Anti-shake method and system for intelligent portable camera

Abstract

The invention discloses an anti-shake method and system for an intelligent portable camera, and belongs to the technical field of camera anti-shake. Camera video frames, cellular positioning data and TP interface feedback data are obtained, the TP interface feedback data are used as a priori condition to fuse the camera video frames and the cellular positioning data to obtain comprehensive motion reference data, the data and the camera video frames are combined to obtain camera motion parameters, and a compensation strategy is generated to complete anti-shake. And meanwhile, the system has the functions of remote interaction, real-time transmission and storage based on the 4G network and voice communication. According to the method, the precision of an EIS anti-shake algorithm is improved through multi-source data fusion, a complex scene is adapted in combination with a motion estimation constraint condition and a dynamic compensation strategy, the problems that a traditional EIS is insufficient in precision and low in processing speed are solved, diversified requirements of miniaturized intelligent equipment are met through multifunctional integration, and the shooting quality and the user experience are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of camera anti - shake, and more particularly to an anti - shake method and system for an intelligent portable camera. Background Art

[0002] With the development of mobile Internet and photography technology, people's demand for portable shooting devices is increasing day by day. Intelligent portable cameras have received wide attention due to their advantages such as small size, portability, and communication functions.

[0003] However, in the actual shooting process, camera shake will cause the captured images or videos to be blurred, seriously affecting the shooting quality. Although the traditional optical image stabilization (OIS) technology can solve the shake problem to a certain extent, it has disadvantages such as high cost and large volume, and is not suitable for small devices such as card - type cameras. The electronic image stabilization (EIS) anti - shake algorithm uses digital signal processing technology to process images to compensate for camera shake. However, the current EIS - based anti - shake methods still have problems such as low accuracy and slow processing speed in actual applications, and cannot meet the user's demand for high - quality shooting.

[0004] Therefore, how to propose an anti - shake method and system for an intelligent portable camera to improve the anti - shake accuracy is an urgent problem to be solved by those skilled in the art. Summary of the Invention

[0005] In view of this, the present invention provides an anti - shake method and system for an intelligent portable camera, which can effectively improve the anti - shake accuracy, improve the quality of captured images and videos, and at the same time meet the miniaturization and intelligentization requirements of card - type intelligent portable cameras.

[0006] To achieve the above object, the present invention adopts the following technical solutions:

[0007] On the one hand, the present invention proposes an anti - shake method for an intelligent portable camera, including:

[0008] Obtaining multi - source data, where the multi - source data includes camera video frames, cellular positioning data, and TP interface feedback data;

[0009] Taking the TP interface feedback data as a prior condition, fusing the camera video frames and the cellular positioning data to obtain comprehensive motion reference data;

[0010] Taking the camera video frames as the original input of the EIS anti - shake algorithm, and combining with the comprehensive motion reference data to obtain camera motion parameters;

[0011] Generating a compensation strategy according to the camera motion parameters, performing compensation transformation on the camera video frames, and completing anti - shake.

[0012] Preferably, taking the data feedback from the TP interface as a prior condition, the camera video frames and the cellular positioning data are fused to obtain comprehensive motion reference data, including:

[0013] Preprocess the camera video frames, the cellular positioning data, and the data feedback from the TP interface;

[0014] The preprocessed data feedback from the TP interface is mapped to a jitter risk level through a machine learning pre-trained model;

[0015] Dynamically adjust the fusion weights of the camera video frames and the cellular positioning data according to the jitter risk level; at the same time, convert the jitter risk level into a motion estimation constraint condition;

[0016] Perform weighted averaging on the preprocessed camera video frames and the cellular positioning data according to the adjusted fusion weights to obtain a comprehensive displacement vector;

[0017] The comprehensive displacement vector and the motion estimation constraint condition together constitute the comprehensive motion reference data.

