Unmanned aerial vehicle identification method, device and equipment and readable storage medium

Abstract

The invention discloses an unmanned aerial vehicle identification method, device and equipment and a readable storage medium, and the method comprises the steps: carrying out the time-frequency conversion of a to-be-identified signal, obtaining conversion data, and judging whether there is an unmanned aerial vehicle image transmission signal in the to-be-identified signal or not based on the conversion data; if yes, determining a frequency domain where the image transmission signal of the unmanned aerial vehicle is located based on the conversion data as a selected frequency domain, processing the conversion data based on the selected frequency domain to obtain processed data, calculating the sum of each row and the sum of each column of the processed data, performing normalization processing, performing splicing to obtain spliced data, and performing data processing on the spliced data; and inputting the spliced data into the model to obtain an unmanned aerial vehicle identification result. According to the method, whether the unmanned aerial vehicle image transmission signals exist in the to-be-identified signals or not is judged, the frequency domain where the unmanned aerial vehicle image transmission signals are located is determined, the data are processed based on the selected frequency domain, the spliced data are finally obtained through calculation, the spliced data are finally processed through the model, the unmanned aerial vehicle identification result is obtained, and identification of the unmanned aerial vehicle is achieved.

Description

Technical Field

[0001] The present application relates to the technical field of unmanned aerial vehicles, and in particular, to a method, device, equipment and readable storage medium for identifying unmanned aerial vehicles. Background Art

[0002] In recent years, the unmanned aerial vehicle industry has developed rapidly and the technology has become increasingly mature. The characteristics of small size, flexible movement and easy operation of unmanned aerial vehicles have made them show great potential in many fields such as civil use and scientific research. For example, in fields such as aerial photography, environmental monitoring, geographical mapping and power line inspection, unmanned aerial vehicles play an important role. However, with the wide application of unmanned aerial vehicle technology, the number of unmanned aerial vehicles has increased rapidly. The improper and illegal use of unmanned aerial vehicles has also led to frequent public safety and privacy security problems, posing a serious threat to society and people's lives, attracting wide attention. Therefore, how to identify unmanned aerial vehicles has been a problem that people have been concerned about. Summary of the Invention

[0003] In view of this, the present application provides a method, device, equipment and readable storage medium for identifying unmanned aerial vehicles, so as to facilitate the identification of unmanned aerial vehicles.

[0004] In order to achieve the above object, the following solutions are proposed:

[0005] A method for identifying an unmanned aerial vehicle, comprising:

[0006] Obtaining a signal to be identified;

[0007] Performing time-frequency conversion on the signal to be identified to obtain conversion data;

[0008] Based on the conversion data, determining whether there is an unmanned aerial vehicle video transmission signal in the signal to be identified;

[0009] If it exists, based on the conversion data, determining the frequency domain where the unmanned aerial vehicle video transmission signal is located, and taking the determined frequency domain as a selected frequency domain;

[0010] Based on the selected frequency domain, processing the conversion data to obtain processed data;

[0011] Calculating the sum of each row and the sum of each column of the processed data respectively, and splicing the sum of each row and the sum of each column after normalization processing to obtain spliced data;

[0012] Inputting the spliced data into an unmanned aerial vehicle identification model to obtain an unmanned aerial vehicle identification result;

[0013] The unmanned aerial vehicle identification model is configured to have the ability to extract the input spliced data to obtain data features, and based on the data features, obtain an unmanned aerial vehicle identification result.

[0014] Optionally, determining whether there is a drone video transmission signal in the signal to be recognized based on the converted data includes:

[0015] Calculating the maximum value of each row in the converted data to obtain a first set;

[0016] Using a clustering algorithm, dividing the first set into two non - overlapping sets;

[0017] Respectively determining the maximum values in the two non - overlapping sets, comparing the two determined maximum values, and determining the smaller maximum value as the noise value;

[0018] Determining the number of data in the converted data that is greater than the noise value as a comparison value;

[0019] Calculating a reference value based on the sampling frequency and sampling time;

[0020] Comparing the comparison value with the reference value;

[0021] If the comparison value is greater than the reference value, there is a drone video transmission signal in the signal to be recognized.

