Multi-mode and data fusion-based number calling voice recognition output system and method

Abstract

The invention relates to the technical field of big data, in particular to a calling voice recognition output system and method based on multi-mode and data fusion. The method comprises the following steps: acquiring voice data, analyzing a voice instruction, searching full-field ADSB information, performing number matching, outputting a matching result, and generating standardized radio voice data; a bilingual system automatic switching algorithm is realized; various standard compatible processing of alphanumeric pronunciation is realized; two calling fuzzy matching functions, a front and back nasal interference processing algorithm and a weight processing method in the mode are realized; identifying a calling number by adopting a multi-source auxiliary means; according to the airline Chinese abbreviation, aircraft number calling identification is carried out; the system comprises an audio processing module, a multi-source auxiliary module, an instruction control and processing module and a number matching module. The voice instruction is analyzed, the whole ADSB information is searched, and a multi-source auxiliary means is adopted to recognize the call number, so that the recognition performance in a complex environment is improved, and the real-time requirement is met.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of big data, and particularly relates to a call sign voice recognition output system and method based on multi-modal and data fusion. BACKGROUND

[0002] Call sign voice recognition technology has important applications in aviation, maritime affairs, emergency communication and other fields, mainly used for automatically identifying and recording call sign information in communication. At present, this technology is mainly based on traditional speech recognition methods (such as hidden Markov model, Gaussian mixture model) and deep learning models (such as recurrent neural network, convolutional neural network), and call sign recognition is achieved through acoustic feature extraction and pattern matching. However, the existing technology still has the following problems:

[0003] Insufficient noise resistance: There are often background noise, electromagnetic interference and speech distortion in actual communication environment, which causes the accuracy of traditional speech recognition model to decrease significantly.

[0004] Similar call signs are easily confused: Some call signs are similar in pronunciation (such as "B" and "D", "M" and "N"), and existing models are difficult to effectively distinguish, which easily causes misrecognition.

[0005] Limited support for low-resource languages: There is a lack of high-quality training data for call sign recognition of small languages or specific industry terms (such as aviation letter interpretation method), and the model has poor generalization.

[0006] High real-time requirement: The existing end-to-end deep learning model has high computational complexity, and it is difficult to realize real-time processing on low-power devices (such as portable radio terminals).

[0007] Therefore, there is an urgent need for a call sign voice recognition method with high robustness and high precision to improve the recognition performance in complex environments and meet the real-time requirements. SUMMARY

[0008] The purpose of the present application is to provide a call sign voice recognition output system and method based on multi-modal and data fusion, aiming to improve the recognition performance in complex environments and meet the real-time requirements.

[0009] To achieve the above purpose, a call sign voice recognition output method based on multi-modal and data fusion is adopted, which includes the following steps:

[0010] Acquire voice data, parse voice instructions and search for all-field ADSB information, perform call sign matching, output matching results, and generate standardized radio voice data;

[0011] Implement a bilingual system automatic switching algorithm;

[0012] Implement multiple standard compatible processing of alphanumeric pronunciation;

[0013] Two call sign fuzzy matching functions are realized, and a front-back nasal sound interference processing algorithm and a weight processing method in this mode are provided.

[0014] Real-time fusion of ADSB data, ACDM data, laser radar data, and video image data is used to identify the call sign, and call sign data is recorded.

[0015] According to the Chinese abbreviation of the airline, the aircraft call sign is identified.

[0016] Among them, in the steps of acquiring voice data, analyzing voice instructions and searching for full-field ADSB information, call sign matching, outputting matching results, and generating standardized radio voice data:

[0017] Confirm the call object, get the call trigger prefix, and output the call sign data;

[0018] Get the call sign data and convert the call sign data into an expression;

[0019] According to the expression, the call sign is matched, and the matching result is output, and the standardized radio voice data is generated.

[0020] Among them, in the steps of acquiring voice data, analyzing voice instructions and searching for full-field ADSB information, call sign matching, outputting matching results, and generating standardized radio voice data:

[0021] In the automatic mode, the recognition mode is switched according to the quality of the call channel.

[0022] Among them, in the steps of call sign matching according to the expression, and outputting the matching result, and generating standardized radio voice data:

[0023] When the expression is an aircraft, search the flight plan and the current full-field ADSB information, and perform call sign matching;

[0024] When the expression is a vehicle, search the vehicle database and the current full-field ADSB information, and perform call sign matching.

