Target posture recognition method, device and medium based on through-wall radar

Abstract

The application provides a target posture recognition method and device based on a through-wall radar and a medium, wherein the method comprises: collecting echo signals corresponding to each target posture and obtaining a feature set of the echo signals corresponding to each target posture; collecting an echo signal corresponding to a to-be-recognized posture and obtaining an image feature of the echo signal corresponding to the to-be-recognized posture; and based on the feature set of each target posture and the image feature of the to-be-recognized posture, the to-be-recognized posture is determined to obtain a posture type corresponding to the to-be-recognized posture. The application can realize automatic recognition of human postures, improve the recognition efficiency, and improve the posture recognition accuracy.

Description

Technical Field

[0001] This invention belongs to the field of attitude recognition technology, specifically relating to a target attitude recognition method, device, and computer storage medium based on through-wall radar. Background Technology

[0002] Human posture recognition, as a crucial component of human perception tasks, aims to identify different target posture types of the human body, such as standing, sitting, and squatting. Most existing human posture recognition methods are based on optical acquisition systems such as cameras. They extract image features from optical images or videos to identify the target human posture. However, optical acquisition systems typically use visible light or infrared light signals for information acquisition. When there are obstacles (such as walls) between the human body and the acquisition device, the light is blocked, preventing the human posture from being properly recognized.

[0003] Through-wall radar systems, due to their strong penetration characteristics and high resolution, can detect targets behind obstacles. However, existing through-wall radar human posture recognition methods usually extract statistical features of human posture based on radar echo signals, such as echo signal intensity and density, and classify human target postures according to these statistical features. This process often relies on manually selected statistical features, which is not only time-consuming and labor-intensive, but also has low recognition accuracy.

[0004] Therefore, how to improve the efficiency and accuracy of human target posture recognition based on through-wall radar has become a technical problem that needs to be solved in this field. Summary of the Invention

[0005] In view of the shortcomings of the prior art, the purpose of this invention is to provide a method, device and computer storage medium for human target posture recognition based on through-wall radar, which solves the problems of low efficiency and low accuracy caused by the need for manual determination of statistical features in the human target posture recognition process of existing through-wall radar.

[0006] To achieve the above and other related objectives, the present invention provides a target attitude recognition method based on through-wall radar in a first aspect, comprising: acquiring echo signals corresponding to each target attitude and obtaining a feature set of the echo signals corresponding to each target attitude; acquiring echo signals corresponding to an attitude to be identified and obtaining image features of the echo signals corresponding to the attitude to be identified; and determining the attitude to be identified based on the feature set of each target attitude and the image features of the attitude to be identified, so as to obtain the attitude type corresponding to the attitude to be identified.

[0007] In one embodiment of the present invention, the step of collecting echo signals corresponding to each target posture further includes: collecting echo signals of experimental personnel of different body types under the same target posture, and performing mean-averaging processing on the echo signals, and using the mean-averaged echo signals as the echo signals corresponding to the target postures.

[0008] In one embodiment of the present invention, the step of obtaining the feature set of the echo signal corresponding to each target attitude includes: converting the echo signal corresponding to each target attitude into a digital signal; converting each digital signal into a corresponding range Doppler image; and extracting feature information related to the target attitude information from each range Doppler image to serve as the feature set corresponding to each target attitude.

[0009] In one embodiment of the present invention, converting the echo signal corresponding to each target attitude into a digital signal includes: using an AD-ADC analog-to-digital converter to convert each echo signal into a corresponding digital signal.

[0010] In one embodiment of the present invention, the step of extracting feature information related to target pose information from each of the range Doppler images includes: extracting feature information of target points from each of the range Doppler images according to a preset feature type, so as to form a feature set of the target pose; wherein, the target point is a feature point trace of the target pose.

[0011] In one embodiment of the present invention, the step of extracting feature information related to target pose information from each of the range Doppler images includes: using a feature extractor to obtain image features from each of the range Doppler images; outputting each of the image features to a classifier for classification, and using the various features obtained after classification as feature information of the target pose.

[0012] In one embodiment of the present invention, obtaining the image features of the echo signal corresponding to the pose to be identified includes: converting the acquired echo signal from an analog signal to a digital signal; converting the digital signal into a range Doppler image; and extracting feature information from the range Doppler image to obtain the image features of the pose to be identified.

[0013] In one embodiment of the present invention, determining the pose to be identified includes: in the feature set of each target pose, using a discriminator to obtain the feature with the highest conformity to the image feature corresponding to the pose to be identified, and taking the target pose corresponding to the feature as the determination result of the pose to be identified.

[0014] In a second aspect, the present invention provides an electronic device, comprising: a processor and a memory; the memory for storing a computer program, and the processor for executing the computer program stored in the memory, so that the electronic device performs any of the target attitude recognition methods based on through-wall radar described above.

