Off-line voice recognition threshold switching method and system based on Bluetooth communication technology, electronic equipment and storage medium
By manually adjusting the speech recognition threshold offline using Bluetooth communication technology, the adaptability of the speech recognition system in different environments is solved, achieving low-cost and efficient speech recognition threshold switching.
Patent Information
- Application Number
- CN202511072382.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-07-31
- Publication Date
- 2025-11-07
AI Technical Summary
Existing speech recognition systems have poor adaptability in different environments, fixed threshold settings cannot meet the diverse needs of users, offline devices cannot switch speech recognition thresholds, and high computing power models increase deployment costs.
In offline mode, the target device and the terminal device establish a connection via Bluetooth communication technology. The terminal device displays an array of recognition thresholds for wake words and command words. Users can manually adjust and update the recognition thresholds of the target device. This lightweight solution uses Bluetooth command switching and a pre-stored recognition threshold array.
It enables efficient switching of speech recognition thresholds in offline mode, reduces equipment costs, improves applicability, and eliminates the need for high-computing-power speech recognition models.
Smart Images
Figure CN120913554A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application belongs to the technical field of speech recognition, and particularly relates to an offline speech recognition threshold switching method and system based on Bluetooth communication technology, an electronic device, and a storage medium. BACKGROUND
[0002] Speech recognition (ASR) refers to the process of converting speech signals into text or instructions, and the speech recognition threshold refers to a set of parameters or standards used to determine whether to trigger or accept the recognition result in a speech recognition system.
[0003] In the prior art, conventional speech recognition systems mostly use fixed threshold settings or threshold online automatic switching settings. Due to different application environments, different noise levels, and different user speaking styles, speech speeds, volumes, and pronunciation clarity, fixed threshold settings have poor environmental adaptability and cannot meet the diverse use needs of users, limiting the application scenarios. Threshold online automatic switching settings rely on the network and cannot be used offline, and require the use of more complex speech recognition models, which requires high computing power for speech recognition models, greatly increasing the deployment cost of the speech recognition system. SUMMARY
[0004] The present application provides an offline speech recognition threshold switching method and system based on Bluetooth communication technology, an electronic device, and a storage medium to solve the technical problem that the speech recognition threshold of conventional speech recognition systems cannot be efficiently switched or can only be switched through a high-power speech recognition model in a networked state in the prior art.
[0005] To solve the above problems, the technical solution of the present application is as follows: an offline speech recognition threshold switching method based on Bluetooth communication technology, comprising the following steps: S1: configuring a target device and a terminal device to establish a communication connection through Bluetooth, enabling the terminal device to send a first control instruction to the target device; S2: the target device stores a wake-up word, a command word, and an identification threshold array corresponding to the wake-up word and the command word used in an offline state, and after receiving the first control instruction, the target device sends a data set containing all wake-up words, command words, and identification threshold arrays corresponding to the wake-up words and command words to the terminal device; S3: The data set of all wake-up words, command words, and the corresponding recognition threshold value arrays of the wake-up words and command words is displayed in a visual manner in the terminal device, and any target wake-up word and / or command word can be manually selected, the recognition threshold value of the target wake-up word and / or command word is adjusted within the corresponding recognition threshold value array range, a second control instruction is generated based on the recognition threshold value adjustment value of the target wake-up word and / or command word, and the terminal device is enabled to send the second control instruction to the target device; S4: After receiving the second control instruction, the target device updates the recognition threshold value of the wake-up word and / or command word in the firmware of the target device.
[0006] Preferably, in S1, the communication connection between the target device and the terminal device is established by Bluetooth, including the following steps: S11: The terminal device is moved to the Bluetooth connection coverage range of the target device, enabling the terminal device and the Bluetooth module of the target device to be initialized, and after initialization, the terminal device scans and discovers the target device and realizes pairing to establish Bluetooth communication connection between the target device and the terminal device.