[0025] Among them, in the step of switching the recognition mode according to the quality of the call channel:

[0026] According to the background noise and signal power, the SNR (signal-to-noise ratio) is calculated, and the recognition mode is switched;

[0027] When the signal-to-noise ratio is high, the recognition mode is single-word replacement mode;

[0028] When the signal-to-noise ratio is medium, the recognition mode is double-word replacement mode;

[0029] When the signal-to-noise ratio is low, the recognition mode is star sound replacement mode.

[0030] In the step of identifying the call sign by using the multi-source auxiliary means of real-time fusion of ADSB data, ACDM data, laser radar data and video image data respectively, and recording the call sign data, the specific process of the ACDM data is as follows:

[0031] The flight call sign in the next two hours is preloaded to control the identification range.

[0032] In the step of identifying the call sign by using the multi-source auxiliary means of real-time fusion of ADSB data, ACDM data, laser radar data and video image data respectively, and recording the call sign data, the specific process of the laser radar data and the video image data is as follows:

[0033] The airport vehicle database information is read every ten minutes to record the call sign;

[0034] The flight information in the next two hours is read every minute to record the call sign;

[0035] The ADSB information is read in real time to record the call sign;

[0036] The image reports the target without a call sign, and records a temporary call sign; wherein the image acquisition sources include a camera and a laser radar;

[0037] When the fuzzy matching result is greater than 1, the call sign of the target has higher matching priority in the visible range.

[0038] In the step of identifying the aircraft call sign according to the Chinese abbreviation of the airline:

[0039] The airline two-digit code and three-digit code are replaced according to the Chinese abbreviation of the airline, and the last four digits of the call sign are connected to complete the call sign identification.

[0040] The application also provides a call sign voice recognition output system based on multi-modal and data fusion, which comprises an audio processing module, a bilingual system switching module, a compatible processing module, a fuzzy matching module, a multi-source auxiliary module, an instruction control and processing module; wherein:

[0041] The audio processing module is used to acquire voice data, analyze voice instructions and search for full-field ADSB information, perform call sign matching, output matching results, and generate standardized radio voice data;

[0042] The bilingual system switching module is used to realize a bilingual system automatic switching algorithm.

[0043] The compatible processing module is used to realize a plurality of standard compatible processing of alphanumeric pronunciation.

[0044] The fuzzy matching module is used to realize two call sign fuzzy matching functions, and a front and rear nasal sound interference processing algorithm and a weight processing method under the mode.

[0045] The multi-source auxiliary module is used to respectively adopt multi-source auxiliary means of real-time fusion of ADSB data, ACDM data, laser radar data and video image data to identify the call sign, and record the call sign data.

[0046] The instruction control and processing module is used to identify the aircraft call sign according to the Chinese abbreviation of the airline.

[0047] The call sign voice recognition output system and method based on multi-modal and data fusion of the present application respectively adopt the audio processing module, the bilingual system switching module, the compatible processing module, the fuzzy matching module, the multi-source auxiliary module, and the instruction control and processing module to perform the following steps: obtaining voice data, analyzing voice instructions and searching for full-field ADSB information, performing call sign matching, outputting matching results, and generating standardized radio voice data; realizing bilingual system automatic switching algorithm; realizing multiple standard compatible processing of alphanumeric pronunciation; realizing two call sign fuzzy matching functions, and a front and rear nasal sound interference processing algorithm and a weight processing method under the mode; respectively adopting multi-source auxiliary means of real-time fusion of ADSB data, ACDM data, laser radar data and video image data to identify the call sign, and recording the call sign data; identifying the aircraft call sign according to the Chinese abbreviation of the airline; and improving the recognition performance in complex environments and meeting the real-time requirements by analyzing voice instructions and searching for full-field ADSB information, and using multi-source auxiliary means to identify the call sign. BRIEF DESCRIPTION OF DRAWINGS

[0048] In order to more clearly illustrate the technical solutions of the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiment or prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and those skilled in the art can also obtain other drawings according to these drawings without creative labor.

[0049] Figure 1 is the step flow chart of the call sign voice recognition output method based on multi-modal and data fusion of the present application.

