Shielding detection method for security radar

Abstract

The invention discloses a shielding detection method for a security radar, and the method comprises the steps: S1, a radar system alternately transmits two waveforms in a time division multiplexing mode in each processing frame period, and the two waveforms are a main detection waveform W1 used for executing a conventional security detection task and a health monitoring waveform W2 specially designed for shielding detection optimization; s2, preprocessing a received signal of the health monitoring waveform W2, and then performing distance dimension FFT (Fast Fourier Transform) processing to obtain a distance-amplitude spectrum; s3, extracting characteristic parameters based on the distance-amplitude spectrum; and S4, fusing the feature parameters based on an evidence theory and making a decision. According to the shielding detection method for the security radar, high-reliability and classifiable shielding detection is realized, the accuracy of shielding detection is improved, the false alarm rate is reduced, dependence on an external sensor is not needed, the system cost is reduced, and integration in an existing radar system is facilitated.

Description

Technical Field

[0001] This invention belongs to the field of security radar technology, specifically relating to a method for detecting obstruction in security radar. Background Technology

[0002] In the field of security monitoring, millimeter-wave radar is widely used in various security scenarios due to its all-weather operation and non-visual detection advantages. However, radars deployed outdoors are susceptible to natural factors such as dust accumulation, rain and snow, frost, and spider webs. They may also be subject to malicious obstruction. These conditions can lead to a decrease in radar detection performance or even complete failure, creating blind spots in security monitoring and posing significant risks to security operations.

[0003] Currently, existing technical solutions for radar obstruction detection have the following main shortcomings: Firstly, detection schemes based on a single signal strength threshold monitor the overall strength of the radar received signal or the amplitude of near-range reflected signals, determining obstruction when the signal exceeds a fixed threshold. This scheme has significant drawbacks. On one hand, it has a high false alarm rate; for example, normal environmental interference such as a bird briefly landing in front of the radar can trigger false alarms. On the other hand, it has poor adaptability; static thresholds cannot cope with environmental changes, easily leading to missed alarms or continuous false alarms.

[0004] Secondly, the multi-sensor data comparison approach compares radar data with data from external sensors such as cameras and laser beams. If other sensors detect a target but the radar does not, radar anomaly is suspected. This approach requires deploying and integrating multiple heterogeneous sensors, increasing system complexity and hardware costs. Furthermore, it cannot effectively identify slight or partial obstructions, resulting in significant detection limitations. Summary of the Invention

[0005] The main objective of this invention is to provide a method for detecting obstruction in security radar, which achieves highly reliable and classifiable obstruction detection, improves the accuracy of obstruction detection, reduces the false alarm rate, and eliminates the need for external sensors, thereby reducing system costs and facilitating integration into existing radar systems.

[0006] To achieve the above objectives, the present invention provides a method for detecting obstruction of security radar, comprising the following steps: Step S1: In each processing frame period, the radar system alternately transmits two waveforms in a time-division multiplexing manner: the main detection waveform W1, which is used to perform routine security detection tasks, and the health monitoring waveform W2, which is optimized for occlusion detection. Step S2: Preprocess the received signal of the health monitoring waveform W2, and then perform range-dimensional FFT processing to obtain the range-amplitude spectrum; Step S3: Extract feature parameters based on the distance-amplitude spectrum. The feature parameters include the surface reflectivity index (SRI), spatial consistency feature (SCC), and temporal evolution feature (TEF). Step S4: Fusing feature parameters based on evidence theory and making decisions: Based on the current measured values ​​of SRI, SCC, and TEF, and using a preset membership function or empirical rule, calculate the degree of support of each feature for each proposition in the recognition framework; and use the Dempster-Shafer evidence fusion rule to fuse the basic probability assignments from SRI, SCC, and TEF values ​​to obtain the combined reliability and similarity for each proposition; based on the fused reliability, output the final detection result according to the preset decision rule, including the occlusion state and occlusion type.

[0007] As a further preferred technical solution to the above technical solution, for step S3: The surface reflectivity index (SRI) is calculated as follows: define a surface region and calculate the sum of signal energy within that region, compare it with the average energy of the background noise region, take the logarithm to obtain the SRI value, and calculate the SRI value for each receiving channel separately. Spatial consistency characteristics (SCC) include SCC coefficient of variation and SCC maximum difference. The time evolution characteristics of the TEF include the short-term slope of the TEF and the long-term stability of the TEF, where: The TEF_short-term slope is a linearly fitted slope of the SRI values ​​of the most recent few frames to reflect rapid changes in occlusion. TEF_Long-term stability is the variance or standard deviation of SRI over a long time window, distinguishing between slow accumulation and sudden events.