[0022] Optionally, determining the frequency domain where the drone video transmission signal is located based on the converted data, and taking the determined frequency domain as the selected frequency domain, includes:

[0023] Calculating the sum of each row in the converted data to obtain a second set;

[0024] Using a clustering algorithm, dividing the second set into two non - overlapping sets;

[0025] Respectively determining the maximum values in the two non - overlapping sets, comparing the two determined maximum values, and taking the set corresponding to the larger maximum value as a preliminary selection set;

[0026] Based on the preliminary selection set, determining the frequency domain where the drone video transmission signal is located, and taking the determined frequency domain as the selected frequency domain.

[0027] Optionally, determining the frequency domain where the drone video transmission signal is located based on the preliminary selection set, and taking the determined frequency domain as the selected frequency domain, includes:

[0028] Calculating the longest continuous sequence of the preliminary selection set;

[0029] Taking the frequency domain where the longest continuous sequence is located as the frequency domain where the drone video transmission signal is located, and taking the determined frequency domain as the selected frequency domain.

[0030] Optionally, performing time - frequency conversion on the signal to be recognized to obtain converted data, includes:

[0031] Process the signal to be recognized through short-time Fourier transform to obtain the transformed data;

[0032] Process the transformed data to obtain the converted data.

[0033] Optionally, the UAV recognition model includes: an input layer, a data feature extraction layer, and a UAV recognition layer connected in series in sequence;

[0034] The training process of the UAV recognition model includes:

[0035] Obtain the spliced training data through the input layer;

[0036] Extract the data features of the training data through the data feature extraction layer;

[0037] Obtain the UAV recognition result based on the data features through the UAV recognition layer;

[0038] Taking the determined UAV recognition result to approach the UAV recognition result corresponding to the training data as the training objective, update the parameters of the UAV recognition model.

[0039] A UAV recognition device includes:

[0040] A signal acquisition module, configured to acquire a signal to be recognized;

[0041] A signal conversion module, configured to perform time-frequency conversion on the signal to be recognized to obtain the converted data;

[0042] A signal judgment module, configured to judge whether there is a UAV video transmission signal in the signal to be recognized based on the converted data; if so, determine the frequency domain where the UAV video transmission signal is located based on the converted data, and use the determined frequency domain as the selected frequency domain;

[0043] A data processing module, configured to process the converted data based on the selected frequency domain to obtain the processed data;

[0044] A data splicing module, configured to calculate the sum of each row and the sum of each column of the processed data respectively, and splice them after normalizing the sum of each row and the sum of each column to obtain the spliced data;

[0045] A UAV recognition module, configured to input the spliced data into a UAV recognition model to obtain a UAV recognition result, and the UAV recognition model is configured to have the ability to extract the input spliced data to obtain data features and obtain a UAV recognition result based on the data features.

[0046] Optionally, the signal judgment module executes a process of judging whether there is a drone video transmission signal in the signal to be recognized based on the converted data, including:

[0047] Calculate the maximum value of each row in the converted data to obtain a first set;

[0048] Use a clustering algorithm to divide the first set into two non-overlapping sets;

[0049] Respectively determine the maximum values in the two non-overlapping sets, compare the two determined maximum values, and determine the smaller maximum value as the noise value;

[0050] Determine the number of data in the converted data that is greater than the noise value as a comparison value;

[0051] Based on the sampling frequency and sampling time, calculate a reference value;

[0052] Compare the comparison value with the reference value;

[0053] If the comparison value is greater than the reference value, there is a drone video transmission signal in the signal to be recognized.

[0054] A drone recognition device includes: a memory and a processor;

[0055] The memory is used to store a program;

[0056] The processor is used to execute the program to implement each step of the drone recognition method as described in any one of the foregoing.

[0057] A readable storage medium stores a computer program, and when the computer program is executed by a processor, each step of the drone recognition method as described in any one of the foregoing is implemented.

