Data Enhancement Method, System and Storage Medium Based on FMCW Millimeter Wave Radar

Abstract

The present invention relates to the field of radar data processing technology, and in particular to a data enhancement method, system, and storage medium based on FMCW millimeter-wave radar. The data enhancement method comprises the following steps: receiving a point cloud data set collected by a millimeter-wave radar; mapping the received point cloud data set into a spatiotemporal coordinate system; performing coordinate compression processing on the point cloud data within the spatiotemporal coordinate system; and superimposing the compressed point cloud data using a spatiotemporal superposition function to form an enhanced signal. The present invention utilizes the characteristics of human body movements, which are slow, continuous, and rarely repeated. By compressing the point cloud coordinates and superimposing them within a spatial matrix, the point cloud generated by the human body can be enhanced, providing the necessary data foundation for subsequent algorithms.

Description

Technical Field

[0001] The present invention relates to the field of radar data processing technology, and in particular to a data enhancement method, system and storage medium based on FMCW millimeter wave radar. Background Art

[0002] FMCW millimeter-wave radar transmits frequency-modulated continuous waves and receives echoes reflected from objects, enabling it to obtain information such as an object's speed, direction, and energy. This information is presented as a radar point cloud. However, in smart home environments, devices such as washing machines and fans also generate radar point clouds. This non-human point cloud information can interfere with the recognition of abnormal human behavior, leading to false positives and missed detections.

[0003] Traditional methods directly use raw point cloud data to judge abnormal human behavior. Since raw point cloud information is relatively sparse in the edge area of radar detection and the noise in the home environment is difficult to completely remove, directly using raw point cloud data to judge falls will lead to a significant reduction in detection accuracy.

[0004] In addition, traditional radar fall detection algorithms do not fully consider the characteristics of human movements when processing point cloud data and lack effective data enhancement methods, further limiting the accuracy and reliability of fall detection. Summary of the Invention

[0005] The purpose of the present invention is to overcome the shortcomings of the prior art and provide a data enhancement method, system and storage medium based on FMCW millimeter wave radar to solve the problem that the existing radar fall detection algorithm lacks effective data enhancement means, which limits the accuracy and reliability of fall detection.

[0006] The technical solution to achieve the above purpose is:

[0007] The present invention provides a data enhancement method based on FMCW millimeter wave radar, comprising the following steps:

[0008] Receive point cloud datasets collected by millimeter-wave radar;

[0009] Mapping the received point cloud data set into a spatiotemporal coordinate system;

[0010] Performing coordinate compression processing on the point cloud data in the spatiotemporal coordinate system;

[0011] The compressed point cloud data is superimposed using the spatiotemporal superposition function to form an enhanced signal.

[0012] The present invention further improves the data enhancement method based on FMCW millimeter wave radar by using the following formula to perform superposition processing on the point cloud data:

[0013] ,

[0014] in, Indicates the number of acquisition windows. Each acquisition window collects a set number of point cloud data. is the weight, Q is the superimposed point cloud dataset, Represents the compressed point cloud dataset, Indicates continuous time The collected point cloud data, express Point cloud dataset mi The three-dimensional coordinates of the points, j =1,2,… mi .

[0015] The present invention further improves the data enhancement method based on FMCW millimeter wave radar by using the following formula to perform coordinate compression processing on the point cloud data:

[0016] Qi =quantize( PCA ( Pi )),

[0017] in, Qi Represents the compressed point cloud data frame, Pi Represents the received point cloud dataset.

[0018] The further improvement of the data enhancement method based on FMCW millimeter wave radar of the present invention is that it also includes: using a clustering algorithm to cluster the enhanced signal to divide it into s Different clustering patterns.

[0019] The present invention also provides a storage medium, on which is stored a program for a data enhancement method based on FMCW millimeter-wave radar. When the program for a data enhancement method based on FMCW millimeter-wave radar is executed by a processor, the steps of the data enhancement method based on FMCW millimeter-wave radar are implemented.

