Method and equipment for realizing double-screen simultaneous display real-time translation machine

Abstract

The invention provides a double-screen simultaneous display real-time translation machine implementation method and device, and the method comprises the following steps: configuring hardware: selecting an SSD2381 chip to drive two touch screen modules, and enabling the two touch screen modules to display the same content; touch screen modules are installed, wherein a rotatable aluminum alloy support is manufactured, and the two touch screen modules are installed on the support. Multi-task processing is achieved through a multi-task processing algorithm, the multi-task processing algorithm can monitor the resource use condition of a chip in real time, resources are dynamically allocated according to the task requirements, it is ensured that all tasks can run efficiently and parallelly through a time slice rotation and priority preemption mechanism, the overall performance is improved, and the method is suitable for large-scale popularization and application. The system bottleneck caused by task waiting is reduced, and the multi-task processing algorithm can also realize load balancing and uniformly distribute tasks to each processing unit of the system, so that the condition that some processing units are overloaded and other processing units are idle is avoided, and the overall performance and efficiency are improved.

Description

Technical Field

[0001] The present invention relates to a method and a device, in particular to a method and a device for realizing a dual-screen simultaneous display real-time translation machine, belonging to the technical field of speech translation. Background Art

[0002] A real-time translator is an electronic device that can instantly translate one language into another. It usually has functions such as speech recognition, text translation, and speech synthesis, which can achieve instant translation of spoken or written language and help people with different language backgrounds communicate. Real-time translators are widely used in travel, business, international conferences, and education.

[0003] When the real-time translator is working, it first converts human voice signals into digital signals that can be processed by computers through speech recognition technology. This process includes microphone acquisition of voice signals, preprocessing of signals (such as noise removal, amplification and filtering), feature extraction (such as spectrum, cepstrum, etc.), and pattern matching, that is, matching the extracted features with the pre-trained voice model, identifying the corresponding words and phrases, and then converting the source language text into the target language text. Through a series of complex technical processing steps, the real-time translator realizes the function of instantly translating one language into another, which greatly promotes international communication and cooperation.

[0004] In the existing technology, the system performance of the dual-screen simultaneous display real-time translation machine often faces bottlenecks. The root cause of this problem is the lack of an efficient multi-tasking processing mechanism. The lack of this mechanism directly leads to the lag of the overall system performance and response speed, which in turn affects the user experience and the practicality of the device: When a dual-screen real-time translator processes multiple tasks simultaneously, such as speech recognition, machine translation, screen display updates, and possible network communications, if there is no reasonable and efficient multi-tasking mechanism to coordinate the execution of these tasks, the system is prone to resource conflicts, task blocking, and other problems. These problems will not only reduce the execution efficiency of a single task, but also cause the entire system to become stuck or crash. Therefore, introducing an efficient multi-tasking mechanism is the key to solving the performance bottleneck of the dual-screen real-time translator system. To this end, a dual-screen real-time translator implementation method and device are proposed. Summary of the invention

[0005] In view of this, the present invention provides a method and device for implementing a dual-screen simultaneous display real-time translation machine to solve or alleviate the technical problems existing in the prior art and at least provide a beneficial option.

[0006] The technical solution of the embodiment of the present invention is implemented as follows: a method for implementing a dual-screen simultaneous display real-time translation machine comprises the following steps: Configure hardware: Select SSD2381 chip to drive two touch screen modules, and the two touch screen modules display the same content; Touch screen module installation: Make a rotatable aluminum alloy bracket and install two touch screen modules on the bracket; Configure the sound collection module: Arrange the microphones according to the orientation of the touch screen module, and use the denoising algorithm to reduce the impact of environmental noise on the sound collection module; Establish multi-tasking mechanism: SSD2381 chip provides task scheduling function, assigns tasks to available processor cores, and implements multi-tasking through multi-tasking algorithm; Real-time translation: Use the LSTM model to achieve real-time translation of sound signals; System testing: functional testing, performance testing and stability testing. Functional testing includes screen display testing, sound collection and translation testing, and user interaction testing. Performance testing includes processing speed testing and power consumption testing. Stability testing includes long-term operation testing and environmental adaptability testing.

[0007] Further preferably, when configuring the hardware, the touch screen module is connected to the SSD2381 chip using Mipi and I2C interfaces, and a driver for the SSD2381 is written to implement screen initialization, image data transmission, and touch event processing functions.

