84results about How to "Increase sampling rate" patented technology

The invention provides an MRI (Magnetic Resonance Imaging) reconstruction method of a subspace diffusion model based on orthogonal decomposition. The MRI reconstruction method comprises the following steps: forming an open source brain data set; a subspace diffusion model based on orthogonal decomposition is constructed, and k-space data is decomposed into subspaces for diffusion, so that data distribution is better captured, and robustness is kept in a noise environment; designing an iterative solution algorithm for extracting prior information in the model; and applying an iterative solution algorithm to the model to test the test data set, and outputting a final restored image. According to the method, the learning of effective information in complex distribution is promoted, the inherent high-dimensional extrapolation challenge is solved, and the calculation cost related to a diffusion model is reduced. Comprehensive experiments on different data sets clearly show that the method is superior to other methods in the aspects of reconstruction speed and quality.

The invention discloses a self-adaptive point cloud sampling adjustment method. The method comprises the following steps: training a feature analysis model; sampling a workpiece at an initial sampling rate to obtain point cloud data of the workpiece, inputting the point cloud data into the feature analysis model to obtain regional features, and dividing a local region based on the regional features; and adjusting the sampling rate of the local region based on the region features. The local structure of the point cloud can be reflected more finely based on the division of the regional features, so that the subsequent sampling rate adjustment is more accurate. The refined division mode is beneficial for reserving important features such as edges and angular points in the point cloud, removing redundant information at the same time, and providing a more valuable data basis for subsequent point cloud processing tasks.

The invention belongs to the technical field of optical imaging, and particularly relates to a multispectral demosaicing method based on a variable lens. Comprising the steps that S1, a demosaicing device is built, and the demosaicing device comprises a variable lens module and a spectrum camera which are sequentially arranged in the light path direction; s2, after the variable lens module is aligned with the to-be-measured target, the variable lens module is adjusted, so that the to-be-measured target is imaged on the spectrum camera for no less than one time, and a multispectral mosaic image corresponding to the number of times of imaging is obtained; and S3, realizing demosaicing processing on the multispectral mosaic image based on different demosaicing algorithms according to a quantitative relationship between the imaging times in the step S2 and the number of filtering channels in a single filtering period of the spectral camera. According to the method, the sampling rate of real data can be increased, the loss of pattern texture information is reduced, and a high-resolution multispectral image can be directly obtained without using an algorithm under an ideal condition.

The invention discloses a sampling-free sample preparation cross-belt coal quality online detection LIBS (laser-induced breakdown spectroscopy) spectrum system which comprises a belt conveying system and an optical detection system, and a laser range finder is arranged on a branch conveying belt; and the laser range finder is used for detecting and recording the height of the coal in the detection area on the branch conveying belt. An optical detection system based on the LIBS technology is directly erected on the coal conveying belt, the problems that the speed of the conveying belt is high and real-time sampling detection cannot be achieved are solved, rapid and real-time cross-belt coal quality quantitative analysis with high sampling rate is achieved, and the detection accuracy is improved for the problem that the spectral stability is affected by the height change of coal on the belt. And the focal length of the LIBS detection system is quickly adjusted through the height of the sample in the detection area recorded by the laser range finder, so that the stability of the spectrum is improved.

The invention discloses a mechanical arm obstacle avoidance path planning method which comprises the following steps: setting a starting point and a target point, adding a starting point tree and terminal point tree set, and initializing collision detection failure times; dynamically adjusting a sampling strategy according to the number of failures, and switching between constraint sampling and random sampling to generate sampling points; nodes are expanded in the starting point tree and the terminal point tree respectively, collision detection is carried out, if no obstacle exists, a corresponding tree set is added, and the number of failures is reset; if the obstacle exists, the failure frequency is increased, and sampling is carried out again; bidirectionally connecting the starting point tree and the ending point tree, finding a path if the starting point tree is communicated with the ending point tree, and ending planning; if the communication path is not communicated, selecting a tree with fewer nodes for expansion, and returning to re-sampling until the communication path is found; according to the method, the path planning efficiency and success rate of the mechanical arm in a complex environment are remarkably improved, and meanwhile the path length and energy consumption are reduced.

