A speech recognition system, device and medium based on cloud computing and edge computing
By combining cloud computing and edge computing in the speech recognition system, using the voice scene recognition model to compare the cloud and local recognition results, selecting the most accurate recognition result output, solving the problem of poor voice interaction experience under poor network, and achieving weak High accuracy recognition and smooth interaction in network environments.
Patent Information
- Application Number
- CN202110686781.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-06-21
- Publication Date
- 2024-11-12
- Estimated Expiration
- 2041-06-21
AI Technical Summary
The existing voice recognition system cannot guarantee the normal progress of voice interaction in the case of poor network, resulting in poor user experience and lack of fine input and sentence breaking processing when switching between the cloud and local recognition engines, resulting in poor interaction.
The speech recognition system based on cloud computing and edge computing is adopted. Voice information is collected through the voice entry processing module, and processed and calculated separately in the cloud and local area. The speech scene recognition judgment module is used to compare the recognition results and confidence through the speech scene recognition model algorithm. , select the most accurate identification result output to ensure that the accuracy and user experience of speech recognition are improved in unstable network environments.
Improve the accuracy and user experience of voice recognition in poor network environments, solve the problem of incomplete information caused by cloud and local switching, and use user habits and voice database upgrades to further improve the accuracy of voice recognition and adapt to weak network environments.
Smart Images

Figure CN113380253B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of speech recognition, and in particular to a speech recognition system, device and medium based on cloud computing and edge computing. Background Art
[0002] Cloud computing is a type of distributed computing, which refers to a computing system that processes data through the network "cloud" to obtain results and returns them to the user. Edge computing refers to providing the nearest computing service close to the source of the data. The goal of speech recognition is to convert the vocabulary content in human speech into computer-readable input, such as binary code or character sequence.
[0003] At present, there are generally two ways of voice recognition. One is to use cloud recognition engine to perform relevant voice recognition. When the network is not good, voice recognition cannot continue, resulting in poor user experience. At this time, the corresponding voice interaction experience is poor. The other way is to use a combination of local recognition engine and cloud recognition engine. When the network is good, cloud recognition is used to perform voice result recognition. When the network is not good, the local recognition engine is used for recognition. The combination of the two systems ensures the normal voice interaction. This type of system cannot guarantee an excellent interactive experience due to the weak local voice recognition ability. In addition, when switching between the two modes, there is no fine input segmentation processing, and there will always be unsmooth interaction. Summary of the invention
[0004] The purpose of the present invention is to provide a speech recognition system based on cloud computing and edge computing, which can adapt to poor network environments, improve the accuracy of speech recognition, and improve user experience.
[0005] Another object of the present invention is to provide an electronic device that can adapt to an environment with poor network conditions, improve the accuracy of voice recognition, and enhance user experience.
[0006] Another object of the present invention is to provide a computer-readable storage medium that can adapt to poor network environments, improve the accuracy of speech recognition, and improve user experience.
[0007] The embodiment of the present invention is achieved as follows:
[0008] In a first aspect, an embodiment of the present application provides a speech recognition system based on cloud computing and edge computing, comprising the following steps:
[0009] Voice input processing module: used to collect voice information;
[0010] A cloud voice processing module is used to process and calculate the voice information through the cloud, and output a first voice recognition result and a first recognition confidence T1 based on a cloud voice database;
[0011] A local edge speech processing module: used to process and calculate the speech information locally, and output a second speech recognition result and a second recognition confidence T2 based on a local edge speech database;
[0012] Speech scene recognition judgment module: used to compare the first speech recognition result and the second speech recognition result through the speech scene recognition model algorithm, and output the first cloud scene result S1 and the first local scene result S2; compare the first recognition confidence T1 and the second recognition confidence T2 through the speech scene recognition model algorithm, and output the first recognition result confidence t1 and the second recognition result confidence t2; judge whether S1=S2 is true, and judge if S1=S2 is true Is it true? If it is true, output the first cloud scene result S1, otherwise perform cloud scene recognition and output the second cloud scene result; when S1=S2, it is judged that Whether it is true, if it is true, output the first local scene result S2, otherwise perform local scene recognition and output the second local scene result.
