognition models requires extensive training with a large amount of sample data, but currently available sample data is limited, and data annotation and classification are challenging. These issues severely impact the monitoring of simulated image transmission signals from FPV drones, resulting in low efficiency and accuracy in signal recognition. Summary of the Invention

[0007] This invention primarily addresses the problem of how to quickly and accurately identify simulated image transmission signals from FPV drones. It discloses an FPV drone identification method and apparatus based on simulated image transmission signals.

[0008] In a first aspect, this invention discloses an FPV unmanned aerial vehicle (UAV) identification method based on analog image transmission signals, comprising:

[0009] S1, Set the acquisition parameter information; the acquisition parameter information includes the acquisition center frequency and acquisition bandwidth;

[0010] S2, based on the collected parameter information, the space electromagnetic wave signal is collected to obtain the collected signal;

[0011] S3, the acquired signal is processed to obtain an identification signal.

[0012] The process of identifying the acquired signal to obtain an identification signal includes:

[0013] S31, The acquired signal is transformed to obtain a transformed signal set;

[0014] S32, Perform feature discrimination processing on the transformed signal set to obtain the initial recognition signal;

[0015] S33, Perform parameter comprehensive discrimination processing on the initial identification signal to obtain the identification signal.

[0016] The transformation processing of the acquired signals to obtain a transformed signal set includes:

[0017] S311, Perform frequency domain transformation on the acquired signal to obtain a frequency domain signal;

[0018] S312, Perform SVMD transformation on the acquired signal to obtain a time-domain feature signal;

[0019] S313, using the frequency domain signal and the time domain characteristic signal, a transformed signal set is constructed.

[0020] The step of performing feature discrimination processing on the transformed signal set to obtain an initial recognition signal includes:

[0021] S321, Perform frequency domain feature statistical processing on the frequency domain signals in the transformed signal set to obtain a set of frequency domain statistical values;

[0022] S322, Perform time-domain feature statistical processing on the time-domain feature signals in the transformed signal set to obtain a set of time-domain statistical values;

[0023] S323, Perform feature matching and discrimination processing on the frequency domain statistical value set and the time domain statistical value set to obtain the initial identification signal.

[0024] The step of performing feature matching and discrimination processing on the frequency domain statistical value set and the time domain statistical value set to obtain the initial identification signal includes:

[0025] S3231, using the set of frequency domain statistical values, a first statistical vector is constructed;

[0026] S3232, using the set of time-domain statistical values, a second statistical vector is constructed;

[0027] S3233, perform difference fusion calculation on the first statistical vector and the second statistical vector with a preset set of standard statistical vectors to obtain the first difference value;

[0028] S3234, determine whether the first difference value is less than a preset first discrimination threshold, and obtain the first discrimination result;

[0029] If the first discrimination result is yes, the acquired signal is determined to be the initial identification signal; if the first discrimination result is no, the identification process ends.

[0030] The expression for the difference fusion calculation is:

[0031] ,

[0032] in, This represents the j-th element of the i-th statistical vector. This represents the j-th element of the i-th statistical standard vector. r 1 indicates the first difference value.

[0033] The step of performing parameter synthesis and discrimination processing on the initial identification signal to obtain the identification signal includes:

[0034] S331, the bandwidth, average power and signal duration of the initial identification signal are statistically obtained;

[0035] S332, determine whether the bandwidth, average power and signal duration simultaneously satisfy bandwidth ≥ AMHz, average signal power ≥ BdB and burst duration ≥ Cms, and obtain the second determination result;

[0036] If the second discrimination result is yes, the initial identification signal is determined to be an identification signal; if the second discrimination result is no, the identification process ends.

[0037] A second aspect of the present invention discloses an FPV unmanned aerial vehicle (UAV) identification device based on analog image transmission signals, the device comprising:

[0038] Memory containing executable program code;

[0039] A processor coupled to the memory;

[0040] The processor calls the executable program code stored in the memory to execute the FPV UAV identification method based on analog image transmission signals.

[0041] In a third aspect of this invention, a computer-storable medium is disclosed, wherein the computer-storable medium stores computer instructions, and when the computer instructions are invoked by a computer, they are used to execute the FPV UAV identification method based on analog image transmission signals.