[0050] Figure 2 is the step flow chart of S100 of the present application.

[0051] Figure 3 is the flowchart of the switching process of the present application.

[0052] Figure 4 is the step flow chart of identifying the call sign by using laser radar data and video image data of the present application.

[0053] Figure 5 is the structural principle diagram of the call sign voice recognition output system based on multi-modal and data fusion of the application.

[0054] Figure 6 is the structural principle diagram of the electronic device of the application.

[0055] 701-audio processing module, 702-bilingual system switching module, 703-compatible processing module, 704-fuzzy matching module, 705-multi-source auxiliary module, 706-instruction control and processing module. DETAILED DESCRIPTION

[0056] The exemplary embodiments will be described in detail herein with reference to the attached drawings. In the following description, the same numbers are used to designate the same elements, unless otherwise indicated. The embodiments described in the following exemplary embodiments are not intended to represent all embodiments consistent with the present application.

[0057] The terminology used in the present application is for the purpose of describing particular embodiments only and is not intended to be limiting of the present application. As used in the description of the application and the appended claims, the singular forms "a", "an" and "the" are intended to include plural forms as well, unless the context clearly indicates otherwise. It will also be understood that the term "and / or" as used herein refers to and encompasses any and all possible combinations of one or more of the associated listed items.

[0058] It will be understood that, although the terms first, second, third, etc. can be used herein to describe various information, these terms are not intended to denote a temporal or chronological order. Rather, these terms are used only to distinguish different sets of information from one another. For example, a first information can be termed a second information, and similarly, a second information can also be termed a first information, without departing from the scope of the present application. As used herein, the word "if' can be construed to mean "when" or "in response to determining" depending on the context.

[0059] Referring to Figures 1-4 The present application provides a call sign voice recognition output method based on multi-modal and data fusion, comprising the following steps:

[0060] S100: acquiring voice data, parsing voice instructions and searching for full-field ADSB information, performing call sign matching, outputting matching results, and generating standardized radio voice data;

[0061] S200: implementing a bilingual system automatic switching algorithm;

[0062] S300: implementing multi-standard compatible processing of alphanumeric pronunciation;

[0063] S400: two call sign fuzzy matching functions are implemented, and the front and rear nasal sound interference processing algorithm and the weight processing method in this mode are implemented;

[0064] S500: multi-source auxiliary means of real-time fusion of ADSB data, ACDM data, laser radar data, and video image data are used to identify the call sign, and the call sign data is recorded;

[0065] S600: aircraft call sign identification is performed according to the Chinese abbreviation of the airline.

[0066] In the embodiment, first, the voice data is acquired, the voice instruction is parsed, the full-field ADSB information is searched, the call sign is matched, the matching result is output, and the standardized radio voice data is generated; then, the bilingual automatic switching algorithm is implemented; then, the multi-standard compatible processing of the alphanumeric pronunciation is implemented; then, two call sign fuzzy matching functions are implemented, and the front and rear nasal sound interference processing algorithm and the weight processing method in this mode are implemented; multi-source auxiliary means of real-time fusion of ADSB data, ACDM data, laser radar data, and video image data are used to identify the call sign, and the call sign data is recorded; finally, aircraft call sign identification is performed according to the Chinese abbreviation of the airline; by parsing the voice instruction and searching the full-field ADSB information, and using multi-source auxiliary means to identify the call sign, the accuracy of call sign matching is improved.

[0067] Among them, in the process of acquiring voice data, parsing voice instructions, searching full-field ADSB information, matching call signs, outputting matching results, and generating standardized radio voice data, the specific process is as follows:

[0068] S101: confirm the call object, acquire the call trigger prefix, and output the call sign data; wherein the call trigger prefix includes the aircraft call sign and the vehicle type;

[0069] S102: acquire the call sign data and convert the call sign data into an expression;

[0070] S103: perform call sign matching according to the expression, output the matching result, and generate standardized radio voice data; when the expression is an aircraft, search the flight plan and the current full-field ADSB information to perform call sign matching; when the expression is a vehicle, search the vehicle database and the current full-field ADSB information to perform call sign matching;

[0071] S104: switch the identification mode according to the channel quality; calculate the SNR (signal-to-noise ratio) according to the background noise and signal power, and switch the identification mode; when the signal-to-noise ratio is high, the identification mode is single-word replacement mode; when the signal-to-noise ratio is medium, the identification mode is double-word replacement mode; when the signal-to-noise ratio is low, the identification mode is star sound replacement mode.

