Environmental noise identification method and system based on FMCW millimeter wave radar, and storage medium

Abstract

The invention relates to the technical field of radar data processing, in particular to an environmental noise recognition method and system based on an FMCW millimeter wave radar and a storage medium, and the recognition method comprises the steps: receiving the detection data of the millimeter wave radar; performing information segmentation on the received detection data based on behavior data to obtain a data set in continuous time; carrying out compression and superposition processing on the data set in continuous time in a space matrix by utilizing a point cloud coordinate compression technology; setting a breakdown value, and when the occurrence frequency of a certain point cloud data in the superposition process exceeds the set breakdown value, calibrating the point cloud data as a breakdown state, thereby completing the recognition of the environmental noise. According to the invention, the accuracy and efficiency of environment noise recognition can be remarkably improved, the problem of sudden noise interference in a home environment can be effectively solved, and the detection stability and accuracy of a radar system in a complex environment are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of radar data processing, and in particular to an environmental noise recognition method, system and storage medium based on FMCW millimeter wave radar. Background Art

[0002] FMCW millimeter wave radar is widely used in human motion capture and environmental perception due to its strong penetration, accurate ranging, and strong anti-interference ability. However, in practical applications, environmental noise has become one of the key factors affecting the quality of radar data. These noises may come from a variety of non-deterministic sources, such as wind blowing wet shower curtains, washing machines running, and sudden operation of equipment such as fans. They will generate a large number of interference signals in the radar detection area. Especially when emergency events such as human falls occur, these noises may cover the key point cloud information, resulting in missed reports or false reports, which seriously affects the reliability of the fall recognition algorithm.

[0003] Traditional methods usually rely on statistical point cloud coordinates, duration, frequency and other information to try to filter out noise, but the effect is limited when faced with complex and changeable environmental noise. Especially those high-frequency and repetitive noises, such as washing machine operation and fan rotation, their characteristics make them difficult to effectively distinguish through simple statistical methods. To address these problems, a technology that can efficiently identify sudden noise is urgently needed to enhance the performance of radar systems in complex environments. Summary of the invention

[0004] The purpose of the present invention is to overcome the defects of the prior art and provide an environmental noise recognition method, system and storage medium based on FMCW millimeter wave radar to solve the problem that the existing noise recognition method has limited effect when facing complex and changeable environmental noise.

[0005] The technical solution to achieve the above purpose is: The present invention provides an environmental noise recognition method based on FMCW millimeter wave radar, comprising the following steps: Receive detection data from millimeter-wave radar; Performing information segmentation on the received detection data based on the behavior data to obtain a data set in a continuous time; Using the compression technology of point cloud coordinates, the data sets in continuous time are compressed and superimposed in the space matrix; The breakdown value is set. When the frequency of occurrence of a certain point cloud data during the superposition process exceeds the set breakdown value, the point cloud data is marked as a breakdown state, thereby completing the identification of environmental noise.

[0006] A further improvement of the environmental noise recognition method based on FMCW millimeter wave radar of the present invention is that when information segmentation is performed on the received detection data based on the behavior data, the consistency of the posture changes in the segmented data is ensured.

[0007] The further improvement of the environmental noise recognition method based on FMCW millimeter wave radar of the present invention is that the data set in the continuous time is compressed in the space matrix using the following formula: Qi =quantize( PCA ( Pt )), in, Qi Represents the compressed point cloud data frame, Pt Represents the continuous time data set obtained by segmentation.

[0008] A further improvement of the environmental noise recognition method based on FMCW millimeter wave radar of the present invention is to perform superposition processing on the compressed data using the following formula: , in, is the superimposed point cloud data, Qi is the matrix after compression and dimensionality reduction.

[0009] The present invention also provides a storage medium, on which is stored a program of an environmental noise identification method based on FMCW millimeter wave radar. When the program of the environmental noise identification method based on FMCW millimeter wave radar is executed by a processor, the steps of the environmental noise identification method based on FMCW millimeter wave radar are implemented.

[0010] The present invention further provides an environmental noise recognition system based on FMCW millimeter wave radar, comprising: A receiving unit, used for receiving detection data of the millimeter wave radar; a segmentation unit connected to the receiving unit, and configured to perform information segmentation on the received detection data based on the behavior data to obtain a data set within a continuous time; A compression unit, connected to the segmentation unit, for compressing the data set in a continuous time in a spatial matrix using a compression technology of point cloud coordinates; The processing unit is connected to the compression unit and is used to perform superposition processing on the compressed data. During the superposition process, it is determined whether the occurrence frequency of a certain point cloud data exceeds the set breakdown value. If so, the point cloud data is marked as a breakdown state, thereby completing the identification of environmental noise.