[0008] As a further preferred technical solution to the above technical solution, step S4, calculating the degree of support of each feature for each proposition in the recognition framework, is specifically implemented as follows: Step S4.1: SRI calculation, where: ; in: ; ; in, For the surface area amplitude, The average energy of the background noise region. For the first Each receiving channel is in the distance unit The amplitude value at that point; This refers to the surface of the radome and the near-space region; This represents the background noise region used to estimate the system's background noise. , Number of receive channels; SRI's BPA formula is: ; ; in, For the preset membership function, , The reliability coefficient is dynamically adjusted based on the signal-to-noise ratio measured by SRI; a set of fundamental propositions. ; Step S4.2: SCC calculation, where: ; ; in, SCC is the coefficient of variation. For SCC_ maximum difference, Standard deviation The mean is denoted as SRI, and the mean is the SRI value for each receiving channel. The SCC's BPA formula is: ; ; ; in, β is the preset membership function, and β is the reliability coefficient of SCC evidence; Step S4.3: TEF calculation, where: The TEF_short-term slope is linearly fitted to the SRI values ​​of the most recent M frames, and the fitting formula is: ; Normalized time; TEF_Long-term stability: Calculate the standard deviation of the detrended SRI series within a window of length L. ; Where L is greater than M; The BPA formula for TEF is: ; ; in, For the preset membership function, is the reliability coefficient of TEF evidence.

[0009] As a further preferred technical solution to the above technical solution, in step S4, the Dempster-Shafer evidence fusion formula is: ; in: The degree of confidence in proposition C is represented by A and B, where C is the set of propositions supported after fusion, and A and B are set variables. A subset of the set, where m1 and m2 represent the degree of support of the two pieces of evidence to be fused. , representing the coefficient of evidence conflict; And set a threshold : like If so, it will be judged as malicious obstruction; Otherwise if If so, it is determined to be natural occlusion; Otherwise if If so, it is considered normal; Otherwise, if the reliability of any proposition is between and If the area between them is considered as suspected obstruction; Otherwise, it is judged as an unknown state.

[0010] As a further preferred technical solution to the above technical solution, the preprocessing in step S2 includes filtering, amplification, and analog-to-digital conversion.

[0011] The beneficial effects of this invention are as follows: 1. High detection accuracy and low false alarm rate: This invention uses an ultra-wideband pulse health monitoring waveform optimized for surface detection, which can accurately capture signal characteristics of the very near area of ​​the radar radome surface; it innovatively defines multi-dimensional features such as surface reflectivity index (SRI), spatial consistency feature (SCC), and temporal evolution feature (TEF). These features are extremely sensitive to obstructions on the radar radome surface, but not sensitive to distant targets that are normally active in the environment; by fusing multi-feature information through Dempster-Shafer evidence theory, it effectively suppresses false alarms caused by environmental interference such as birds and temporary approaching objects, significantly improving the accuracy and reliability of obstruction detection, while also enabling the classification and identification of obstruction types.

[0012] 2. Self-contained, low cost, and easy to integrate: This method is based entirely on the radar's own transmit and receive signals for processing, without relying on any external sensors such as cameras or lidar. This avoids the increased system complexity and cost caused by multi-sensor deployment and fusion. Furthermore, it does not require major modifications to the radar hardware and can be implemented in existing radar systems simply through software or firmware upgrades. It is easy to integrate and has a wide range of applications. Attached Figure Description

[0013] Figure 1 This is a flowchart illustrating the present invention. Detailed Implementation

[0014] The following description is intended to disclose the present invention and enable those skilled in the art to implement it. The preferred embodiments described below are merely examples, and other obvious variations will occur to those skilled in the art. The basic principles of the invention defined in the following description can be applied to other embodiments, modifications, improvements, equivalents, and other technical solutions that do not depart from the spirit and scope of the invention.

[0015] In the preferred embodiments of the present invention, those skilled in the art should note that the radar systems and the like involved in the present invention can be considered as prior art.

[0016] Preferred embodiment.

[0017] like Figure 1 As shown, this invention discloses a method for detecting obstruction in security radar, characterized by comprising the following steps: Step S1: Within each processing frame period, the radar system alternately transmits two waveforms in a time-division multiplexing manner: the main detection waveform W1, used for performing routine security detection tasks (such as target detection and tracking), and the health monitoring waveform W2, specifically optimized for occlusion detection. W1 waveform: A linear frequency modulated continuous wave is used, with parameters optimized for routine security detection tasks. Bandwidth B1 = 200MHz, chirp duration Tc = 1100µs, used for detecting and tracking moving targets at medium to long distances.