[0007] Preferably, in S2, after the target device sends the data set containing all wake-up words, command words, and the corresponding recognition threshold value arrays of the wake-up words and command words to the terminal device, the following steps are further included: S21: When the target device and the terminal device use Bluetooth communication connection with Bluetooth technology below Bluetooth 6.0, the terminal device obtains the measured value of the received signal strength when it communicates with the target device by Bluetooth, and based on the reference value of the received signal strength in the ideal state, the terminal device calculates the distance between the terminal device and the target device, and the calculation method is: Wherein, d1 is the distance between the target device and the terminal device, RSSI0 is the reference value of the received signal strength under a known fixed distance d0, RSSI is the measured value of the received signal strength, and n is the environmental attenuation factor; S22: When the target device and the terminal device use Bluetooth communication connection with Bluetooth technology of Bluetooth 6.0 and above, the terminal device sends a group of initial sine wave signals to the target device, the target device sends the initial sine wave signals back to the terminal device after receiving the initial sine wave signals to form return sine wave signals, and the terminal device calculates the distance d1 between the terminal device and the target device based on the phase difference value between the initial sine wave signals and the return sine wave signals. Or, the terminal device sends a set of test data packets to the target device, and the target device reversely sends the test data packets to the terminal device after receiving the test data packets, and the terminal device calculates the distance d1 between the terminal device and the target device based on the total transmission time of the test data packets. S23: According to the voice recognition threshold-device distance algorithm, the terminal device provides the ideal recognition threshold of the target device at the current location, and the calculation method of the voice recognition threshold-device distance algorithm is as follows: Wherein, is the ideal recognition threshold of the target device, , , is the calibration coefficient.
[0008] Preferably, the recognition threshold of the wake-up word and / or the command word in the firmware of the target device is updated in S4, including the following steps: S41: The target device reads the target wake-up word and / or the command word and the recognition threshold adjustment value of the target wake-up word and / or the command word recorded in the second control instruction, and traverses the original wake-up word and / or the command word stored in the firmware to obtain the original recognition threshold of the original wake-up word and / or the command word corresponding to the target wake-up word and / or the command word. It is judged whether the recognition threshold adjustment value of the target wake-up word and / or the command word is within the range of the recognition threshold array of the original wake-up word and / or the command word. If it is correct, the recognition threshold of the original wake-up word and / or the command word in the firmware is adjusted.
[0009] Preferably, the offline voice recognition threshold switching method based on Bluetooth communication technology further includes the following steps: S51: After the target device updates the recognition threshold of the wake-up word and / or the command word, a set of test audio is provided by the user, and the test audio contains the wake-up word and / or the command word adjusted at this time; S52: After the target device receives the test audio, the test audio is adjusted in volume, the volume adjustment range is ±10% of the original volume of the test audio, and a plurality of sets of first simulation audio are generated; S53: Adjust the speech speed of the test audio, the speech speed adjustment range is ±10% of the original speech speed of the test audio, and a plurality of sets of second simulation audio are generated; S54: Fuzzing adjustment is performed on the test audio, the test audio is mixed with white noise, and the white noise ratio does not exceed 5% of the original corpus of the test audio, and a plurality of sets of third simulation audio are generated; S55: The target device performs speech recognition processing on the first analog audio, the second analog audio and the third analog audio respectively based on the adjusted recognition threshold, obtains speech recognition content, compares the speech recognition content with the adjusted wake-up word and / or command word content, calculates the accuracy of the speech recognition result, and sends the accuracy of the speech recognition result to the terminal device for display.
[0010] Preferably, the offline speech recognition threshold switching method based on Bluetooth communication technology further comprises the following steps: S6: The user enters audio data of a new wake-up word and / or command word and text data of the new wake-up word and / or command word through the terminal device, and sets an identification threshold array of the new wake-up word and / or command word. The terminal device sends the audio data, text data and identification threshold array of the new wake-up word and / or command word to the target device through Bluetooth technology, and writes them into the firmware of the target device.
[0011] Preferably, the terminal device comprises a lightweight application program, a mobile application program or a browser control page applied to a mobile client.