[0011] A further improvement of the environmental noise recognition system based on FMCW millimeter wave radar of the present invention is that the segmentation unit ensures the continuity of posture changes in the segmented data when performing information segmentation on the detection data.

[0012] A further improvement of the environmental noise identification system based on FMCW millimeter wave radar of the present invention is that the compression unit compresses the data set in the continuous time in the spatial matrix using the following formula: Qi =quantize( PCA ( Pt )), in, Qi Represents the compressed point cloud data frame, Pt Represents the continuous time data set obtained by segmentation.

[0013] A further improvement of the environmental noise recognition system based on FMCW millimeter wave radar of the present invention is that the processing unit performs superposition processing on the compressed data using the following formula: , in, is the superimposed point cloud data, Qi is the matrix after compression and dimensionality reduction.

[0014] The beneficial effects of the environmental noise identification method, system and storage medium based on FMCW millimeter wave radar of the present invention are: The present invention reduces the amount of data by compressing the data while retaining key information, and increases the density of the data by superimposing the data. It also makes high-frequency repetitive noise signals gradually stand out under continuous superposition until a breakdown state is reached, at which time the noise characteristics become extremely distinct and easy to separate from the background noise.

[0015] The present invention can significantly improve the accuracy and efficiency of environmental noise recognition, effectively solve the problem of sudden noise interference in home environments, and improve the detection stability and accuracy of radar systems in complex environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 The figure is a flow chart of the environmental noise recognition method based on FMCW millimeter wave radar of the present invention.

[0017] Figure 2 This is a system diagram of the environmental noise recognition system based on FMCW millimeter wave radar of the present invention.

[0018] Figure 3 This is a point cloud accumulation visualization grid diagram generated by the environmental noise identification method based on FMCW millimeter wave radar and the interference source in the system of the present invention.

[0019] Figure 4 It is a grid diagram after the noise point cloud generated by the interference source in the environmental noise identification method based on FMCW millimeter wave radar and the system of the present invention is accumulated to the breakdown value. DETAILED DESCRIPTION

[0020] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments.

[0021] See also Figure 1 The present invention provides an environmental noise recognition method, system and storage medium based on FMCW millimeter wave radar, which aims to improve the stability and accuracy of radar-based human behavior recognition algorithms by efficiently and accurately identifying and filtering out non-deterministic noise in the environment, especially in key applications such as human fall detection. The environmental noise recognition method, system and storage medium based on FMCW millimeter wave radar of the present invention are described below in conjunction with the accompanying drawings.

[0022] See also Figure 2 , shows the system diagram of the environmental noise identification system based on FMCW millimeter wave radar of the present invention. Figure 2 , the environmental noise recognition system based on FMCW millimeter wave radar of the present invention is described.

[0023] like Figure 2 As shown, the environmental noise identification system based on FMCW millimeter wave radar of the present invention includes a receiving unit 21, a segmentation unit 22, a compression unit 23 and a processing unit 24, wherein the segmentation unit 22 is connected to the receiving unit 21, the compression unit 23 is connected to the segmentation unit 22, and the processing unit 24 is connected to the compression unit 23; the receiving unit 21 is used to receive the detection data of the millimeter wave radar; the segmentation unit 22 is used to perform information segmentation on the received detection data based on the behavior data to obtain a data set in a continuous time; the compression unit 23 is used to compress the data set in a continuous time in a space matrix using the compression technology of the point cloud coordinates; the processing unit 24 is used to perform superposition processing on the compressed data, and in the superposition process, it is determined whether the occurrence frequency of a certain point cloud data exceeds the set breakdown value, and if so, the point cloud data is calibrated as a breakdown state, thereby completing the identification of environmental noise.

[0024] The millimeter wave radar of the present invention is used for human fall detection in a home environment. The millimeter wave radar performs real-time detection of the home environment to obtain detection data, and the detection data includes the characteristics of human activity information. However, some non-deterministic noises often appear in the home environment, such as the noise generated by the wind in the bathroom blowing the wet shower curtain, the operation of the washing machine, and the fan originally blocked by the human body being directly exposed to the radar detection range after falling. These noises will cover the point cloud information generated by the human fall process, resulting in missed reports. The environmental noise recognition system of the present invention is used to realize the non-deterministic noise that appears in the radar detection data. The receiving unit 21 of the present invention is connected to the millimeter wave radar and can receive the detection data of the millimeter wave radar in real time.

[0025] Furthermore, the segmentation unit 22 is used to synchronize the detection data of the millimeter wave radar and perform information segmentation based on the collected behavior data, wherein the behavior data is human activity information, such as human activity information from sitting to standing and human activity information of walking, etc. Specifically, the human activity information can be manually marked so that the segmentation unit 22 can identify the corresponding human activity information during segmentation.