[0020] The present invention further provides a data enhancement system based on FMCW millimeter wave radar, comprising:

[0021] A receiving unit, used to receive a point cloud data set collected by a millimeter-wave radar;

[0022] a coordinate mapping unit connected to the receiving unit, and configured to map the point cloud data set received by the receiving unit into a spatiotemporal coordinate system;

[0023] a compression unit, connected to the coordinate mapping unit, for performing coordinate compression processing on the point cloud data in the spatiotemporal coordinate system;

[0024] The processing unit is connected to the compression unit and is used to perform superposition processing on the compressed point cloud data using a spatiotemporal superposition function to form an enhanced signal.

[0025] A further improvement of the data enhancement system based on FMCW millimeter wave radar of the present invention is that the processing unit performs superposition processing on the point cloud data using the following formula:

[0026] ,

[0027] in, Indicates the number of acquisition windows. Each acquisition window collects a set number of point cloud data. is the weight, Q is the superimposed point cloud dataset, Represents the compressed point cloud dataset, Indicates continuous time The collected point cloud data, express Point cloud dataset mi The three-dimensional coordinates of the points, j =1,2,… mi .

[0028] A further improvement of the data enhancement system based on FMCW millimeter wave radar of the present invention is that the compression unit performs coordinate compression processing on the point cloud data using the following formula:

[0029] Qi =quantize( PCA ( Pi )),

[0030] in, Qi Represents the compressed point cloud data frame, Pi Represents the received point cloud dataset.

[0031] The data enhancement system based on FMCW millimeter wave radar of the present invention is further improved in that it also includes a clustering unit connected to the processing unit for clustering the enhanced signal using a clustering algorithm to divide it into s Different clustering patterns.

[0032] The beneficial effects of the data enhancement method, system, and storage medium based on FMCW millimeter wave radar of the present invention are:

[0033] The present invention utilizes the characteristics of human body movements being slow, continuous and rarely repeated. By compressing the point cloud coordinates and superimposing them in a spatial matrix, the point cloud generated by the human body can be strengthened, providing the necessary data basis for subsequent algorithms.

[0034] The data enhancement of the present invention can adapt to the user environment and actively enhance the human body point cloud, providing more accurate and reliable data support for the detection of abnormal human behavior. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 The figure is a flow chart of the data enhancement method based on FMCW millimeter wave radar of the present invention.

[0036] Figure 2 This is a system diagram of the data enhancement system based on FMCW millimeter wave radar of the present invention.

[0037] Figure 3 This is a data enhancement method based on FMCW millimeter wave radar in the present invention and a point cloud enhancement information diagram after the human body is stationary in the system. DETAILED DESCRIPTION

[0038] The present invention will be further described below with reference to the accompanying drawings and specific embodiments.

[0039] See Figure 1 The present invention provides a data enhancement method, system, and storage medium based on FMCW millimeter-wave radar. These methods primarily leverage the slow, continuous, and rarely repetitive nature of human motion. By compressing point cloud coordinates and superimposing them within a spatial matrix, they enhance the point cloud data generated by the human body, providing the necessary data foundation for subsequent algorithms. The following describes the data enhancement method, system, and storage medium based on FMCW millimeter-wave radar, along with the accompanying figures.

[0040] See Figure 2 , shows the system diagram of the data enhancement system based on FMCW millimeter wave radar of the present invention. Figure 2 , the data enhancement system based on FMCW millimeter wave radar of the present invention is described.

[0041] like Figure 2 As shown, the data enhancement system based on FMCW millimeter wave radar of the present invention includes a receiving unit 21, a coordinate mapping unit 22, a compression unit 23 and a processing unit 24, wherein the receiving unit 21 is connected to the coordinate mapping unit 22, the coordinate mapping unit 22 is connected to the compression unit 23, and the compression unit 23 is connected to the processing unit 24, the receiving unit 21 is used to receive the point cloud data set collected by the millimeter wave radar; the coordinate mapping unit 22 is used to map the point cloud data set received by the receiving unit 21 into a space-time coordinate system; the compression unit 23 is used to perform coordinate compression processing on the point cloud data in the space-time coordinate system; the processing unit 24 is used to perform superposition processing on the compressed point cloud data using a space-time superposition function to form an enhanced signal.