[0058] As can be seen from the above technical solutions, a method, device, equipment, and readable storage medium for drone recognition provided by an embodiment of the present application obtain a signal to be recognized, perform time-frequency conversion on the signal to be recognized to obtain conversion data, and based on the conversion data, determine whether there is a drone video transmission signal in the signal to be recognized. If so, based on the conversion data, determine the frequency domain where the drone video transmission signal is located, use the determined frequency domain as the selected frequency domain, process the conversion data based on the selected frequency domain to obtain processed data, calculate the sum of each row and the sum of each column of the processed data respectively, and splice them after normalizing the sum of each row and the sum of each column to obtain spliced data. Input the spliced data into a drone recognition model to obtain a drone recognition result. The drone recognition model is configured to have the ability to extract the input spliced data to obtain data features and obtain a drone recognition result based on the data features. The present application determines whether there is a drone video transmission signal in the signal to be recognized, determines the frequency domain where the drone video transmission signal is located, processes the data based on the selected frequency domain, finally calculates and obtains the spliced data, and finally uses the model to process the spliced data to obtain a drone recognition result, realizing the recognition of drones. Further, after determining that there is a drone video transmission signal in the signal to be recognized, by determining the frequency domain where the drone video transmission signal is located and processing the conversion data based on the frequency domain, interference from other data can be excluded to a certain extent, making drone recognition more accurate. Description of the Drawings

[0059] To more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only the embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on the provided drawings.

[0060] Figure 1 It is a flowchart of a method for drone recognition provided by an embodiment of the present application;

[0061] Figure 2 It is a spectrogram obtained by processing a signal to be recognized provided by an embodiment of the present application;

[0062] Figure 3 It is a display diagram of a data splicing method provided by an embodiment of the present application;

[0063] Figure 4 It is a schematic structural diagram of a device for drone recognition provided by an embodiment of the present application;

[0064] Figure 5Hardware structure block diagram of a drone identification device disclosed in an embodiment of the present application. Detailed implementation manners

[0065] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present application without creative efforts shall fall within the protection scope of the present application.

[0066] Figure 1 A flowchart of a drone identification method provided in an embodiment of the present application may include the following steps:

[0067] Step S100, acquire a signal to be identified.

[0068] Specifically, software can be used to collect the signal. During the collection process, the sampling frequency r is set to be greater than 40 MHz, the signal is received with a fixed gain, and according to the preset sampling time t, the signal to be identified is acquired.

[0069] Step S101, perform time-frequency conversion on the signal to be identified to obtain conversion data.

[0070] Specifically, the size of each sampling window of the signal to be identified is s, and the number of sampling times w can be calculated according to the following formula:

[0071] w = r × t / s

[0072] After performing time-frequency conversion on the signal to be identified, conversion data x(i, j) is obtained, where 0 ≤ i ≤ s and 0 ≤ j ≤ w.

[0073] Step S102, based on the conversion data, determine whether there is a drone video transmission signal in the signal to be identified.

[0074] Specifically, based on the conversion data, determine whether there is a drone video transmission signal in the signal to be identified. If there is a drone video transmission signal in the signal to be identified, then execute step S103. If there is no drone video transmission signal in the signal to be identified, then return to step S100.

[0075] Step S103, based on the conversion data, determine the frequency domain where the drone video transmission signal is located, and use the determined frequency domain as the selected frequency domain.

[0076] Specifically, based on the conversion data x(i, j), the frequency domain where the drone video transmission signal is located can be determined as the selected frequency domain. After determining that there is a drone video transmission signal in the signal to be identified, by determining the frequency domain where the drone video transmission signal is located, the signal in the determined frequency domain can be analyzed.

[0077] Step S104: Process the converted data based on the selected frequency domain to obtain the processed data.

[0078] Specifically, after determining the selected frequency domain, the converted data can be processed based on the selected frequency domain to obtain the processed data, and the processed data is associated with the selected frequency domain.

[0079] Step S105: Calculate the sum of each row and the sum of each column of the processed data respectively, and splice them after normalizing the sum of each row and the sum of each column to obtain the spliced data.

[0080] Step S106: Input the spliced data into the UAV recognition model to obtain the UAV recognition result.

[0081] Specifically, the UAV recognition model is configured to have the ability to extract the input spliced data to obtain data features, and based on the data features, obtain the UAV recognition result. Among them, the UAV recognition result can include data such as the type and quantity of UAVs.