[0042] In a fourth aspect of this invention, an information data processing terminal is disclosed, which is used to implement the FPV UAV identification method based on analog image transmission signals.

[0043] The beneficial effects of this invention are as follows:

[0044] This invention significantly improves the recognition accuracy and robustness of simulated image transmission signals from FPV drones by fusing frequency and time domain features. It extracts frequency domain statistics (such as harmonic mean and flow values) through frequency domain transformation and time domain feature signals through SVMD transformation, constructing a multi-dimensional feature vector. This vector more comprehensively characterizes the intrinsic properties of CVBS signals and effectively distinguishes noise and interference signals. For example, even in low signal-to-noise ratio environments, this invention maintains high recognition reliability through the fusion calculation of the differences between frequency and time domain statistics.

[0045] This invention reduces reliance on dedicated hardware circuits and achieves unified identification of analog signals of different standards through software algorithms, thereby lowering equipment costs and complexity. For example, it eliminates the need to design independent circuits for different standards such as PAL, SECAM, and NTSC; signal matching can be completed simply by using a preset set of standard statistical vectors, improving the system's flexibility and scalability. This feature is particularly suitable for portable UAV monitoring equipment, facilitating miniaturization and low-power design.

[0046] This invention optimizes the real-time performance of signal processing by using a pre-defined set of standard statistical vectors and a difference fusion calculation mechanism. The SVMD transform, combined with the rapid calculation of frequency domain statistics, avoids the complex iterative process of traditional VMD parameter optimization. Simultaneously, the mathematical expression for difference fusion calculation employs efficient numerical computation, significantly reducing computational latency and meeting the real-time recognition requirements of UAVs in high-speed flight. For example, in racing scenarios, this invention can complete signal recognition within milliseconds, ensuring timely response to control commands.

[0047] The feature matching discrimination mechanism of this invention effectively enhances anti-interference capabilities. By setting multi-dimensional difference thresholds, it can accurately identify CVBS signals in complex environments such as multipath propagation, occlusion, and co-channel interference. For example, when the signal undergoes phase distortion due to multipath effects, this invention can still make a comprehensive discrimination by calculating the difference in frequency domain statistics (c1) and the energy difference in time domain features (c2), avoiding misjudgment. This characteristic enables this invention to maintain stable recognition performance in complex environments such as urban canyons and dense building clusters. Attached Figure Description

[0048] Figure 1 This is a flowchart illustrating the implementation of the method of the present invention. Detailed Implementation

[0049] To better understand the content of this invention, an embodiment is provided here.

[0050] Figure 1 This is a flowchart illustrating the implementation of the method of the present invention.

[0051] In FPV UAV analog image transmission systems, the CVBS composite video baseband signal serves as the core carrier for analog video transmission, and its modulation performance directly determines image quality, transmission delay, and anti-interference capability. Compared to digital image transmission systems, CVBS-based analog image transmission naturally possesses low latency advantages due to the elimination of complex compression and decompression processes, making it particularly suitable for real-time visual feedback in high-speed maneuvering scenarios. However, analog signals are susceptible to noise and multipath interference in wireless channels, thus requiring optimized modulation mechanisms to balance signal fidelity and transmission reliability. This section will systematically explain the generation and modulation principles of CVBS signals and their key technological implementation in analog image transmission links, based on the application requirements of FPV UAVs.

[0052] CVBS (Composite Video Broadcast Signal) is an analog signal standard widely used in the field of video transmission. It combines luminance (Y), chrominance (C), and synchronization signals into a single electrical signal for transmission, making it compatible with most traditional video equipment and a crucial foundation for video image transmission by FPV drones.

[0053] In a first aspect, this invention discloses an FPV unmanned aerial vehicle (UAV) identification method based on analog image transmission signals, comprising:

[0054] S1, Set the acquisition parameter information; the acquisition parameter information includes the acquisition center frequency and acquisition bandwidth;

[0055] S2, based on the collected parameter information, the space electromagnetic wave signal is collected to obtain the collected signal;

[0056] S3, perform identification processing on the acquired signal to obtain an identification signal;

[0057] The identification signal is used to characterize the identification result of the FPV drone.