[0018] W2 waveform: Employs a wide-bandwidth linear frequency modulated continuous wave with parameters optimized for occlusion detection tasks. Bandwidth B2 = 2 GHz, chirp duration Tc2 = 10 µs, to provide extremely high range resolution, enabling clear separation of reflected signals from extremely thin attachments on the radome surface across the range spectrum.

[0019] The W1 and W2 waveforms are transmitted alternately in the time domain at a predetermined ratio (9:1) to ensure that the main detection function is dominant, while periodic health monitoring is inserted.

[0020] Step S2: Preprocess the received signal of the health monitoring waveform W2, and then perform range-dimensional FFT processing to obtain the range-amplitude spectrum; Step S3: Extract feature parameters based on the distance-amplitude spectrum. The feature parameters include the surface reflectivity index (SRI), spatial consistency feature (SCC), and temporal evolution feature (TEF). Step S4: Based on evidence theory, fuse feature parameters and make decisions: Based on the current measured values ​​of SRI, SCC, and TEF, and based on the preset membership function or empirical rules, calculate the support level (BPA) of each feature for each proposition in the recognition framework; and use the Dempster-Shafer evidence fusion rule to fuse the basic probability assignments from SRI, SCC, and TEF values ​​to obtain the combined reliability and similarity for each proposition; based on the fused reliability, output the final detection result according to the preset decision rules. The detection result includes occlusion status and occlusion type.

[0021] Specifically, for step S3: The surface reflectivity index (SRI) is calculated as follows: Define a surface region (a very narrow interval near zero) and calculate the sum of signal energy within this region. Compare this sum with the average energy of the background noise region and take the logarithm to obtain the SRI value. Calculate the SRI value for each receiving channel separately. Spatial consistency characteristics (SCC) include SCC coefficient of variation and SCC maximum difference. The time evolution characteristics of the TEF include the short-term slope of the TEF and the long-term stability of the TEF, where: The TEF_short-term slope is a linearly fitted slope of the SRI values ​​of the most recent few frames (the most recent 10-20 frames) to reflect rapid changes in occlusion. TEF_long-term stability is the variance or standard deviation of SRI over a long time window, distinguishing between slow accumulation (such as dust) and sudden events (such as mud splash).

[0022] More specifically, in step S4, calculating the degree of support of each feature for each proposition in the recognition framework is implemented as follows: Step S4.1: SRI calculation, where: ; in: ; ; in, For the surface area amplitude, The average energy of the background noise region. For the first Each receiving channel is in the distance unit The amplitude value at that point; This refers to the surface of the radome and the near-space region; This represents the background noise region used to estimate the system's background noise. , Number of receive channels; SRI's BPA formula is: ; ; in, For the preset membership function, , For reliability coefficient (0 < ≤1), dynamically adjusted based on the signal-to-noise ratio measured by SRI; basic proposition set (recognition framework). ; Step S4.2: SCC calculation, where: ; ; in, SCC is the coefficient of variation. For SCC_ maximum difference, Standard deviation The mean is denoted as SRI, and the mean is the SRI value for each receiving channel. The SCC's BPA formula is: ; ; ; in, β is the preset membership function, and β is the reliability coefficient of SCC evidence (0 < β ≤ 1). Step S4.3: TEF calculation, where: The TEF_short-term slope is linearly fitted to the SRI values ​​of the most recent M frames, and the fitting formula is: ; Normalized time; TEF_Long-term stability: Calculate the standard deviation of the detrended SRI series within a window of length L. ; Where L is greater than M; The BPA formula for TEF is: ; ; in, For the preset membership function, The reliability coefficient of TEF evidence (0 < ≤1).

[0023] Furthermore, in step S4, the Dempster-Shafer evidence fusion formula is: ; in: The degree of confidence in proposition C is represented by A and B, where C is the set of propositions supported after fusion, and A and B are set variables. A subset of , where m1 and m2 are the support levels (BPA) of the two pieces of evidence to be fused. , representing the coefficient of evidence conflict; And set a threshold : like If so, it will be judged as malicious obstruction; Otherwise if If so, it is determined to be natural occlusion; Otherwise if If so, it is considered normal; Otherwise, if the reliability of any proposition is between and If the area between them is considered as suspected obstruction; Otherwise, it is judged as an unknown state.

[0024] Furthermore, the preprocessing in step S2 includes filtering, amplification, and analog-to-digital conversion.