[0012] Based on the same concept, the present application also provides an offline speech recognition threshold switching system based on Bluetooth communication technology for executing the offline speech recognition threshold switching method based on Bluetooth communication technology as described in any one of the above embodiments, comprising a target device and a terminal device. The target device is provided with a first Bluetooth connection module for establishing a communication connection with an external device through Bluetooth; a data storage module for writing, storing and updating wake-up words, command words and identification threshold arrays corresponding to the wake-up words and command words respectively; and a speech recognition module for performing speech result recognition on external audio according to the stored wake-up words, command words and identification threshold arrays corresponding to the wake-up words and command words respectively. The terminal device is provided with a second Bluetooth connection module for establishing a communication connection with an external device through Bluetooth; a data display module for displaying the wake-up words, command words and identification threshold arrays corresponding to the wake-up words and command words through a visual method; and a data adjustment module for adjusting the identification threshold of the target wake-up word and / or command word.
[0013] Based on the same concept, the present application also provides an electronic device comprising a processor and a memory, wherein the memory stores machine executable instructions executable by the processor, and the processor executes the machine executable instructions to implement the offline speech recognition threshold switching method based on Bluetooth communication technology as described in any one of the above embodiments.
[0014] Based on the same concept, the application also provides a computer readable storage medium, which stores computer executable instructions, when the computer executable instructions are called and executed by a processor, the computer executable instructions cause the processor to implement the offline voice recognition threshold switching method based on Bluetooth communication technology according to any one of the above.
[0015] Compared with the prior art, the application has the following advantages and positive effects: The application provides an offline voice recognition threshold switching method, system, electronic device and storage medium based on Bluetooth communication technology, in an offline state, the target device and the terminal device establish a communication connection through Bluetooth, then the target device sends data containing all wake-up words, command words, and recognition thresholds corresponding to the wake-up words and command words to the terminal device through Bluetooth communication, the terminal device can adjust the recognition threshold of the wake-up word and / or command word through manual editing, and resend the adjusted value to the target device, and finally update the recognition threshold of the wake-up word and / or command word in the firmware of the target device. Through the application, efficient switching of the voice recognition threshold of the target device can be realized in an offline state, and the target device does not need to increase additional computing power. BRIEF DESCRIPTION OF DRAWINGS
[0016] Fig. 1 The application provides a flowchart of an offline voice recognition threshold switching method based on Bluetooth communication technology. Fig. 2 The application provides a structure diagram of an offline voice recognition threshold switching system based on Bluetooth communication technology. DETAILED DESCRIPTION
[0017] The application provides an offline voice recognition threshold switching method, system, electronic device and storage medium based on Bluetooth communication technology, which will be further described in detail below in combination with the drawings and specific embodiments. The advantages and features of the application will be clearer according to the following description and claims.
[0018] First embodiment Referring to Figs. 1-2 The embodiment provides an offline voice recognition threshold switching method based on Bluetooth communication technology, which is used to realize efficient, simple and controllable switching of the voice recognition threshold in the offline state of the device, including the following steps: S1: configure the target device and the terminal device to establish a communication connection through Bluetooth, and enable the terminal device to send a first control instruction to the target device.
[0019] In the embodiment, the target device refers to a device to be controlled with voice recognition and control function, such as a smart lamp, a smart sound box, etc. with voice recognition and control function. Different use environments (such as home, office, outdoor or in-car) where the target device is located have different background noise levels, and the user's speaking manner, volume, speed and pronunciation clarity are different. The target device with a fixed voice recognition threshold may perform well in some application environments, but may cause misrecognition or missed recognition in other application environments. Therefore, the target device needs to have a voice recognition threshold adjustment function for different application scenarios. In order to further reduce the cost of the target device and improve the applicability of the target device, the target device in the embodiment can have a passive voice recognition threshold adjustment function in an offline state without the help of a high-power voice recognition model.
[0020] The terminal device refers to a device that has a function of adjusting the voice recognition threshold of the target device.
