Environmental Noise Recognition Method, System and Storage Medium Based on FMCW Millimeter-Wave Radar

Abstract

The present invention relates to the technical field of radar data processing, and particularly to an environmental noise recognition method, system and storage medium based on an FMCW millimeter-wave radar. The recognition method includes: receiving detection data of the millimeter-wave radar; performing information segmentation on the received detection data based on behavior data to obtain a data set within continuous time; using the compression technology of point cloud coordinates to perform compression and superposition processing on the data set within continuous time in a spatial matrix; setting a breakdown value, and when the occurrence frequency of a certain point cloud data exceeds the set breakdown value during the superposition process, then marking the point cloud data as a breakdown state, thereby completing the recognition of environmental noise. The present invention can significantly improve the accuracy and efficiency of environmental noise recognition, can effectively solve the problem of sudden noise interference in the home environment, and improves the detection stability and accuracy of the radar system in complex environments.

Description

Technical Field

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

[0002] Due to its strong penetration, accurate ranging, strong anti-interference ability and other characteristics, FMCW millimeter-wave radar is widely used in human motion capture and environmental perception. However, in actual 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 the wind blowing a wet shower curtain, the operation of a washing machine, and the sudden operation of devices such as a fan. They will generate a large number of interference signals in the radar detection area. Especially when an emergency such as a person falling occurs, these noises may cover up key point cloud information, resulting in missed reports or false reports, seriously affecting the reliability of the fall recognition algorithm.

[0003] Traditional methods usually rely on statistics of information such as the coordinates, duration, and frequency of point clouds to try to filter out noise, but in the face of complex and variable environmental noise, the effect is limited. Especially those noises with high frequency and strong repeatability, such as the operation of a washing machine and the rotation of a fan, their characteristics make it difficult to effectively distinguish them through simple statistical methods. To address these problems, there is an urgent need for a technology that can efficiently identify sudden noises to enhance the performance of the radar system 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 an FMCW millimeter-wave radar, so as to solve the problem that the existing noise recognition methods have limited effects in the face of complex and variable environmental noise.

[0005] The technical solution for achieving the above purpose is as follows:

[0006] The present invention provides an environmental noise recognition method based on an FMCW millimeter-wave radar, including the following steps:

[0007] Receiving the detection data of the millimeter-wave radar;

[0008] Performing information segmentation on the received detection data based on behavior data to obtain a data set within a continuous time;

[0009] Using the compression technology of point cloud coordinates, performing compression and superposition processing on the data set within a continuous time in a spatial matrix;

[0010] Setting a breakdown value. When the occurrence frequency of a certain point cloud data exceeds the set breakdown value during the superposition process, the point cloud data is marked as the breakdown state, thus completing the recognition of environmental noise.

[0011] A further improvement of the environmental noise recognition method based on the FMCW millimeter-wave radar of the present invention lies in ensuring the coherence of attitude changes in the segmented data when performing information segmentation on the received detection data based on behavioral data.

[0012] A further improvement of the environmental noise recognition method based on the FMCW millimeter-wave radar of the present invention lies in performing compression processing on the data set within continuous time in the spatial matrix by using the following formula:

[0013] Qi =quantize( PCA ( Pti )),

[0014] where, Qi represents the compressed point cloud data frame, Pti represents the data set within continuous time obtained by segmentation.

[0015] A further improvement of the environmental noise recognition method based on the FMCW millimeter-wave radar of the present invention lies in performing superposition processing on the compressed data by using the following formula:

[0016] ,

[0017] where, is the superposed point cloud data, Qi is the matrix after compression and dimensionality reduction.

[0018] The present invention also provides a storage medium, on which a program of an environmental noise recognition method based on the FMCW millimeter-wave radar is stored. When the program of the environmental noise recognition method based on the FMCW millimeter-wave radar is executed by a processor, the steps of the environmental noise recognition method based on the FMCW millimeter-wave radar are implemented.

[0019] The present invention further provides an environmental noise recognition system based on the FMCW millimeter-wave radar, including:

[0020] a receiving unit for receiving detection data of the millimeter-wave radar;

[0021] a segmentation unit connected to the receiving unit for performing information segmentation on the received detection data based on behavioral data to obtain a data set within continuous time;

[0022] a compression unit connected to the segmentation unit for performing compression processing on the data set within continuous time in the spatial matrix by using the compression technology of point cloud coordinates;

[0023] A processing unit, connected to the compression unit, is configured to perform superposition processing on the data after compression processing. During the superposition process, it determines whether the occurrence frequency of a certain point cloud data exceeds a set breakdown value. If so, it labels the point cloud data as the breakdown state, thereby completing the identification of environmental noise.