The invention discloses a TGD-OFDR / OCC-BOTDR integrated multi-parameter sensing system and a working method thereof, and relates to the technical field of optical fiber sensors. The multi-parameter sensing system comprises a laser, an optical fiber coupler, an electro-optical modulator, an optical fiber circulator, an optical fiber Bragg grating, an acoustic optical modulator, an erbium-doped optical fiber amplifier, a random scrambler, a piezoelectric ceramic tube, a heating vessel, a delay optical fiber, a photoelectric balance detector, a filter and detector, a data acquisition card, a polarization controller, a direct-current power supply and a microwave source. An arbitrary waveform generator and the like. According to the invention, multiple parameters can be represented at the same time, the measuring speed is fast, and the sampling rate of the system can be effectively improved under the condition that sensing parameters are the same; the method has the advantage of high spatial resolution; the sensor has the advantages of high fusion and low cost, can realize simultaneous measurement of vibration and temperature / strain through single sensing, and is high in fusion and effective in cost control.

The invention relates to a weak current signal detection system and method based on a fluxgate, and the system comprises an excitation module, a range selection module, a power amplification module, a fluxgate magnetic core coil, a sampling resistor, a differential amplification module and a data processing module which are connected in sequence, the magnetic core current reaches saturation current; the range selection module is used for adjusting the gain of the power amplification module; the power amplification module is used for amplifying the excitation signal; the fluxgate magnetic core coil is used for sensing current to be measured; the sampling resistor is used for sampling saturation current of a magnetic core coil of the fluxgate; the differential amplification module is used for amplifying saturation current of a magnetic core coil of the fluxgate; and the data processing module performs time difference calculation on the amplified saturation current through resampling, and outputs a weak current detection result. Compared with the prior art, the hardware cost can be reduced while the measurement precision is improved.

The invention provides a physical information long-short-term memory network site seismic response prediction method based on kinematics derivative constraint and data folding. The method comprises the following steps: collecting and preprocessing bedrock and ground surface seismic oscillation data; displacement, speed and acceleration supervision labels are constructed through surface acceleration integration; constructing a long-short-term memory network containing a data folding and unfolding module; physical constraint loss is constructed by utilizing kinematics derivative relations between displacement and speed and between speed and acceleration, and a network is trained by combining an adaptive weight strategy; and outputting earth surface displacement, speed and acceleration time history by adopting the trained network. According to the method, under the condition that constitutive parameters of a soil layer are not explicitly given, physical consistency constraint can be applied to the field seismic response prediction process, the long-time program train training burden is reduced through sequence folding processing, and the method is suitable for field seismic response prediction.

The invention discloses a step frequency continuous wave ground penetrating radar imaging device and a rapid imaging method, and belongs to the technical field of microwave ground penetrating imaging. The vector network analyzer is adopted, the radio frequency power amplifier is carried, the ultra-wideband directional antenna is combined, signals of a step frequency continuous wave system are adopted, the frequency range is 700 MHz to 3 GHz, the bandwidth is large, meanwhile, the large transmitting power can be achieved, and meanwhile the requirements for the detection depth and the detection resolution are met; according to the method, direct wave interference is eliminated by adopting a singular value decomposition algorithm, then the effective signal strength is improved based on an automatic gain algorithm, and finally clutter interference is eliminated by adopting a block matching algorithm, so that the signal-to-noise ratio is improved; according to the method, an F-K migration imaging algorithm is developed, processed data are converted to a frequency wavenumber domain through Fourier transform, wavenumber domain data are filled before fast inverse Fourier transform, and the transverse space sampling rate and the imaging speed are increased.

The invention discloses a new energy transmission line inter-phase fault distance measurement method and system, and the method comprises the steps: starting an element to act after a fault occurs, and inputting low-voltage ride-through control and complex-frequency signal injection control; fault phase selection is realized according to a voltage phase selection element, and fault phase voltages and currents under two characteristic frequencies are extracted through filtering; if the fault is a two-phase short-circuit fault, extracting and replacing the fault phase-to-phase voltage, and replacing the transition resistor with Rf / 2; if the fault is a three-phase short-circuit fault, the phasors of the phase voltage and current difference of any two phases are brought in; according to the equivalent circuit under the characteristic frequency, a frequency domain parameter equation about the fault distance is obtained; respectively splitting a real part and an imaginary part of the frequency domain parameter equation of the fault distance to obtain a real coefficient equation set about the fault distance; when the non-power-frequency electrical quantity is extracted, the real coefficient equation set is expanded into an overdetermined equation with the equation number of 4N, and the solution of the overdetermined equation in the least square meaning is solved; and obtaining a distance measurement result of the new energy transmission line.