[0082] As can be seen from the above technical solution, a UAV recognition method provided by an embodiment of the present application includes: obtaining a signal to be recognized, performing time-frequency conversion on the signal to be recognized to obtain converted data, based on the converted data, determining whether there is a UAV video transmission signal in the signal to be recognized, if so, based on the converted data, determining the frequency domain where the UAV video transmission signal is located, using the determined frequency domain as the selected frequency domain, based on the selected frequency domain, processing the converted data to obtain the processed data, calculating the sum of each row and the sum of each column of the processed data respectively, and splicing them after normalizing the sum of each row and the sum of each column to obtain the spliced data, inputting the spliced data into the UAV recognition model to obtain the UAV recognition result, and the UAV recognition model is configured to have the ability to extract the input spliced data to obtain data features, and based on the data features, obtain the UAV recognition result. The present application realizes the recognition of UAVs by determining whether there is a UAV video transmission signal in the signal to be recognized, determining the frequency domain where the UAV video transmission signal is located, processing the data based on the selected frequency domain, finally calculating to obtain the spliced data, and finally using the model to process the spliced data to obtain the UAV recognition result. Further, after determining that there is a UAV video transmission signal in the signal to be recognized, by determining the frequency domain where the UAV video transmission signal is located and processing the converted data based on the frequency domain, the interference of other data can be excluded to a certain extent, making the UAV recognition more accurate.

[0083] In some embodiments of the present application, step S101: Performing time-frequency conversion on the signal to be recognized to obtain converted data may include:

[0084] S11. Process the signal to be recognized through short-time Fourier transform to obtain the transformed data.

[0085] Specifically, process the signal to be recognized through short-time Fourier transform, and each transformed data x f (i, j) is obtained.

[0086] S12. Process the transformed data to obtain the converted data.

[0087] Specifically, the transformed data can be processed through the following formula to obtain the converted data x(i, j):

[0088] x(i, j) = 10 log 10 (x f (i, j) 2 )

[0089] where 0 ≤ i ≤ s, 0 ≤ j ≤ w, s is the size of each sampling window, w is the number of sampling times, Figure 2 is the spectrogram obtained after processing the signal to be recognized.

[0090] In some embodiments of the present application, step S102. Based on the converted data, determine whether there is a UAV video transmission signal in the signal to be recognized, which may include the following steps:

[0091] S21. Calculate the maximum value of each row in the converted data to obtain the first set.

[0092] Specifically, the maximum value of each row in the converted data x(i, j) can be calculated according to the following formula:

[0093] x smax (si) = max{x(i, j)|i = si, 0 ≤ j ≤ w}

[0094] where 0 ≤ si ≤ s, si ≠ s / 2, s is the size of each sampling window, w is the number of sampling times.

[0095] S22. Use the clustering algorithm to divide the first set into two non-overlapping sets.

[0096] Specifically, the clustering algorithm K-means can be used to divide the first set x smax (si) into two non-overlapping sets, and at this time k = 2 for the clustering algorithm K-means.

[0097] S23. Determine the maximum values in the two non-overlapping sets respectively, compare the two determined maximum values, and determine the smaller maximum value as the noise value.

[0098] Specifically, from the two disjoint sets obtained in S22, find the maximum value in each set respectively, compare the maximum values corresponding to the two sets, and determine the smaller maximum value as the noise value.

[0099] S24. Determine the number of data in the converted data that is greater than the noise value as the comparison value.

[0100] Specifically, compare the determined noise value with the data in the converted data, determine the number of data greater than the noise value, and use this number as the comparison value a.

[0101] S25. Calculate the reference value based on the sampling frequency and sampling time.

[0102] Specifically, the reference value b can be calculated according to the following formula using the sampling frequency r and the sampling time t:

[0103] b = λ × r × t

[0104] Specifically, λ is a coefficient, and λ can be selected as 0.1.

[0105] S26. Compare the comparison value with the reference value.

[0106] Specifically, compare the comparison value a and the reference value b obtained above. If the comparison value a is greater than the reference value b, it can be considered that there is a drone video transmission signal in the signal to be recognized.