[0018] Preferably, taking the camera video frames as the original input of the EIS anti-shake algorithm, combining with the comprehensive motion reference data to obtain camera motion parameters, including:

[0019] Perform feature point matching between adjacent camera video frames, and calculate the preliminary camera motion parameters according to the matched feature point pairs;

[0020] Judge whether the preliminary motion parameters meet the motion estimation constraint conditions;

[0021] If it is satisfied, the preliminary motion parameters are weighted averaged with the comprehensive displacement vector and then combined with the global trend prediction result based on the Kalman filter to obtain the final camera motion parameters;

[0022] If not, activate the reinforcement compensation strategy and / or enable the backup strategy.

[0023] Preferably, the motion estimation conditions include translation / rotation threshold conditions and feature point matching quality conditions;

[0024] If the translation / rotation threshold condition is not met, activate the reinforcement compensation strategy;

[0025] If the feature point matching quality condition is not met, enable the backup strategy.

[0026] If the translation / rotation threshold condition and the feature point matching quality condition are not met, activate the reinforcement compensation strategy and enable the backup strategy at the same time.

[0027] Preferably, the enhanced compensation strategy includes: expanding the compensation range; adopting a bicubic interpolation algorithm; increasing the weight of cellular positioning data;

[0028] The backup strategy includes reusing historical motion parameters or a global motion priority mode;

[0029] Reusing historical motion parameters includes skipping the local preliminary motion parameters of the current frame and directly using the average motion parameters of the previous three frames as temporary parameters;

[0030] The global motion priority mode includes generating compensation parameters for compensation only based on the displacement converted from cellular positioning and the global trend prediction result based on the Kalman filter.

[0031] Preferably, an anti-shake method for an intelligent portable camera further includes:

[0032] Establishing a connection with a remote terminal, receiving and parsing remote control instructions; based on the feedback data of the TP interface, transmitting the working status to the remote terminal in real time to form a two-way interactive control;

[0033] Performing real-time remote transmission and storage of the camera video frames after anti-shake compensation based on the 4G network, and simultaneously implementing a voice call function.

[0034] On the other hand, the present invention also proposes an anti-shake system for an intelligent portable camera for implementing the above anti-shake method, including:

[0035] A data acquisition module for acquiring multi-source data, where the multi-source data includes camera video frames, cellular positioning data, and TP interface feedback data;

[0036] A data processing module for fusing the camera video frames and the cellular positioning data with the feedback data of the TP interface as a prior condition to obtain comprehensive motion reference data;

[0037] A motion parameter acquisition module for using the camera video frames as the original input of the EIS anti-shake algorithm and combining the comprehensive motion reference data to obtain camera motion parameters;

[0038] A compensation transformation module for generating a compensation strategy according to the camera motion parameters, performing compensation transformation on the camera video frames, and completing anti-shake.

[0039] A remote interaction module for establishing a connection with a remote terminal, receiving and parsing remote control instructions; based on the feedback data of the TP interface, transmitting the working status to the remote terminal in real time to form a two-way interactive control;

[0040] A multi-functional communication storage module is used to perform real-time remote transmission and storage of the camera video frames with anti-shake compensation based on the 4G network, and at the same time implement the voice call function.

[0041] As can be seen from the above technical solutions, compared with the prior art, the present invention discloses a method and system for anti-shake of an intelligent portable camera. By obtaining the camera video frames, cellular positioning data, and TP interface feedback data, the camera video frames and cellular positioning data are fused with the TP interface feedback data as the prior condition to obtain comprehensive motion reference data. Combining this data with the camera video frames to obtain the camera motion parameters and generate a compensation strategy to complete anti-shake, and at the same time having functions of remote interaction, real-time transmission and storage based on the 4G network, and voice call. The present invention uses multi-source data fusion to improve the accuracy of the EIS anti-shake algorithm, combines motion estimation constraint conditions with dynamic compensation strategies to adapt to complex scenarios, solves the problems of insufficient accuracy and slow processing speed of traditional EIS, and meets the diverse needs of small intelligent devices through multi-functional integration, improving the shooting quality and user experience. Description of the Drawings

[0042] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only the embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained according to the provided drawings.

[0043] Figure 1 It is the flowchart of the method provided by the present invention;

[0044] Figure 2 It is the system architecture diagram provided by the present invention. Detailed Embodiments

[0045] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0046] On the one hand, referring to Figure 1 , the embodiments of the present invention provide a method for anti-shake of an intelligent portable camera, including the following steps:

[0047] S1. Obtain multi-source data, and the multi-source data includes camera video frames, cellular positioning data, and TP interface feedback data.