[0024] A further improvement of the environmental noise identification system based on the FMCW millimeter-wave radar of the present invention lies in that when the segmentation unit performs information segmentation on the detection data, it ensures the coherence of the attitude change in the segmented data.

[0025] A further improvement of the environmental noise identification system based on the FMCW millimeter-wave radar of the present invention lies in that the compression unit uses the following formula to perform compression processing on the data set within continuous time in the spatial matrix:

[0026] Qi =quantize( PCA ( Pti )),

[0027] where, Qi represents the compressed point cloud data frame, Pti represents the data set within continuous time obtained by segmentation.

[0028] A further improvement of the environmental noise identification system based on the FMCW millimeter-wave radar of the present invention lies in that the processing unit uses the following formula to perform superposition processing on the data after compression processing:

[0029] ,

[0030] where, is the superposed point cloud data, Qi is the matrix after compression and dimensionality reduction.

[0031] The beneficial effects of the environmental noise identification method, system, and storage medium based on the FMCW millimeter-wave radar of the present invention are as follows:

[0032] By performing compression processing on the data, the present invention reduces the data volume while retaining key information, performs superposition processing on the data to increase the data density, and also makes the high-frequency repetitive noise signals gradually prominent under continuous superposition until reaching a breakdown state. At this time, the noise characteristics become extremely distinct and are easy to separate from the background noise.

[0033] The present invention can significantly improve the accuracy and efficiency of environmental noise identification, 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. Description of the Drawings

[0034] Figure 1This is a flowchart of the method for identifying environmental noise based on an FMCW millimeter-wave radar according to the present invention.

[0035] Figure 2 This is a system diagram of the system for identifying environmental noise based on an FMCW millimeter-wave radar according to the present invention.

[0036] Figure 3 This is a visualization grid map of the point cloud accumulation generated by the interference source in the method and system for identifying environmental noise based on an FMCW millimeter-wave radar according to the present invention.

[0037] Figure 4 This is a grid map after the noise point cloud generated by the interference source in the method and system for identifying environmental noise based on an FMCW millimeter-wave radar according to the present invention accumulates to the breakdown value. Detailed implementation manners

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

[0039] Refer to Figure 1 The present invention provides a method, system, and storage medium for identifying environmental noise based on an FMCW millimeter-wave radar, aiming to efficiently and accurately identify and filter out non-deterministic noise in the environment, improve the stability and accuracy of the human behavior recognition algorithm based on radar, and is particularly significant in key applications such as human fall detection. The method, system, and storage medium for identifying environmental noise based on an FMCW millimeter-wave radar according to the present invention will be described below in conjunction with the accompanying drawings.

[0040] Refer to Figure 2 which shows a system diagram of the system for identifying environmental noise based on an FMCW millimeter-wave radar according to the present invention. The system for identifying environmental noise based on an FMCW millimeter-wave radar according to the present invention will be described below in conjunction with Figure 2 for the description.

[0041] As Figure 2 shown, the system for identifying environmental noise based on an FMCW millimeter-wave radar according to the present invention includes a receiving unit 21, a segmentation unit 22, a compression unit 23, and a processing unit 24. Among them, 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 behavior data to obtain a data set within a continuous time period. The compression unit 23 is used to perform compression processing on the data set within a continuous time period in the spatial matrix by using the compression technology of point cloud coordinates. The processing unit 24 is used to perform superposition processing on the compressed data. During the superposition process, it is judged whether the occurrence frequency of a certain point cloud data exceeds a set breakdown value. If so, the point cloud data is marked as the breakdown state, thus completing the identification of environmental noise.

[0042] The millimeter-wave radar of the present invention is used for detecting human body falls in a home environment. The millimeter-wave radar performs real-time detection on the home environment to obtain detection data, which includes human activity information characteristics. However, in a home environment, there are often some non-deterministic noises, such as the noise generated when the wind in the bathroom blows the wet shower curtain, the operation of the washing machine, and the fan that was originally blocked by the human body is directly exposed within the radar detection range after a fall. These noises will cover up the point cloud information generated during the human body fall process, resulting in missed reports. The environmental noise recognition system of the present invention is used to identify the non-deterministic noises 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.

[0043] Furthermore, the segmentation unit 22 is used to synchronize the detection data of the millimeter-wave radar in time and perform information segmentation based on the collected behavior data, where the behavior data is human activity information, such as the activity information of a human body from a sitting position to a standing position and the activity information of a human body walking. Specifically, the human activity information can be manually marked to facilitate the segmentation unit 22 to identify the corresponding human activity information during segmentation.