[0042] The FMCW millimeter-wave radar of the present invention is used for human fall detection in a home environment. When the radar wave hits an object, it generates an echo. The echo of the FMCW millimeter-wave radar contains information such as speed, direction, and energy. This information is called a radar point cloud. Washing machines, fans, and other devices in the home environment also generate point cloud information. The presence of this interference information can affect the identification of abnormal human behavior. The data enhancement system of the present invention can adapt to the user environment and actively enhance the human point cloud data. The original point cloud information is relatively sparse, especially in the edge area of radar detection, which is prone to false positives and missed positives. Traditional radar fall algorithms directly use the original point cloud for judgment, resulting in a significant reduction in their detection accuracy. The data enhancement system of the present invention mainly utilizes the characteristics of human body movements, which are slow, continuous, and rarely repeated. By compressing the point cloud coordinates and superimposing them within a spatial matrix, the point cloud generated by the human body can be enhanced, providing the necessary data foundation for subsequent algorithms.

[0043] In a specific embodiment of the present invention, the FMCW millimeter wave radar collects a point cloud data set in a continuous time, and the receiving unit of the present invention is connected to the FMCW millimeter wave radar for receiving the point cloud data set collected by the FMCW millimeter wave radar. Specifically, the millimeter wave radar collects a point cloud data set in a continuous time. t1,t2,…,tn The point cloud datasets collected in the project are: Pt1, Pt2,…,Ptn .in Pti ={( xj,yj,zj,Ij )| j = 1,2,…,mi}, indicating the three-dimensional coordinate information (x, y, z) of mi points collected at time ti and other information I (Such as intensity, speed, etc.). The FMCW millimeter wave radar is used to capture the changes in human body movements in time series, and the received data set P ={ Pt1, Pt2,…Ptn} for subsequent analysis.

[0044] In a specific embodiment of the present invention, the coordinate mapping unit 22 of the present invention is used to perform time-space coordinate mapping, and each time point ti Point cloud data Pti Mapping to a unified space-time coordinate system ensures that different Pti can be analysed within a common framework.

[0045] Furthermore, the compression unit 23 performs coordinate compression processing on the point cloud data using the following formula:

[0046] Qi =quantize( PCA ( Pi )),

[0047] in, QiRepresents the compressed point cloud data frame, Pi Represents the received point cloud dataset, i Indicates the number of point cloud datasets, PCA It represents the algorithm used for dimensionality reduction compression, and quantize represents the algorithm used to compress the amount of data, such as sampling method.

[0048] The coordinate compression processing of point cloud data in the spatiotemporal coordinate system reduces the complexity of data storage and processing by reducing data redundancy and retaining key features, while retaining sufficient information to support subsequent analysis.

[0049] The compressed point cloud data is represented as: Q1, Q2, ..., QN}.

[0050] Furthermore, the processing unit 24 performs overlay processing on the point cloud data using the following formula:

[0051] ,

[0052] in, Indicates the number of acquisition windows. Each acquisition window collects a set number of point cloud data. is the weight, adjusted according to the distance between frames, Q is the superimposed point cloud dataset, Represents the compressed point cloud dataset, Indicates continuous time The collected point cloud data, express Point cloud dataset mi The three-dimensional coordinates of the points, j =1,2,… mi .

[0053] Use the spacetime superposition function: S:P′ → Q ,in Q is the superimposed point cloud dataset, P′ is the compressed point cloud data; function S adjusts weights based on the frequency and characteristics of each time frame. These weights determine the contribution of each point cloud data point in the final overlay result. Specifically, S uses the given weights and inter-frame feature information to integrate compressed point cloud data collected at multiple time points, generating a point cloud dataset that represents a more comprehensive or holistic representation.

[0054] Furthermore, the data enhancement system of the present invention further comprises a clustering unit connected to the processing unit, and the clustering unit is used to utilize a clustering algorithm G:Q → {C1,C2,…,Cs}, cluster the enhanced signal to divide it intos Different clustering patterns Ck The clustering algorithm is used to group point clouds according to their spatial distribution and temporal continuity.