[0048] Among them, the camera video frames are RGB video frames collected by the camera module at a frame rate of 30 - 60fps, with a resolution supporting 1080p / 720p. The light flux is controlled by the camera shutter switch and converted into digital signals by the main board image sensor.

[0049] The cellular positioning data is accessed to the 4G cellular network through the SIM card small board, and location information such as longitude, latitude, and timestamp is obtained using the base station signal. The positioning frequency is synchronized with the video frame rate (10fps in this embodiment), and the original data contains noise and signal drift.

[0050] TP interface feedback data: The TP assembly (touch screen) real - time collects touch coordinates, pressure values, and operation frequencies (swiping, long - pressing), and transmits them to the main board through the circuit connected by the TP back glue, reflecting the user's holding state (one - hand / two - hands) and operation intensity.

[0051] S2. Using the TP interface feedback data as a prior condition, fuse the camera video frames and cellular positioning data to obtain comprehensive motion reference data, including:

[0052] S21. Pre - process the camera video frames, cellular positioning data, and TP interface feedback data.

[0053] The pre - processing of camera video frames includes grayscale conversion and feature extraction. Specifically:

[0054] Use Gray = 0.299R + 0.587G + 0.114B to convert to a grayscale image to reduce the amount of calculation;

[0055] Use the ORB algorithm to detect FAST corner points, generate BRIEF descriptors, extract 500 - 1000 feature points per frame, and output the set of feature point coordinates {(u i ,v i )}.

[0056] The pre - processing of cellular positioning data is as follows:

[0057] Median filtering: Take the median of 5 consecutive positioning data to filter out outliers.

[0058] Scale conversion: Convert the meter - level displacement to pixel - level equivalent displacement according to the camera internal parameters (focal length f, field of view FOV). The formula is:

[0059] The pre - processing of TP interface feedback data includes:

[0060] Data cleaning: Remove invalid touches (false touches with a duration < 5ms) and retain valid operation data.

[0061] S22. The pre - processed TP interface feedback data is mapped to the jitter risk level through a machine - learning pre - trained model.

[0062] In this embodiment, the TP interface feedback data, including touch pressure, operation frequency, etc., is mapped into a jitter risk level, which is divided into high-risk and low-risk levels, through a pre-trained random forest model.

[0063] S23. Dynamically adjust the fusion weights of the camera video frames and the cellular positioning data according to the jitter risk level; at the same time, convert the jitter risk level into a motion estimation constraint condition.

[0064] Adjust the fusion weights α (camera weight) and β (cellular weight) of the camera video frames and the cellular positioning data according to the jitter risk level, satisfying α + β = 1:

[0065] In the case of high risk, give priority to trusting the details of the video frame. At this time, α = 0.7 and β = 0.3.

[0066] In the case of low risk, combine the macroscopic trend of cellular positioning. At this time, α = 0.4 and β = 0.6.

[0067] The pixel-level displacement reflects the jitter of the local details of the image (tiny hand shakes), and the meter-level displacement reflects the overall movement of the camera (low-frequency displacement during walking). The combination of the two can distinguish "effective movement" (the user actively moves the camera composition) from "invalid jitter" (unintentional random vibration), avoiding image distortion caused by excessive compensation of traditional EIS.

[0068] Furthermore, the motion estimation conditions include translation / rotation threshold conditions and feature point matching quality conditions;

[0069] The translation / rotation threshold conditions are: Set the maximum allowable translation amount (in this embodiment, set to high risk ≤ 15 pixels, low risk ≤ 30 pixels) and rotation angle (high risk ≤ 10°, low risk ≤ 20°) between adjacent frames according to the risk level.

[0070] The feature point matching quality conditions are: Require the number of matching points ≥ 50, and the RANSAC inlier rate ≥ 50% (inliers are valid matching points that conform to the homography model).

[0071] S24. Perform weighted averaging on the preprocessed camera video frames and cellular positioning data according to the adjusted fusion weights to obtain a comprehensive displacement vector.