[0044] After the segmentation unit 22 segments the data, it screens the segmented data to ensure the integrity of the retained human activity information characteristics.

[0045] Still further, when the segmentation unit 22 performs information segmentation on the detection data, it ensures the coherence of the posture changes in the segmented data. The coherence of the posture changes means that the segmented data is the continuous data of a certain action of the human body.

[0046] The data set within the continuous time segmented by the segmentation unit 22 can be expressed as: Pti ={( xj , yj , zj , Ij )∣ j =1,2,…, mi}, where within the continuous time t1 , t2 ,…, tn The collected data is expressed as: Pt1 , Pt2 ,…, Ptn, represents the three-dimensional coordinate information ( ti ) and other information mi (such as intensity, speed, etc.) of the data collected from x, y, z to I at time

[0047] In a specific embodiment of the present invention, the compression unit 23 performs compression processing on the data set within continuous time in the spatial matrix using the following formula:

[0048] Qi =quantize( PCA ( Pti )),

[0049] wherein, Qi represents the compressed point cloud data frame, { Q1, Q2, … QN}, Pti represents the data set within continuous time obtained by segmentation.

[0050] The compression unit 23 of the present invention uses the compression technology of point cloud coordinates to perform compression processing on the segmented data in the spatial matrix, aiming to reduce the data volume while retaining key information.

[0051] In a specific embodiment of the present invention, the processing unit 24 performs superposition processing on the compressed data using the following formula:

[0052] ,

[0053] wherein, is the superposed point cloud data, Qi is the matrix after compression and dimensionality reduction.

[0054] Furthermore, a breakdown value T is set as a threshold to process the superposed data. When the occurrence frequency of a certain point cloud data during superposition exceeds the threshold T, it is considered that the point reaches the "breakdown" state. The formula is:

[0055] if Si > T, then Si ∈ thresholding 。

[0056] This process can not only enhance the data density, but also make the high-frequency repeated noise signals (i.e., the main interference sources) gradually prominent under continuous superposition until reaching a "breakdown" critical point. At this time, the noise characteristics become extremely distinct and are easy to separate from the background noise.

[0057] The set breakdown value T of the present invention can be obtained by conversion according to the detection data of the millimeter-wave radar or can 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.

[0058] Example, as Figure 3 and Figure 4 shown, Figure 3 shows the cumulative visualization grid map of the point cloud generated by the interference source, Figure 4The grid map shows the noise point cloud generated by the interference source after accumulating to the breakdown value. It can be seen that the noise characteristics become extremely distinct and are easy to separate from the background noise.

[0059] The environmental noise recognition method of the present invention can be called the cumulative breakdown method, which can significantly improve the accuracy and efficiency of environmental noise recognition. 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.

[0060] The present invention also provides an environmental noise recognition method based on an FMCW millimeter-wave radar. The following is an explanation of this environmental noise recognition method.

[0061] As Figure 1 shown, the environmental noise recognition method based on an FMCW millimeter-wave radar of the present invention includes the following steps:

[0062] Execute step S11 to receive the detection data of the millimeter-wave radar; then execute step S12;

[0063] Execute step S12 to perform information segmentation on the received detection data based on behavioral data to obtain a data set within continuous time; then execute step S13;

[0064] Execute step S13 to use the compression technology of point cloud coordinates to perform compression and superposition processing on the data set within continuous time in the spatial matrix; then execute step S14;

[0065] Execute step S14 to set a breakdown value. When the occurrence frequency of a certain point cloud data during the superposition process exceeds the set breakdown value, then mark this point cloud data as the breakdown state, thus completing the recognition of environmental noise.

[0066] In a specific embodiment of the present invention, when performing information segmentation on the received detection data based on behavioral data, ensure the coherence of the attitude changes in the segmented data.

[0067] Specifically, synchronize the recorded data (i.e., the detection data of the millimeter-wave radar) in time, perform information segmentation based on the collected behavioral data, and then screen the segmented data to ensure the integrity of the human activity information features retained. Note that the coherence feature of the attitude changes in the data must be retained during cutting.

[0068] Obtain the data: Pti ={( xj , yj , zj , Ij )∣ j =1,2,…, mi}, where within continuous time t1, t2 ,…, tn The collected data is represented as: Pt1 , Pt2 ,…, Ptn, indicating the three-dimensional coordinate information ( ti from which the data is collected to mi ), as well as other information x, y, z ), such as intensity, speed, etc. I (such as intensity, speed, etc.).

