CN-122017782-A - Shielding detection method of security radar

CN122017782ACN 122017782 ACN122017782 ACN 122017782ACN-122017782-A

Abstract

The invention discloses a shielding detection method of a security radar, which comprises the following steps of S1, alternately transmitting two waveforms in a time division multiplexing mode in each processing frame period by a radar system, namely a main detection waveform W1 for executing a conventional security detection task and a health monitoring waveform W2 specially designed for shielding detection, S2, preprocessing a received signal of the health monitoring waveform W2, then carrying out distance dimension FFT processing to obtain a distance-amplitude spectrum, S3, extracting characteristic parameters based on the distance-amplitude spectrum, and S4, fusing the characteristic parameters and making a decision based on an evidence theory. The shielding detection method for the security radar disclosed by the invention realizes high-reliability and sortable shielding detection, improves the accuracy of shielding detection, reduces the false alarm rate, does not need to depend on an external sensor, reduces the system cost, and is convenient to integrate in the existing radar system.

Inventors

LI JIAN
LU YUMIN
ZHU XINEN

Assignees

浙江芯力微电子股份有限公司

Dates

Publication Date: 20260512
Application Date: 20260211

Claims (5)

1. The shielding detection method of the security radar is characterized by comprising the following steps of: step S1, the radar system alternately transmits two waveforms in a time division multiplexing mode in each processing frame period, wherein the waveforms are a main detection waveform W1 for executing a conventional security detection task and a health monitoring waveform W2 specially designed for shielding detection optimization; Step S2, preprocessing a received signal of the health monitoring waveform W2, and then performing distance dimension FFT processing to obtain a distance-amplitude spectrum; step S3, extracting characteristic parameters based on a distance-amplitude spectrum, wherein the characteristic parameters comprise a surface reflectivity index SRI, a spatial consistency characteristic SCC and a time evolution characteristic TEF; The method comprises the steps of S4, fusing characteristic parameters based on an evidence theory, making a decision, namely calculating the support degree of each characteristic to each proposition in an identification framework based on a preset membership function or experience rule according to the current measured values of an SRI value, an SCC value and a TEF value, fusing basic probability distribution from the SRI value, the SCC value and the TEF value by utilizing a Dempster-Shafer evidence fusion rule to obtain comprehensive credibility and plausibility of each proposition after combination, and outputting a final detection result including a shielding state and a shielding type according to the fused credibility and a preset decision rule.
2. The shielding detection method of a security radar according to claim 1, wherein for step S3: Defining a surface area, calculating the sum of signal energy in the area, comparing the sum with the average energy of a background noise area, taking logarithm to obtain SRI values, and respectively calculating the SRI value of each receiving channel; the spatial consistency feature SCC comprises an SCC_variation coefficient and an SCC_maximum difference; The time evolution feature TEF includes TEF short-term slope and TEF long-term stability, wherein: the TEF short-term slope is a linear fit slope of the SRI values of the last several 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 slow accumulation from sudden events.
3. The occlusion detection method of claim 2, wherein in step S4, calculating the support degree of each feature to each proposition in the recognition frame is implemented as follows: Step S4.1, SRI calculation, wherein: ; Wherein: ; ; Wherein, the For the amplitude of the surface area, For the average energy of the background noise region, Is the first The receiving channels are in distance units Amplitude values at; representing the radome surface and the very near space region; Representing a background noise region for estimating a system noise floor; , for the number of receive channels; the BPA formula for SRI is: ; ; Wherein, the Is a preset membership function, and the method is characterized in that, , Dynamic adjustment of signal-to-noise ratio according to SRI measurement for reliability coefficient ; Step S4.2, SCC calculation, wherein: ; ; Wherein, the For the SCC _ coefficient of variation, For the SCC _ maximum difference, Is the standard deviation of the two-dimensional image, The average value is the SRI value of each receiving channel; the BPA formula for SCC is: ; ; ; Wherein, the Beta is the reliability coefficient of the SCC evidence, which is a preset membership function; step S4.3, TEF calculation, wherein: the TEF_short term slope is linearly fitted to the SRI value of the nearest M frames, and the fitting formula is: ; Is normalized time; tef_long term stability within a window of length L, standard deviation after SRI sequence detrending was calculated: ; wherein L is greater than M; the BPA formula for TEF is: ; ; Wherein, the Is a preset membership function, and the method is characterized in that, Is the reliability coefficient of the TEF evidence.
4. The method for detecting occlusion of a security radar according to claim 3, wherein in step S4, a Dempster-Shafer evidence fusion formula is: ; Wherein: To trust the proposition C, C is the proposition set supported after fusion, A, B is the set variable M1, m2 are the degrees of support of the two pieces of evidence to be fused, Representing evidence conflict coefficients; And set a threshold value : If it is Judging that the vehicle is ill-intentioned and shielded; Otherwise if not Judging that the shielding is natural; Otherwise if not Judging that the vehicle is normal; otherwise, if the credibility of any proposition is between And (3) with If yes, judging that the vehicle is suspected to be blocked; otherwise, determining the unknown state.
5. The method for detecting the occlusion of a security radar according to claim 1, wherein the preprocessing in the step S2 includes filtering, amplifying and analog-to-digital conversion.