[0107] In some embodiments of the present application, step S103. Based on the converted data, determine the frequency domain where the drone video transmission signal is located, and use the determined frequency domain as the selected frequency domain, which may include:

[0108] S31. Calculate the sum of each row in the converted data to obtain the second set.

[0109] Specifically, the maximum value of each row in the converted data x(i, j) can be calculated according to the following formula:

[0110]

[0111] where 0 ≤ si ≤ s, s is the size of each sampling window, and w is the number of sampling times.

[0112] S32. Use the clustering algorithm to divide the second set into two disjoint sets.

[0113] Specifically, the clustering algorithm K-means can be used to divide the first set x ssum (si) into two disjoint sets, where k = 2 for the clustering algorithm K-means at this time.

[0114] S33. Determine the maximum values in two non - intersecting sets respectively, compare the two determined maximum values, and use the set corresponding to the larger maximum value as the pre - selected set.

[0115] Specifically, from the two non - intersecting sets obtained in S22, find the maximum value in each set respectively, compare the maximum values corresponding to the two sets, and use the set corresponding to the larger maximum value as the pre - selected set.

[0116] S34. Based on the pre - selected set, determine the frequency domain where the UAV video transmission signal is located, and use the determined frequency domain as the selected frequency domain.

[0117] Specifically, after determining the pre - set set, the frequency domain where the UAV video transmission signal is located can be predicted based on the pre - selected set, and the predicted and determined frequency domain is used as the selected frequency domain.

[0118] In some embodiments of the present application, S34. Based on the pre - selected set, determine the frequency domain where the UAV video transmission signal is located, and use the determined frequency domain as the selected frequency domain, which may include:

[0119] S41. Calculate the longest continuous sequence of the pre - selected set.

[0120] Specifically, the pre - selected set si obtained above t is the range where the UAV video transmission signal may exist. For the pre - selected set si t find the longest continuous sequence seq. Among them, the process of finding the longest continuous sequence seq for the pre - selected set si t can be to select two data si1 and si2 in the pre - selected set si t If |si1 - si2|≤con, then si1 and si2 are considered continuous, where con > 1, and con can be set to 5.

[0121] S42. Determine the frequency domain where the longest continuous sequence is located as the frequency domain where the UAV video transmission signal is located, and use the determined frequency domain as the selected frequency domain.

[0122] Specifically, the frequency domain where the longest continuous sequence seq obtained in the above steps is located can be considered as the frequency domain where the UAV video transmission signal is located. Among them, seqf is the head of seq and seql is the tail of seq.

[0123] Furthermore, the bandwidth bw = r / s×(seql - seqf) and the center frequency point cf = (seql + seqf) / 2×r / s - r / 2 + tlo can be determined, where tlo is the local oscillator frequency point set by the software radio.

[0124] Based on this, in step S104, the converted data is processed based on the selected frequency domain to obtain processed data, which may include: using the head seqf and tail seql of the longest continuous sequence seq in the selected frequency domain to process the converted data x(i, j) according to the following formula to obtain the processed data x l (i, j):

[0125]

[0126] During the above processing, the data within the selected frequency domain remains unchanged, and the data outside the selected frequency domain is considered not to be the UAV signal, so all are changed to the minimum value, so that the data outside the selected frequency domain has an impact on subsequent calculations.

[0127] In addition, during the process of processing the converted data based on the selected frequency domain to obtain the processed data, the data in the selected frequency domain can also be extracted separately, and only the extracted data is used for subsequent calculations.

[0128] Based on this, in step S105, the sum of each row and the sum of each column of the processed data are calculated respectively, and the sum of each row and the sum of each column are normalized and then spliced to obtain the spliced data, which may include: calculating the sum of each row of the processed data x l (i, j) x lssum (si):

[0129]

[0130] Among them, 0 ≤ si ≤ s, s is the size of each sampling window, and w is the number of sampling times.

[0131] Calculate the sum of each column of the processed data x according to the following formula l (i, j) x lwsum (wj):

[0132]

[0133] Among them, 0 ≤ wj ≤ w, s is the size of each sampling window, and w is the number of sampling times.

