A dual window radar pulse detection method and system

Abstract

The application discloses a kind of double window radar pulse detection method and system, mainly solve the prior art in complex electromagnetic environment radar pulse signal is susceptible to non-stationary noise and multi-source interference, leading to unstable detection and inaccurate pulse start-stop position positioning problem.The implementation scheme includes: the energy sequence is constructed to the energy transformation and standardization processing of received radar signal;Forward and backward window energy ratio is calculated in short window, combined with energy change rate and statistical characteristics to construct joint decision function, obtain pulse candidate interval;According to the energy variation characteristics, the length of long window is adaptively adjusted, and the consistency analysis, continuity and stability verification of candidate interval are carried out under the condition of long window, to realize interference rejection;Finally, the short window detection result and long window verification result are fused, to determine the pulse start and end position, the present application can effectively improve weak pulse detection sensitivity, enhance complex background anti-interference ability, improve pulse boundary positioning precision, improve detection stability, can be used for radar signal processing and target identification.

Description

Technical Field

[0001] This invention belongs to the field of radar and communication signal processing technology, specifically relating to a dual-window radar pulse detection method and system, which can be used for radar signal preprocessing and effective pulse extraction under non-stationary noise and multi-source interference conditions. Background Technology

[0002] With the continuous development of radar technology and electronic reconnaissance technology, the types of radiation sources in the electromagnetic space are becoming increasingly diverse, and the signal density is significantly increasing. Different signal systems exhibit high complexity in terms of frequency domain distribution, time domain structure, and modulation characteristics. In practical applications, radar echoes or reconnaissance received signals usually exist in the form of burst pulses, characterized by short pulse duration, unstable repetition intervals, and discrete signal energy distribution. Furthermore, they are often superimposed on non-stationary noise, ground clutter, and various man-made interferences.

[0003] In such complex electromagnetic environments, achieving reliable detection of radar pulse signals, especially accurate positioning of the pulse's start and end points, has become a critical issue in the signal preprocessing stage. The pulse detection results directly affect the accuracy of subsequent parameter estimation, radiation source identification, feature extraction, and signal sorting, and are also a crucial foundation for electromagnetic data construction and intelligent analysis model training.

[0004] Patent application CN202410249293.0 discloses a radar signal detection method based on an improved detection strategy. This method constructs a sliding window to perform local statistical analysis on the received signal and combines this with an adaptive threshold decision mechanism to detect the target signal. This method improves the environmental adaptability of the detection to some extent by estimating the local noise background and dynamically adjusting the detection threshold. However, this method is essentially still a detection framework based on statistical modeling, and its detection performance depends on the stability of the noise statistical characteristics. When the electromagnetic environment exhibits non-stationary characteristics or sudden interference exists, the local statistical characteristics are prone to shift, leading to inaccurate threshold estimation and thus increasing the false alarm rate or causing missed detections. Furthermore, this method mainly relies on a single statistical feature for decision-making, lacking a deep characterization of the internal structural features of the signal, making it difficult to effectively distinguish between pulse signals and complex interference.

[0005] Patent application CN201910589144.8 discloses a radar pulse detection method based on sliding window energy detection. This method divides the signal into segments using a sliding window, calculates local energy changes, and detects the pulse signal based on energy abrupt changes. However, this method typically uses a fixed window length, lacking the ability to adaptively adjust to pulse signals of different scales. When the window length is small, the detection results are easily affected by random noise disturbances; when the window length is large, it may reduce the accuracy of pulse boundary positioning or even cause short pulse signals to be submerged. Furthermore, this method mainly relies on a single energy feature for decision-making, lacking multi-feature joint analysis and candidate interval verification mechanisms. Its anti-interference capability is limited in complex electromagnetic environments, and its detection stability needs improvement. Summary of the Invention

[0006] The purpose of this invention is to address the shortcomings of existing technologies by proposing a dual-window radar pulse detection method to improve the accuracy and stability of radar pulse detection in non-stationary and complex electromagnetic environments.

[0007] The technical approach to achieving the objective of this invention is as follows: By constructing an adaptive pulse detection mechanism based on a dual-window energy ratio, the energy of short-window and long-window signals is jointly modeled, and the energy ratio is used to characterize the relative energy evolution of the signal, thereby achieving stable detection of the effective pulse interval and reducing dependence on noise statistical characteristics; by constructing a multi-scale adaptive adjustment and joint decision mechanism, the window scale is dynamically adjusted according to local energy changes, and features such as energy change trends and local stability are integrated. At the same time, the detection results are constrained and screened by combining candidate interval consistency verification and interference discrimination processes, thereby achieving accurate positioning of pulse intervals and interference suppression.