[0015] In a third aspect, the present invention provides a computer storage medium storing a computer program, the computer program being executed by a processor using either the cross-modal retrieval model training method described above or the target attitude recognition method based on through-wall radar as described above.

[0016] As described above, the target posture recognition method, device, and computer storage medium based on through-wall radar provided by the present invention acquire the echo signal of the target posture, and perform signal conversion and feature extraction on the echo signal to obtain the feature set corresponding to each target posture. Based on the feature set of each target posture, the posture to be identified is determined as the posture type corresponding to the feature with the highest conformity, thereby realizing the automatic recognition of human posture, which not only improves the efficiency of recognition, but also improves the accuracy of posture recognition. Attached Figure Description

[0017] Figure 1 The diagram shows a flowchart of an implementation of the target attitude recognition method based on through-wall radar of the present invention.

[0018] Figure 2 The diagram shows a flowchart of step S100 in a specific embodiment of the present invention. Detailed Implementation

[0019] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that, unless otherwise specified, the following embodiments and features described therein can be combined with each other.

[0020] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Therefore, the drawings only show the components related to the present invention and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0021] To address the problems existing in the prior art, this invention first provides a target attitude recognition method based on through-wall radar, which utilizes through-wall detection radar equipment to collect echo signals from the through-wall radar through a wall.

[0022] Please see Figure 1 The diagram shows a flowchart of the target attitude recognition method based on through-wall radar provided in an embodiment of the present invention.

[0023] like Figure 1 As shown, the target attitude recognition method based on through-wall radar includes the following steps:

[0024] S100: Collect echo signals corresponding to each target attitude and obtain the feature set of the echo signals corresponding to each target attitude;

[0025] In one specific embodiment, step 100, when specifically executed, is as follows: Figure 2 As shown, it includes the following sub-steps:

[0026] S101, acquire the echo signals of each target's attitude;

[0027] Specifically, in an environment with obstacles, the experimenters sequentially pose different targets; for the k-th type of target pose, the radar sends a single electromagnetic wave signal and receives the corresponding echo signal; the echo signal is collected and used as the echo signal corresponding to the k-th type of target pose.

[0028] Optionally, if there are several experimental personnel and the different experimental personnel have different body types, step S101 further includes averaging the echo signals collected from different experimental personnel under the same target posture to obtain new echo signals.

[0029] S102, convert the echo signal into a range Doppler image;

[0030] Specifically, the echo signal of the target attitude is converted into a corresponding digital signal, that is, the digital signal corresponding to the echo signal is obtained; spatial distribution information and Doppler information are extracted from the digital signal to obtain the corresponding range Doppler image, that is, the range Doppler image corresponding to the echo signal is obtained.

[0031] Optionally, the implementation of converting the echo signal of the target attitude into a corresponding digital signal includes:

[0032] The echo signals are converted into corresponding digital signals using an AD-ADC analog-to-digital converter.

[0033] S103, extract the feature information of the range Doppler image as a feature set of the target posture corresponding to the echo signal;

[0034] According to the preset feature type, the feature information of the target point is extracted from the distance Doppler image, and each feature information is stored as a feature set of the target pose.

[0035] The feature information is a feature vector; the target point is the feature point trace of the target pose, used to characterize the point that distinguishes the target pose from other target poses.

[0036] In one implementation example, the preset feature types include distance, angle, and amplitude.

[0037] Optionally, the target pose feature information can also be obtained in the following ways:

[0038] A feature extractor is used to obtain image features related to the target pose information in the distance Doppler image; each image feature is output to a classifier for classification, and the various features obtained after classification are used as the feature information of the target pose.

[0039] S200, acquire the echo signal of the posture to be identified, and obtain the image features of the echo signal corresponding to the posture to be identified;

[0040] Specifically, for a single human posture to be identified, the echo signal of the human posture is first acquired by through-wall radar, and the acquired echo signal is converted from analog signal to digital signal; then the digital signal is converted into a range Doppler image by a converter; feature information in the range Doppler image is extracted to obtain the image features of the posture to be identified.

[0041] The methods for converting the acquired echo signal from analog to digital and for converting the digital signal into a range Doppler image are the same as those in step S102 above, and will not be repeated here.

[0042] Similarly, the specific implementation method for extracting feature information from the distance Doppler image is the same as that in step S103 above, and will not be repeated here.

[0043] S300, based on the feature set corresponding to each target pose and the image features of the pose to be identified, the pose to be identified is determined to obtain the pose type corresponding to the pose to be identified.

[0044] Specifically, in each of the stored target pose feature sets, a discriminator is used to obtain the target pose with the highest feature conformity to the pose to be identified, and this target pose is used as the determination result corresponding to the pose to be identified, thereby determining the pose type corresponding to the pose to be identified.

[0045] Optionally, the discriminator may employ existing image data comparison software.