[0008] Further preferably, when the touch screen module is installed, a rotating shaft is installed on the aluminum alloy bracket to allow the touch screen module to rotate freely, and a locking knob for fixing the touch screen module is provided on the bracket.

[0009] Further preferably, when configuring the sound collection module, a microphone with low noise, high sensitivity and wide-band response characteristics is selected; The denoising algorithm comprises the following steps: Signal preprocessing: The collected noisy speech signal is framed, the length of each frame is selected to be 20-30 milliseconds, the frame shift is 8-11 milliseconds, and each frame signal is windowed; Noise estimation: During the inactive period of the speech signal, the power spectral density of the background noise is estimated using the LMS algorithm; Adaptive filtering: Using the estimated noise characteristics, an adaptive filter is designed, which is used to dynamically adjust its parameters according to the input noisy speech signal; Spectral subtraction: Perform Fourier transform on the adaptively filtered speech signal to obtain a spectrum representation. Based on the noise estimation result, the estimated noise spectrum is subtracted from the spectrum of the speech signal. Inverse Fourier transform and reconstruction: Perform inverse Fourier transform on the processed spectrum to obtain the denoised speech signal in the time domain, and perform overlap-addition or overlap-save processing on the inverse transformed signal to restore the continuity of the original signal; Post-processing: Post-process the denoised speech signal, including removing the residual DC component and performing gain adjustment.

[0010] Further preferably, when establishing a multi-tasking mechanism, the SSD2381 chip is divided into multiple independent control areas, each of which is responsible for a specific function; Using an asynchronous programming model, tasks do not block the entire system while waiting for I / O operations or other resources; The multi-tasking algorithm comprises the following steps: Task division and priority setting: The functions of the dual-screen real-time translator are divided into multiple independent task modules, including speech recognition tasks, machine translation tasks, and display control tasks. A priority is set for each task based on the importance and real-time requirements of the task; Task queue management: Create a task queue to store tasks to be executed. Each task in the task queue contains task ID, priority, and execution function information. The task queue is sorted according to the priority of the task, and high-priority tasks are executed first; Round-robin scheduling: A fixed time slice is assigned to each task. The size of the time slice is set according to the processing capability of the chip and the complexity of the task. The chip executes the tasks in the queue in turn according to the time slice rotation method. Priority preemption mechanism: Based on the time slice round-robin scheduling, a priority preemption mechanism is introduced. When a high-priority task arrives, if the time slice of the currently executed task has not been used up, the current task will be interrupted immediately, and the high-priority task will be inserted to the front of the queue and executed. The interrupted task will continue to execute within its remaining time slice, or the time slice will be reallocated as needed; Task synchronization and communication: Design synchronization and communication mechanisms between tasks. For tasks that require frequent interaction, design dedicated data buffers or shared memory areas to store and access data. Resource management and optimization: Monitor the resource usage of the chip, including CPU usage and memory occupancy. For tasks that consume a lot of resources, asynchronous execution is used to implement them.

[0011] Further preferably, during real-time translation, the LSTM model is used to implement real-time translation of sound signals, which specifically includes the following steps: Dataset preparation: Collect and prepare multilingual conversation datasets for training LSTM models; Model training: Use the deep learning framework to train the LSTM model and optimize the model parameters; Model evaluation and optimization: Evaluate model performance through the test set and optimize the model based on the evaluation results; Real-time translation: The trained LSTM model is embedded into the SSD2381 chip to achieve real-time translation of sound signals.

[0012] Further preferably, the functional test specifically includes: Screen display test: test the display effect of the screen at different angles and lighting conditions; Sound collection and translation test: Test the sound collection effect of the microphone, and the translation accuracy and speed of the real-time translation module; User interaction testing: testing the usability and responsiveness of the user interface; The performance test specifically includes: Processing speed test: Test the processing speed of the device under different loads; Power consumption test: Test the power consumption of the device and optimize the power management strategy.

[0013] Further preferably, the stability test specifically includes: Long-term operation test: Let the device run for a long time to observe whether any abnormalities occur; Environmental adaptability test: Test the stability and reliability of the equipment in environments with different temperatures, humidity, and electromagnetic interference.

[0014] A dual-screen simultaneous display real-time translation device, comprising a first touch screen, a second touch screen, an SSD2381 chip and a digital microphone; The digital microphones are installed inside the first touch screen and inside the second touch screen respectively, and the signal end of the SSD2381 chip is connected to the signal end of the first touch screen and the signal end of the second touch screen respectively.