[0058] The process of identifying the acquired signal to obtain an identification signal includes:

[0059] S31, The acquired signal is transformed to obtain a transformed signal set;

[0060] S32, Perform feature discrimination processing on the transformed signal set to obtain the initial recognition signal;

[0061] S33, Perform parameter comprehensive discrimination processing on the initial identification signal to obtain the identification signal;

[0062] The transformation processing of the acquired signals to obtain a transformed signal set includes:

[0063] S311, Perform frequency domain transformation on the acquired signal to obtain a frequency domain signal;

[0064] S312, Perform SVMD transformation on the acquired signal to obtain a time-domain feature signal;

[0065] S313, using the frequency domain signal and the time domain characteristic signal, a transformed signal set is constructed;

[0066] The step of performing feature discrimination processing on the transformed signal set to obtain an initial recognition signal includes:

[0067] S321, Perform frequency domain feature statistical processing on the frequency domain signals in the transformed signal set to obtain a set of frequency domain statistical values;

[0068] S322, Perform time-domain feature statistical processing on the time-domain feature signals in the transformed signal set to obtain a set of time-domain statistical values;

[0069] S323, Perform feature matching and discrimination processing on the frequency domain statistical value set and the time domain statistical value set to obtain the initial identification signal;

[0070] The time-domain statistical information set includes mean, variance, mode, and average rise time;

[0071] The frequency domain statistical information set includes the median value, harmonic mean, flow rate value, and average derivative of the frequency domain signal.

[0072] The average rise time is the average rise time of all amplitudes of the acquired signal.

[0073] The harmonic mean is the average of all harmonic frequencies of the frequency domain signal;

[0074] The fluidity value is the square root of the variance of the first derivative of the frequency domain signal divided by the variance of the signal.

[0075] The average derivative is the average value of the derivatives at each point in the frequency domain signal;

[0076] The step of performing feature matching and discrimination processing on the frequency domain statistical value set and the time domain statistical value set to obtain the initial identification signal includes:

[0077] S3231, using the set of frequency domain statistical values, a first statistical vector is constructed; the first statistical vector includes the median value, harmonic mean, flow value, and average derivative;

[0078] S3232, using the set of time-domain statistical values, a second statistical vector is constructed; the second statistical vector includes the mean, variance, mode, and average rise time.

[0079] S3233, perform difference fusion calculation on the first statistical vector and the second statistical vector with a preset set of standard statistical vectors to obtain the first difference value;

[0080] S3234, determine whether the first difference value is less than a preset first discrimination threshold, and obtain the first discrimination result;

[0081] S3235, when the first discrimination result is yes, the collected signal is determined to be the initial identification signal; when the first discrimination result is no, the identification process ends.

[0082] The expression for the difference fusion calculation is:

[0083] ,

[0084] in, This represents the j-th element of the i-th statistical vector. This represents the j-th element of the i-th statistical standard vector. r 1 indicates the first difference value;

[0085] This invention introduces a tangent function to nonlinearly map the relative differences between statistical vector elements. This maintains sensitivity to subtle changes when differences are small, and amplifies the discriminative power of the tangent function when differences are large, effectively avoiding the problem of small differences being masked or large differences being overcompressed. Simultaneously, by multiplying the sum of differences between two statistical vectors (frequency domain and time domain) and taking the square root, it achieves the organic fusion of multi-dimensional feature differences. This preserves the independence of frequency and time domain features while strengthening their correlation through the product relationship, ensuring that significant differences in one dimension are not masked by weak differences in another, thus improving the ability to distinguish complex interference signals. For example, when there is a slight shift in frequency domain features but significant differences in time domain features, this expression can amplify the overall difference through the product operation, preventing the interference signal from being misidentified as the target signal.

[0086] The step of performing feature matching and discrimination processing on the frequency domain statistical value set and the time domain statistical value set to obtain the initial identification signal includes:

[0087] S32301, The first difference value is obtained by performing a first difference calculation on the set of frequency domain statistical values ​​and the acquired signal, the standard CVBS signal and the first statistical standard vector;

[0088] S32302, The time-domain statistical value set and the acquired signal are compared with the standard CVBS signal and the second statistical standard vector to perform a second difference calculation to obtain a second difference value;

[0089] S32303, determine whether the first difference value and the second difference value meet the matching conditions, and obtain the third discrimination result;

[0090] S32304, if the third discrimination result is yes, determine that the collected signal is the initial identification signal; if the third discrimination result is no, end the identification process;

[0091] The first difference calculation process includes:

[0092] ,

[0093] Where c1 is the first difference value. In order to collect signals, For standard CVBS signals, The Frobenius norm value of the difference vector between the first statistical vector and the first statistical standard vector constructed for the set of frequency domain statistical values. This represents the convolution operation, where T is the length of the acquired signal.