[0046] To address the problems existing in the prior art, the present invention also provides an electronic device in a second aspect, the electronic device comprising: a processor, a memory, a transceiver, a communication interface, and a system bus; the memory and the communication interface are connected to the processor and the transceiver through the system bus and complete mutual communication; the memory is used to store computer programs; the communication interface is used to communicate with other devices; and the processor and the transceiver are used to run the computer programs, causing the processing device to execute the various steps in the target attitude recognition method based on through-wall radar as described above.

[0047] The processors mentioned above can be general-purpose processors, including central processing units (CPUs), network processors (NPs), etc.; they can also be digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0048] Furthermore, in a third aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when invoked by a processor, implements the various steps in the target attitude recognition method based on through-wall radar as described above.

[0049] A computer-readable storage medium can be a tangible device capable of holding and storing instructions used by an instruction execution device. Computer-readable storage media can be, for example, (but not limited to) electrical storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of computer-readable storage media include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital multifunction disc (DVD), memory sticks, floppy disks, and mechanical encoding devices.

[0050] The computer-readable program described herein can be downloaded from a computer-readable storage medium to various computing / processing devices, or downloaded via a network, such as the Internet, a local area network, a wide area network, and / or a wireless network, to an external computer or external storage device. A network adapter card or network interface in each computing / processing device receives computer-readable program instructions from the network and forwards these instructions to the computer-readable storage medium in the respective computing / processing device.

[0051] In summary, the target posture recognition method, device, and computer storage medium provided by this invention, based on through-wall radar, acquires echo signals of target postures using through-wall radar. By performing signal conversion and feature extraction on the echo signals, a feature set corresponding to each target posture is obtained. Based on the feature set of each target posture, the posture to be identified is determined as the posture type corresponding to the feature with the highest conformity, thereby achieving automated recognition of human postures. This not only improves recognition efficiency but also enhances recognition accuracy. Furthermore, during data extraction, only one extraction of human posture data is required, eliminating the need for multiple extractions. This reduces workload while further improving recognition efficiency and maintaining good performance. It effectively solves the problems of existing technologies that require manually setting statistical features of target postures and manually judging posture categories, leading to time-consuming and labor-intensive judgment processes and low recognition accuracy.

[0052] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the invention. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in the present invention should still be covered by the claims of the present invention.

Claims

1. A target attitude recognition method based on through-wall radar, characterized in that, include: Collect echo signals corresponding to the attitudes of each target and obtain the feature set of the echo signals corresponding to the attitudes of each target; The step of obtaining the feature set of the echo signal corresponding to each target posture includes: converting the echo signal corresponding to each target posture into a digital signal; converting each digital signal into a corresponding range Doppler image; and extracting feature information related to the target posture information from each range Doppler image as the feature set corresponding to each target posture. Acquire the echo signal corresponding to the posture to be identified, and obtain the image features of the echo signal corresponding to the posture to be identified; The step of obtaining the image features of the echo signal corresponding to the pose to be identified includes: converting the acquired echo signal from an analog signal to a digital signal; converting the digital signal into a range Doppler image; and extracting feature information from the range Doppler image to obtain the image features of the pose to be identified. Based on the feature set of each target pose and the image features of the pose to be identified, the pose to be identified is determined to obtain the pose type corresponding to the pose to be identified; The step of determining the pose to be identified includes: in the feature set of each target pose, using a discriminator to obtain the feature with the highest conformity to the image feature corresponding to the pose to be identified, and taking the target pose corresponding to the feature as the determination result of the pose to be identified.

2. The target attitude recognition method based on through-wall radar according to claim 1, characterized in that, The process of collecting echo signals corresponding to the attitudes of each target also includes: Echo signals from experimenters of different body types under the same target posture were collected, and the echo signals were averaged. The averaged echo signals were then used as the echo signals corresponding to the target posture.

3. The target attitude recognition method based on through-wall radar according to claim 1, characterized in that, The process of converting the echo signals corresponding to each target attitude into digital signals includes: The echo signals are converted into corresponding digital signals using an analog-to-digital converter.

4. The target attitude recognition method based on through-wall radar according to claim 1, characterized in that, The extraction of feature information related to target pose information from each of the range Doppler images includes: According to the preset feature type, feature information of the target point is extracted from each of the distance Doppler images to form the feature set of the target pose; wherein, the target point is the feature point trace of the target pose.

5. The target attitude recognition method based on through-wall radar according to claim 1, characterized in that, The extraction of feature information related to target pose information from each of the range Doppler images includes: Image features in each of the distance Doppler images are obtained using a feature extractor; The image features are output to a classifier for classification, and the various features obtained after classification are used as the feature information of the target pose.

6. An electronic device, characterized in that, include: Processor and memory; The memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory to cause the electronic device to perform the target attitude recognition method based on through-wall radar as described in any one of claims 1 to 5.

7. A computer storage medium storing a computer program, characterized in that, The computer program is executed by a processor using the target attitude recognition method based on through-wall radar as described in any one of claims 1 to 5.

Images