[0025] It is worth mentioning that the technical features such as radar systems involved in this patent application should be regarded as prior art. The specific structure, working principle, and possible control methods and spatial arrangement of these technical features can be adopted using conventional choices in the field, and should not be regarded as the inventive point of this patent. This patent will not be further elaborated in detail.

[0026] For those skilled in the art, modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this invention should be included within the protection scope of this invention.

Claims

1. A method for detecting obstruction in security radar, characterized in that, Includes the following steps: Step S1: In each processing frame period, the radar system alternately transmits two waveforms in a time-division multiplexing manner: the main detection waveform W1, which is used to perform routine security detection tasks, and the health monitoring waveform W2, which is optimized for occlusion detection. Step S2: Preprocess the received signal of the health monitoring waveform W2, and then perform range-dimensional FFT processing to obtain the range-amplitude spectrum; Step S3: Extract feature parameters based on the distance-amplitude spectrum. The feature parameters include the surface reflectivity index (SRI), spatial consistency feature (SCC), and temporal evolution feature (TEF). Step S4: Based on evidence theory, fuse feature parameters and make decisions: Based on the current measured values ​​of SRI, SCC, and TEF, and based on the preset membership function or empirical rules, calculate the degree of support of each feature for each proposition in the recognition framework; and use the Dempster-Shafer evidence fusion rule to fuse the basic probability assignments from SRI, SCC, and TEF values ​​to obtain the combined reliability and similarity of each proposition after fusion. Based on the reliability of the fusion, the final detection result is output according to the preset decision rules. The detection result includes the occlusion status and occlusion type.

2. The method for detecting obstruction of security radar according to claim 1, characterized in that, For step S3: The surface reflectivity index (SRI) is calculated as follows: define a surface region and calculate the sum of signal energy within that region, compare it with the average energy of the background noise region, take the logarithm to obtain the SRI value, and calculate the SRI value for each receiving channel separately. Spatial consistency characteristics (SCC) include SCC coefficient of variation and SCC maximum difference. The time evolution characteristics of the TEF include the short-term slope of the TEF and the long-term stability of the TEF, where: The TEF_short-term slope is a linearly fitted slope of the SRI values ​​of the most recent few frames to reflect rapid changes in occlusion. TEF_Long-term stability is the variance or standard deviation of SRI over a long time window, distinguishing between slow accumulation and sudden events.

3. The method for detecting obstruction of security radar according to claim 2, characterized in that, In step S4, calculating the degree of support of each feature for each proposition in the recognition framework is specifically implemented as follows: Step S4.1: SRI calculation, where: ； in: ； ； in, For the surface area amplitude, The average energy of the background noise region. For the first Each receiving channel is in the distance unit The amplitude value at that point; This refers to the surface of the radome and the near-space region; This represents the background noise region used to estimate the system's background noise. , Number of receive channels; SRI's BPA formula is: ； ； in, For the preset membership function, , The reliability coefficient is dynamically adjusted based on the signal-to-noise ratio measured by SRI; a set of fundamental propositions. ; Step S4.2: SCC calculation, where: ； ； in, SCC is the coefficient of variation. For SCC_ maximum difference, Standard deviation The mean is denoted as SRI, and the mean is the SRI value for each receiving channel. The SCC's BPA formula is: ； ； ； in, β is the preset membership function, and β is the reliability coefficient of SCC evidence; Step S4.3: TEF calculation, where: The TEF_short-term slope is linearly fitted to the SRI values ​​of the most recent M frames, and the fitting formula is: ; Normalized time; TEF_Long-term stability: Calculate the standard deviation of the detrended SRI series within a window of length L. ； Where L is greater than M; The BPA formula for TEF is: ； ； in, For the preset membership function, is the reliability coefficient of TEF evidence.

4. The method for detecting obstruction of a security radar according to claim 3, characterized in that, In step S4, the Dempster-Shafer evidence fusion formula is: ； in: The degree of confidence in proposition C is represented by A and B, where C is the set of propositions supported after fusion, and A and B are set variables. A subset of the set, where m1 and m2 represent the degree of support of the two pieces of evidence to be fused. , representing the coefficient of evidence conflict; And set a threshold : like If so, it will be judged as malicious obstruction; Otherwise if If so, it is determined to be natural occlusion; Otherwise if If so, it is considered normal; Otherwise, if the reliability of any proposition is between and If the area between them is considered as suspected obstruction; Otherwise, it is judged as an unknown state.

5. The method for detecting obstruction of a security radar according to claim 1, characterized in that, The preprocessing in step S2 includes filtering, amplification, and analog-to-digital conversion.

Images