[0008] Based on the above ideas, the technical solution of the present invention includes:

[0009] 1. A dual-window radar pulse detection method, characterized in that it comprises:

[0010] (1) Collect radar signals acquired by radar receiving equipment, perform energy transformation processing on the radar signals, construct an energy sequence characterizing the change of signal power over time, and perform standardization processing on it;

[0011] (2) Perform short window sliding on the energy sequence, calculate the forward window energy and backward window energy at the current time respectively, construct the short window energy ratio sequence, and calculate the energy change rate and statistical characteristics within the local window range under the short window scale;

[0012] (3) Based on the energy ratio, energy change rate and statistical characteristics, establish a short window joint decision function for cross-scale energy evolution consistency to detect the location of signal energy abrupt change, and compare the joint decision function with the preset decision threshold to obtain the short window detection candidate interval;

[0013] (4) Calculate the rate of change and statistical variance of the energy sequence within the local window range under the long window scale, and adaptively adjust the length of the long window so that the length of the long window changes dynamically with the signal characteristics;

[0014] (5) Under the adaptive long window condition, the energy sequence is reprocessed with a sliding window, and the forward window energy and backward window energy are calculated respectively to form a long window energy ratio sequence;

[0015] (6) Based on the long window energy ratio sequence, the corresponding energy change rate and statistical variance, establish a long window joint decision function to make a decision on the long window, and perform consistency verification and interference discrimination on the short window detection pulse candidate interval according to the decision result to obtain the interval that the long window verification passes;

[0016] (7) Perform an intersection operation on the two intervals of the short window and the long window, and retain only the pulse interval that simultaneously satisfies the short window detection condition and the long window consistency verification condition to obtain the start position and end position of the final pulse signal.

[0017] Furthermore, in step (6), a decision is made on the long window based on the decision function, and the short window is determined based on the decision result.

[0018] The window detection pulse candidate interval is used for consistency verification and interference discrimination, including:

[0019] 6a) The output of the long window joint decision function With preset decision threshold Compare and obtain the decision sequence :in, This indicates that the current state is considered a valid signal. This indicates that the condition is either background noise or non-pulse interference.

[0020] 6b) Based on the start and end times of the candidate interval, extract the decision result subsequence within the corresponding time range from the decision sequence, calculate the ratio of the number of valid state sampling points in the candidate interval to the total number of sampling points, and compare this ratio with a preset ratio threshold: if the percentage of valid states is greater than the preset ratio threshold, mark the candidate interval as a consistent candidate interval and execute 6c); otherwise, remove the candidate interval.

[0021] 6c) Within the consensus candidate interval, extract those that satisfy... The continuous sampling point sequence constitutes a continuous state interval, and its corresponding continuous duration is compared with a preset duration threshold: if the continuous duration is greater than the preset duration threshold, the interval is marked as a continuous candidate interval and 6d is executed; otherwise, the candidate interval is determined to be an invalid pulse interval and is removed.

[0022] 6d) Within the continuous candidate interval, calculate the degree of deviation of the energy sequence corresponding to the candidate interval from the local mean, and compare it with a preset stability threshold: if the degree of deviation is less than the preset stability threshold, the interval is used as the long window verification pass interval; otherwise, the candidate interval is determined to be an invalid pulse interval and is removed.

[0023] 2. A dual-window radar pulse detection system, characterized in that it comprises:

[0024] The signal acquisition and preprocessing module is used to acquire radar signals and perform energy transformation and standardization to output a standardized energy sequence.

[0025] The short window feature extraction module is used to slide the energy sequence through a short window, calculate the forward and backward window energies, construct a short energy ratio sequence, and extract the rate of change and statistical features.

[0026] The short window decision module is used to construct a decision function based on the short window energy ratio sequence, energy change rate and statistical characteristics, and compare its output with a preset threshold to obtain the short window candidate interval.

[0027] The long window parameter adaptive module is used to adaptively adjust the length of the long window based on the energy change rate and statistical characteristics;

[0028] The long window feature extraction module is used to slide the energy sequence through a long window, calculate the forward and backward window energies, and construct a long window energy ratio sequence.

[0029] The long window decision and verification module is used to construct the decision function and make decisions, while performing consistency verification and interference discrimination on the short window candidate intervals to obtain the long window verification passed intervals;

[0030] The fusion decision module is used to perform intersection calculations on the short window candidate interval and the long window verification pass interval to obtain the pulse start and end positions.

[0031] Compared with the prior art, the present invention has the following advantages:

[0032] Firstly, by performing energy conversion and standardization processing on the radar received signal, the present invention unifies the form of signal power expression, thereby reducing the dependence on the stability of noise statistical characteristics and maintaining good detection stability and robustness even in non-stationary electromagnetic environments.

[0033] Secondly, this invention constructs the energy ratio of the forward window and the backward window to characterize the relative evolution of signal energy, rather than relying on absolute energy or noise statistical modeling. This effectively avoids the performance degradation problem of traditional threshold estimation methods when noise distribution changes, and improves the adaptability to complex backgrounds.