[0072] For the displacement of the feature points of the preprocessed video frame Δfeat = (Δx feat , Δy feat ) and the equivalent displacement of cellular positioning Δcell = (Δx cell , Δy cell ) are weighted averaged to obtain a comprehensive displacement vector Δfusion:

[0073] Δfusion = α·Δfeat + β·Δcell。

[0074] S25. The combined displacement vector and the motion estimation constraint conditions together constitute the combined motion reference data, and the final combined motion reference data is (Δfusion, constraint conditions).

[0075] S3. Use the camera video frame as the original input of the EIS anti-shake algorithm, and combine the combined motion reference data to obtain the camera motion parameters, including:

[0076] S31. Perform feature point matching between adjacent camera video frames, and calculate the preliminary motion parameters of the camera according to the matched feature point pairs.

[0077] Use the BFMatcher algorithm to match the feature points of adjacent frames, calculate the Hamming distance to filter the matching pairs, and eliminate the mismatched points through the RANSAC algorithm to retain the inlier set I.

[0078] Based on the inliers I, solve the homography matrix H by the least squares method, and decompose it to obtain the preliminary motion parameters: translation amount (Δx local , Δy local ), rotation angle θ local .

[0079] S32. Determine whether the preliminary motion parameters meet the motion estimation constraint conditions;

[0080] If satisfied (the preliminary parameters are within the threshold and the matching quality meets the standard), then the preliminary motion parameters are weighted averaged with the combined displacement vector and then combined with the global trend prediction result based on the Kalman filter to obtain the final camera motion parameters, specifically including:

[0081] Fusion of the preliminary parameters and the combined displacement vector: Weightedly average the preliminary motion parameters (Δx local , Δy local , θ local ) with the combined displacement vector. Taking Δx local as an example, the weighting formula is as follows:

[0082] Δx = α·Δx local + (1 - α)·Δx fusion ; α is determined by the jitter risk level.

[0083] Introduce the predicted trend value of the Kalman filter Smooth the fused parameters as follows to obtain the final camera motion parameters.

[0084]

[0085] λ is the trend correction coefficient, which is dynamically adjusted according to the historical motion stability, and the range is 0 to 0.3.

[0086] If not satisfied, activate the enhanced compensation strategy and / or enable the backup strategy.

[0087] If the translation / rotation threshold condition is not satisfied, activate the enhanced compensation strategy, including:

[0088] Expand the compensation range: During normal compensation, only 5% of the video frame edge may be processed. After activating the enhanced compensation strategy, the compensation range is expanded to 20% of the edge to handle larger amplitude jitters.

[0089] Adopt a more advanced interpolation algorithm: Switch from the bilinear interpolation algorithm with relatively small computational complexity to the bicubic interpolation algorithm with larger computational complexity but better compensation effect, reducing the edge blurring phenomenon that may occur after large displacement compensation.

[0090] Increase the weight of cellular positioning data: Increase the weight of cellular positioning data in the calculation of motion parameters from 30% to 60%, give priority to referring to the macro motion trend, and avoid misjudging the user's active movement as jitter.

[0091] If the feature point matching quality condition is not satisfied, enable the backup strategy, including:

[0092] Reuse historical motion parameters: Skip the local preliminary motion parameters of the current frame and directly use the average motion parameters of the previous 3 frames as temporary parameters to maintain the basic anti-shake function.

[0093] Global motion priority mode: Generate compensation parameters only based on the displacement converted from cellular positioning and the global trend prediction result based on Kalman filter, reducing the dependence on the feature point matching result of the current frame.

[0094] If the translation / rotation threshold condition and the feature point matching quality condition are not satisfied, activate the enhanced compensation strategy and enable the backup strategy simultaneously.

[0095] When the translation and rotation exceed the threshold and the feature point matching quality does not meet the standard, the camera faces a more complex jitter situation. At this time, more aggressive compensation measures and more reliable parameter calculation methods need to be taken simultaneously. Specifically, on the one hand, expand the compensation range, adopt an advanced interpolation algorithm, and increase the weight of cellular positioning data; on the other hand, reuse historical motion parameters or adopt the global motion priority mode to ensure effective anti-shake effect in complex situations.