[0055] The data enhancement system of this invention effectively enhances raw point cloud data. This enhanced point cloud data not only increases data density and information content, but also provides a richer and more accurate data foundation for subsequent fall detection algorithms. On this basis, the fall detection algorithm is optimized to increase its processing speed and recognition accuracy.

[0056] like Figure 3 The figure shows the point cloud enhancement information generated by the data enhancement system of the present invention after the human body is at rest. The data enhancement system and method of the present invention can achieve the best judgment effect when the human body is at rest. The data of the human body during movement is a scattered point cloud and is only used for data buffering. After the human body is at rest, the enhanced point cloud data serves as the data input for the algorithm, further improving the accuracy and reliability of fall detection.

[0057] The data enhancement method and system of the present invention can adapt to the user environment and actively enhance the human body point cloud, providing more accurate and reliable data support for the detection of abnormal human behavior.

[0058] The present invention also provides a data enhancement method based on FMCW millimeter wave radar, which is described below.

[0059] like Figure 1 As shown, the data enhancement method of the present invention includes the following steps:

[0060] Execute step S11 to receive a point cloud data set collected by a millimeter-wave radar; then execute step S12;

[0061] Execute step S12 to map the received point cloud data set into a spatiotemporal coordinate system; then execute step S13;

[0062] Execute step S13 to perform coordinate compression processing on the point cloud data in the spatiotemporal coordinate system; then execute step S14;

[0063] Step S14 is executed to perform superposition processing on the compressed point cloud data using a spatiotemporal superposition function to form an enhanced signal.

[0064] First, perform multi-frame data aggregation:

[0065] Assume that the radar is in continuous time t1,t2,…,tn The point cloud datasets collected in are:

[0066] Pt1, Pt2,…, Ptn,

[0067] in Pti ={( xj,yj,zj,Ij )∣ j =1,2,…,mi} means at time ti Collected mi The three-dimensional coordinates (x, y, z) of each point and other information I (Such as intensity, speed, etc.).

[0068] This process captures the changes in human motion in time series and provides a dataset P={Pt1,Pt2,…,Ptn} For subsequent analysis.

[0069] Then perform space-time coordinate mapping:

[0070] Each time point ti Point cloud data Pti , mapped to a unified space-time coordinate system to ensure different Pti can be analysed within a common framework.

[0071] The coordinate compression processing of point cloud data in the spatiotemporal coordinate system reduces the complexity of data storage and processing by reducing data redundancy and retaining key features, while retaining sufficient information to support subsequent analysis.

[0072] Furthermore, the following formula is used to perform coordinate compression on the point cloud data:

[0073] Qi =quantize( PCA ( Pi )),

[0074] in, Qi Represents the compressed point cloud data frame, Pi Represents the received point cloud dataset, i Indicates the number of point cloud datasets, PCA It represents the algorithm used for dimensionality reduction compression, and quantize represents the algorithm used to compress the amount of data, such as sampling method.

[0075] Assume that the compressed point cloud data is represented as:

[0076] {Q1,Q2,…,QN}

[0077] Each of these Qi Represents the compressed point cloud data frame.

[0078] In a specific embodiment of the present invention,

[0079] Use the spacetime superposition function: S:P′ → Q ,in Q It is the superimposed point cloud dataset.

[0080] The following formula is used to overlay the point cloud data:

[0081] ,

[0082] in, Indicates the number of acquisition windows. Each acquisition window collects a set number of point cloud data. is the weight, adjusted according to the distance between frames, Q is the superimposed point cloud dataset, Represents the compressed point cloud dataset, Indicates continuous time The collected point cloud data, express Point cloud dataset mi The three-dimensional coordinates of the points, j =1,2,… mi .

[0083] Furthermore, it also includes: using clustering algorithm G:Q → {C1,C2,…,Cs}, cluster the enhanced signal to divide it into s Different clustering patterns Ck .