[0026] After segmenting the data, the segmentation unit 22 screens the segmented data to ensure the integrity of the retained human activity information characteristics.

[0027] Furthermore, when segmenting the detection data, the segmentation unit 22 ensures the continuity of posture changes in the segmented data. The continuity of posture changes means that the segmented data is continuous data of a certain action of the human body.

[0028] The data set in the continuous time obtained by segmentation unit 22 can be expressed as: Pt ={( xj , yj , zj , Ij )∣ j =1,2,…, mi}, where continuous time t1 , t2 ,…, tn The collected data is represented as: Pt1 , Pt2 ,…, Ptn, Indicates the time from ti Collect data from mi The three-dimensional coordinate information ( x,y,z ) and other information I (such as intensity, speed, etc.).

[0029] In a specific implementation of the present invention, the compression unit 23 compresses the data set in the continuous time in the space matrix using the following formula: Qi =quantize( PCA ( Pt )), in, Qi Represents the compressed point cloud data frame, { Q1, Q2, ... QN}, Pt Represents the continuous time data set obtained by segmentation.

[0030] The compression unit 23 of the present invention utilizes the compression technology of point cloud coordinates to compress the segmented data in a spatial matrix, aiming to reduce the amount of data while retaining key information.

[0031] In a specific implementation of the present invention, the processing unit 24 performs superposition processing on the compressed data using the following formula: , in, is the superimposed point cloud data, Qi is the matrix after compression and dimensionality reduction.

[0032] Furthermore, a breakdown value T is set as a threshold to process the superimposed data. When the frequency of occurrence of a point cloud data in the superposition process exceeds the threshold T, it is considered that the point has reached the "breakdown" state. The formula is: if Si>T, then Si ∈ thresholding .

[0033] This process not only enhances the density of data, but also makes high-frequency repetitive noise signals (i.e. the main source of interference) gradually stand out under continuous superposition until a critical point of "breakdown" is reached. At this time, the noise characteristics become extremely distinct and easy to separate from the background noise.

[0034] The set breakdown value T of the present invention can be converted according to the detection data of the millimeter wave radar, and can also be obtained through machine learning. When the detection data of the millimeter wave radar is sufficient, the set breakdown value T tends to be stable.

[0035] Examples, such as Figure 3 and Figure 4 As shown, Figure 3 The point cloud accumulation visualization grid diagram generated by the interference source is shown. Figure 4 The grid diagram shows the noise point cloud generated by the interference source after it accumulates to the breakdown value. It can be seen that the noise characteristics become extremely clear and easy to separate from the background noise.

[0036] The environmental noise identification method of the present invention can be called the cumulative breakdown method, which can significantly improve the accuracy and efficiency of environmental noise identification. This invention can effectively solve the problem of sudden noise interference in the home environment and improve the detection stability and accuracy of the radar system in complex environments.

[0037] The present invention also provides an environmental noise recognition method based on FMCW millimeter wave radar, and the environmental noise recognition method is described below.

[0038] like Figure 1 As shown, the environmental noise recognition method based on FMCW millimeter wave radar of the present invention includes the following steps: Execute step S11 to receive the detection data of the millimeter wave radar; then execute step S12; Execute step S12, perform information segmentation on the received detection data based on the behavior data to obtain a data set within a continuous time; then execute step S13; Execute step S13, use the compression technology of point cloud coordinates to compress and overlay the data set in the continuous time in the space matrix; then execute step S14; Execute step S14 to set a breakdown value. When the frequency of occurrence of a certain point cloud data during the superposition process exceeds the set breakdown value, the point cloud data is marked as a breakdown state, thereby completing the identification of environmental noise.

[0039] In a specific embodiment of the present invention, when information segmentation is performed on the received detection data based on the behavior data, the consistency of posture changes in the segmented data is ensured.

[0040] Specifically, the recorded data (i.e., the detection data of the millimeter-wave radar) is time-synchronized, information segmentation is performed based on the collected behavioral data, and then the segmented data is screened to ensure the integrity of the retained human activity information features. Note that the coherence features of the posture changes in the data must be retained during segmentation.

[0041] Get the data: Pt ={( xj , yj , zj , Ij )∣ j =1,2,…, mi}, where continuous time t1 , t2 ,…, tn The collected data is represented as: Pt1 , Pt2 ,…, Ptn, Indicates the time from ti Collect data from mi The three-dimensional coordinate information ( x,y, z ) and other information I (such as intensity, speed, etc.).