[0096] S4. Generate a compensation strategy based on the camera motion parameters, perform compensation transformation on the camera video frame, and complete anti-shake.

[0097] According to the final motion parameters (Δx final , Δy final , θ final) Perform an inverse transformation on the current frame:

[0098] Translation compensation: Adjust the image position through the image translation matrix ;

[0099] Rotation compensation: Correct the rotational jitter through the rotation matrix ;

[0100] Edge filling: Use bilinear interpolation or bicubic interpolation to fill the edge area after transformation, keeping the image size consistent.

[0101] Furthermore, an anti-shake method for an intelligent portable camera in this embodiment further includes:

[0102] Establish a connection with a remote terminal, receive and parse remote control instructions; based on the feedback data of the TP interface, transmit the working status to the remote terminal in real time to form a two-way interactive control.

[0103] The remote interaction establishes a TCP / UDP connection through a 4G module (SIM card small board), receives JSON format instructions from the remote terminal (mobile phone / PC), parses them to control the camera shutter switch, adjusts the anti-shake parameters, and transmits the device status (battery power, storage capacity, signal strength) in real time.

[0104] Based on the 4G network, perform real-time remote transmission and storage of the camera video frames after anti-shake compensation, and at the same time implement the voice call function.

[0105] Specifically, during real-time transmission, first encode the anti-shake video frames through H.264 (bit rate 1 - 4 Mbps), encapsulate them with the voice data collected by the microphone (Opus encoding, bit rate 16 kbps) into the FLV format, and push the stream to the live server or upload it to the cloud through the RTMP protocol.

[0106] The video / images are named according to the time stamp and stored in the built-in eMMC or external USB storage, supporting circular recording (overwriting the earliest file) and AES-128 encryption of sensitive data.

[0107] During a voice call, the microphone signal at the collection end is processed by AEC (echo cancellation) and ANS (adaptive noise reduction) and then encoded into the PCM format, and real-time call is realized through VoIP technology; after decoding at the receiving end, it is played through the speaker, supporting volume adjustment by sliding the TP interface, and the video stream is displayed in the picture-in-picture mode during the call.

[0108] On the other hand, as Figure 2 shown, an anti-shake system for an intelligent portable camera is also proposed in an embodiment of the present invention for implementing the above anti-shake method, including:

[0109] A data acquisition module for acquiring multi-source data, where the multi-source data includes camera video frames, cellular positioning data, and TP interface feedback data;

[0110] A data processing module for fusing the camera video frames and the cellular positioning data with the TP interface feedback data as a prior condition to obtain comprehensive motion reference data;

[0111] A motion parameter acquisition module for using the camera video frames as the original input of the EIS anti-shake algorithm and combining with the comprehensive motion reference data to obtain camera motion parameters;

[0112] A compensation transformation module for generating a compensation strategy according to the camera motion parameters, performing compensation transformation on the camera video frames, and completing anti-shake.

[0113] A remote interaction module for establishing a connection with a remote terminal, receiving and parsing remote control instructions; based on the TP interface feedback data, real-time transmitting the working state back to the remote terminal to form a two-way interactive control;

[0114] A multi-functional communication and storage module for performing real-time remote transmission and storage of the camera video frames after anti-shake compensation based on the 4G network, and simultaneously implementing a voice call function.

[0115] In this specification, the various embodiments are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. The same or similar parts among the various embodiments can be referred to each other. For the devices disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple, and the relevant parts can be referred to the description of the method part.

[0116] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be obvious to those skilled in the art. The general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. An anti-shake method for an intelligent portable camera, characterized in that, Including: Obtain multi-source data, where the multi-source data includes camera video frames, cellular positioning data, and TP interface feedback data; Using the TP interface feedback data as a prior condition, fuse the camera video frames and the cellular positioning data to obtain comprehensive motion reference data; Take the camera video frames as the original input of the EIS anti-shake algorithm, and combine with the comprehensive motion reference data to obtain camera motion parameters; Generate a compensation strategy according to the camera motion parameters, perform compensation transformation on the camera video frames, and complete anti-shake.