Description

Shielding detection method of security radar Technical Field The invention belongs to the technical field of security radar, and particularly relates to a shielding detection method of security radar. Background In the security monitoring field, the millimeter wave radar is widely applied to various security scenes by virtue of all-weather working capacity and non-visual detection advantages. However, the surface of the radome is susceptible to natural factors such as dust accumulation, rain and snow coverage, frost condensation, spider web adhesion and the like, and meanwhile, artificial malicious shielding can be possibly encountered, and the conditions can lead to radar detection performance reduction and even complete failure, form security monitoring blind areas and bring great hidden danger to security work. At present, the following deficiencies mainly exist in the prior art scheme for radar shielding detection: First, based on the detection scheme of a single signal intensity threshold, the detection scheme is used for monitoring the overall intensity of a radar received signal or the amplitude of a short-distance reflection signal, and judging that the radar received signal is shielded when the detection scheme exceeds a fixed threshold. The scheme has obvious defects that on one hand, the false alarm rate is high, normal environment interference such as short stay of a bird in front of a radar triggers false alarm, on the other hand, the adaptability is poor, the static threshold cannot cope with environment change, and the problem of missing report or continuous false alarm is easily caused. And secondly, based on a scheme of multi-sensor data comparison, external sensors such as cameras and laser correlation are compared with radar data, and when other sensors detect targets and the radar does not detect targets, radar abnormality is suspected. The scheme needs to deploy and fuse various heterogeneous sensors, not only increases the complexity and hardware cost of the system, but also can not effectively identify slight shielding or local shielding conditions, and has larger detection limitation. Disclosure of Invention The invention mainly aims to provide a shielding detection method of a security radar, which realizes high-reliability and sortable shielding detection, improves the accuracy of shielding detection, reduces the false alarm rate, does not need to depend on an external sensor, reduces the system cost and is convenient to integrate in the existing radar system. In order to achieve the above purpose, the invention provides a shielding detection method of a security radar, comprising the following steps: step S1, the radar system alternately transmits two waveforms in a time division multiplexing mode in each processing frame period, wherein the waveforms are a main detection waveform W1 for executing a conventional security detection task and a health monitoring waveform W2 specially designed for shielding detection optimization; Step S2, preprocessing a received signal of the health monitoring waveform W2, and then performing distance dimension FFT processing to obtain a distance-amplitude spectrum; step S3, extracting characteristic parameters based on a distance-amplitude spectrum, wherein the characteristic parameters comprise a surface reflectivity index SRI, a spatial consistency characteristic SCC and a time evolution characteristic TEF; The method comprises the steps of S4, fusing characteristic parameters based on an evidence theory, making a decision, namely calculating the support degree of each characteristic to each proposition in an identification framework based on a preset membership function or experience rule according to the current measured values of an SRI value, an SCC value and a TEF value, fusing basic probability distribution from the SRI value, the SCC value and the TEF value by utilizing a Dempster-Shafer evidence fusion rule to obtain comprehensive credibility and plausibility of each proposition after combination, and outputting a final detection result including a shielding state and a shielding type according to the fused credibility and a preset decision rule. As a further preferable embodiment of the above-described embodiment, for step S3: Defining a surface area, calculating the sum of signal energy in the area, comparing the sum with the average energy of a background noise area, taking logarithm to obtain SRI values, and respectively calculating the SRI value of each receiving channel; the spatial consistency feature SCC comprises an SCC_variation coefficient and an SCC_maximum difference; The time evolution feature TEF includes TEF short-term slope and TEF long-term stability, wherein: the TEF short-term slope is a linear fit slope of the SRI values of the last several 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 slow accumulation from su