[0013] In some embodiments of the present invention, the voice input processing module is used to upload the collected voice information to the cloud and locally respectively after noise elimination and human voice detection.
[0014] In some embodiments of the present invention, the voice input processing module is used to determine whether there is a network, and if so, upload the voice information to the cloud and the local respectively; if not, upload the voice information to the local.
[0015] In some embodiments of the present invention, the speech scene recognition judgment module is used to determine whether the first recognition confidence T1 is much greater than the second recognition confidence T2. If so, local scene recognition is performed and a second local scene result is output; otherwise, S1=S2 is determined.
[0016] In some embodiments of the present invention, the cloud-based voice processing module is used to process and calculate the voice information through the cloud, including recognizing the voice information through a cloud-based voice database; the local edge voice processing module is used to process and calculate the voice information locally, including recognizing the voice information through a local voice database.
[0017] In some embodiments of the present invention, the voice scene recognition judgment module is used to perform intent scene positioning on the first voice recognition result to obtain the first cloud scene result S1, and to perform intent scene positioning on the second voice recognition result to obtain the first local scene result S2.
[0018] In some embodiments of the present invention, the speech scene recognition judgment module is used to determine whether the first local scene result S2 is empty. When the first local scene result S2 is empty, the cloud scene recognition is performed and the second cloud scene result is output. Otherwise, the judgment Whether it is established, if it is established, the first cloud scene result S1 is output, otherwise, local scene recognition is performed and the second local scene result is output.
[0019] In some embodiments of the present invention, the above-mentioned speech recognition system based on cloud computing and edge computing also includes a speech interaction module: used to output speech interaction information based on output result feedback.
[0020] In a second aspect, an embodiment of the present application provides an electronic device, comprising: a memory for storing one or more programs; a processor; when the one or more programs are executed by the processor, a system as described in any one of the first aspects above is implemented.
[0021] In a third aspect, an embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements a system as described in any one of the first aspects.
[0022] Compared with the prior art, the embodiments of the present invention have at least the following advantages or beneficial effects:
[0023] In a first aspect, an embodiment of the present application provides a speech recognition system based on cloud computing and edge computing, comprising the following steps:
[0024] Voice input processing module: used to collect voice information;
[0025] A cloud voice processing module is used to process and calculate the voice information through the cloud, and output a first voice recognition result and a first recognition confidence T1 based on a cloud voice database;
[0026] A local edge speech processing module: used to process and calculate the speech information locally, and output a second speech recognition result and a second recognition confidence T2 based on a local edge speech database;
[0027] Speech scene recognition judgment module: used to compare the first speech recognition result and the second speech recognition result through the speech scene recognition model algorithm, and output the first cloud scene result S1 and the first local scene result S2; compare the first recognition confidence T1 and the second recognition confidence T2 through the speech scene recognition model algorithm, and output the first recognition result confidence t1 and the second recognition result confidence t2; judge whether S1=S2 is true, and judge if S1=S2 is true Is it true? If it is true, output the first cloud scene result S1, otherwise perform cloud scene recognition and output the second cloud scene result; when S1=S2, it is judged that Whether it is true, if it is true, output the first local scene result S2, otherwise perform local scene recognition and output the second local scene result.
[0028] For the first aspect: the embodiment of the present application collects voice information, and processes and calculates the voice information using the cloud and the local respectively, so that it can be used when the network is unstable, and improve the user's experience when interacting through voice; by outputting the first voice recognition result and the first recognition confidence based on the cloud voice database, by outputting the second voice recognition result and the second recognition confidence based on the local voice database, the voice database can be used to recognize the content of the voice information, and the confidence is used to reflect the accuracy of the current voice recognition, and the first voice recognition result and the second voice recognition result are compared by the voice scene recognition model algorithm, and the accuracy of the cloud scene result and the local scene result are outputted. After judging the accuracy of the two according to the logical relationship, the cloud or local scene is selected to execute the corresponding result according to the accuracy, which can improve the accuracy of voice recognition; by comparing the voice information obtained by the cloud and the local through the voice recognition result and the confidence, the problem of incomplete information caused by the switching between the cloud and the local can be solved in a weak network environment such as the subway. The present invention uses the voice scene recognition model algorithm to compare the voice recognition results of the cloud scene and the local scene, and can improve the accuracy of voice recognition by using user habits; using the cloud and the local scene to output multiple recognition results of different voice information, it can be used for upgrading the voice database, and further accurately recognize the voice by using user habits. The present invention can adapt to weak network environments, improve the accuracy of voice recognition, and thus enhance the user's experience of using voice interaction.