[0034] Third, this invention integrates energy ratio, energy change rate and statistical characteristics at a short window scale to establish a joint decision mechanism for cross-scale energy evolution consistency, thereby realizing the comprehensive detection of multiple features at signal abrupt change locations. Compared with existing methods that rely on only a single energy feature, it can more accurately distinguish between real pulses and random noise disturbances.

[0035] Fourth, by introducing statistical analysis and adaptive adjustment mechanism for window length under long window scale, this invention enables the window scale to change dynamically with signal characteristics, thereby ensuring the accuracy of pulse boundary positioning while taking into account noise resistance, and solving the problem that it is difficult to balance detection accuracy and stability under different pulse scales with a fixed window length.

[0036] Fifth, by constructing a joint decision function under a long window scale and performing consistency verification and interference discrimination on the candidate intervals obtained by short window detection, this invention achieves constraint and screening of detection results, effectively suppresses false alarms and false detections, and significantly improves anti-interference capability in complex electromagnetic environments.

[0037] Sixth, by intersecting and fusing the short-window detection results and the long-window verification results, this invention retains only the pulse intervals that simultaneously meet the multi-scale decision conditions, thereby achieving collaborative constraints between detection and verification, improving the accuracy of pulse start and end position positioning and the reliability of the overall detection results. Attached Figure Description

[0038] Figure 1 This is a flowchart illustrating the implementation of the dual-window radar pulse detection method of the present invention;

[0039] Figure 2 This is a block diagram of the dual-window radar pulse detection system of the present invention. Detailed Implementation

[0040] The embodiments of the present invention will be described in detail below with reference to the accompanying drawings:

[0041] Example 1: Dual-window radar pulse detection method.

[0042] This embodiment focuses on processing burst radar signals received in complex electromagnetic environments, where the signals are superimposed with non-stationary noise and multi-source interference. By constructing an energy ratio-driven multi-feature joint decision mechanism, combined with adaptive window adjustment and cross-scale consistency verification, stable detection and precise positioning of pulse signals are achieved. (Refer to...) Figure 1 The implementation steps of this example include the following:

[0043] Step 1: Radar signal preprocessing and energy sequence construction.

[0044] 1.1) The radiation source emits space electromagnetic signals, and the radar receiving equipment collects the space electromagnetic signals to obtain the raw time-domain signal containing target echoes, noise, and interference components. :

[0045] ,

[0046] in, Indicates the target echo signal. Indicates environmental noise. Indicates interference signal;

[0047] 1.2) For different types of radar signals, corresponding energy calculation methods are used to convert the original time-domain signal... Converted into an energy sequence characterizing the change in signal power :

[0048] If the radar signal is a real-valued signal, then by analyzing the original time-domain signal... Performing a squaring operation yields the corresponding energy value sequence. :

[0049] ;

[0050] If the radar signal is an IQ signal, then by analyzing the in-phase components at the same time... Orthogonal components The signal energy is calculated by squaring the signals separately and then summing the results. The energy sequence is represented as follows:

[0051] ;

[0052] 1.3) Regarding the energy sequence Standardization is performed to reduce the impact of different signal amplitude ranges on the detection results and improve the stability of subsequent pulse detection processes. The formula is as follows:

[0053] ,

[0054] in, This represents the standardized energy sequence. Represents the original energy sequence. This represents the mean. It represents the standard deviation.

[0055] Step 2: Short window energy ratio and statistical feature extraction.

[0056] After completing the energy calculation and normalization of the radar signal, in order to analyze energy changes based on a local time scale and detect the location of energy abrupt changes, a short-window sliding process needs to be introduced on the energy sequence for local feature extraction and ratio calculation. The implementation includes the following:

[0057] 2.1) Set the short window length The energy sequence is processed using a sliding window method, and all energy values ​​within the window are accumulated to obtain the forward window energy. :

[0058] ,

[0059] in, This represents the standardized energy sequence. For sequence indices, the value range is from arrive ;

[0060] 2.2) At the current moment, the energy values ​​within the window are accumulated to obtain the backward window energy. :

[0061] ;

[0062] 2.3) Based on the aforementioned forward window energy With backward window energy The energy ratio sequence was calculated. :

[0063] ,

[0064] in, To prevent constants with a denominator of zero;

[0065] 2.4) At the current moment, for the energy sequence Perform differential operations to obtain the rate of energy change. This is used to characterize the instantaneous change characteristics of signal energy, and its formula is:

[0066] ,

[0067] 2.5) Let the length of the local sliding window under the short window scale be... And calculate the local mean of the energy sequence within this local window. The formula for representing the average energy level during that time period is:

[0068] ,

[0069] in, This represents the standardized energy sequence; For sequence indices, their value ranges from arrive ;

[0070] 2.6) Within the local window range, the deviation of the energy sequence is statistically calculated based on the local mean to obtain the local variance. It is used to characterize the intensity of energy fluctuations, and its formula is:

[0071] .