[0021] S2: The target device stores the wake-up words, command words used in the offline state, and the recognition threshold arrays corresponding to the wake-up words and command words, respectively. After receiving the first control instruction, the target device sends a data set containing all the wake-up words, command words, and the recognition threshold arrays corresponding to the wake-up words and command words, respectively, to the terminal device.
[0022] The recognition threshold array refers to the dynamic adjustable range of the recognition threshold corresponding to the wake-up word and the command word.
[0023] S3: The data set of all the wake-up words, command words, and the recognition threshold arrays corresponding to the wake-up words and command words, respectively, is displayed in a visual manner in the terminal device, and any target wake-up word and / or command word can be manually selected. The recognition threshold of the target wake-up word and / or command word is adjusted within the corresponding recognition threshold array range. The second control instruction is generated based on the recognition threshold adjustment value of the target wake-up word and / or command word, so that the terminal device can send the second control instruction to the target device.
[0024] In the embodiment, the wake-up words, command words stored in the target device, and the recognition threshold corresponding to the wake-up words and command words, respectively, can be visually displayed on the display interface of the terminal device after the target device and the terminal device are connected by Bluetooth. The user can manually select one or more groups of wake-up words and / or command words to adjust the recognition threshold of the wake-up words and / or command words. In one embodiment, the recognition threshold array is displayed as a slider bar on the terminal device, and a slider is arranged on the slider bar. The user adjusts the recognition threshold by manually sliding the slider. In another embodiment, the recognition threshold array is represented in the form of percentage on the terminal device, and the user inputs the percentage value to adjust the recognition threshold.
[0025] S4: After the target device receives the second control instruction, the recognition threshold of the wake-up word and / or the command word in the firmware of the target device is updated, thereby realizing the function of switching the recognition threshold of the target device for one or more groups of wake-up words and / or command words.
[0026] In summary, the embodiment provides an offline voice recognition threshold switching method based on Bluetooth communication technology. The target device and the terminal device are connected by Bluetooth. First, the target device sends the wake-up word and command word information to the terminal device by Bluetooth. The terminal device displays the wake-up word and command word information of the target device and manually adjusts the recognition threshold of the wake-up word and / or the command word in the terminal device. Finally, the terminal device sends the adjusted recognition threshold value of the wake-up word and / or the command word to the target device by Bluetooth, realizing data updating of the target device, i.e., voice recognition threshold switching function. Through this embodiment, the target device and the terminal device can be used in offline state without network connection. The target device does not need to be configured with a high-performance voice recognition model with voice recognition threshold dynamic adjustment function. The lightweight scheme of Bluetooth instruction switching and pre-stored recognition threshold array can realize the voice recognition threshold switching function of the target device.
[0027] Next, the specific steps and implementation functions of the offline voice recognition threshold switching method based on Bluetooth communication technology provided by the embodiment will be further described in detail. Preferably, in one embodiment, the communication connection between the target device and the terminal device is established by Bluetooth in S1, including the following steps: S11: Move the terminal device to the Bluetooth connection coverage range of the target device to enable the Bluetooth modules of the terminal device and the target device to be initialized. After initialization, the terminal device scans and discovers the target device and realizes pairing to establish Bluetooth communication connection between the target device and the terminal device.
[0028] In this embodiment, the Bluetooth modules of the terminal device and the target device are first initialized to ensure that the terminal device and the target device are successfully connected and data is safely transmitted. Meanwhile, there may be multiple target devices in the same use environment. In this embodiment, the multiple target devices are real-time signals that can be connected. The terminal device finds the target device to be connected through space scanning and realizes single pairing.