[0015] Further preferably, the digital microphone is used to collect voice data of the user, and the voice data is sent to the SSD2381 chip; The SSD2381 chip is used to realize real-time translation of sound signals using the LSTM model, and the translated data is displayed through the first touch screen and the second touch screen respectively.

[0016] The embodiment of the present invention has the following advantages due to the adoption of the above technical solution: 1. The present invention provides a task scheduling function through the SSD2381 chip, assigns tasks to available processor cores, realizes multi-tasking through a multi-tasking processing algorithm, can monitor the resource usage of the chip in real time, and dynamically allocate resources according to the needs of the tasks. Through the time slice rotation and priority preemption mechanism, it is ensured that each task can run efficiently and in parallel, thereby improving the overall performance. The scheduling mechanism in the algorithm can arrange the execution order according to the priority of the task, and high-priority tasks can obtain processing resources first, thereby ensuring the timely execution of key tasks and reducing system bottlenecks caused by task waiting. In addition, the multi-tasking processing algorithm can also achieve load balancing, evenly assign tasks to various processing units of the system, avoid the situation where some processing units are overloaded while other processing units are idle, and improve the overall performance and efficiency.

[0017] 2. The present invention reduces the impact of environmental noise on the sound collection module through a denoising algorithm, and can adaptively adjust the filtering parameters according to the changes in environmental noise, so as to maintain a good denoising effect in different noise environments. At the same time, it combines the advantages of adaptive filtering and spectral subtraction, and can effectively remove background noise and residual noise, improve the signal-to-noise ratio and intelligibility of speech signals. Moreover, the computational complexity of the denoising algorithm is moderate, and it can achieve fast processing in a real-time translation machine to meet the needs of real-time communication.

[0018] The above summary is for illustrative purposes only and is not intended to be limiting in any way. In addition to the illustrative aspects, embodiments and features described above, further aspects, embodiments and features of the present invention will be readily apparent by reference to the accompanying drawings and the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0020] Figure 1 A flowchart of a method for implementing a dual-screen simultaneous display real-time translation machine according to the present invention; Figure 2 is a flowchart of the steps of the denoising algorithm of the present invention; Figure 3 is a flow chart of the steps of the multi-tasking algorithm of the present invention; Figure 4 This is a principle structure diagram of a dual-screen simultaneous display real-time translation device of the present invention; Figure 5The present invention is a schematic diagram of the working process of a dual-screen simultaneous display real-time translation device. DETAILED DESCRIPTION

[0021] In the following, only some exemplary embodiments are briefly described. As those skilled in the art will appreciate, the described embodiments may be modified in various ways without departing from the spirit or scope of the present invention. Therefore, the drawings and descriptions are considered to be exemplary and non-restrictive in nature.

[0022] The embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0023] like Figure 1-Figure 3 As shown, an embodiment of the present invention provides a method for implementing a dual-screen simultaneous display real-time translation machine, comprising the following steps: Configure the hardware: Select SSD2381 chip to drive two touch screen modules. The two touch screen modules display the same content. By clicking the touch screen with your fingers, you can control the content displayed on the screen. Confirm the interface compatibility between the touch screen module and the SSD2381 chip, use the Mipi and I2C interfaces to connect the touch screen module to the SSD2381 chip, write the SSD2381 driver, and implement screen initialization, image data transmission, and touch event processing functions. The SSD2381 chip has a built-in high-performance 4-core 64-bit ARM Cortex-A55 core processor with a main frequency of up to 1.5GHz. This high-performance processor core can ensure that the translator performs well in processing complex translation tasks and provides fast and accurate translation results.

[0024] Touch screen module installation: A rotatable aluminum alloy bracket is made by using aluminum alloy material to reduce the overall weight. Two touch screen modules are installed on the bracket. A rotating shaft is installed on the aluminum alloy bracket to allow the touch screen module to rotate freely. A locking knob for fixing the touch screen module is provided on the bracket to ensure that the touch screen module can be stably fixed after adjusting the angle.