[0069] In a specific embodiment of the present invention, the following formula is used to compress the data set in the continuous time in the spatial matrix:

[0070] Qi =quantize( PCA ( Pti )),

[0071] where Qi represents the compressed point cloud data frame, Pti represents the data set in the continuous time obtained by segmentation.

[0072] The purpose of compressing the data is to reduce the data volume while retaining the key information.

[0073] Furthermore, the following formula is used to perform superposition processing on the compressed data:

[0074] ,

[0075] where is the superposed point cloud data, Qi is the matrix after compression and dimensionality reduction.

[0076] Still further, a breakdown value T is set as the threshold to process the superposed data. When the occurrence frequency of a certain point cloud data during superposition exceeds the threshold T, it is considered that the point reaches the "breakdown" state. The formula is:

[0077] if Si > T, then Si ∈ thresholding .

[0078] This process can not only enhance the data density, but also make the high-frequency repetitive noise signals (i.e., the main interference sources) gradually become prominent under continuous superposition until reaching a "breakdown" critical point, at which time the noise characteristics become extremely distinct and are easy to separate from the background noise.

[0079] The present invention also provides a storage medium, on which a program for the method of identifying environmental noise based on an FMCW millimeter-wave radar is stored, and the steps of the method of identifying environmental noise based on an FMCW millimeter-wave radar implemented when the program for the method of identifying environmental noise based on an FMCW millimeter-wave radar is executed by a processor.

[0080] The present invention has been described in detail above in conjunction with the embodiments with reference to the drawings. Those of ordinary skill 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 to the present invention, and the scope of the present invention will be defined by the scope defined in the appended claims.

Claims

1. An environmental noise recognition method based on an FMCW millimeter-wave radar, characterized in that, It includes the following 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 within continuous time; Using the compression technology of point cloud coordinates to perform compression and superposition processing on the data set within continuous time in the spatial matrix; Setting a breakdown value. When the occurrence frequency of a certain point cloud data exceeds the set breakdown value during the superposition process, the point cloud data is marked as the breakdown state, thus completing the identification of environmental noise; wherein, When performing information segmentation on the received detection data based on behavior data, ensure the coherence of posture changes in the segmented data and screen the segmented data to ensure the integrity of the retained human activity information features.

2. The method for identifying environmental noise based on an FMCW millimeter-wave radar according to claim 1, characterized in that, Using the following formula to perform compression processing on the data set within continuous time in the spatial matrix: Qi = quantize(PCA(Pti)), where Qi represents the compressed point cloud data frame, and Pti represents the data set within continuous time obtained by segmentation.

3. The environmental noise recognition method based on FMCW millimeter wave radar according to claim 1, characterized in that, Using the following formula to perform superposition processing on the compressed data: where S is the superposed point cloud data, and Qi is the matrix after compression and dimensionality reduction.

4. A storage medium, characterized in that, A program for the environmental noise identification method based on the FMCW millimeter-wave radar is stored on the storage medium. When the program for the environmental noise identification method based on the FMCW millimeter-wave radar is executed by a processor, it realizes the steps of the environmental noise identification method based on the FMCW millimeter-wave radar as described in any one of claims 1 to 3.

5. An environmental noise recognition system based on an FMCW millimeter-wave radar, characterized in that, It includes: A receiving unit for receiving the detection data of the millimeter-wave radar; A segmentation unit connected to the receiving unit for performing information segmentation on the received detection data based on behavior data to obtain a data set within continuous time; when the segmentation unit performs information segmentation on the detection data, it ensures the coherence of posture changes in the segmented data and screens the segmented data to ensure the integrity of the retained human activity information features; A compression unit connected to the segmentation unit for using the compression technology of point cloud coordinates to perform compression processing on the data set within continuous time in the spatial matrix; A processing unit connected to the compression unit for performing superposition processing on the data after compression processing. During the superposition process, it determines whether the occurrence frequency of a certain point cloud data exceeds the set breakdown value. If so, the point cloud data is marked as the breakdown state, thus completing the identification of environmental noise.

6. The environmental noise recognition system based on the FMCW millimeter-wave radar according to claim 5, characterized in that The compression unit uses the following formula to perform compression processing on the data set within continuous time in the spatial matrix: Qi = quantize(PCA(Pti)), where Qi represents the compressed point cloud data frame, and Pti represents the data set within continuous time obtained by segmentation.

7. The environmental noise recognition system based on the FMCW millimeter-wave radar according to claim 5, characterized in that, The processing unit uses the following formula to perform superposition processing on the data after compression processing: where S is the superposed point cloud data, and Qi is the matrix after compression and dimensionality reduction.

Images