[0094] The first difference calculation process effectively extracts the temporal similarity features between the acquired signal and the standard CVBS signal through convolution. The convolution process is equivalent to sliding matching of the signals, which can suppress the influence of short-term noise on the overall similarity judgment. Simultaneously, a phase adjustment term is introduced. (in This is the Frobenius norm of the frequency domain statistical vector difference, which can dynamically compensate for signal phase shifts caused by differences in frequency domain characteristics. It is particularly suitable for scenarios where phase distortion of FPV drone image transmission signals is caused by changes in transmission distance, obstruction, etc. For example, when the signal experiences phase shift due to multipath propagation, this phase adjustment term can... Real-time computation enables adaptive phase calibration, ensuring that the convolution results accurately reflect the essential similarity of the signals and improving recognition stability under frequency domain distortion.

[0095] The expression for the second difference calculation is:

[0096] ,

[0097] Where c2 is the second difference value. , and , which are the first, third, and fourth elements of the difference vector between the second statistical vector and the second statistical standard vector constructed from the set of time-domain statistical values, respectively.

[0098] The expression for calculating the second difference in this invention achieves dynamic adaptation to differences in time-domain statistical characteristics by adjusting the weighted average of the time-domain error in the numerator. Corresponding to the difference in means, Corresponding to the difference in average rise time, the introduction of these two factors allows the calculation process to automatically focus on the feature dimensions with the most significant differences in the time domain, enhancing sensitivity to signal waveform distortion. The denominator is obtained through... (Corresponding to the mode difference) Normalizing the standard signal energy effectively eliminates the impact of signal strength fluctuations (such as signal attenuation caused by changes in UAV distance) on difference calculation, ensuring that the c2 value can stably reflect the essential differences in waveforms under different signal strengths. For example, when the signal strength decreases due to power attenuation but the waveform remains unchanged, normalization can avoid misjudging it as an increase in difference. When the waveform is distorted due to interference, the weighted error term can significantly amplify the c2 value, improving the accuracy of time-domain feature discrimination.

[0099] The matching conditions are:

[0100] ,

[0101] ,

[0102] ,

[0103] Where b1, b2 and b3 are preset difference thresholds, and a1 and a2 are preset first standard difference values ​​and second standard difference values, respectively.

[0104] The preset set of standard statistical vectors is obtained by performing the above transformation on the standard CVBS signal and then statistically analyzing the frequency domain statistical value set and the time domain statistical value set of the transformed signal. The first statistical standard vector is constructed using the frequency domain statistical value set, and the second statistical standard vector is constructed using the time domain statistical value set.

[0105] The step of performing parameter synthesis and discrimination processing on the initial identification signal to obtain the identification signal includes:

[0106] S331, the bandwidth, average power and signal duration of the initial identification signal are statistically obtained;

[0107] S332, determine whether the bandwidth, average power and signal duration simultaneously satisfy bandwidth ≥ AMHz, average signal power ≥ BdB and burst duration ≥ Cms, and obtain the second determination result;

[0108] S333, if the second determination result is yes, determine the initial identification signal as an identification signal; if the second determination result is no, end the identification process;

[0109] The values ​​of A, B, and C can be 4, 20, and 9, respectively.

[0110] The parameter comprehensive discrimination processing of this invention further improves the reliability of recognition. By performing secondary discrimination on the initial recognition signal, combined with preset standard difference values ​​(a1, a2) and product constraint conditions (…), the reliability of recognition is further improved. This invention can effectively eliminate abnormal signals and false matches, reducing the false alarm rate. For example, in scenarios where multiple drones operate simultaneously, this invention can accurately distinguish target signals from other interference signals, ensuring the accuracy of monitoring results.

[0111] The SVMD stands for Successive Variational Mode Decomposition;

[0112] The acquisition of space electromagnetic wave signals based on the acquired parameter information can be achieved using a software radio platform or a communication signal receiver.