[0025] Configure the sound collection module: The microphones are arranged according to the orientation of the touch screen module, and the noise reduction algorithm is used to reduce the impact of environmental noise on the sound collection module. The denoising algorithm can adaptively adjust the filter parameters according to the changes in environmental noise, thereby maintaining a good denoising effect in different noise environments. At the same time, combining the advantages of adaptive filtering and spectrum subtraction, it can effectively remove background noise and residual noise, improve the signal-to-noise ratio and intelligibility of speech signals, and the computational complexity of the denoising algorithm is moderate, which can achieve fast processing in the real-time translator to meet the needs of real-time communication. Among them, the denoising algorithm includes the following steps: Signal preprocessing: The collected noisy speech signal is divided into frames, the length of each frame is selected to be 20-30 milliseconds, the frame shift is 8-11 milliseconds, and each frame signal is windowed to reduce the discontinuity of the frame edge; Noise estimation: During the inactive period of the speech signal, the LMS algorithm is used to estimate the power spectral density of the background noise. The accuracy of noise estimation is crucial to the subsequent denoising effect, so it is necessary to ensure that noise samples are accurately identified and extracted during the inactive period; Adaptive filtering: Using the estimated noise characteristics, an adaptive filter is designed. The filter is used to dynamically adjust its parameters according to the input noisy speech signal to remove the background noise to the greatest extent. The output of the adaptive filter will be the speech signal after preliminary denoising. Spectral subtraction: Perform Fourier transform on the adaptively filtered speech signal to obtain a spectrum representation. Based on the noise estimation result, the estimated noise spectrum is subtracted from the spectrum of the speech signal to achieve further denoising. The LMS algorithm is an adaptive filtering algorithm that minimizes the mean square value of the output error by iteratively adjusting the filter coefficients; In order to avoid the introduction of musical noise, the spectrum after subtraction can be smoothed, such as using techniques such as spectral weighting or nonlinear transformation; Inverse Fourier transform and reconstruction: Perform inverse Fourier transform on the processed spectrum to obtain the denoised speech signal in the time domain, and perform overlap-addition or overlap-save processing on the inverse transformed signal to restore the continuity of the original signal; Post-processing: Post-process the denoised speech signal, including removing the residual DC component and adjusting the gain to ensure the sound quality and intelligibility of the output signal; When configuring the sound collection module, choose a microphone with low noise, high sensitivity, and wide-band response characteristics.

[0026] Establish a multi-tasking mechanism: The SSD2381 chip provides task scheduling functions, assigning tasks to available processor cores and implementing multitasking through multitasking algorithms; The multi-tasking algorithm can monitor the resource usage of the chip in real time and dynamically allocate resources according to the needs of the task. Through the time slice rotation and priority preemption mechanism, it ensures that each task can run efficiently and in parallel, improving the overall performance. The scheduling mechanism in the algorithm can arrange the execution order according to the priority of the task. High-priority tasks can obtain processing resources first, thereby ensuring the timely execution of key tasks and reducing system bottlenecks caused by task waiting. In addition, the multi-tasking algorithm can also achieve load balancing and evenly distribute tasks to each processing unit of the system, avoiding the situation where some processing units are overloaded while other processing units are idle, thereby improving the overall performance and efficiency. When establishing a multi-tasking mechanism, the SSD2381 chip is divided into multiple independent control areas, each of which is responsible for a specific function; Using an asynchronous programming model, tasks do not block the entire system while waiting for I / O operations or other resources; The multi-tasking algorithm includes the following steps: Task division and priority setting: The functions of the dual-screen real-time translator are divided into multiple independent task modules, including speech recognition tasks, machine translation tasks, and display control tasks. A priority is set for each task according to the importance and real-time requirements of the task. Among them, speech recognition and display control tasks are set as the first and second priorities respectively to ensure real-time performance and user experience; Task queue management: Create a task queue to store tasks to be executed. Each task in the task queue contains task ID, priority, and execution function information. The task queue is sorted according to the priority of the task, and high-priority tasks are executed first; Time slice round-robin scheduling: A fixed time slice is allocated to each task. The size of the time slice is set according to the processing power of the chip and the complexity of the task. The chip executes the tasks in the queue in turn according to the time slice round-robin method. When the time slice of a task is used up, if the task is not completed, it will be put back to the end of the queue and wait for the next scheduling; Priority preemption mechanism: Based on the time slice round-robin scheduling, a priority preemption mechanism is introduced. When a high-priority task arrives, if the time slice of the currently executed task has not been used up, the current task will be interrupted immediately, and the high-priority task will be inserted to the front of the queue and executed. The interrupted task will continue to execute within its remaining time slice, or the time slice will be reallocated as needed; Task synchronization and communication: Design synchronization and communication mechanisms between tasks to ensure data consistency and collaboration between tasks. For example, synchronization primitives such as semaphores and message queues can be used to achieve synchronization and communication between tasks. For tasks that require frequent interaction, design dedicated data buffers or shared memory areas to store and access data. Resource management and optimization: monitor the resource usage of the chip, including CPU usage and memory occupancy. For tasks that consume a lot of resources, asynchronous execution is used to implement them. Fault recovery and fault tolerance: Design fault recovery and fault-tolerant mechanisms to deal with abnormal situations that may occur during task execution. For example, use retry mechanisms, backup tasks and other strategies to ensure the reliable execution of tasks. For critical tasks, design redundant backup and fault switching mechanisms to improve system reliability and stability.