[0029] In a second aspect, an embodiment of the present application provides an electronic device, comprising: a memory for storing one or more programs; a processor; when the one or more programs are executed by the processor, a system as described in any one of the first aspects above is implemented.
[0030] In a third aspect, an embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements a system as described in any one of the first aspects.
[0031] Regarding the second to third aspects: the principles of the embodiments of the present application are the same as those of the first aspect and need not be described again here. BRIEF DESCRIPTION OF THE DRAWINGS
[0032] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.
[0033] Figure 1 Schematic diagram of the principle of a speech recognition system based on cloud computing and edge computing in an embodiment of the present invention Figure 1 ;
[0034] Figure 2 The process diagram of the speech recognition system based on cloud computing and edge computing in the embodiment of the present invention is as follows Figure 1 ;
[0035] Figure 3 The process diagram of the speech recognition system based on cloud computing and edge computing in the embodiment of the present invention is as follows Figure 2 . Implementation
[0036] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. The components of the embodiments of the present application described and shown in the drawings here can be arranged and designed in various different configurations.
[0037] Therefore, the following detailed description of the embodiments of the present application provided in the accompanying drawings is not intended to limit the scope of the present application for which protection is sought, but merely represents selected embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without creative work are within the scope of protection of the present application.
[0038] In the description of this application, it is also necessary to explain that, unless otherwise clearly specified and limited, the terms "disposed" and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be an indirect connection through an intermediate medium, or it can be the internal communication of two elements. For ordinary technicians in this field, the specific meanings of the above terms in this application can be understood according to specific circumstances.
[0039] In conjunction with the accompanying drawings, some embodiments of the present application are described in detail below. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other. Example
[0040] See also Figure 1 to Figure 3 , Figure 1 to Figure 3 The following is a schematic diagram of the principle and process of a speech recognition system based on cloud computing and edge computing provided by an embodiment of the present application. The speech recognition system based on cloud computing and edge computing includes a speech input processing module: used to collect speech information;
[0041] A cloud voice processing module is used to process and calculate the voice information through the cloud, and output a first voice recognition result and a first recognition confidence T1 based on a cloud voice database;
[0042] A local edge speech processing module: used to process and calculate the speech information locally, and output a second speech recognition result and a second recognition confidence T2 based on a local edge speech database;
[0043] Speech scene recognition judgment module: used to compare the first speech recognition result and the second speech recognition result through the speech scene recognition model algorithm, and output the first cloud scene result S1 and the first local scene result S2; compare the first recognition confidence T1 and the second recognition confidence T2 through the speech scene recognition model algorithm, and output the first recognition result confidence t1 and the second recognition result confidence t2; judge whether S1=S2 is true, and judge if S1=S2 is true Is it true? If it is true, output the first cloud scene result S1, otherwise perform cloud scene recognition and output the second cloud scene result; when S1=S2, it is judged that Whether it is true, if it is true, output the first local scene result S2, otherwise perform local scene recognition and output the second local scene result.
[0044] The voice input processing module can be used to collect voice information emitted by people when voice controlling smart devices. The collected device can be a handset, which is sent to the cloud through the network for recognition, and sent locally through electrical signals for recognition. The cloud voice processing module processes and calculates the voice information through the cloud, which can be obtained by using the cloud voice database, wherein the cloud voice database stores the voice content and confidence of multiple recognized voice information. Optionally, the voice content and confidence stored in the cloud voice database are obtained through a cloud voice data model, wherein the cloud voice data model is trained by a machine through multiple groups of voice data. And each group of voice data includes voice information and voice content and confidence for identifying voice information. Thus, different voice information is entered into the cloud voice data model, which can realize the recognition of the content of the voice information to obtain the voice recognition result and the confidence of the recognition result in the scene. After the local edge voice processing module processes and calculates the voice information locally, it can be obtained through the local scene voice database. Among them, the local scene voice database can be obtained through the local voice data model. In addition, the local scene voice database is the same as the cloud scene voice database, and the local voice data model is the same as the cloud voice data model, and there is no need to repeat the description here. Among them, the confidence value reflects the accuracy of the current speech recognition, and is related to factors such as the current network status, packet error rate, and speech algorithm recognition rate. The higher the value, the more accurate the speech recognition.