[0072] Step 3: Construct the short-window joint decision function to determine the pulse candidate interval;

[0073] After extracting the energy ratio and statistical features under a short window scale, a joint decision function needs to be constructed to determine the location of abrupt changes in signal energy, and the implementation includes the following:

[0074] 3.1) Construct a joint short-window decision function based on energy ratio, energy change rate, and local statistical characteristics:

[0075] ,

[0076] in, This represents a comprehensive judgment value regarding the trend and stability of signal energy changes at the current moment within a short window scale. Represents the short-window energy ratio sequence. Represents the rate of change of energy. Represents local variance. , , These are three weighting coefficients with different values.

[0077] 3.2) Adaptively set the rising and falling thresholds based on the mean and standard deviation of the joint decision function, and then set the joint decision function... The pulse start and end positions are determined by comparing the pulse with the decision threshold.

[0078] When the value of the joint decision function When the value exceeds the rising threshold, that moment is determined to be the pulse start position;

[0079] When the value of the joint decision function When the value is less than the falling threshold, that moment is determined to be the end of the pulse.

[0080] 3.3) Based on the time corresponding to the pulse start position and the pulse end position, the pulse candidate interval is obtained.

[0081] Step 4: Long window adaptive adjustment.

[0082] After completing the joint decision and pulse candidate interval determination under the short window scale, in order to adapt to the changing characteristics of different signal segments and improve the stability and sensitivity of detection, it is necessary to dynamically adjust the window length over a long time scale. The implementation includes the following:

[0083] 4.1) At the current moment, for the energy sequence Perform differential operations to obtain the rate of energy change. This is used to characterize the instantaneous change characteristics of signal energy, and its formula is:

[0084] ;

[0085] 4.2) Let the length of the local sliding window under the long window scale be... And calculate the local mean of the energy sequence within this local window. The formula for representing the average energy level during that time period is:

[0086] ,

[0087] in, This represents the standardized energy sequence. For sequence indices, the value range is from arrive ;

[0088] 4.3) Within the local window range, the deviation of the energy sequence is statistically calculated based on the local mean to obtain the local variance. It is used to characterize the intensity of energy fluctuations, and its formula is:

[0089] ,

[0090] in, This represents the standardized energy sequence. For sequence indices, the value range is from arrive ;

[0091] 4.4) Based on the mean and variance of the energy sequence, adaptively set the energy change rate threshold and local variance threshold, and then set the change rate... and local variance For long window length Adaptive adjustment:

[0092] When the rate of energy change is greater than the energy rate of change threshold or the local variance is greater than the local variance threshold, the length of the long window is reduced to improve the response sensitivity to sudden signals.

[0093] When the rate of energy change is less than the energy rate of change threshold and the local variance is less than the local variance threshold, the length of the long window is increased to enhance the statistical stability of the stationary interval, thereby achieving an adaptive balance between detection sensitivity and stability.

[0094] Step 5: Long window feature extraction.

[0095] After completing the adaptive adjustment of the long window length, in order to further characterize the energy change characteristics of the signal over a long time scale and provide feature support for subsequent consistency verification, it is necessary to perform sliding window processing on the energy sequence under the adaptive long window condition to extract long window features. The implementation includes the following:

[0096] 5.1) Set the length of the long window The energy sequence is processed using a sliding window method with this long window at the current time. Calculate the forward window energy :

[0097] ,

[0098] in, This represents the standardized energy sequence. For sequence indices, the value range is from arrive ;

[0099] 5.2) Calculate the backward window energy at the current time. :

[0100] ,

[0101] 5.3) Based on the forward window energy and the backward window energy, calculate the long window energy ratio sequence. :

[0102] ,

[0103] in, To prevent constants with a denominator of zero.

[0104] Step 6: Consistency verification and interference suppression based on long window.

[0105] After extracting the long window features, in order to constrain the stability and suppress interference of the candidate intervals obtained by the short window detection, it is necessary to construct a decision function based on the long window features, and to screen and verify the candidate intervals through a multi-level decision mechanism. The implementation includes the following:

[0106] 6.1) Construct a long-window joint decision function based on energy ratio, energy change rate, and local statistical characteristics:

[0107] ,

[0108] in, This represents a comprehensive judgment value regarding the trend and stability of signal energy changes at the current moment within a long window scale. Represents a long-window energy ratio sequence. Represents the rate of change of energy. This represents the local variance of the long window. , , These are the weighting coefficients;

[0109] 6.2) The output of the long window joint decision function With preset decision threshold Compare and obtain the decision sequence :

[0110] ,

[0111] in, This indicates that the current state is considered a valid signal. This indicates that the condition is either background noise or non-pulse interference.