[0027] Real-time translation: The LSTM model is used to achieve real-time translation of sound signals, which includes the following steps: Dataset preparation: Collect and prepare multilingual conversation datasets for training LSTM models; Model training: Use the deep learning framework to train the LSTM model and optimize the model parameters; Model evaluation and optimization: Evaluate model performance through the test set and optimize the model based on the evaluation results; Real-time translation: The trained LSTM model is embedded into the SSD2381 chip to achieve real-time translation of sound signals; The LSTM model is a long short-term memory artificial neural network model.

[0028] System testing: perform functional testing, performance testing and stability testing; Functional testing includes screen display testing, sound collection and translation testing, and user interaction testing. Screen display test: Test the display effect of the screen at different angles and lighting conditions to ensure that users can get a clear visual experience in different environments, including: Angle test: Use a viewing angle measuring instrument to measure the changes in screen brightness and color parameters at different angles to determine the viewing angle range of the screen. For example, observe the content displayed on the screen at vertical, horizontal, 45-degree, 60-degree angles, etc. to check whether the brightness and color of the image change significantly. Light test: In a dark room, use a brightness meter to measure the brightness of the screen in full black and full white states, as well as the brightness changes under different light conditions. At the same time, observe the color reproduction and contrast performance of the screen under different light conditions. Sound collection and translation test: Test the sound collection effect of the microphone, the translation accuracy and speed of the real-time translation module, including: Sound collection test: Use professional audio testing equipment to record sound samples of the microphone at different distances, volumes, and speech speeds, and analyze their clarity, fidelity, and noise suppression capabilities; Translation test: Prepare voice samples in multiple languages, translate them through the real-time translation module, and compare the accuracy and speed of the translation results with the standard answers. At the same time, invite testers with multilingual backgrounds to participate in the test to verify the performance of the translation module in different language environments; User interaction testing: Test the usability and responsiveness of the user interface to ensure that users can complete operations easily and quickly, including: Usability testing: Invite users from different backgrounds to participate in the test, observe their operating habits, difficulties and problems when using the user interface, and collect their feedback; Response speed test: Use professional performance testing tools to simulate user operations and measure the response speed of the user interface, such as measuring the time difference between clicking a button and updating the interface, and the smoothness of sliding the screen; Performance testing includes processing speed testing and power consumption testing, among which: Processing speed test: Test the processing speed of the device under different loads to ensure that real-time translation and image processing can proceed smoothly, including: Load testing: By simulating multiple users accessing the system at the same time, performing operations such as translation and image processing, measuring the system's response time, throughput and other performance indicators, using professional load testing tools to simulate user behavior and generate test reports; Benchmark test: Under certain software, hardware and network environments, a certain number of virtual users are simulated to run one or more services. The test results are used as baseline data. In the subsequent tuning or system evaluation process, the same business scenarios are run and the test results are compared to determine whether the tuning is effective or to provide decision-making data for system selection. Power consumption test: Test the power consumption of the device and optimize the power management strategy to reduce energy consumption, including: Power consumption measurement: Use power consumption measurement equipment to measure the power consumption of the system in different working states, such as the changes in power consumption when in standby mode, running the translation module, performing image processing, etc. Power management policy test: By adjusting the system's power management policy, such as reducing screen brightness and closing unnecessary background applications, observe the reduction in power consumption and verify the effectiveness of the policy; Stability testing includes long-term operation testing and environmental adaptability testing. Long-term operation test: Let the device run for a long time to observe whether any abnormal conditions occur, including: Continuous operation test: Let the device run continuously for 72 hours to observe whether there are any abnormal situations, such as crashes and freezes. Use monitoring tools to record the device's operating status and error information. Log analysis: Analyze the log files of the device during operation to find potential problems and abnormal behaviors, and analyze the log information before the system crashes to determine the cause of the crash and the solution; Environmental adaptability test: Test the stability and reliability of the equipment in different temperature, humidity and electromagnetic interference environments, including: Temperature test: Place the device in different temperature environments, including high temperature, low temperature, and sudden temperature change conditions, and observe the operating status and performance changes of the device. A temperature test chamber can be used to simulate different temperature environments; Humidity test: Place the device in a high humidity environment to observe the moisture resistance and stability of the system, and use humidity test equipment to simulate a high humidity environment; Electromagnetic interference test: Place the device in a strong electromagnetic interference environment, such as electromagnetic pulses and radio frequency interference, to observe the device's anti-interference ability and stability. Electromagnetic interference test equipment can be used to simulate a strong electromagnetic interference environment.