The invention provides a double-antenna sea surface height monitoring method based on GNSS-IR, and belongs to the technical field of GNSS-IR sea surface height monitoring. The method comprises the following steps: establishing a double-antenna GNSS-IR monitoring device; respectively extracting direct and interference signal observation values; performing direct and reflected signal separation on the signal-to-noise ratio data under the low elevation angle and the high elevation angle respectively; reflecting surface height characteristic values in the low-elevation-angle reflection component and the high-elevation-angle reflection component are calculated respectively; and error elimination is carried out on the height of the inverted reflecting surface, and the inversion precision is calculated by comparing with tide station data. According to the method, the limitation of a traditional GNSS-IR sea surface height monitoring method on the elevation angle is made up, the mode that only low elevation angle data can be used in an original method is expanded to the mode that high and low elevation angle data can be used for inversion, and the sampling rate of sea surface height inversion is greatly increased in the mode that observation data are segmented according to the elevation angle; the method has the characteristics of high observation time resolution, low cost and the like.

The invention discloses an unsteady-state excitation noise testing method of a composite dielectric gate photosensitive detector, and belongs to the technical field of noise testing of imaging detection devices. In the unsteady state excitation noise test system of the composite dielectric gate photosensitive detector, the composite dielectric gate photosensitive detector to be tested is connected outwards through four ports of a wafer test probe station; three ports of the current and voltage signal sampling analyzer are connected to a drain electrode, a substrate and a source electrode of the composite dielectric gate photosensitive detector through the wafer test probe station, and a fourth port of the current and voltage signal sampling analyzer is connected to a trigger input port of the unsteady state excitation function signal generator; a signal output port of the unsteady state excitation function signal generator is connected to a control gate electrode of the photosensitive detector through the wafer test probe station; and original experimental data obtained after the hardware part is tested is stored in the current and voltage signal sampling analyzer. According to the invention, the noise data of a specific pixel device in the chip can be obtained from the level of a device level.

The invention provides a robot global path planning method and system based on probability interval division, and relates to the technical field of robot path planning. Comprising the following steps: calculating an interval sampling probability based on an obstacle based on an obstacle area and an obstacle total area in each equally divided interval; determining the positions of the starting point and the ending point, and an equal division interval containing the starting point and the ending point, and setting a position-based intra-interval sampling probability of the equal division interval containing the starting point and the ending point; combining the obstacle-based intra-interval sampling probability and the position-based intra-interval sampling probability to obtain an intra-area sampling probability; on the basis of an RRT * algorithm, nodes are selected according to the sampling probability in the region, sampling is carried out on the equal division interval in a variable step size mode, and an initial path is generated through iteration; and optimizing the generated initial path to obtain a global path of the robot. According to the method, the speed of generating the initial path by the growth random tree can be accelerated, and the purposiveness and directivity of the algorithm are enhanced.

The invention belongs to the technical field of path planning and robot motion planning, and particularly relates to a multi-strategy fusion improved RRT * sampling planning method and device and a storage medium, and the method comprises the steps: constructing a dual-search tree through a starting point and a target point in a three-dimensional working space, and selecting a to-be-expanded search tree through a random mode; halton-Bridge mixed sampling is adopted to generate sampling points, so that the global coverage uniformity is improved, and the narrow channel sampling capability is enhanced; calculating the dynamic step length of the exponential decay based on the local obstacle proportion; adopting a hierarchical guidance expansion strategy, preferentially expanding towards a target, introducing an improved artificial potential field to guide expansion when the expansion fails, and taking the bottom with random expansion; performing RRT * neighborhood father node reselection and reconnection optimization on the new node; greedy pruning, self-adaptive interpolation and cubic B-spline smoothing are carried out on a path after the double trees are connected, and an executable smoothing track is output; according to the scheme, planning efficiency, success rate and path quality are improved.