[0029] Preferably, in one embodiment, after the target device sends the data set containing all wake-up words, command words, and recognition threshold arrays corresponding to the wake-up words and command words to the terminal device in S2, the following steps are further included: S21: When the target device and the terminal device are connected by Bluetooth communication with Bluetooth technology below Bluetooth 6.0, the terminal device acquires a measured value of a received signal strength when the terminal device and the target device are connected by Bluetooth communication, and calculates a distance between the terminal device and the target device based on a reference value of the received signal strength in an ideal state. The calculation method is as follows: wherein d1 is the distance between the terminal device and the target device, RSSI0 is the reference value of the received signal strength at a known fixed distance d0, RSSI is the measured value of the received signal strength, and n is an environmental attenuation factor. S22: When the target device and the terminal device are connected by Bluetooth communication with Bluetooth technology of and above Bluetooth 6.0, the terminal device sends a group of initial sine wave signals to the target device, the target device sends the initial sine wave signals back to the terminal device after receiving the initial sine wave signals to form return sine wave signals, and the terminal device calculates the distance d1 between the terminal device and the target device based on a phase difference between the initial sine wave signals and the return sine wave signals. Alternatively, the terminal device sends a group of test data packets to the target device, the target device sends the test data packets back to the terminal device after receiving the test data packets, and the terminal device calculates the distance d1 between the terminal device and the target device based on a total transmission time of the test data packets. S23: According to the voice recognition threshold-device distance algorithm, the terminal device provides an ideal recognition threshold of the target device at the current location of the terminal device. The calculation method of the voice recognition threshold-device distance algorithm is as follows: wherein, is the ideal recognition threshold of the target device, , , is a calibration coefficient. In the voice recognition threshold-device distance algorithm, the influence of sound propagation distance on acoustic characteristics is equivalent to signal attenuation. Therefore, the essence of the voice recognition threshold-device distance algorithm is that the farther the sound propagation distance, the lower the required recognition threshold, and the threshold of voice recognition is relaxed to avoid missing detection of wake-up words and / or command words.
[0030] In the voice recognition process of the target device, the distance from the sound source to the target device is one of the key factors for evaluating the voice recognition threshold, so in this embodiment, a way of automatically calculating the distance between the target device and the terminal device based on Bluetooth technology is provided. For Bluetooth communication connection below Bluetooth 6.0 technology, the signal strength value is used to judge the distance between the target device and the terminal device, and for Bluetooth communication connection above Bluetooth 6.0 technology, the phase difference value of the sine wave signal or the transmission time of the test data packet can be used to judge the distance between the target device and the terminal device. After knowing the distance between the target device and the terminal device, the target device will give the ideal recognition threshold suitable for the target device under the current distance, and send the ideal recognition threshold to the terminal device, providing a reference for the user to adjust the recognition threshold value of the target device under the current position.
[0031] Preferably, in one embodiment, the recognition threshold of the wake-up word and / or command word in the firmware of the target device is updated in S4, including the following steps: S41: The target device reads the target wake-up word and / or command word recorded in the second control instruction and the recognition threshold adjustment value of the target wake-up word and / or command word, and queries the original wake-up word and / or command word stored in the firmware to obtain the original recognition threshold of the original wake-up word and / or command word corresponding to the target wake-up word and / or command word. It is judged whether the recognition threshold adjustment value of the target wake-up word and / or command word is within the range of the recognition threshold array of the original wake-up word and / or command word. If it is correct, the recognition threshold of the original wake-up word and / or command word in the firmware is adjusted. If it is not correct, an error information is displayed through the terminal device.
[0032] Preferably, in one embodiment, the offline voice recognition threshold switching method based on Bluetooth communication technology further includes the following steps: S51: After updating the recognition threshold of the wake-up word and / or command word, the target device provides a set of test audio provided by the user, which contains the wake-up word and / or command word adjusted at this time; S52: After receiving the test audio, the target device adjusts the volume of the test audio, and the volume adjustment range is ±10% of the original volume of the test audio, and generates a plurality of groups of first simulated audio; S53: Adjust the speech speed of the test audio, and the speech speed adjustment range is ±10% of the original speech speed of the test audio, and generate a plurality of groups of second simulated audio; S54: Adjust the fuzzification of the test audio, mix the test audio with white noise, and the white noise ratio does not exceed 5% of the original corpus of the test audio, and generate a plurality of groups of third simulated audio; S55: The target device performs speech recognition processing on the plurality of groups of first, second, and third simulated audio based on the adjusted recognition threshold, respectively, obtains speech recognition content, compares the speech recognition content with the adjusted wake-up word and / or command word content, calculates the accuracy of the speech recognition result, and sends the accuracy of the speech recognition result to the terminal device for display.