[0112] 6.3) Based on the start and end times of the candidate interval, extract the decision result subsequence within the corresponding time range from the decision sequence, calculate the ratio of the number of valid state sampling points in the candidate interval to the total number of sampling points, and obtain the percentage of valid states. :

[0113] ,

[0114] in, Indicates the start time of the candidate interval. Indicates the end time of the candidate interval;

[0115] 6.4) Compare this percentage with a preset percentage threshold:

[0116] If the percentage of valid states If the ratio exceeds a preset threshold, the candidate interval is marked as a consistency candidate interval, and step 6.5 is executed.

[0117] Otherwise, the candidate interval is removed;

[0118] 6.5) Within the consistency candidate interval, extract those that satisfy... A continuous sequence of sampling points is used to form a continuous state interval, and the corresponding continuous duration is compared with a preset duration threshold:

[0119] If the continuous duration exceeds a preset duration threshold, the interval is marked as a continuous candidate interval, and step 6.6 is executed.

[0120] Otherwise, the candidate interval is determined to be an invalid pulse interval and is discarded;

[0121] 6.6) Within a continuous candidate interval, calculate the normalized energy sequence corresponding to the candidate interval relative to the local mean. degree of deviation :

[0122] ,

[0123] , ,

[0124] in, This represents the normalized energy sequence corresponding to the continuous candidate interval. and These represent the start and end times of the continuous candidate interval, respectively.

[0125] 6.7) Degree of deviation Compare with a preset stability threshold:

[0126] If the deviation is less than the preset stability threshold, this interval will be used as the long-window verification pass interval.

[0127] Otherwise, the candidate interval is determined to be an invalid pulse interval and is discarded.

[0128] Step 7: Detection results fusion and output.

[0129] After completing the long-window consistency verification and candidate interval screening, to further improve the reliability of the detection results, the short-window detection results and the long-window verification results are subjected to consistency fusion processing based on time overlap to determine the final effective pulse interval. The implementation includes the following:

[0130] 7.1) The pulse candidate intervals obtained by short window detection are compared with the intervals that have passed long window verification one by one in time alignment to obtain the intervals that overlap in time;

[0131] 7.2) The interval is taken as the effective pulse interval, and its start time and end time are taken as the start position and end position of the final radar pulse signal, respectively, to complete the dual-window radar pulse detection.

[0132] It should be noted that the step numbers in the above examples and the reference numerals in the claims are only for the purpose of clearly and completely describing the embodiments of the present invention and for ease of understanding, and their order is not limited.

[0133] Example 2: Dual-window radar pulse detection system.

[0134] Reference Figure 2 This example includes: a signal acquisition and preprocessing module 1, a short window feature extraction module 2, a short window decision module 3, a long window parameter adaptation module 4, a long window feature extraction module 5, a long window decision and verification module 6, and a fusion decision module 7. The long window decision and verification module 6 includes: a decision function construction submodule 61, a decision sequence generation submodule 62, a candidate interval matching submodule 63, a consistency analysis submodule 64, a continuity verification submodule 65, and a stability discrimination submodule 66.

[0135] The working principle of the entire system is as follows:

[0136] The signal acquisition and preprocessing module 1 is used to acquire the original radar signal collected by the radar receiving device, perform energy transformation processing on it, construct an energy sequence characterizing the signal power change over time, and perform standardization processing on the energy sequence to obtain a uniform scale energy sequence, and output it to the short window feature extraction module 2, the long window parameter adaptive module 4, and the long window feature extraction module 5 respectively.

[0137] The short window feature extraction module 2 is used to perform short window sliding processing on the standardized energy sequence output by the signal acquisition and preprocessing module 1, calculate the forward window energy and backward window energy at the current time, construct the short window energy ratio sequence, and calculate statistical features such as energy change rate, local mean and local variance within the local range of the short window and output them to the short window decision module 3.

[0138] The short window decision module 3 is used to construct a joint decision function for cross-scale energy evolution consistency based on the short window energy ratio, energy change rate and local statistical features output by the short window feature extraction module 2. By comparing the decision function with a preset threshold, the candidate interval for short window detection pulse is obtained, and the candidate interval is output to the long window decision and verification module 6 and the fusion decision module 7 respectively.

[0139] The long window parameter adaptive module 4 is used to calculate the energy change rate and statistical variance of the standardized energy sequence output by the signal acquisition and preprocessing module 1, and adaptively adjust the long window length according to the energy change rate and statistical variance, so that the long window length changes dynamically with the signal characteristics, and outputs the adaptive long window parameters to the long window feature extraction module 5.