The invention provides a sensor array data transmission method which is applied to an FPGA, the FPGA is provided with a PDM digital interface and a DVP digital interface, multichannel PDM signal streams output by a digital sensor array are received through the PDM digital interface, the multichannel PDM signal streams are converted into multichannel PCM data streams in a synchronous digital modulation mode, and the multichannel PCM data streams are transmitted to the FPGA through the DVP digital interface. Dynamic fragmentation compression processing is carried out on the multi-channel PCM data flow through an equal probability sampling method; the method comprises the following steps: receiving a video data stream output by an image sensor through a DVP digital interface, and compressing the video data stream through a H264-like video data compression method; the multi-channel PCM data stream is subjected to combined coding, the coded PCM data stream is interspersed in the compressed video data stream through an interpolation method, and the coded PCM data stream and the compressed video data stream keep time synchronization; and packaging and uploading the interspersed and mixed data flow to a cloud server through a gigabit network communication protocol.

The invention discloses an online monitoring system and method for the radiation status of a radioactive material storage container, and relates to the technical field of nuclear radiation monitoring. The online monitoring system for the radiation status of a radioactive material storage container comprises a detector, an FPGA chip, a communication module, a server and a user terminal; the detector is used to output a pixel matrix containing a radiation response signal and the position information of a target storage container; the FPGA chip is used to characterize the pixel matrix output by the detector as the radiation dose rate of the radioactive material; the server is used to store and display real-time radiation dose rate data and detector position data on the one hand, and to adjust the operating parameters of the FPGA chip on the other hand, and to send real-time radiation dose rate data to the outside on the other hand. The present invention adopts a non-destructive detection method to monitor the status of a radioactive material storage container online, and can inform the user through the user terminal at the first time of a nuclear leak.

The invention relates to a method (30) for adjusting a sampling rate of an observation camera (28) for observing a vehicle occupant (24), comprising the following steps: providing observation data of the observation camera (28); identifying a blink event of the vehicle occupant (24) by evaluating the observation data, the start and end of the blink event being identified; according to the invention, an adjustment of the sampling rate of the observation camera (28) is determined as a function of the detected blink event, the sampling rate being increased if the start of the blink event of the vehicle occupant (24) is detected, and the sampling rate being reduced if the end of the blink event of the vehicle occupant (24) is detected.

This invention provides a method, apparatus, device, and readable storage medium for improving audio generation quality, relating to the field of speech conversion and recognition technology. The method includes acquiring low-sampling-rate audio, a target audio sampling rate, and an audio processing model; calculating an initial high-sampling-rate audio based on the low-sampling-rate audio, the target audio sampling rate, and the audio preprocessing mathematical model; calculating the target audio time-domain signal and target audio wavelet coefficients based on the initial high-sampling-rate audio and an audio signal reconstruction mathematical model; and solving the mathematical model based on the target audio time-domain signal, the target audio wavelet coefficients, and a preset fused audio signal mathematical model to obtain the target high-sampling-rate audio. This invention uses discrete wavelet transform to capture multi-scale details of the audio signal and reconstructs high-sampling-rate audio by combining the time-domain and frequency-domain information of the audio signal, further improving the overall generation quality of the audio.

This invention relates to the field of audio circuit hardware technology, and in particular to a low-latency ARC circuit configuration method and system based on I2S audio switching. External video data is received through an ARC circuit compatibility interface and input to a video conversion chip for data separation, achieving synchronous separation of I2S audio signals and MIPI video signals. The separated I2S audio signals are then simulated and sent to the host SOC and external I / O terminals respectively by a first-stage audio switching chip. Through link delay analysis and instantaneous delay judgment, the audio path is dynamically configured to achieve switching control between the internal processing path and the direct path. A second-stage audio switching chip realizes multi-channel audio multiplexing and direct transmission simulation, achieving a low-latency, highly compatible ARC audio output scheme while ensuring audio and video timing alignment. This invention can significantly reduce the direct transmission latency of ARC audio while ensuring audio and video synchronization and system compatibility, and improves the flexibility and reliability of audio path configuration.

The invention provides a data sampling device of a control system, the control system and a data sampling method, and the system comprises a sampling module which comprises a plurality of sampling channels; the control module is used for sequentially triggering the plurality of sampling channels to start sampling in each triggering period; the trigger time delay between the adjacent triggered sampling channels is less than the sampling period of the sampling channels. The method comprises the following steps: in each triggering period, sequentially triggering a plurality of sampling channels to start sampling; wherein the trigger time delay between the adjacent triggered sampling channels is smaller than the sampling period of the sampling channels; and sequentially receiving data output by the plurality of sampling channels. According to the technical scheme, the sampling delay is reduced through parallel interval sampling of the multiple sampling channels, the method can be applied to a high-frequency control system, and the control effect is improved.