[0134] The sum of each row x l (i, j) x lssum (si) and the sum of each column x lwsum (wi) of the processed data are normalized and spliced to obtain the spliced data x lsum , and the splicing method can be as Figure 3 shown.

[0135] In some embodiments of the present application, the UAV recognition model may include: an input layer, a data feature extraction layer, and a UAV recognition layer connected in series. The training process of the UAV recognition model may include the following steps:

[0136] S51. Obtain the spliced training data through the input layer.

[0137] S52. Extract the data features of the training data through the data feature extraction layer.

[0138] S53. Obtain the UAV recognition result based on the data features through the UAV recognition layer.

[0139] S54. Take the determined UAV recognition result approaching the UAV recognition result corresponding to the training data as the training objective, and update the parameters of the UAV recognition model.

[0140] Next, a UAV recognition device provided by an embodiment of the present application will be described. The UAV recognition device described below can be correspondingly referred to the UAV recognition method described above.

[0141] Refer to Figure 4 as shown in Figure 4 which is a schematic structural diagram of a UAV recognition device provided by an embodiment of the present application. The UAV recognition device may include:

[0142] A signal acquisition module 10, configured to acquire a signal to be recognized;

[0143] A signal conversion module 20, configured to perform time-frequency conversion on the signal to be recognized to obtain converted data;

[0144] A signal judgment module 30, configured to judge whether there is a UAV video transmission signal in the signal to be recognized based on the converted data; if so, determine the frequency domain where the UAV video transmission signal is located based on the converted data, and use the determined frequency domain as the selected frequency domain;

[0145] A data processing module 40, configured to process the converted data based on the selected frequency domain to obtain processed data;

[0146] A data splicing module 50, configured to calculate the sum of each row and the sum of each column of the processed data respectively, and splice them after normalizing the sum of each row and the sum of each column to obtain spliced data;

[0147] A UAV recognition module 60, configured to input the spliced data into a UAV recognition model to obtain a UAV recognition result. The UAV recognition model is configured to have the ability to extract the input spliced data to obtain data features, and obtain a UAV recognition result based on the data features.

[0148] As can be seen from the above technical solution, an unmanned aerial vehicle (UAV) identification device provided in an embodiment of the present application includes: a signal acquisition module 10 that acquires a signal to be identified; a signal conversion module 20 that performs time-frequency conversion on the signal to be identified to obtain conversion data; a signal determination module 30 that, based on the conversion data, determines whether there is a UAV video transmission signal in the signal to be identified. If so, based on the conversion data, it determines the frequency domain where the UAV video transmission signal is located, and uses the determined frequency domain as the selected frequency domain; a data processing module 40 that processes the conversion data based on the selected frequency domain to obtain processed data; a data splicing module 50 that calculates the sum of each row and the sum of each column of the processed data respectively, and splices them after normalizing the sum of each row and the sum of each column to obtain spliced data; a UAV identification module 60 that inputs the spliced data into a UAV identification model to obtain a UAV identification result. The UAV identification model is configured to have the ability to extract the input spliced data to obtain data features, and based on the data features, obtain a UAV identification result. In the present application, by determining whether there is a UAV video transmission signal in the signal to be identified, determining the frequency domain where the UAV video transmission signal is located, processing the data based on the selected frequency domain, finally calculating to obtain the spliced data, and finally using the model to process the spliced data to obtain a UAV identification result, the identification of the UAV is realized. Further, after determining that there is a UAV video transmission signal in the signal to be identified, by determining the frequency domain where the UAV video transmission signal is located and processing the conversion data based on the frequency domain, the interference of other data can be excluded to a certain extent, making the UAV identification more accurate.

[0149] In some embodiments of the present application, the process by which the signal determination module 30 determines whether there is a UAV video transmission signal in the signal to be identified based on the conversion data may include:

[0150] Calculating the maximum value of each row in the conversion data to obtain a first set;

[0151] Using a clustering algorithm to divide the first set into two non-overlapping sets;

[0152] Respectively determining the maximum values in the two non-overlapping sets, comparing the two determined maximum values, and determining the smaller maximum value as the noise value;

[0153] Determining the number of data in the conversion data that is greater than the noise value as a comparison value;

[0154] Calculating a reference value based on the sampling frequency and the sampling time;

[0155] Comparing the comparison value with the reference value;

[0156] If the comparison value is greater than the reference value, there is a UAV video transmission signal in the signal to be identified.