[0029] like Figure 4-Figure 5 As shown, an embodiment of the present invention provides a dual-screen simultaneous display real-time translation device, including a first touch screen, a second touch screen, an SSD2381 chip and a digital microphone; The digital microphones are installed inside the first touch screen and inside the second touch screen respectively. The signal end of the SSD2381 chip is connected to the signal end of the first touch screen and the signal end of the second touch screen respectively. The dual-screen display device is realized by using the SSD2381 chip. Then, the digital microphones of the two screens are used to pick up the voices of people speaking different languages. The artificial intelligence algorithm is used to realize the automatic recognition of languages ​​and the automatic translation of voices into text. Then, the two output texts are superimposed and displayed on the two screens, so that two people speaking different languages ​​can communicate face to face without obstacles. The translation machine can be used in ports, scenic spots, shops, foreign-related hotels and other places. When guests who speak different languages ​​come, as long as the guest says a sentence, the system can automatically identify the guest's language within one second; then the clerk and the customer can communicate seamlessly based on the translation machine.

[0030] In one embodiment, a digital microphone is used to collect the user's voice data, and the voice data is sent to the SSD2381 chip; The SSD2381 chip is used to realize real-time translation of sound signals using the LSTM model. The translated data is displayed through the first touch screen and the second touch screen respectively. The two touch screens are driven simultaneously by SSD2381. The two touch screens display the same content. The content displayed on the screen can be controlled by clicking the touch screen with your fingers.

[0031] When in use: the two screens are displayed in different directions, and the user can adjust the display direction and angle within a certain range. There are two sets of microphones next to each screen, which can collect voice input from user A and user B respectively according to the screen direction. The sound collected from both sides is detected and analyzed in real time through the long short-term memory network algorithm. The system can automatically identify the language types on both sides, translate the conversation process into two texts and display them on the screen at the same time. People on both sides of the screen can see the translated real-time conversation record and communicate without obstacles.

[0032] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of various changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A method for implementing a dual-screen simultaneous display real-time translation machine, characterized in that: The following steps are involved: Configure hardware: Select SSD2381 chip to drive two touch screen modules, and the two touch screen modules display the same content; Touch screen module installation: Make a rotatable aluminum alloy bracket and install two touch screen modules on the bracket; Configure the sound collection module: Arrange the microphones according to the orientation of the touch screen module and reduce the ambient noise through the denoising algorithm; Establish multi-tasking mechanism: SSD2381 chip provides task scheduling function, assigns tasks to available processor cores, and implements multi-tasking through multi-tasking algorithm; Real-time translation: Use the LSTM model to achieve real-time translation of sound signals; System testing: functional testing, performance testing and stability testing. Functional testing includes screen display testing, sound collection and translation testing, and user interaction testing. Performance testing includes processing speed testing and power consumption testing. Stability testing includes long-term operation testing and environmental adaptability testing.

2. A method for implementing a dual-screen simultaneous display real-time translation machine according to claim 1, characterized in that: When configuring the hardware, use the Mipi and I2C interfaces to connect the touch screen module to the SSD2381 chip and write the driver for the SSD2381.

3. The method for implementing a dual-screen simultaneous display real-time translation machine according to claim 1, characterized in that: When the touch screen module is installed, a rotating shaft is installed on the aluminum alloy bracket, and the touch screen module rotates freely. A locking knob for fixing the touch screen module is provided on the bracket.