[0140] The long window feature extraction module 5 is used to perform long window sliding processing on the standardized energy sequence output by the signal acquisition and preprocessing module 1 according to the long window parameters output by the long window parameter adaptive module 4. At the current time, the forward window energy and backward window energy are calculated respectively to construct the long window energy ratio sequence. The energy change rate and local statistical variance are calculated within the long window range and output to the long window decision and verification module 6.

[0141] The long-window decision and verification module 6 is used to perform consistency verification and interference suppression processing on the candidate interval based on the long-window energy ratio, energy change rate, and local statistical variance output by the long-window feature extraction module 5, and the short-window candidate interval input by the short-window decision module 3, wherein:

[0142] The decision function construction submodule 61 is used to construct a long-window joint decision function based on the long-window energy ratio, energy change rate and local variance, and input the decision result into the decision sequence generation submodule 62.

[0143] The decision sequence generation submodule 62 is used to compare the long window joint decision function with a preset threshold, generate a long window decision sequence, and input it into the candidate interval matching submodule 63.

[0144] The candidate interval matching submodule 63 is used to extract the corresponding time range of the decision subsequence from the long window decision sequence according to the start and end times of the short window candidate interval, and input the decision subsequence to the consistency analysis submodule 64;

[0145] The consistency analysis submodule 64 is used to calculate the proportion of valid states within the candidate interval and compare it with a preset proportion threshold to filter out the consistent candidate intervals and output them to the continuity verification submodule 65.

[0146] The continuity verification submodule 65 is used to extract a continuous valid state sequence within the consistency candidate interval, and compare the continuous duration with a preset duration threshold to filter out the continuous candidate interval, and output it to the stability discrimination submodule 66.

[0147] The stability discrimination submodule 66 is used to calculate the degree of deviation of the energy sequence from the local mean within the continuous candidate interval, and compare it with the preset stability threshold to obtain the long window verification pass interval output to the fusion decision module 7;

[0148] The fusion decision module 7 is used to perform time domain intersection operation on the short window candidate interval output by the short window decision module 3 and the long window verification pass interval output by the long window decision and verification module 6, and retain only the interval that simultaneously satisfies the short window detection condition and the long window consistency verification condition to obtain the start position and end position of the final pulse signal.

[0149] It should be noted that the above functional modules can be implemented, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, they can be implemented, in whole or in part, as program instruction products. A program instruction product includes one or a set of program instructions. When the program instructions are loaded and executed on a computer, the described process or function is generated, in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The program instructions can be stored in a computer-readable and writable storage medium, or transferred from one computer's readable and writable storage medium to another.

[0150] In this embodiment, the direct coupling or communication connection between the modules can be achieved through indirect coupling or communication connection via interfaces, devices, or modules. The functional modules and sub-modules in this embodiment can dynamically reside within a single processing unit, or each module can exist physically independently, or two or more modules can dynamically reside within a single processing unit. When these dynamic components are implemented as software functional modules and sold or used as independent products, they can also be stored in a computer-readable and writable storage medium. This storage medium can be a memory, disk, or optical disc, etc.

[0151] The above description is merely two specific embodiments of the present invention and does not constitute any limitation on the present invention. Obviously, those skilled in the art, having understood the idea and spirit of the present invention, can make various modifications and substitutions. For example, the substituted modifications include, but are not limited to, the following:

[0152] In addition to the energy calculation method based on the square of amplitude or the sum of squares of IQ components and the standardization processing method adopted in this embodiment, the radar signal preprocessing and energy sequence construction process can also use envelope detection, power spectrum estimation or other energy characterization methods for energy construction, and can be implemented using minimum-maximum normalization, mean normalization or adaptive normalization methods;

[0153] In addition to the forward and backward window energy ratio construction method and the differential change rate and local statistics extraction method used in this embodiment, the short window feature extraction process can also be implemented by using single-sided sliding window energy statistics, cumulative energy change, local entropy features or other combinations of time-domain statistical features.

[0154] In addition to the weighted fusion decision function based on energy ratio, energy change rate and local variance used in this embodiment, the short-window joint decision process can also be implemented using adaptive threshold decision, double threshold hysteresis decision or classification decision method based on machine learning model.

[0155] In addition to adjusting the window length based on the energy change rate and statistical variance in this embodiment, the long window adaptive adjustment process can also be dynamically adjusted based on signal-to-noise ratio estimation, spectral feature changes or other statistical stability indicators.

[0156] In addition to constructing a joint decision function using energy ratio, rate of change, and statistical variance, and performing step-by-step screening in combination with effective state proportion, continuity constraint, and stability constraint, the long-window consistency verification and interference suppression process can also be implemented using other multi-feature fusion decision mechanisms, probabilistic decision methods, or interval screening methods based on learning models.