The invention discloses a demodulation algorithm for signals in radio frequency direct sampling, which comprises the following steps of: designing a multiphase local oscillator in an FPGA (Field Programmable Gate Array), carrying out multiphase frequency mixing on the multiphase local oscillator and a sampling signal, and carrying out extraction and low-pass filtering processing to realize digital demodulation so as to obtain a baseband signal; the method has the beneficial effects that under the implementation of the algorithm, a local oscillation source, a frequency mixing circuit, a first intermediate-frequency filter circuit, a second intermediate-frequency filter circuit, a medium-frequency amplifying circuit and the like are omitted in a receiving channel, so that the size of the receiving channel is greatly reduced; due to the design of small size and light weight, the TACAN airborne equipment can be installed on certain types with limited weight and size; after most of devices of receiving channels are omitted, the hardware cost is greatly reduced; in the algorithm, the signal bandwidth is realized through the design of a digital filter, and in order to adapt to the DME / P and ADS-B functions, only the bandwidth of a low-pass filter in software needs to be changed.

The invention discloses an image compression method, device and equipment, and relates to the technical field of image processing, and the method comprises the steps: obtaining a sampling region in a to-be-compressed image; wherein the sampling area comprises a fixation area and an interval sampling area; acquiring a sampling mode corresponding to each sampling area; wherein the sampling rate of the fixation area in the sampling area is maximum; sampling the pixel points in each sampling area according to the sampling mode corresponding to each sampling area to obtain a sampling compression image corresponding to the to-be-compressed image; by utilizing the characteristic that the fixation point of the eyes of the user is not in the whole image but focuses on a certain area, the area outside the fixation area can be compressed, so that the compression transmission of the image is realized on the basis of not influencing the watching effect of the head-mounted display equipment as far as possible, the image transmission and coding and decoding delay is reduced, and the user experience is improved. And the user experience is improved.

The invention relates to a data set decomposition-based large language model training method, which comprises the following steps of: performing token processing on an original document in a training data set to obtain a one-dimensional token sequence and calculating an original length; performing binary decomposition on the original length, splitting the token sequence into sub-token sequences, storing the sub-token sequences in buckets according to the length, and adding an original document traceability identifier; calculating a single sampling number of each effective bucket, and configuring a sampling probability; executing random sampling according to the sampling probability, and performing homologous constraint splicing to generate a training batch; and inputting the training batch into the to-be-trained large language model, and performing fine adjustment on the sampling probability adjustment rule and the training hyper-parameter according to the results of the conventional reference performance evaluation, the long context capability evaluation and the training efficiency evaluation. By means of the large language model training method based on data set decomposition, cross-document attention interference can be reduced, calculation waste can be reduced, and training efficiency can be improved.

The present invention provides a trajectory similarity matching method based on low-precision random sampling data, comprising: preprocessing the original data to obtain target data; judging the trajectory points of different targets in the target data and segmenting the trajectory points belonging to different targets to obtain the trajectories of different targets; finding two different targets and pairing all the trajectories under the different targets; processing the two paired trajectories under the two different targets to obtain two similar trajectory segments of the two trajectories; calculating the effective distance, effective length and effective trajectory points of the two trajectories according to the two similar trajectory segments; calculating the similarity of the two trajectories according to the effective distance, effective length and effective trajectory points of the two trajectories; traversing and calculating to obtain the similarity of all trajectories under all targets; and sorting the similarity of all trajectories under all targets. The present invention can realize trajectory similarity matching of low-precision random sampling data.

The invention discloses an enhanced imaging method for a microwave ESM sparse scanning system, and the method comprises the steps: obtaining the amplitude and phase information of a part of radiation sources on a far-field sampling plane, and obtaining sparse sampling data; on the basis, an image reconstruction problem is converted into a sparse optimization problem on a gradient domain by using total variation regularization, the gradient of the image in a smooth region approaches to zero by applying gradient constraint, and a relatively large gradient value is only reserved at a significant feature position. And an alternating direction multiplier method is further adopted to carry out iterative solution on the total variation minimization problem, and complete amplitude and phase information is gradually reconstructed by decomposing and optimizing a target and alternately updating variables. And dynamically adjusting the sampling rate according to the SSIM difference of the reconstructed image between adjacent iterations, and stopping sampling in advance when the SSIM value is higher than the quality threshold and the change tends to be stable. And finally, transforming the reconstructed amplitude and phase information through a far and near field inversion algorithm to obtain a near field diffraction pattern of the radiation source, and completing high-precision positioning and imaging of the radiation source.

