阅读说明：本技术 身份识别方法、终端设备及计算机存储介质 (Identity recognition method, terminal device and computer storage medium ) 是由 刘建华 周安福 马华东 杨宁 唐海 张治� 于 2019-07-15 设计创作，主要内容包括：本申请实施例公开了身份识别方法、终端设备及计算机存储介质,包括：获取反射信号,其中反射信号是信号传输设备在发射无线信号至处于移动状态的目标对象后接收到的反射信号；对反射信号进行处理,得到目标对象的点云数据；点云数据被输入至经训练的残差网络模型对以识别目标对象的身份。本申请实施例可提高身份认证的可靠性。(The embodiment of the application discloses an identity recognition method, terminal equipment and a computer storage medium, which comprise the following steps: acquiring a reflection signal, wherein the reflection signal is received by a signal transmission device after transmitting a wireless signal to a target object in a moving state; processing the reflected signal to obtain point cloud data of a target object; the point cloud data is input to a trained pair of residual network models to identify the identity of the target object. The embodiment of the application can improve the reliability of identity authentication.)

An identity recognition method, comprising:

acquiring a reflection signal, wherein the reflection signal is received by a signal transmission device after transmitting a wireless signal to a target object in a moving state;

processing the reflection signal to obtain point cloud data of the target object;

the point cloud data is input to a trained residual network model to identify the identity of the target object.

The method of claim 1, wherein the wireless signal transmitted by the signal transmission device is millimeter waves and the reflected signal received by the signal transmission device is millimeter waves.

The method of claim 1, wherein the processing the reflected signal to obtain point cloud data of the target object comprises:

and processing the reflected signal according to the Doppler effect or frequency modulation continuous wave principle to obtain the attribute information of at least one surface energy point.

The method of claim 3, wherein before the point cloud data is input to the trained residual network model, further comprising:

acquiring identification information of at least one surface energy point, wherein the identification information is used for indicating that a first surface energy point in the at least one surface energy point belongs to the same frame;

taking attribute information of surface energy points of a plurality of continuous frames as a point cloud sequence to obtain at least one point cloud sequence;

the point cloud data is input to a trained residual network model, comprising:

each of the at least one point cloud sequence is input to a trained residual network model.

The method of claim 4, wherein before the obtaining at least one point cloud sequence by using the attribute information of the surface energy points of the consecutive frames as a point cloud sequence, the method further comprises:

when the number of first surface energy points belonging to the same frame is smaller than a preset number threshold, adding second surface energy points, wherein the sum of the number of the second surface energy points and the number of the first surface energy points reaches the preset number threshold, the identification information of the second surface energy points is the same as the identification information of the first surface energy points, and the attribute information of the second surface energy points is obtained by performing arithmetic mean operation on the attribute information of the first surface energy points;

and deleting a third surface energy point in the first surface energy points when the number of the first surface energy points belonging to the same frame is greater than a preset number threshold, wherein the sum of the number of the surface energy points except the third surface energy point in the first surface energy points reaches the preset number threshold.

The method of claim 4, wherein when the point cloud data is input to the trained residual network model, the trained residual network model performs feature extraction on target attribute information of surface energy points included in the point cloud sequence to obtain attribute features of target attributes indicated by the target attribute information, wherein the target attribute information is any attribute information of the surface energy points;

and processing the attribute characteristics of each attribute to obtain the gait characteristics.

The method of claim 6, wherein the trained residual network model performs feature extraction on target attribute information of surface energy points included in the point cloud sequence to obtain attribute features of target attributes indicated by the target attribute information, and the method comprises:

performing convolution operation on target attribute information of the surface energy points contained in the point cloud sequence by using a first preset convolution layer to obtain a first convolution operation result;

performing maximum pooling operation on the first convolution operation result to obtain a first operation result;

performing convolution operation on the first operation result by using a second preset convolution layer to obtain a second convolution operation result;

performing convolution operation on the second convolution operation result by using the second preset convolution layer to obtain a third convolution operation result;

performing convolution operation on the third convolution operation result by using the second preset convolution layer to obtain a fourth convolution operation result;

performing convolution operation on the fourth convolution operation result by using the second preset convolution layer to obtain a fifth convolution operation result;

performing residual operation on the third convolution operation result and the first operation result to obtain a second operation result;

performing residual operation on the fifth convolution operation result and the third convolution operation result to obtain a third operation result;

and carrying out average pooling operation on the second operation result and the third operation result to obtain the attribute characteristics of the target attribute.

The method of claim 6, wherein processing the attribute characteristics of each attribute to obtain gait characteristics comprises:

connecting the attribute features of each attribute information aiming at any point cloud sequence to obtain the connected attribute features;

and classifying at least one connected attribute feature by using a preset full connection layer to obtain the gait feature.

The method of claim 3, wherein the attribute information for each of the at least one surface energy points comprises: a moving speed of the surface energy point moving to the direction of the signal transmission device, a distance between the surface energy point and the signal transmission device, an angle between the reflected signal and the signal transmission device, a signal-to-noise ratio, a position of the target object, a physical size of the target object, and a point density.

The method of claim 3, wherein said processing said reflected signal based on Doppler effects to obtain attribute information for at least one surface energy point comprises:

obtaining Doppler red shift and Doppler blue shift of the reflected signal;

and calculating the moving speed of the at least one surface energy point to move towards the direction of the signal transmission equipment according to the Doppler red shift and the Doppler blue shift.

The method of claim 3, wherein processing the reflected signal according to frequency modulated continuous wave principles to obtain attribute information of at least one surface energy point comprises:

acquiring a difference value between the receiving time of the signal transmission equipment for receiving the reflected signal and the transmitting time of the signal transmission equipment for transmitting the wireless signal;

acquiring the frequency difference of the target object in a static state and the Doppler frequency shift of the target object in a moving state;

obtaining the frequency difference of the target object in a moving state according to the frequency difference of the target object in a static state and the Doppler frequency shift;

and multiplying the frequency difference of the target object in the moving state by the difference value to obtain the distance between the at least one surface energy point and the signal transmission equipment.

The method of claim 1, wherein the processing the reflected signal to obtain point cloud data of the target object comprises:

acquiring the phase of a receiving antenna of the signal transmission equipment;

and calculating the angle between the reflected signal and the signal transmission equipment according to the phase of the receiving antenna.

The method of claim 1, wherein the processing the reflected signal to obtain point cloud data of the target object comprises:

acquiring the output power of the reflected signal and the output power of noise received by the signal transmission equipment in the process of receiving the reflected signal;

and calculating to obtain the signal-to-noise ratio according to the output power of the reflected signal and the output power of the noise.

A terminal device, characterized in that it comprises means for implementing an identification method according to any of claims 1-13.

A terminal device, characterized in that the terminal device comprises a processor and a memory, the processor being coupled with the memory,

the memory configured to store instructions;

the processor is configured to acquire a reflected signal, wherein the reflected signal is a reflected signal received by the signal transmission device after transmitting a wireless signal to a target object in a moving state; processing the reflection signal to obtain point cloud data of the target object; the point cloud data is input to a trained residual network model to identify the identity of the target object.

A computer storage medium, characterized in that it stores a computer program or instructions which, when executed by a processor, cause the processor to carry out the identification method according to any one of claims 1-13.