[0072] "Star" pronunciation replaces the letter pronunciation in the call sign, based on the function of multi-source auxiliary correction, directly realizes fuzzy matching. Among them, without letter segmentation mapping:

[0073] Currently there is an aircraft call sign "CA1234" on the taxiway, the star sound call should be "star zero two three four", at this time, because it is an aircraft call sign, it is almost all letter + number combination, a "star" sound is equivalent to two letters. If it is a three-letter call sign, such as "CCA5678", then the star sound call is "star five six seven eight", a "star" is equivalent to three letters. For the special call sign "3U8888", the star sound reading should be "three stars eight eight eight".

[0074] For vehicle call signs, there is no aircraft rule, the following principles need to be adopted:

[0075] The call sign starts with "civil aviation" in Chinese, which is generally simplified as "MH" on the system, and is directly omitted when calling vehicles

[0076] Through "star", "double star" and "multi-star", one letter, two letters or multiple letters are matched, such as "MHY305", the star sound reading is "star three hole five", "MHAB123", the star sound reading is "double star zero two three", and there is basically no "multi-star" reading.

[0077] Among them, with letter segmentation mapping:

[0078] As an extension of star sound design, the single letter replacement in a specific range is realized through "empty" (unvoiced), which needs to be checked against the mapping table. Only when "empty" is followed by a reasonable number can the correct mapping of letters be completed. For example, "MHAB123", the reading should be "empty 3 empty 8 zero two three", and for example, "MHW507", the reading should be "empty 4 five hole turn".

[0079] In the above mapping: when calling an aircraft, the trigger prefix is "call sign"; when calling a vehicle, the trigger prefix is "vehicle" for vehicles of uncertain type, and the prefix is "truck", "tractor", "fire truck" and the like for vehicles of specific type.

[0080] Read the call sign according to the rules, and identify each letter and number in turn to convert it into a reasonable regular expression.

[0081] If it is an aircraft, search the preprocessed flight plan and current field ADSB information; if it is a vehicle, search the vehicle database and current field ADSB information for call sign matching.

[0082] In the case of multiple matching results, sort them according to the identification results, and after voice prompt, the user confirms the selection.

[0083] In the identification mode switching:

[0084] High SNR uses single-character replacement, with high efficiency and stability;

[0085] Medium SNR uses double-character replacement, with stability as the priority;

[0086] Low SNR uses star replacement, with anti-interference as the priority.

[0087] Calculate SNR, where:

[0088] Signal Power is:

[0089] Signal Power = sum(abs(Signal).^2) / length(Signal);

[0090] Noise Power is:

[0091] Noise Power = sum(abs(Noise).^2) / length(Noise);

[0092] SNR is:

[0093] SNR = 10*log10(Signal Power / Noise Power);

[0094] When the channel quality is excellent (SNR > 25 dB), automatically switch to single-character mode, and use the display device to display the current mode "single-character / automatic"; when the channel quality is average (SNR < 25 dB), automatically switch to double-character mode, and use the display device to display the current mode "double-character / automatic". The switching process is shown in Figure 3 .

[0095] At the same time:

[0096] Whether in star mode or not, star rules can be used in "double-character mode" or "single-character mode";

[0097] Double-character and single-character mode rules cannot be used at the same time;

[0098] "Star" and "empty" replacement rules can be used at the same time.

[0099] Among them, in the multi-source auxiliary means of real-time fusion ADSB data, ACDM data, laser radar data and video image data respectively to identify the call sign, and record the call sign data:

[0100] The real-time fused ADSB data shortens the spatial recognition range and improves the recognition rate. The ADSB messages broadcasted by the aircraft are received in real time by ADSB receiving equipment (such as ground stations or airborne receivers). These messages contain key information such as the position, speed, altitude, flight number (call sign), and aircraft model of the aircraft. The received ADSB messages exist in various encoding formats and need to be decoded to extract structured information fields such as flight number (call sign), latitude and longitude, and altitude. The decoded data is checked to eliminate incorrect or invalid messages (such as checksum errors, format abnormalities, etc.). A timestamp is added to each ADSB data to ensure the timeliness and consistency of the data. The latitude and longitude coordinates in the ADSB message are converted to the coordinate system used internally by the system (such as UTM or local Cartesian coordinate system) for subsequent processing. A dynamic aircraft status table is maintained to record the latest position, speed, altitude, and other information of each aircraft. Each time a new ADSB message is received, the system updates the status of the corresponding aircraft based on the flight number (call sign). If other data sources (such as radar data, flight plan data, etc.) are accessed simultaneously, the ADSB data needs to be associated and fused with other data. For example:

[0101] Flight number matching: match the flight number in the ADSB message with the flight number in the flight plan database to confirm the identity of the aircraft.

[0102] Position fusion: if the same aircraft is monitored by both ADSB and radar, the system will perform weighted fusion based on the data quality (such as accuracy, timeliness) of both to generate more accurate position information.

[0103] Full-field ADSB information update: update the fused ADSB data to the global aircraft status table to ensure that the latest aircraft information can be obtained.

[0104] The specific process of ACDM data is to preload the flight call sign within two hours and reasonably control the recognition range.

[0105] The specific process of laser radar data and video image data is as follows:

[0106] S531: read the airport vehicle database information every ten minutes, and include the call sign;

[0107] S532: read the flight information within two hours every minute, and include the call sign;

[0108] S533: read the ADSB information in real time, and include the call sign;

[0109] S534: report the target without a call sign on the image, and include the temporary call sign; the image sources include cameras and laser radars;

[0110] S535: When there are more than one fuzzy matching results, the target's call sign has a higher matching priority within the visible range.

[0111] Among them, targets without call signs reported by video and lidar (ended or not started), or towed targets created by apron management, include temporary call signs; video and lidar lidar data and video image data are assisted to deal with special situations without call signs; the temporary call sign generation rule is: "T" + time and minute (2 digits) + [geographic coordinates or device ID (3 digits)] to generate the call sign.

[0112] In the voice command, if the target is within visual range, just read out the call sign; to match all call signs, add the prefix "global" to the voice command.

[0113] Among them, when identifying aircraft call signs based on the Chinese abbreviation of the airline:

[0114] Replace the two- and three-character airline codes with the Chinese abbreviation of the airline, and then connect to the last four digits of the call sign to complete call sign identification. For example:

[0115] Example 1:

[0116] There is a flight "CZ3456" on the interface, and the voice command uses: "call sign" + "China Southern Airlines" + "3456" for identification.

[0117] Example 2:

[0118] There is a flight "SC4568" on the interface, and the voice command uses: "call sign" + "Shandong Airlines" + "four five six eight" for identification.

[0119] It can support dual-mode operation of ICAO call standards and localized rules:

[0120] International mode: strictly follow ICAO Annex 10 voice standards.

[0121] Domestic mode: Compatible with China Civil Aviation special situation call abbreviation standards (such as "dong, yao, guai", etc.).

[0122] Dynamic vocabulary switching: Automatically select the standard based on the aircraft registration number prefix (B-→China, N-→ICAO)

[0123] Use the prefix "callsign" to directly enter the international mode, and the prefix "callsign" to directly enter the domestic mode. The background dual language recognition is enabled at the same time, without manual operation.

[0124] Corresponding to the aforementioned embodiment of the call sign voice recognition and output method based on multimodality and data fusion, the present application also provides an embodiment of a call sign voice recognition and output system based on multimodality and data fusion.

[0125] Figure 5 is a multi-modal and data fusion based call sign voice recognition output system block diagram according to an exemplary embodiment. Referring to Figure 5 , the system can include: an audio processing module 701, a bilingual system switching module 702, a compatible processing module 703, a fuzzy matching module 704, a multi-source auxiliary module 705, an instruction control and processing module 706; wherein:

[0126] The audio processing module 701 is configured to acquire voice data, parse voice instructions and search for full-field ADSB information, perform call sign matching, output matching results, and generate standardized radio voice data.

[0127] The bilingual system switching module 702 is configured to implement a bilingual system automatic switching algorithm.

[0128] The compatible processing module 703 is configured to implement a variety of standard compatible processing of alphanumeric pronunciation.