[0113] This invention also discloses an FPV unmanned aerial vehicle (UAV) identification method based on analog image transmission signals, comprising:

[0114] First, starting with a center frequency of 5650MHz, a bandwidth of 20MHz is set (to be determined based on receiver performance), and FFT and STFT calculations are performed to achieve spectrum monitoring.

[0115] Then, the time-domain characteristics of the signal, such as the average burst interval d and the average duration t, as well as the frequency-domain characteristics, such as the bandwidth B, the average signal power p, and the peak interval n, are matched with the signal.

[0116] Next, it is determined whether the above time-frequency parameters meet the thresholds, such as bandwidth B ≥ 4MHz, average signal power ≥ 20dB (statistical value), and burst duration ≥ 9ms. If the above conditions are met, it can be determined to be an analog image transmission signal. If the above conditions are not met, the next center frequency is set for frequency sweeping.

[0117] Finally, based on the above rules, the result is used to output a signal detection list. The specific process is as follows: Figure 1 As shown.

[0118] A second aspect of the present invention discloses an FPV unmanned aerial vehicle (UAV) identification device based on analog image transmission signals, the device comprising:

[0119] Memory containing executable program code;

[0120] A processor coupled to the memory;

[0121] The processor calls the executable program code stored in the memory to execute the FPV UAV identification method based on analog image transmission signals.

[0122] In a third aspect of this invention, a computer-storable medium is disclosed, wherein the computer-storable medium stores computer instructions, and when the computer instructions are invoked by a computer, they are used to execute the FPV UAV identification method based on analog image transmission signals.

[0123] In a fourth aspect of this invention, an information data processing terminal is disclosed, which is used to implement the FPV UAV identification method based on analog image transmission signals.

[0124] The above description is merely an embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of the present invention should be included within the scope of the claims of the present invention.

Claims

1. A method for identifying FPV unmanned aerial vehicles (UAVs) based on analog image transmission signals, characterized in that, include: S1, Set the data acquisition parameters; The collected parameter information includes the collection center frequency and the collection bandwidth; S2, based on the collected parameter information, the space electromagnetic wave signal is collected to obtain the collected signal; S3, perform identification processing on the acquired signal to obtain an identification signal, including: S31, the acquired signal is transformed to obtain a transformed signal set, including: S311, Perform frequency domain transformation on the acquired signal to obtain a frequency domain signal; S312, Perform SVMD transformation on the acquired signal to obtain a time-domain feature signal; S313, using the frequency domain signal and the time domain characteristic signal, a transformed signal set is constructed; S32, perform feature discrimination processing on the transformed signal set to obtain an initial identification signal, including: S321, Perform frequency domain feature statistical processing on the frequency domain signals in the transformed signal set to obtain a set of frequency domain statistical values; S322, Perform time-domain feature statistical processing on the time-domain feature signals in the transformed signal set to obtain a set of time-domain statistical values; S323, Perform feature matching and discrimination processing on the frequency domain statistical value set and the time domain statistical value set to obtain an initial identification signal, including: S3231, using the set of frequency domain statistical values, a first statistical vector is constructed; S3232, using the set of time-domain statistical values, a second statistical vector is constructed; S3233, perform difference fusion calculation on the first statistical vector and the second statistical vector with a preset set of standard statistical vectors to obtain the first difference value; S3234, determine whether the first difference value is less than a preset first discrimination threshold, and obtain the first discrimination result; If the first determination result is yes, the collected signal is determined to be the initial identification signal; if the first determination result is no, the identification process ends. S33, Perform parameter comprehensive discrimination processing on the initial identification signal to obtain the identification signal.

2. The FPV UAV identification method based on analog image transmission signals as described in claim 1, characterized in that, The expression for the difference fusion calculation is: ， in, This represents the j-th element of the i-th statistical vector. This represents the j-th element of the i-th statistical standard vector. r 1 indicates the first difference value.

3. An FPV unmanned aerial vehicle (UAV) identification device based on analog image transmission signals, characterized in that, The device includes: Memory containing executable program code; A processor coupled to the memory; The processor calls the executable program code stored in the memory to execute the FPV UAV identification method based on analog image transmission signals as described in any one of claims 1 to 2.

Images