[0157] In addition to the intersection operation method of short window detection results and long window verification results used in this embodiment, the detection result fusion process can also be achieved by weighted fusion, confidence fusion or rule-based decision fusion methods.

[0158] However, all the modifications and substitutions based on the ideas of this invention are within the scope of protection of the claims of this invention.

Claims

1. A dual-window radar pulse detection method, characterized in that, include: (1) Collect radar signals acquired by radar receiving equipment, perform energy transformation processing on the radar signals, construct an energy sequence characterizing the change of signal power over time, and perform standardization processing on it; (2) Perform short window sliding on the energy sequence, calculate the forward window energy and backward window energy at the current time respectively, construct the short window energy ratio sequence, and calculate the energy change rate and statistical characteristics within the local window range under the short window scale; (3) Based on the energy ratio, energy change rate and statistical characteristics, establish a short window joint decision function for cross-scale energy evolution consistency to detect the location of signal energy abrupt change, and compare the joint decision function with the preset decision threshold to obtain the short window detection candidate interval; (4) Calculate the rate of change and statistical variance of the energy sequence within the local window range under the long window scale, and adaptively adjust the length of the long window so that the length of the long window changes dynamically with the signal characteristics; (5) Under the adaptive long window condition, the energy sequence is reprocessed with a sliding window, and the forward window energy and backward window energy are calculated respectively to form a long window energy ratio sequence; (6) Based on the long window energy ratio sequence, the corresponding energy change rate and statistical variance, establish a long window joint decision function to make a decision on the long window, and perform consistency verification and interference discrimination on the short window detection pulse candidate interval according to the decision result to obtain the interval that the long window verification passes; (7) Perform an intersection operation on the two intervals of the short window and the long window, and retain only the pulse interval that simultaneously satisfies the short window detection condition and the long window consistency verification condition to obtain the start position and end position of the final pulse signal.

2. The method according to claim 1, characterized in that, The energy transformation processing of the radar signal in (1) to construct an energy sequence characterizing the signal power change over time includes: 1a) Perform sampling-point energy calculation processing on the collected raw radar signal to convert the time-domain signal amplitude into the corresponding instantaneous energy value in order to obtain an energy sequence that reflects the change of signal power over time; 1b) For different types of radar signals, corresponding energy calculation methods are used to characterize the signal energy: When the radar signal is a real-value signal, the corresponding energy value is obtained by squaring the signal amplitude. When the radar signal is an IQ complex signal, the signal energy is calculated by summing the squares of the in-phase and quadrature components at the same time, thereby obtaining a unified form of energy sequence expression.

3. The method according to claim 1, characterized in that, The energy sequence is standardized in (1) using the following formula: ， in, This represents the standardized energy sequence. Represents the original energy sequence. This represents the mean. It represents the standard deviation.

4. The method according to claim 1, characterized in that, In step (2), the forward window at the current time is calculated respectively. Using the aperture energy and the backward window energy, a short window energy ratio sequence is constructed, including: 2a) Set the short window length according to the pulse width range and sampling rate of the radar signal. The standardized energy sequence is divided into two parts, a forward window and a backward window, with the current time as the center, to establish a time reference relationship for energy changes. 2b) The energy sequence is accumulated or weighted within the forward window range to obtain the forward window energy that represents the energy change trend after the current moment, which is used to characterize the future energy change trend; 2c) The energy sequence is accumulated or weighted within the backward window range to obtain the backward window energy, which represents the energy distribution before the current time and is used to characterize the historical energy accumulation characteristics. 2d) Calculate the short window energy ratio based on the numerical relationship between the forward window energy and the backward window energy, and arrange them in chronological order to form an energy ratio sequence. Furthermore, by introducing a preset constant to avoid the case where the denominator is zero, the energy ratio can stably reflect the degree of abrupt changes in signal energy.

5. The method according to claim 1, characterized in that, The calculation of energy change rate and statistical characteristics within a local window range under a short window scale in (2) includes: 2a) At the current time, perform a difference operation on the energy sequence to obtain the rate of energy change. This is used to characterize the instantaneous change characteristics of signal energy, and its formula is: ； 2b) Let the length of the local sliding window under the short window scale be... And calculate the local mean of the energy sequence within this local window. The formula for representing the average energy level during that time period is: ， in, This represents the standardized energy sequence; For sequence indices, their value ranges from arrive ; 2c) Within the local window, the deviation of the energy sequence is statistically calculated based on the local mean to obtain the local variance. It is used to characterize the intensity of energy fluctuations, and its formula is: 。 6. The method according to claim 1, characterized in that: The short-window joint decision function for cross-scale energy evolution consistency established in (3) is expressed as follows: ， The long window joint decision function established in (6) has the following formula: ， in, This represents a comprehensive judgment value regarding the trend and stability of signal energy changes at the current moment within a short window scale. Represents the short-window energy ratio sequence. Represents the rate of change of energy. Represents local variance. , , These are three weighting coefficients with different values; This represents a comprehensive judgment value regarding the trend and stability of signal energy changes at the current moment within a long window scale. Represents a long-window energy ratio sequence. This represents the local variance of a long window.