[0045] The speech scene recognition judgment module compares the recognition results obtained locally and in the cloud, where the speech scene recognition model algorithm includes multiple groups of recognition data, each group of data includes speech information and cloud content and local content for marking the content of the speech information. The local content is obtained through different speech information in the local scene, and the cloud content is obtained through different speech information in the cloud scene, thereby further improving the accuracy of speech information recognition. The speech scene recognition model algorithm compares the cloud and local speech recognition results to obtain a recognition result with a higher similarity to the speech scene recognition model algorithm, and by comparing the cloud and local speech recognition results, the recognition result that is more in line with the semantic scene after comparison is screened out. Logical operations are performed based on the local and cloud recognition results and confidence levels, the preferred solution for the cloud or local is selected, and the final debugged result is output.
[0046] In detail, the recognized voice information can be used for further training of the above models, and can also be used in the field of voice interaction to provide accurate feedback on the voice information input by the user, thereby improving the user experience when using voice interaction.
[0047] In some embodiments of the present invention, the voice input processing module is used to upload the collected voice information to the cloud and locally respectively after noise elimination and human voice detection.
[0048] In detail, the collected speech is noise-eliminating or voice-enhancing by means of noise elimination and voice recognition, and the processed results are input into the local and cloud for initialization, thereby improving the accuracy of semantic recognition of voice information.
[0049] In some embodiments of the present invention, the voice input processing module is used to determine whether there is a network, and if so, upload the voice information to the cloud and the local respectively; if not, upload the voice information to the local.
[0050] In the above step (1), it is determined whether there is a network, so that when there is a network, the voice information is uploaded to the cloud for recognition, and when there is no network, it is uploaded to the local, and the local voice recognition can be used to provide voice feedback according to the content, so as to adapt to the use without a network and facilitate the backup of voice information. The uploaded local can be stored in the server's memory, and the server can upload it to the cloud server through the gateway.
[0051] In some embodiments of the present invention, the speech scene recognition judgment module is used to determine whether the first recognition confidence T1 is much greater than the second recognition confidence T2. If so, local scene recognition is performed and a second local scene result is output; otherwise, S1=S2 is determined.
[0052] Specifically, it is determined whether the first recognition confidence T1 is greater than the second recognition confidence T2. If it is much greater, local scene recognition is performed and the result after execution is output. Otherwise, it proceeds to the next step, that is, it is determined whether S1=S2. When the local confidence is much greater than the cloud confidence, the local scene is selected for recognition first, and the recognition result is directly output, which can obtain a more accurate recognition result.
[0053] In some embodiments of the present invention, the cloud-based voice processing module is used to process and calculate the voice information through the cloud, including recognizing the voice information through a cloud-based voice database; the local edge voice processing module is used to process and calculate the voice information locally, including recognizing the voice information through a local voice database.
[0054] In detail, the voice information is processed through the cloud by recognizing the voice information through the cloud voice database, thereby obtaining the voice recognition result. The same is true for the local processing of voice information. The cloud and the local use big data to recognize the voice separately, and then use big data again in the comparison process to obtain the rules of the two recognizing the same voice information, thereby obtaining the final recognition result, greatly improving the accuracy of the recognition result. The process of recognizing voice information through the voice database has been described in detail above, and there is no need to repeat it here.
[0055] In some embodiments of the present invention, the voice scene recognition judgment module is used to perform intent scene positioning on the first voice recognition result to obtain the first cloud scene result S1, and to perform intent scene positioning on the second voice recognition result to obtain the first local scene result S2.