[0129] The fuzzy matching module 704 is configured to implement two call sign fuzzy matching functions, and a front and rear nasal sound interference processing algorithm and a weight processing method in this mode.

[0130] The multi-source auxiliary module 705 is configured to use multi-source auxiliary means of real-time fusion of ADSB data, ACDM data, laser radar data, and video image data to recognize call signs, and record call sign data.

[0131] The instruction control and processing module 706 is configured to recognize aircraft call signs according to airline Chinese abbreviations.

[0132] In the embodiment, the audio processing module 701 acquires voice data, parses voice instructions and searches for full-field ADSB information, performs call sign matching, outputs matching results, and generates standardized radio voice data; the bilingual system switching module 702 implements a bilingual system automatic switching algorithm; the compatible processing module 703 implements a variety of standard compatible processing of alphanumeric pronunciation; the fuzzy matching module 704 implements two call sign fuzzy matching functions, and a front and rear nasal sound interference processing algorithm and a weight processing method in this mode; the multi-source auxiliary module 705 uses multi-source auxiliary means of real-time fusion of ADSB data, ACDM data, laser radar data, and video image data to recognize call signs, and records call sign data; the instruction control and processing module 706 recognizes aircraft call signs according to airline Chinese abbreviations; by parsing voice instructions and searching for full-field ADSB information, and using multi-source auxiliary means to recognize call signs, the recognition performance in complex environments is improved, and real-time requirements are met.

[0133] As to the system in the above-mentioned embodiments, the specific manner in which the various modules perform operations has been described in detail in the embodiments related to the method, and will not be described in detail here.

[0134] For the system embodiments, since they basically correspond to the method embodiments, the relevant parts are referred to the part of the method embodiments. The above-described device embodiments are only schematic, wherein the units shown as separate components can or can not be physically separate, and the components shown as units can or can not be physical units, i.e., can be located in one place or distributed on multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the present application. Those skilled in the art can understand and implement without creative labor.

[0135] Correspondingly, the present application also provides an electronic device, comprising: one or more processors; a memory for storing one or more programs; when the one or more programs are executed by the one or more processors, the one or more processors implement the high-trust learning resource security sharing method based on big data as described above. As Figure 6 As shown in the figure, a hardware structure diagram of an electronic device provided by the embodiment of the present application is provided. In addition to the processor, the memory and the network interface shown in the figure, the electronic device provided by the embodiment of the present application can also include other hardware according to the actual function of the electronic device. Figure 6 As shown in the figure, a hardware structure diagram of an electronic device provided by the embodiment of the present application is provided. In addition to the processor, the memory and the network interface shown in the figure, the electronic device provided by the embodiment of the present application can also include other hardware according to the actual function of the electronic device.

[0136] Correspondingly, the present application also provides a computer readable storage medium, which stores computer instructions, and the instructions are executed by a processor to implement the high-trust learning resource security sharing method based on big data as described above. The computer readable storage medium can be an internal storage unit of any data processing capable device, such as a hard disk or a memory. The computer readable storage medium can also be an external storage device, such as a plug-in hard disk, a smart media card (SMC), an SD card, a flash card, etc. Further, the computer readable storage medium can include both the internal storage unit and the external storage device. The computer readable storage medium is used to store the computer program and other programs and data required by the data processing capable device, and can also be used to temporarily store data that has been output or will be output.

[0137] Other embodiments of the application will be apparent to those skilled in the art from consideration of the specification and practice of the application disclosed herein. It is intended that the application embrace any and all variations, modifications, and adaptations of the application described herein, which are within the scope of the general inventive concept and include those expressly described herein, as well as other combinations of features, functions, and concepts included in the present disclosure or that are inherent in this field of technology.

[0138] It is to be understood that the application is not limited to the precise construction here described and as shown in the drawings, and that various modifications and changes can be made by those skilled in the art without departing from the scope of the application.

Claims

1. A call sign speech recognition and output method based on multimodality and data fusion, characterized in that: The steps include: Acquire voice data, parse voice commands, search for full-field ADSB information, perform call sign matching, output matching results, and generate standardized radio voice data; Implement bilingual automatic switching algorithm; Achieve multi-standard compatible processing of alphanumeric pronunciation; Implement two call sign fuzzy matching functions, front and back nasal sound interference processing algorithm and weight processing method under this mode; Call sign recognition is performed using multi-source auxiliary means that fuse ADSB data, ACDM data, lidar data, and video image data in real time, and the call sign data is recorded; Identify the aircraft call sign based on the Chinese abbreviation of the airline.