[0033] In this embodiment, after the user switches the speech recognition threshold of a certain wake-up word or command word in the target device, the target device can be optionally subjected to a speech recognition test, that is, a group of test audio containing the adjusted wake-up word or command word is output to the target device. After receiving the test audio, the target device expands the test audio from multiple dimensions, including adjusting the volume, speed, and fuzziness of the test audio, thereby simulating the audio differences output by the user at different times and in different states. Finally, the target device performs speech recognition on a plurality of groups of simulated audio based on the adjusted recognition threshold, and calculates the accuracy of the speech recognition result. If the accuracy is low, the user can be reminded by the terminal device that the adjusted recognition threshold is low under the current application environment and interval distance.
[0034] Preferably, in an embodiment, the offline speech recognition threshold switching method based on Bluetooth communication technology further comprises the following steps: S6: The user enters audio data and text data of the new wake-up word and / or command word through the terminal device, and sets an array of recognition thresholds of the new wake-up word and / or command word. The terminal device sends the audio data, text data, and array of recognition thresholds of the new wake-up word and / or command word to the target device through Bluetooth technology, and writes them into the firmware of the target device.
[0035] In this embodiment, in addition to adjusting the recognition threshold of the originally stored wake-up word and / or command word in the target device, new audio data, text data, and array of recognition thresholds of the new wake-up word and / or command word can also be entered into the target device through Bluetooth, thereby improving the convenience of data loading in the target device.
[0036] Preferably, in an embodiment, the terminal device includes a lightweight application program (such as a WeChat mini-program) applied to a mobile client, a mobile application program (Android / iOS App), or a browser control page.
[0037] Second embodiment Based on the same concept, the present application also provides an offline speech recognition threshold switching system based on Bluetooth communication technology, which is used to perform the offline speech recognition threshold switching method based on Bluetooth communication technology as described in any one of the first embodiments, and includes a target device and a terminal device.
[0038] The target device is provided with a first Bluetooth connection module for establishing a communication connection with an external device through Bluetooth; a data storage module for writing, storing and updating the wake-up word, the command word and the identification threshold value corresponding to the wake-up word and the command word respectively; and a voice recognition module for performing voice result recognition on external audio according to the stored wake-up word, the command word and the identification threshold value corresponding to the wake-up word and the command word respectively.
[0039] The terminal device is provided with a second Bluetooth connection module for establishing a communication connection with an external device through Bluetooth; a data display module for displaying the wake-up word, the command word and the identification threshold value corresponding to the wake-up word and the command word through a visual method; and a data adjustment module for adjusting the identification threshold value of the target wake-up word and / or the command word.
[0040] The functions of the constituent units in the above-mentioned off-line voice recognition threshold switching system based on Bluetooth communication technology correspond to the steps in the above-mentioned off-line voice recognition threshold switching method embodiment based on Bluetooth communication technology, and the functions and implementation processes will not be repeated here.
[0041] The embodiment also provides an electronic device including a processor and a memory, the memory storing machine executable instructions capable of being executed by the processor, and the processor executes the machine executable instructions to implement the above-mentioned off-line voice recognition threshold switching method based on Bluetooth communication technology.
[0042] The embodiment also provides a computer readable storage medium storing computer executable instructions, and the computer executable instructions, when invoked and executed by a processor, cause the processor to implement the steps of the above-mentioned off-line voice recognition threshold switching method based on Bluetooth communication technology.
[0043] In summary, the present application provides an off-line voice recognition threshold switching method, system, electronic device and storage medium based on Bluetooth communication technology, which enables the target device and the terminal device to establish a communication connection through Bluetooth in an offline state, and then the target device sends data containing all wake-up words, command words and identification threshold values corresponding to the wake-up words and the command words to the terminal device through Bluetooth communication. The terminal device can adjust the identification threshold values of the wake-up words and / or the command words through manual editing, and resend the adjusted identification threshold values to the target device. Finally, the identification threshold values of the wake-up words and / or the command words are updated in the firmware of the target device. Through the present application, efficient switching of the voice recognition threshold of the target device can be realized in an offline state, and the target device does not need to increase additional computing power.