7. The method according to claim 1, characterized in that, The calculation of the rate of change and statistical variance of the energy sequence within a local window range under a long window scale in (4) includes: 4a) At the current time, perform a difference operation on the energy sequence to obtain the rate of energy change. This is used to characterize the instantaneous change characteristics of signal energy, and its formula is: ， 4b) Let the length of the local sliding window under the long window scale be... And calculate the local mean of the energy sequence within this local window. The formula for representing the average energy level during that time period is: ， in, This represents the standardized energy sequence. For sequence indices, the value range is from arrive ; 4c) Within the local window, the deviation of the energy sequence is statistically calculated based on the local mean to obtain the local variance. It is used to characterize the intensity of energy fluctuations, and its formula is: ， in, This represents the standardized energy sequence. For sequence indices, the value range is from arrive .

8. The method according to claim 1, characterized in that, In step (6), the long window is judged according to the decision function, and the consistency verification and interference discrimination of the short window detection pulse candidate interval are performed according to the decision result, including: 6a) The output of the long window joint decision function With preset decision threshold Compare and obtain the decision sequence : ， in, This indicates that the current state is considered a valid signal. This indicates that the condition is either background noise or non-pulse interference. 6b) Based on the start and end times of the candidate interval, extract the decision result subsequence within the corresponding time range from the decision sequence, calculate the ratio of the number of valid state sampling points in the candidate interval to the total number of sampling points, and compare this ratio with a preset ratio threshold: If the proportion of valid states is greater than the preset proportion threshold, then the candidate interval is marked as a consistency candidate interval, and 6c is executed. Otherwise, the candidate interval is removed; 6c) Within the consensus candidate interval, extract those that satisfy... A continuous sequence of sampling points is used to form a continuous state interval, and the corresponding continuous duration is compared with a preset duration threshold: If the continuous duration exceeds a preset duration threshold, the interval is marked as a continuous candidate interval, and execution is performed for 6 days. Otherwise, the candidate interval is determined to be an invalid pulse interval and is discarded; 6d) Within a continuous candidate interval, calculate the deviation of the energy sequence corresponding to the candidate interval from the local mean, and compare it with a preset stability threshold: If the deviation is less than the preset stability threshold, this interval will be used as the long-window verification pass interval. Otherwise, the candidate interval is determined to be an invalid pulse interval and is discarded.

9. A dual-window radar pulse detection system, characterized in that, include: The signal acquisition and preprocessing module is used to acquire radar signals and perform energy transformation and standardization to output a standardized energy sequence. The short window feature extraction module is used to slide the energy sequence through a short window, calculate the forward and backward window energies, construct a short energy ratio sequence, and extract the rate of change and statistical features. The short window decision module is used to construct a decision function based on the short window energy ratio sequence, energy change rate and statistical characteristics, and compare its output with a preset threshold to obtain the short window candidate interval. The long window parameter adaptive module is used to adaptively adjust the length of the long window based on the energy change rate and statistical characteristics; The long window feature extraction module is used to slide the energy sequence through a long window, calculate the forward and backward window energies, and construct a long window energy ratio sequence. The long window decision and verification module is used to construct the decision function and make decisions, while performing consistency verification and interference discrimination on the short window candidate intervals to obtain the long window verification passed intervals; The fusion decision module is used to perform intersection calculations on the short window candidate interval and the long window verification pass interval to obtain the pulse start and end positions.

10. The system according to claim 10, characterized in that, The long-window decision and verification module includes: The decision function construction submodule is used to construct a long-window joint decision function based on the long-window energy ratio sequence, energy change rate, and statistical characteristics to characterize the energy change trend and stability of the signal over a long time scale. Decision sequence generation submodule: used to compare the output of the joint decision function with a preset threshold to obtain a decision sequence that reflects the signal state; Candidate interval matching submodule: used to extract the decision result subsequence within the corresponding time range from the decision sequence based on the start and end times of the pulse candidate interval obtained by short window detection; Consistency Analysis Submodule: Used to perform consistency analysis on the decision result subsequence, by calculating the proportion of valid state sampling points and comparing it with a preset proportion threshold to filter consistency candidate intervals; The continuity verification submodule is used to extract a sequence of continuous valid state sampling points within the consistency candidate interval and compare the continuous duration with a preset duration threshold to filter the continuity candidate interval. Stability discrimination submodule: It is used to calculate the degree of deviation of the energy sequence corresponding to the continuous candidate interval from the local mean, and compare it with the preset stability threshold to complete the effective interval screening.

Images