[0042] In a specific implementation of the present invention, the following formula is used to compress the data set in the continuous time in the space matrix: Qi =quantize( PCA ( Pt )), in, Qi Represents the compressed point cloud data frame, Pt Represents the continuous time data set obtained by segmentation.

[0043] The purpose of compressing data is to reduce the amount of data while retaining key information.

[0044] Furthermore, the compressed data is superimposed using the following formula: , in, is the superimposed point cloud data, Qi is the matrix after compression and dimensionality reduction.

[0045] Furthermore, a breakdown value T is set as a threshold to process the superimposed data. When the frequency of occurrence of a point cloud data in the superposition process exceeds the threshold T, it is considered that the point has reached the "breakdown" state. The formula is: if Si>T, then Si ∈ thresholding .

[0046] This process not only enhances the density of data, but also makes high-frequency repetitive noise signals (i.e. the main source of interference) gradually stand out under continuous superposition until a critical point of "breakdown" is reached. At this time, the noise characteristics become extremely distinct and easy to separate from the background noise.

[0047] The present invention also provides a storage medium, on which is stored a program of an environmental noise identification method based on FMCW millimeter wave radar, and the steps of the environmental noise identification method based on FMCW millimeter wave radar are implemented when the program of the environmental noise identification method based on FMCW millimeter wave radar is executed by a processor.

[0048] The present invention is described in detail above in conjunction with the embodiments of the accompanying drawings. A person skilled in the art can make various variations of the present invention according to the above description. Therefore, certain details in the embodiments should not constitute a limitation of the present invention, and the scope of protection of the present invention shall be defined by the scope of the attached claims.

Claims

1. A method for identifying environmental noise based on FMCW millimeter wave radar, characterized in that: The steps include: Receive detection data from millimeter-wave radar; Performing information segmentation on the received detection data based on the behavior data to obtain a data set in a continuous time; Using the compression technology of point cloud coordinates, the data sets in continuous time are compressed and superimposed in the space matrix; The breakdown value is set. When the frequency of occurrence of a certain point cloud data during the superposition process exceeds the set breakdown value, the point cloud data is marked as a breakdown state, thereby completing the identification of environmental noise.

2. The environmental noise identification method based on FMCW millimeter wave radar according to claim 1, characterized in that: When performing information segmentation on the received detection data based on the behavior data, the consistency of posture changes in the segmented data is ensured.

3. The environmental noise identification method based on FMCW millimeter wave radar according to claim 1, characterized in that: The following formula is used to compress the data set in continuous time in the spatial matrix: Qi =quantize( PCA ( Pt )), in, Qi Represents the compressed point cloud data frame, Pt Represents the continuous time data set obtained by segmentation.

4. The environmental noise identification method based on FMCW millimeter wave radar according to claim 1, characterized in that: The compressed data is superimposed using the following formula: ， in, is the superimposed point cloud data, Qi is the matrix after compression and dimensionality reduction.

5. A storage medium, characterized in that: The storage medium stores a program of an environmental noise identification method based on FMCW millimeter-wave radar, and when the program of the environmental noise identification method based on FMCW millimeter-wave radar is executed by a processor, the steps of the environmental noise identification method based on FMCW millimeter-wave radar as described in any one of claims 1 to 4 are implemented.

6. An environmental noise recognition system based on FMCW millimeter wave radar, characterized in that: include: A receiving unit, used for receiving detection data of the millimeter wave radar; a segmentation unit connected to the receiving unit, and configured to perform information segmentation on the received detection data based on the behavior data to obtain a data set within a continuous time; A compression unit, connected to the segmentation unit, for compressing the data set in a continuous time in a spatial matrix using a compression technology of point cloud coordinates; The processing unit is connected to the compression unit and is used to perform superposition processing on the compressed data. During the superposition process, it is determined whether the occurrence frequency of a certain point cloud data exceeds the set breakdown value. If so, the point cloud data is marked as a breakdown state, thereby completing the identification of environmental noise.

7. The environmental noise recognition system based on FMCW millimeter wave radar as claimed in claim 6, characterized in that: When performing information segmentation on the detection data, the segmentation unit ensures the continuity of posture changes in the segmented data.

8. The environmental noise recognition system based on FMCW millimeter wave radar as claimed in claim 6, characterized in that: The compression unit compresses the data set in the continuous time in the space matrix using the following formula: Qi =quantize( PCA ( Pt )), in, Qi Represents the compressed point cloud data frame, Pt Represents the continuous time data set obtained by segmentation.

9. The environmental noise recognition system based on FMCW millimeter wave radar as claimed in claim 6, characterized in that: The processing unit performs superposition processing on the compressed data using the following formula: ， in, is the superimposed point cloud data, Qi is the matrix after compression and dimensionality reduction.

Images