[0056] The use of intention scene positioning to process recognition results is a prior art and is not specifically limited here. Optionally, the intention scene recognition system based on voice information can convert the voice information into text through multimodal input and perform multimodal input conversion on the multimodal input. Intent recognition is performed based on the converted text, and each obtained intent is scored. Intent recognition can be performed in a traditional way; the voice information and each candidate voice information are similarly calculated; the calculated similarity is used as another score, and weighted addition and re-arrangement are performed with the score obtained in the traditional way, and the candidate intent with the highest score is used as the final output intent. The multimodal input of the voice information can be in the form of audio, image, text, etc. Intent scene positioning can enhance the processing of scene results and further improve the accuracy of scene results.
[0057] In some embodiments of the present invention, the speech scene recognition judgment module is used to determine whether the first local scene result S2 is empty. When the first local scene result S2 is empty, the cloud scene recognition is performed and the second cloud scene result is output. Otherwise, the judgment Whether it is established, if it is established, the first cloud scene result S1 is output, otherwise, local scene recognition is performed and the second local scene result is output.
[0058] Specifically, when the first local scene result does not exist, the cloud scene recognition is performed and the result is output, otherwise it is determined whether the accuracy of the cloud is high, otherwise the local scene is switched to output the result. The above second cloud scene result is the result after automatic correction, which improves the accuracy of speech recognition technology.
[0059] In some embodiments of the present invention, the above-mentioned speech recognition system based on cloud computing and edge computing also includes a speech interaction module: used to output speech interaction information based on output result feedback.
[0060] In detail, feedback is provided based on the output results in the cloud or locally. The output content may be voice communication or the result of a search based on voice information, and may be applied to speech recognition technology in the field of artificial intelligence.
[0061] The present invention can smoothly complete the interactive experience of users in specific scenarios by switching between local voice recognition and cloud voice recognition, and can meet the recognition of voice information with more content under unstable network conditions; by comparing local and cloud recognition results, the optimal result output is selected to improve the accuracy of the recognition result. Example
[0062] An embodiment of the present application provides an electronic device, comprising: a memory for storing one or more programs; a processor; when the one or more programs are executed by the processor, the system described in the above-mentioned embodiment 1 is implemented.
[0063] The memory, processor and communication interface are electrically connected to each other directly or indirectly to realize data transmission or interaction. For example, these elements can be electrically connected to each other through one or more communication buses or signal lines. The memory can be used to store software programs and modules, such as program instructions / modules corresponding to the speech recognition system based on cloud computing and edge computing provided in the embodiment of the present application, and the processor executes various functional applications and data processing by executing the software programs and modules stored in the memory. The communication interface can be used to communicate signaling or data with other node devices.
[0064] Among them, the memory can be, but is not limited to, random access memory (RAM), read only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable read-only memory (EEPROM), etc.
[0065] The processor can be an integrated circuit chip with signal processing capabilities. The processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.
[0066] In the embodiments provided in the present application, it should be understood that the disclosed devices and systems can also be implemented in other ways. The device embodiments described above are merely schematic. For example, the flowcharts in the accompanying drawings show the possible architecture, functions and operations of the devices, systems and computer program products according to the multiple embodiments of the present application. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a part of a code, and the module, a program segment or a part of a code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the flowchart, and the combination of boxes in the flowchart, can be implemented with a dedicated hardware-based system that performs a specified function or action, or can be implemented with a combination of dedicated hardware and computer instructions.
[0067] In addition, the functional modules in the various embodiments of the present application may be integrated together to form an independent part, or each module may exist separately, or two or more modules may be integrated to form an independent part.
[0068] If the functions are implemented in the form of software function modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the system described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.
[0069] In summary, the embodiments of the present application provide a speech recognition system, device and medium based on cloud computing and edge computing:
[0070] The embodiment of the present application collects voice information, and processes and calculates the voice information using the cloud and the local respectively, so that it can be used when the network is unstable, and improves the user's experience when interacting through voice; by outputting the first voice recognition result and the first recognition confidence based on the cloud voice database, by outputting the second voice recognition result and the second recognition confidence based on the local voice database, the voice database can be used to recognize the content of the voice information, and the confidence is used to reflect the accuracy of the current voice recognition, and the first voice recognition result and the second voice recognition result are compared by the voice scene recognition model algorithm, and the accuracy of the cloud scene result and the local scene result are outputted. After judging the accuracy of the two according to the logical relationship, the cloud or local scene is selected to execute the corresponding result according to the accuracy, which can improve the accuracy of voice recognition; by comparing the voice information obtained by the cloud and the local through the voice recognition result and the confidence, the problem of incomplete information caused by the switching between the cloud and the local can be solved in a weak network environment such as a subway. The present invention uses the voice scene recognition model algorithm to compare the voice recognition results of the cloud scene and the local scene, and can improve the accuracy of voice recognition by using user habits; using the cloud and the local scene to output multiple recognition results of different voice information, it can be used for upgrading the voice database, and further accurately recognize the voice by using user habits. The present invention can adapt to weak network environments, improve the accuracy of voice recognition, and thus enhance the user's experience of using voice interaction.