2. The anti-shake method of an intelligent portable camera according to claim 1, wherein, Using the TP interface feedback data as a prior condition, fuse the camera video frames and the cellular positioning data to obtain comprehensive motion reference data, including: Preprocess the camera video frames, the cellular positioning data, and the TP interface feedback data; The preprocessed TP interface feedback data is mapped to a jitter risk level through a machine learning pre-trained model; Dynamically adjust the fusion weights of the camera video frames and the cellular positioning data according to the jitter risk level; at the same time, convert the jitter risk level into a motion estimation constraint condition; Perform weighted averaging on the preprocessed camera video frames and the cellular positioning data according to the adjusted fusion weights to obtain a comprehensive displacement vector; The comprehensive displacement vector and the motion estimation constraint condition together constitute comprehensive motion reference data.

3. The anti-shake method of an intelligent portable camera according to claim 2, wherein Take the camera video frames as the original input of the EIS anti-shake algorithm, and combine with the comprehensive motion reference data to obtain camera motion parameters, including: Perform feature point matching between adjacent camera video frames, and calculate the preliminary motion parameters of the camera according to the matched feature point pairs; Judge whether the preliminary motion parameters meet the motion estimation constraint conditions; If they meet, the preliminary motion parameters are weighted averaged with the comprehensive displacement vector and then combined with the global trend prediction result based on the Kalman filter to obtain the final camera motion parameters; If they do not meet, activate the enhanced compensation strategy and / or enable the backup strategy.

4. The anti-shake method of an intelligent portable camera according to claim 3, characterized in that, The motion estimation conditions include translation / rotation threshold conditions and feature point matching quality conditions; If the translation / rotation threshold conditions are not met, activate the enhanced compensation strategy; If the feature point matching quality conditions are not met, enable the backup strategy; If the translation / rotation threshold conditions and the feature point matching quality conditions are not met, activate the enhanced compensation strategy and enable the backup strategy at the same time.

5. The anti-shake method of an intelligent portable camera according to claim 4, characterized in that, The enhanced compensation strategy includes: expanding the compensation range; adopting the bicubic interpolation algorithm; increasing the weight of the cellular positioning data; The backup strategy includes reusing historical motion parameters or the global motion priority mode; The reuse of historical motion parameters includes skipping the local preliminary motion parameters of the current frame and directly using the average motion parameters of the previous 3 frames as temporary parameters; The global motion priority mode includes generating compensation parameters only based on the displacement converted by the cellular positioning and the global trend prediction result based on the Kalman filter for compensation.

6. The anti-shake method of an intelligent portable camera according to claim 1, characterized in that Also including: Establish a connection with the remote terminal, receive and parse remote control instructions; Based on the TP interface feedback data, transmit the working status to the remote terminal in real time to form two-way interactive control; Perform real-time remote transmission and storage of the camera video frames with anti-shake compensation based on the 4G network, and at the same time implement the voice call function.

7. An anti-shake system for an intelligent portable camera, characterized in that, It includes: A data acquisition module for acquiring multi-source data, where the multi-source data includes camera video frames, cellular positioning data, and TP interface feedback data; A data processing module for fusing the camera video frames and the cellular positioning data with the TP interface feedback data as a prior condition to obtain comprehensive motion reference data; A motion parameter acquisition module for using the camera video frames as the original input of the EIS anti-shake algorithm and combining the comprehensive motion reference data to obtain camera motion parameters; A compensation transformation module for generating a compensation strategy according to the camera motion parameters, performing compensation transformation on the camera video frames, and completing anti-shake.

8. The anti-shake system of an intelligent portable camera according to claim 7, characterized in that, It further includes: A remote interaction module for establishing a connection with a remote terminal, receiving and parsing remote control instructions; Based on the TP interface feedback data, real-time transmit the working status back to the remote terminal to form a two-way interactive control; A multi-functional communication storage module for performing real-time remote transmission and storage of the camera video frames with anti-shake compensation based on the 4G network, and at the same time implementing the voice call function.

Images