[0044] The embodiments of the present application are explained in detail above with reference to the accompanying drawings, but the present application is not limited to the above-described embodiments. Even if various changes are made to the present application, if the changes fall within the scope of the claims of the present application and equivalents thereof, they are still within the protective scope of the present application.
Claims
1. A Bluetooth communication technology-based off-line voice recognition threshold switching method, characterized by, Comprising the following steps: S1: configuring the target device and the terminal device to establish a communication connection through Bluetooth, enabling the terminal device to send a first control instruction to the target device; S2: the target device stores the wake-up word, command word used in offline state, and the identification threshold array corresponding to the wake-up word and command word, respectively, after the target device receives the first control instruction, it sends a data set containing all wake-up words, command words, and identification threshold arrays corresponding to the wake-up words and command words to the terminal device; S3: the data set of all wake-up words, command words, and identification threshold arrays corresponding to the wake-up words and command words is displayed in a visual manner in the terminal device, and any target wake-up word and / or command word can be manually selected, the identification threshold of the target wake-up word and / or command word is adjusted within the corresponding identification threshold array range, a second control instruction is generated based on the identification threshold adjustment value of the target wake-up word and / or command word, and the terminal device sends the second control instruction to the target device; S4: after the target device receives the second control instruction, the identification threshold of the wake-up word and / or command word in the firmware of the target device is updated.
2. The method of claim 1, wherein the Bluetooth communication technology-based off-line voice recognition threshold switching method is characterized by, In S1, the target device and the terminal device are configured to establish a communication connection through Bluetooth, comprising the following steps: S11: move the terminal device to the Bluetooth connection coverage range of the target device, enable the terminal device and the Bluetooth module of the target device to initialize, after initialization, the terminal device scans and discovers the target device and realizes pairing, and establishes Bluetooth communication connection between the target device and the terminal device.
3. The threshold switching method for off-line speech recognition based on Bluetooth communication technology according to claim 1, wherein, After the target device sends the data set containing all wake-up words, command words, and identification threshold arrays corresponding to the wake-up words and command words to the terminal device in S2, the following steps are further included: S21: when the target device and the terminal device use Bluetooth communication connection with Bluetooth technology below 6.0, the terminal device obtains the measured value of the received signal strength when it communicates with the target device through Bluetooth, based on the reference value of the received signal strength in the ideal state, the terminal device calculates the distance between the terminal device and the terminal device, and the calculation method is: Wherein, d1 is the distance between the target device and the terminal device, RSSI0 is the reference value of the received signal strength under the known fixed distance d0, RSSI is the measured value of the received signal strength, and n is the environmental attenuation factor; S22: when the target device and the terminal device use Bluetooth communication connection with Bluetooth technology of 6.0 and above, the terminal device sends a group of initial sine wave signals to the target device, the target device sends the initial sine wave signal back to the terminal device after receiving it, forming a return sine wave signal, and the terminal device calculates the distance d1 between the terminal device and the terminal device based on the phase difference between the initial sine wave signal and the return sine wave signal. Or, the terminal device sends a set of test data packets to the target device, and the target device reversely sends the test data packets to the terminal device after receiving the test data packets. The terminal device calculates the distance d1 between the terminal device and the target device based on the total transmission time of the test data packets. S23: According to the voice recognition threshold-device distance algorithm, the terminal device provides the ideal recognition threshold of the target device at the current location. The calculation method of the voice recognition threshold-device distance algorithm is as follows: wherein, is the ideal recognition threshold for the target device, , , is a calibration coefficient.