[0157] In some embodiments of the present application, the signal judgment module 30 executes a process of determining the frequency domain where the drone video transmission signal is located based on the conversion data and using the determined frequency domain as the selected frequency domain, which may include:

[0158] Calculate the sum of each row in the conversion data to obtain a second set;

[0159] Use a clustering algorithm to divide the second set into two non - overlapping sets;

[0160] Determine the maximum values in the two non - overlapping sets respectively, compare the two determined maximum values, and use the set corresponding to the larger maximum value as the pre - selected set;

[0161] Based on the pre - selected set, determine the frequency domain where the drone video transmission signal is located, and use the determined frequency domain as the selected frequency domain.

[0162] In some embodiments of the present application, the signal judgment module 30 executes a process of determining the frequency domain where the drone video transmission signal is located based on the pre - selected set and using the determined frequency domain as the selected frequency domain, which may include:

[0163] Calculate the longest continuous sequence of the pre - selected set;

[0164] Determine the frequency domain where the longest continuous sequence is located as the frequency domain where the drone video transmission signal is located, and use the determined frequency domain as the selected frequency domain.

[0165] In some embodiments of the present application, the signal conversion module 20 executes a process of performing time - frequency conversion on the signal to be recognized to obtain conversion data, which may include:

[0166] Process the signal to be recognized through short - time Fourier transform to obtain transformed data;

[0167] Process the transformed data to obtain conversion data.

[0168] The embodiments of the present application further provide a drone recognition device, Figure 5 shows the hardware structure block diagram of the training device of the lane line annotation model. Refer to Figure 5 , the hardware structure of the drone recognition device may include: at least one processor 1, at least one communication interface 2, at least one memory 3, and at least one communication bus 4;

[0169] In the embodiments of the present application, the number of the processor 1, the communication interface 2, the memory 3, and the communication bus 4 is at least one, and the processor 1, the communication interface 2, and the memory 3 complete communication with each other through the communication bus 4;

[0170] The processor 1 may be a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present invention, etc.;

[0171] The memory 3 may include a high-speed RAM memory, and may also include a non-volatile memory, such as at least one disk memory;

[0172] Among them, the memory stores a program, and the processor can call the program stored in the memory. The program is used to: implement each processing flow in the foregoing UAV recognition method.

[0173] The embodiment of the present application also provides a storage medium, which can store a program suitable for execution by a processor. The program is used to: implement each processing flow in the foregoing UAV recognition method.

[0174] Finally, it should also be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements, but also includes other elements not expressly listed, or further includes elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "including a..." does not exclude the existence of additional identical elements in the process, method, article or device including the element.

[0175] The various embodiments in this specification are described in a progressive manner. The key point of each embodiment is to illustrate the differences from other embodiments. The various embodiments can be combined with each other, and the same or similar parts can be referred to each other.

[0176] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be obvious to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to these embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for identifying an unmanned aerial vehicle, characterized in that include: Acquire a signal to be identified; Performing time-frequency conversion on the signal to be identified to obtain conversion data; Based on the conversion data, determining whether there is a drone image transmission signal in the signal to be identified; If it exists, then based on the conversion data, determine the frequency domain where the drone image transmission signal is located, and use the determined frequency domain as the selected frequency domain; Based on the selected frequency domain, processing the converted data to obtain processed data; The sum of each row and the sum of each column of the processed data are calculated respectively, and the sum of each row and the sum of each column are normalized and then spliced ​​to obtain spliced ​​data; The spliced ​​data is input into the drone recognition model to obtain the drone recognition result; The drone identification model is configured to have the ability to extract the input spliced ​​data, obtain data features, and obtain drone identification results based on the data features.