4. The method for implementing a dual-screen simultaneous display real-time translation machine according to claim 1, characterized in that: The denoising algorithm is used to reduce the impact of environmental noise on the sound collection module; The specific steps include: Signal preprocessing: The collected noisy speech signal is framed, the length of each frame is selected to be 20-30 milliseconds, the frame shift is 8-11 milliseconds, and each frame signal is windowed; Noise estimation: During the inactive period of the speech signal, the power spectral density of the background noise is estimated using the LMS algorithm; Adaptive filtering: Using the estimated noise characteristics, an adaptive filter is designed, which is used to dynamically adjust its parameters according to the input noisy speech signal; Spectral subtraction: Perform Fourier transform on the adaptively filtered speech signal to obtain a spectrum representation. Based on the noise estimation result, the estimated noise spectrum is subtracted from the spectrum of the speech signal. Inverse Fourier transform and reconstruction: Perform inverse Fourier transform on the processed spectrum to obtain the denoised speech signal in the time domain, and perform overlap-addition or overlap-save processing on the inverse transformed signal to restore the continuity of the original signal; Post-processing: Post-process the denoised speech signal, including removing the residual DC component and performing gain adjustment.

5. The method for implementing a dual-screen simultaneous display real-time translation machine according to claim 1, characterized in that: When establishing a multi-tasking mechanism, the SSD2381 chip is divided into multiple independent control areas, each of which is responsible for a specific function; Using an asynchronous programming model, tasks do not block the entire system while waiting for I / O operations or other resources; The multi-tasking algorithm comprises the following steps: Task division and priority setting: The functions of the dual-screen real-time translator are divided into multiple independent task modules, including speech recognition tasks, machine translation tasks, and display control tasks. A priority is set for each task based on the importance and real-time requirements of the task; Task queue management: Create a task queue to store tasks to be executed. Each task in the task queue contains task ID, priority, and execution function information. The task queue is sorted according to the priority of the task, and high-priority tasks are executed first; Round-robin scheduling: A fixed time slice is assigned to each task. The size of the time slice is set according to the processing capability of the chip and the complexity of the task. The chip executes the tasks in the queue in turn according to the time slice rotation method. Priority preemption mechanism: Based on the time slice round-robin scheduling, a priority preemption mechanism is introduced. When a high-priority task arrives, if the time slice of the currently executed task has not been used up, the current task will be interrupted immediately, and the high-priority task will be inserted to the front of the queue and executed. The interrupted task will continue to execute within its remaining time slice, or the time slice will be reallocated as needed; Task synchronization and communication: Design synchronization and communication mechanisms between tasks. For tasks that require frequent interaction, design dedicated data buffers or shared memory areas to store and access data. Resource management and optimization: Monitor the resource usage of the chip, including CPU usage and memory occupancy. For tasks that consume a lot of resources, asynchronous execution is used to implement them.

6. The method for implementing a dual-screen simultaneous display real-time translation machine according to claim 1, characterized in that: During real-time translation, the LSTM model is used to achieve real-time translation of sound signals, which specifically includes the following steps: Dataset preparation: Collect and prepare multilingual conversation datasets for training LSTM models; Model training: Use the deep learning framework to train the LSTM model and optimize the model parameters; Model evaluation and optimization: Evaluate model performance through the test set and optimize the model based on the evaluation results; Real-time translation: The trained LSTM model is embedded into the SSD2381 chip to achieve real-time translation of sound signals.

7. The method for implementing a dual-screen simultaneous display real-time translation machine according to claim 1, characterized in that: The functional test specifically includes: Screen display test; Sound collection and translation test; User interaction testing; The performance test specifically includes: Processing speed test; Power consumption test.

8. The method for implementing a dual-screen simultaneous display real-time translation machine according to claim 1, characterized in that: The stability test specifically includes: Long running tests; Environmental adaptability test.

9. A dual-screen simultaneous display real-time translation device, applied to a dual-screen simultaneous display real-time translation method as claimed in any one of claims 1 to 8, characterized in that: Includes a first touch screen, a second touch screen, an SSD2381 chip and a digital microphone; The digital microphones are installed inside the first touch screen and inside the second touch screen respectively, and the signal end of the SSD2381 chip is connected to the signal end of the first touch screen and the signal end of the second touch screen respectively.

10. The dual-screen simultaneous display real-time translation device according to claim 9, characterized in that: The digital microphone is used to collect the user's voice data, and the voice data is sent to the SSD2381 chip; The SSD2381 chip is used to realize real-time translation of sound signals using the LSTM model, and the translated data is displayed through the first touch screen and the second touch screen respectively.

Images