4. The method of claim 1, wherein the Bluetooth communication technology-based off-line voice recognition threshold switching method is characterized by, In S4, the recognition threshold of the wake-up word and / or the command word in the firmware of the target device is updated, including the following steps: S41: The target device reads the target wake-up word and / or the command word and the recognition threshold adjustment value of the target wake-up word and / or the command word recorded in the second control instruction, and traverses the original wake-up word and / or the command word stored in the firmware to obtain the original recognition threshold of the original wake-up word and / or the command word corresponding to the target wake-up word and / or the command word. It is judged whether the recognition threshold adjustment value of the target wake-up word and / or the command word is within the range of the recognition threshold array of the original wake-up word and / or the command word. If it is correct, the recognition threshold of the original wake-up word and / or the command word in the firmware is adjusted.
5. The method of claim 1, wherein the Bluetooth communication technology-based off-line voice recognition threshold switching method is characterized by, Further comprising the following steps: S51: After the target device updates the recognition threshold of the wake-up word and / or the command word, the user provides a set of test audio, which contains the adjusted wake-up word and / or the command word; S52: After the target device receives the test audio, the test audio is adjusted in volume, the volume adjustment range is ±10% of the original volume of the test audio, and a plurality of first simulation audios are generated; S53: The test audio is adjusted in speech speed, the speech speed adjustment range is ±10% of the original speech speed of the test audio, and a plurality of second simulation audios are generated; S54: The test audio is adjusted to be blurred, and the test audio is mixed with white noise, the proportion of white noise is not more than 5% of the original corpus of the test audio, and a plurality of third simulation audios are generated; S55: The target device performs voice recognition processing on a plurality of first simulation audios, second simulation audios and third simulation audios based on the adjusted recognition threshold, respectively, obtains voice recognition content, compares the voice recognition content with the content of the adjusted wake-up word and / or the command word, calculates the accuracy of the voice recognition result, and the target device sends the accuracy of the voice recognition result to the terminal device for display.
6. The method of claim 1, wherein the Bluetooth communication technology-based off-line voice recognition threshold switching method is characterized by, Further comprising the following steps: S6: The user enters the audio data and the text data of the new wake-up word and / or the command word through the terminal device, and sets the recognition threshold array of the new wake-up word and / or the command word. The terminal device sends the audio data, text data and recognition threshold array of the new wake-up word and / or the command word to the target device through Bluetooth technology, and writes them into the firmware of the target device.
7. The method of claim 1, wherein the Bluetooth communication technology-based off-line voice recognition threshold switching method is characterized by, The terminal device includes a lightweight application program, a mobile application program or a browser control page applied to a mobile client.
8. A Bluetooth communication technology-based off-line voice recognition threshold switching system, characterized by, A target device and a terminal device are used to implement the Bluetooth communication technology-based offline voice recognition threshold switching method according to any one of claims 1-7. The target device is provided with a first Bluetooth connection module, which is used to establish a communication connection with an external device through Bluetooth; a data storage module, which is used to write, store and update wake-up words, command words and the recognition threshold arrays corresponding to the wake-up words and the command words respectively; and a voice recognition module, which is used to perform voice result recognition on external audio according to the stored wake-up words, command words and the recognition threshold arrays corresponding to the wake-up words and the command words respectively. The terminal device is provided with a second Bluetooth connection module, which is used to establish a communication connection with an external device through Bluetooth; a data display module, which is used to display the wake-up words, command words and the recognition threshold arrays corresponding to the wake-up words and the command words through a visual mode; and a data adjustment module, which is used to adjust the recognition threshold of a target wake-up word and / or a command word.
9. An electronic device, comprising: A processor and a memory are included, the memory stores machine executable instructions that can be executed by the processor, and the processor executes the machine executable instructions to implement the Bluetooth communication technology-based offline voice recognition threshold switching method according to any one of claims 1-7.
10. A computer-readable storage medium, characterized in that, The computer readable storage medium stores computer executable instructions, and when the computer executable instructions are called and executed by a processor, the computer executable instructions cause the processor to implement the Bluetooth communication technology-based offline voice recognition threshold switching method according to any one of claims 1-7.