[0084] The present invention also provides a storage medium, on which is stored a program for a data enhancement method based on an FMCW millimeter-wave radar, and the steps of the data enhancement method based on an FMCW millimeter-wave radar are implemented when the program for the data enhancement method based on an FMCW millimeter-wave radar is executed by a processor.

[0085] The present invention has been described in detail above with reference to the embodiments of the accompanying drawings. A person skilled in the art can make various modifications to the present invention based on the above description. Therefore, certain details in the embodiments should not be construed as limiting the present invention. The scope of protection of the present invention shall be determined by the scope defined in the appended claims.

Claims

1. A data enhancement method based on FMCW millimeter wave radar, characterized in that: The steps include: Receive point cloud datasets collected by millimeter-wave radar; Mapping the received point cloud data set into a spatiotemporal coordinate system; Taking advantage of the fact that human body movements are slow, continuous and rarely repeated, coordinate compression processing is performed on the point cloud data in the space-time coordinate system. The formula is as follows: Qi=quantize(PCA(Pi)), Among them, Qi represents the compressed point cloud data frame, and Pi represents the received point cloud data set; Use the spatiotemporal superposition function to superimpose the compressed point cloud data to form an enhanced signal, which serves as the data basis for optimizing the fall detection algorithm; The following formula is used to overlay the point cloud data: Among them, q represents the number of acquisition windows, and a set number of point cloud data are collected in each acquisition window. ω is the weight, and Q is the superimposed point cloud data set. Represents the compressed point cloud dataset, Represents the continuous time t k ′ collected point cloud data, (x′ j ,y′ j ,z′ j )express The 3D coordinates of mi points in the point cloud dataset, j = 1, 2, ...mi; Use the space-time overlay function to adjust the weights based on the frequency and characteristics between each time frame. These weights are used to determine the contribution of the point cloud data at each time point to the final overlay result.

2. The data enhancement method based on FMCW millimeter wave radar according to claim 1, characterized in that: The method further includes: utilizing a clustering algorithm to cluster the enhanced signal to divide it into s different clustering modes.

3. A storage medium, characterized in that: The storage medium stores a program of a data enhancement method based on FMCW millimeter-wave radar. When the program of the data enhancement method based on FMCW millimeter-wave radar is executed by a processor, the steps of the data enhancement method based on FMCW millimeter-wave radar as described in any one of claims 1 to 2 are implemented.

4. A data enhancement system based on FMCW millimeter wave radar, characterized in that: include: A receiving unit, used to receive a point cloud data set collected by a millimeter-wave radar; a coordinate mapping unit connected to the receiving unit, and configured to map the point cloud data set received by the receiving unit into a spatiotemporal coordinate system; The compression unit is connected to the coordinate mapping unit and is used to perform coordinate compression processing on the point cloud data in the space-time coordinate system by taking advantage of the fact that human body movements are slow, continuous and rarely repeated. The formula is as follows: Qi=quantize(PCA(Pi)), Among them, Qi represents the compressed point cloud data frame, and Pi represents the received point cloud data set; a processing unit connected to the compression unit, configured to perform a spatiotemporal superposition process on the compressed point cloud data to form an enhanced signal, which serves as a data basis for optimizing the fall detection algorithm; The processing unit performs overlay processing on the point cloud data using the following formula: Among them, q represents the number of acquisition windows, and a set number of point cloud data are collected in each acquisition window. ω is the weight, and Q is the superimposed point cloud data set. Represents the compressed point cloud dataset, Represents the continuous time t k ′ collected point cloud data, (x′ j ,y′ j ,z′ j )express The 3D coordinates of mi points in the point cloud dataset, j = 1, 2, ...mi; Use the space-time overlay function to adjust the weights based on the frequency and characteristics between each time frame. These weights are used to determine the contribution of the point cloud data at each time point to the final overlay result.

5. The data enhancement system based on FMCW millimeter wave radar according to claim 4, characterized in that: It also includes a clustering unit connected to the processing unit, and configured to perform clustering division on the enhanced signal using a clustering algorithm to divide the enhanced signal into s different clustering modes.

Images