The application provides a symbol interval length measurement method and device of a low carrier frequency CPFSK signal and an electronic device, acquires additional phase information through IQ demodulation and generates a modulation absolute frequency curve, accurately determines a symbol interval starting point and an ending point by using a waveform intermediate value, and realizes measurement of an instantaneous symbol interval length; secondly, when the carrier frequency is lower than a preset threshold, the demodulation center frequency is offset from the carrier frequency to a target frequency, the maximum analysis bandwidth is expanded to twice the offset center frequency, thereby breaking through the limitation of twice the carrier frequency bandwidth, improving the sampling rate and time resolution capability. Thus, under the condition of a low carrier frequency, a high-resolution symbol interval length measurement result can still be obtained.

The invention discloses a circuit breaker stroke curve test system and method, the test system comprises a stable light source, an emission collimator, a circuit breaker moving part, a receiving collimator, an optical power acquisition loop and a processing unit, the output of the stable light source is connected to the tail part of the emission collimator through an optical fiber, the emission collimator and the receiving collimator are oppositely arranged, and the optical power acquisition loop is connected with the processing unit. A gap for movement of a moving part of the circuit breaker is reserved between the transmitting collimator and the receiving collimator; the tail part of the receiving collimator is connected to an optical power acquisition loop through an optical fiber; and the optical power acquisition loop transmits data to a processing unit. According to the test system and method, the online test of the state of the circuit breaker can be realized. The sensor is not mechanically or electrically connected with a circuit breaker operating mechanism, and has the advantages of high sampling rate, strong anti-interference capability, small installation size and the like.

The present invention discloses a method, system, device and storage medium for adjusting the sampling rate of a touch screen, the method comprising: creating a curve class, the curve class comprising a source point, an end point and a control point, the source point, the end point and the control point are used to update the curve trajectory, the control point is used to record the trajectory points generated by touching the touch screen; determining the Euclidean distance between the current control point and the previous control point, and determining the number of sampling points according to the Euclidean distance; determining the information of the sampling points according to the curve trajectory and the number of sampling points. The present invention determines the number of sampling points on the current curve trajectory by determining the Euclidean distance between adjacent control points, and completes the sampling of the curve trajectory. These sampling points cannot be recorded by the system native API, thereby improving the sampling rate of the Android device and ensuring the authenticity and effectiveness of the sampling points. The present invention can be widely used in the field of terminal technology.

The utility model discloses a sampling device for water conservancy ecological resource monitoring, which comprises a main body unit, the main body unit comprises a bottom plate, the upper end face of the bottom plate is fixedly connected with a vertical block, the vertical block is provided with a moving assembly, the upper end of the moving assembly is provided with a mounting plate, the mounting plate is fixedly provided with a stepping motor, and the stepping motor is connected with the vertical block; the output end of the stepping motor penetrates through the mounting plate and is fixedly connected with a round sleeve, the round sleeve is fixedly connected with a connecting plate, the connecting plate is fixedly provided with a water suction pump, the input end of the water suction pump penetrates through the connecting plate and is fixedly connected with a vertical pipe, and the lower end of the vertical pipe is fixedly connected with a sampling head. According to the lake water sampling device, the rotating motor is started, the output end of the rotating motor drives the first gear to rotate, the first gear drives the second circular plate to rotate through the first gear and the rotating rod, and then the collecting barrel is driven to rotate, so that multiple parts of lake water can be sampled at a time, multiple times of sampling are not needed, the sampling time is shortened, and the sampling rate is increased.

The present invention discloses a low-cost, distributed RFID-based human motion recognition method and system. The method includes sending an excitation signal and a coding signal through a transmitter; the excitation signal is used to activate an RFID tag; the coding signal is used to instruct the RFID tag to send a reflection signal; the reflection signal is received by at least one receiver; the reflection signal has different phases after being reflected by different human motions; after eliminating the carrier frequency offset, self-interference and random phase of the reflection signal, the phase information in the reflection signal is extracted; the phase information is input into a trained classifier to obtain the human motion corresponding to the phase information output by the classifier; the classifier stores the mapping relationship between phase information and human motion. The solution provided by the present invention can reduce costs, increase signal sampling rate, and thus improve the detection accuracy of the system.