[0071] The above description is only the preferred embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.
[0072] It will be apparent to those skilled in the art that the present application is not limited to the details of the exemplary embodiments described above, and that the present application can be implemented in other specific forms without departing from the spirit or essential features of the present application. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the present application is defined by the appended claims rather than the above description, and it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims be included in the present application. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.
Claims
1. A speech recognition system based on cloud computing and edge computing, characterized in that: Comprising: A voice input processing module: for collecting voice information; A cloud voice processing module: for processing and calculating the voice information through the cloud, and outputting a first voice recognition result and a first recognition confidence level T1 based on a cloud voice database; A local edge voice processing module: for processing and calculating the voice information locally, and outputting a second voice recognition result and a second recognition confidence level T2 based on a local edge voice database; A voice scene recognition and judgment module: for comparing the first voice recognition result and the second voice recognition result through a voice scene recognition model algorithm, and outputting a first cloud scene result S1 and a first local scene result S2; comparing the first recognition confidence level T1 and the second recognition confidence level T2 through a voice scene recognition model algorithm, and outputting a first recognition result confidence level t1 and a second recognition result confidence level t2; judging whether S1 = S2 holds, when S1 = S2 holds, judging whether T1 * t1 > T2 * t2 holds, when it holds, outputting the first cloud scene result S1, otherwise performing cloud scene recognition and outputting a second cloud scene result; when S1 = S2 does not hold, judging whether T1 * t1 < T2 * t2 holds, when it holds, outputting the first local scene result S2, otherwise performing local scene recognition and outputting a second local scene result; The voice scene recognition and judgment module is used to judge whether the first recognition confidence level T1 is much greater than the second recognition confidence level T2, when it is much greater, performing local scene recognition and outputting a second local scene result, otherwise judging whether S1 = S2 holds; The voice scene recognition and judgment module is used to perform intent scene positioning on the first voice recognition result to obtain the first cloud scene result S1, and perform intent scene positioning on the second voice recognition result to obtain the first local scene result S2; The voice scene recognition and judgment module is used to judge whether the first local scene result S2 is empty, when the first local scene result S2 is empty, performing cloud scene recognition and outputting the second cloud scene result, otherwise judging whether T1 * t1 > T2 * t2 holds, when it holds, outputting the first cloud scene result S1, otherwise performing local scene recognition and outputting a second local scene result.
2. A speech recognition system based on cloud computing and edge computing as claimed in claim 1, characterized in that: The voice input processing module is used to upload the collected voice information to the cloud and local respectively after noise cancellation and voice detection.
3. A speech recognition system based on cloud computing and edge computing as claimed in claim 1, characterized in that: The voice input processing module is used to judge whether there is a network, when there is a network, uploading the voice information to the cloud and local respectively, when there is no network, uploading the voice information to the local.
4. A speech recognition system based on cloud computing and edge computing as claimed in claim 1, characterized in that: The cloud voice processing module is used to process and calculate the voice information through the cloud, including recognizing the voice information through a cloud voice database; the local edge voice processing module: for processing and calculating the voice information locally, including recognizing the voice information through a local voice database.
5. The speech recognition system based on cloud computing and edge computing as claimed in claim 1, characterized in that: It further includes a voice interaction module: for feedback and output of voice interaction information according to the output result.
6. An electronic device, characterized in that: Comprising: A memory, for storing one or more programs; Processor; when the one or more programs are executed by the processor, the system according to any one of claims 1 to 5 is implemented.
7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the system according to any one of claims 1 to 5 is implemented.