2. The call sign speech recognition and output method based on multimodality and data fusion according to claim 1, characterized in that: In the steps of acquiring voice data, parsing voice commands, searching for full-field ADSB information, performing call sign matching, outputting matching results, and generating standardized radio voice data: Confirm the call object, obtain the call trigger prefix, and output the call sign data; Get call sign data and convert the call sign data into expressions; Call sign matching is performed according to the expression, and the matching results are output to generate standardized radio voice data.

3. The call sign speech recognition and output method based on multimodality and data fusion as claimed in claim 2, characterized in that: In the steps of acquiring voice data, parsing voice commands, searching for full-field ADSB information, performing call sign matching, outputting matching results, and generating standardized radio voice data: In automatic mode, the recognition mode is switched according to the call channel quality.

4. The call sign speech recognition and output method based on multimodality and data fusion as claimed in claim 2, characterized in that: In the step of matching call signs according to expressions, outputting matching results, and generating standardized radio voice data: When the expression is an aircraft, the flight plan and current full-field ADSB information are searched for call sign matching; When the expression is a vehicle, the vehicle database and the current full-field ADSB information are searched for call sign matching.

5. The call sign speech recognition and output method based on multimodality and data fusion as claimed in claim 3, characterized in that: In the step of switching the identification mode according to the call channel quality: Calculate SNR (signal-to-noise ratio) based on background noise and signal power, and switch recognition mode; When the signal-to-noise ratio is high, the recognition mode is the single-word replacement mode; When the signal-to-noise ratio is medium, the recognition mode is the double-word replacement mode; When the signal-to-noise ratio is low, the recognition mode is the star-sound replacement mode.

6. The call sign speech recognition and output method based on multimodality and data fusion according to claim 1, wherein: In the steps of using real-time fusion of ADSB data, ACDM data, lidar data, and video image data to identify call signs and record call sign data, the specific process of ACDM data is as follows: Preload flight call signs for the next two hours to control the recognition range.

7. The call sign speech recognition and output method based on multimodality and data fusion according to claim 1, wherein: In the steps of identifying call signs by using a multi-source auxiliary method of real-time fusion of ADSB data, ACDM data, lidar data, and video image data, and recording the call sign data, the specific process of the lidar data and video image data is as follows: Read the airport vehicle database information every ten minutes and record the call signs; Read flight information within two hours every one minute and record call signs; Read ADSB information in real time and record call signs; Images are reported for targets without call signs, with temporary call signs included; the image acquisition sources include cameras and lidar; When there are more than one fuzzy matching results, the target's call sign has a higher matching priority within the visible range.

8. The call sign speech recognition and output method based on multimodality and data fusion as claimed in claim 1, characterized in that: In the steps of identifying the aircraft call sign based on the Chinese abbreviation of the airline: Replace the two- and three-character airline codes with the airline's Chinese abbreviation, then connect to the last four digits of the call sign to complete call sign identification.

9. A call sign speech recognition and output system based on multimodal and data fusion, applied to the call sign speech recognition and output method based on multimodal and data fusion as claimed in claim 1, characterized in that: It includes audio processing module, bilingual switching module, compatibility processing module, fuzzy matching module, multi-source auxiliary module, and instruction control and processing module; among them: The audio processing module is used to obtain voice data, analyze voice commands and search for full-field ADSB information, perform call sign matching, output matching results, and generate standardized radio voice data; The bilingual switching module is used to implement a bilingual automatic switching algorithm; The compatible processing module is used to implement multiple standard compatible processing of alphanumeric pronunciation; The fuzzy matching module is used to implement two call sign fuzzy matching functions, a front and back nasal sound interference processing algorithm, and a weight processing method under this mode; The multi-source auxiliary module is used to identify call signs by using multi-source auxiliary means of real-time fusion of ADSB data, ACDM data, lidar data and video image data, and record call sign data; The command control and processing module is used to identify the aircraft call sign according to the Chinese abbreviation of the airline.

Images