2. The method according to claim 1, characterized in that The determining, based on the conversion data, whether there is a drone image transmission signal in the signal to be identified includes: Calculate the maximum value of each row in the converted data to obtain a first set; Using a clustering algorithm, dividing the first set into two disjoint sets; Determine the maximum values ​​in the two disjoint sets respectively, compare the two determined maximum values, and determine the smaller maximum value as the noise value; determining the number of data in the converted data that is greater than the noise value as a comparison value; Based on the sampling frequency and sampling time, the reference value is calculated; comparing the comparison value with the reference value; If the comparison value is greater than the reference value, then there is a drone image transmission signal in the signal to be identified.

3. The method according to claim 1, wherein The method of determining the frequency domain where the drone image transmission signal is located based on the conversion data, and using the determined frequency domain as the selected frequency domain, includes: Calculating the sum of each row in the converted data to obtain a second set; Using a clustering algorithm, dividing the second set into two disjoint sets; Determine the maximum values ​​in the two disjoint sets respectively, compare the two determined maximum values, and use the set corresponding to the larger maximum value as the pre-selected set; Based on the pre-selected set, the frequency domain in which the drone image transmission signal is located is determined, and the determined frequency domain is used as the selected frequency domain.

4. The method according to claim 3, wherein The determining, based on the preselected set, the frequency domain where the drone image transmission signal is located, and using the determined frequency domain as the selected frequency domain, includes: Calculating the longest continuous sequence of the preselected set; The frequency domain where the longest continuous sequence is located is determined as the frequency domain where the drone image transmission signal is located, and the determined frequency domain is used as the selected frequency domain.

5. The method according to claim 1, characterized in that The step of performing time-frequency conversion on the signal to be identified to obtain conversion data includes: Processing the signal to be identified by short-time Fourier transform to obtain transformed data; The transformed data is processed to obtain conversion data.

6. The method according to any one of claims 1-5, characterized in that The drone identification model comprises: an input layer, a data feature extraction layer, and a drone identification layer which are cascaded in sequence; The training process of the drone recognition model includes: Obtaining the spliced ​​training data through the input layer; Extracting data features of the training data through the data feature extraction layer; Through the UAV recognition layer, based on the data features, obtain the UAV recognition result; Taking the determined UAV recognition result approaching the UAV recognition result corresponding to the training data as the training objective, update the parameters of the UAV recognition model.

7. An unmanned aerial vehicle recognition device, characterized in that, Including: A signal acquisition module, configured to acquire a signal to be recognized; A signal conversion module, configured to perform time-frequency conversion on the signal to be recognized to obtain conversion data; A signal judgment module, configured to judge whether there is a UAV video transmission signal in the signal to be recognized based on the conversion data; If it exists, based on the conversion data, determine the frequency domain where the UAV video transmission signal is located, and use the determined frequency domain as the selected frequency domain; A data processing module, configured to process the conversion data based on the selected frequency domain to obtain processed data; A data splicing module, configured to calculate the sum of each row and the sum of each column of the processed data respectively, and splice them after normalizing the sum of each row and the sum of each column to obtain spliced data; A UAV recognition module, configured to input the spliced data into a UAV recognition model to obtain a UAV recognition result, and the UAV recognition model is configured to have the ability to extract the input spliced data to obtain data features and obtain a UAV recognition result based on the data features.

8. The device according to claim 7, characterized in that, The process of the signal judgment module executing the judgment of whether there is a UAV video transmission signal in the signal to be recognized based on the conversion data includes: Calculate the maximum value of each row in the conversion data to obtain a first set; Use a clustering algorithm to divide the first set into two non-overlapping sets; Determine the maximum values in the two non-overlapping sets respectively, compare the determined two maximum values, and determine the smaller maximum value as the noise value; Determine the number of data in the conversion data that is greater than the noise value as a comparison value; Calculate a reference value based on the sampling frequency and sampling time; Compare the comparison value with the reference value; If the comparison value is greater than the reference value, there is a UAV video transmission signal in the signal to be recognized.

9. An unmanned aerial vehicle identification device, characterized in that Including: A memory and a processor; The memory is used to store a program; The processor is used to execute the program to implement each step of the UAV recognition method according to any one of claims 1-6.

10. A readable storage medium, on which a computer program is stored, characterized in that, When the computer program is executed by the processor, each step of the UAV recognition method according to any one of claims 1-6 is implemented.

Images