A method of detecting seismic analysis features and generating an alert

Abstract

The application discloses a method for detecting vibration analysis characteristics and generating an alarm, comprising collecting and extracting original data of a half-cycle waveform of a current vibration sensor until the number of data is sufficient to start preprocessing, performing preprocessing on the collected data according to the frequency adopted to obtain absolute values every time the original data is collected; performing data rolling F t2 second processing to obtain rising edge length data and falling edge length data; then judging whether the effective waveform meets the condition or not, and saving the waveform data if the condition is met; judging whether the number and period of the effective waveform meet the alarm requirement according to the saved waveform data, and performing alarm processing if the requirement is met. The application judges by processing the rising edge length data and the falling edge length data to eliminate invalid waveforms, reduce or even eliminate misjudgment, improve the accuracy, and save the time and manpower cost of police response.

Description

Technical Field

[0001] This invention belongs to the field of buried vibration detection and alarm technology, specifically relating to a method for detecting vibration analysis characteristics and generating an alarm. Background Technology

[0002] Given the current complex border security situation, early warning for border security requires a series of effective technical means. In important locations such as military restricted areas, military airports, nuclear power plants, and hazardous chemical plants, corresponding products are also needed to provide early warning protection to prevent unauthorized intruders. In recent years, geological disasters have occurred frequently due to extreme weather, making early warning for earthquake-prone areas a focus of emergency management departments and local governments at all levels.

[0003] However, existing vibration analysis methods mostly rely solely on amplitude to determine whether an object is approaching. This method often fails to distinguish between valid and invalid waveforms, leading to false alarms and wasting time and manpower for emergency response. Summary of the Invention

[0004] To address the problems mentioned in the background art, the present invention provides a method for detecting vibration analysis characteristics and generating an alarm, thereby solving the problem that the prior art is prone to misjudgment, causing alarm failures and wasting time and human resources for emergency response.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] A method for detecting vibration analysis characteristics and generating an alarm includes the following steps:

[0007] S1: Enable ADC acquisition to collect and extract the raw data of the current half-cycle waveform of the vibration sensor;

[0008] S2: Preprocess data. Collect raw data until the number of collections is sufficient to enable the judgment of valid waveforms. Each time raw data is collected, the collected raw data is preprocessed once. The number of loops is used to ensure that the preprocessing is only performed once before the next collection, so as to obtain the absolute value of removing excess jitter, low frequency data and AC data.

[0009] The steps for preprocessing data in S2 include:

[0010] S2.1: Calculate the average value of the raw data by rolling to smooth the collected data and remove unnecessary jitter;

[0011] S2.2: Perform a rolling calculation of the average value based on the frequency F. t1 Secondary difference, filtering low-frequency data;

[0012] S2.3: Take the absolute value of the difference data from the previous step according to the frequency F, filter the AC data, and calculate F in a rolling manner when the corresponding number of times is collected in cover_cnt. t3 The sum of the second largest absolute values, Data_Sum, Data_Sum1;

[0013] Where F represents the frequency of current data processing;

[0014] F t1 This indicates the number of processing steps at the current processing frequency: 1.

[0015] F t3 This indicates the number of processing steps at the current processing frequency: 3.

[0016] cover_cnt represents the number of data collections;

[0017] S3: For each collection of raw data, the absolute value obtained is processed twice by data rolling to obtain the rising edge length data and the falling edge length data;

[0018] The data rolling processing steps in S3 include:

[0019] S3.1: Find F t2 The maximum absolute value within a given number of iterations;

[0020] S3.2: Rolling calculation of F t3 The average of the maximum values ​​within a range is used to smooth the waveform data. The preprocessed data Data_Sum and Data_Sum1 are then incremented by the corresponding number of acquisitions plus F. t2 The maximum number of times, minus the maximum number of times the previous collection was collected, and then divided by the average number of times;

[0021] S3.3: Find the F of the average value t3 The difference between the current sampling count and the baseline is checked. If the difference is greater than the baseline, the rising edge flag Rising_Flg is incremented by one, and the falling edge flag Fall_Flg is cleared. If Rising_Flg equals 1, the current sampling count cover_cnt is assigned to cover_cntrising, where cover_cntrising represents the sampling count corresponding to the current rising edge, and Fall_Length = cover_cnt - cover_cnt fall Get the width of the falling edge;

[0022] S3.4: If the number is less than the baseline, the falling edge flag Fall_Flg is incremented by one, and the rising edge flag Rising_Flg is cleared to zero. If Fall_Flg equals 1, the current sampling count cover_cnt is assigned to cover_cnt. fall cover_cntfall This indicates the number of acquisitions corresponding to the current falling edge, and Rising_Length = cover_cnt - cover_cntrising; This gives the rising edge width.

[0023] Among them, F t2 This indicates the number of processing steps at the current processing frequency: 2.

[0024] S4: Based on the rising edge length data and falling edge length data, determine whether the valid waveform meets the conditions. If it does, save the waveform data.

[0025] The specific judgment method in S4 is as follows: if the waveform rising edge length Rising_Length is greater than the predetermined length, the falling edge length Fall_Length is greater than the predetermined length, and the total waveform length Rising_Length + Fall_Length is greater than the predetermined length, then the current waveform is determined to be a valid waveform, and its waveform start position is recorded as cover_cnt - Rising_Length - Fall_Length, waveform length Rising_Length + Fall_Length, and waveform Wave_cnt count is incremented by 1;

[0026] S5: Determine whether the number and period of valid waveforms meet the alarm requirements based on the saved waveform data. If they do, perform alarm processing; otherwise, process the valid waveforms and return to S1.

[0027] Preferably, the waveform output in S1 fluctuates around 125, which is the middle value between 0 and 255. If no vibration is detected, the data fluctuates slightly around 125.

[0028] Preferably, the judgment method in S5 is as follows: determine whether the number of valid waveform data in the waveform data reaches 5, find the maximum value of the last five valid waveform periods max_data, if the difference between the five adjacent waveform periods is less than the average period of the five waveforms, and the maximum period is less than a predetermined value, then the alarm condition is met; if one of the conditions is not met, then remove the first valid waveform data currently recorded and return to S1 to continue collecting data.

[0029] Preferably, after an alarm is triggered, the source of the vibration is determined to be a person, vehicle, or animal based on the frequency of the five most recent valid waveforms, thereby generating a corresponding alarm.

[0030] Preferably, the raw data in S1 is amplified and excess precision is removed to output data with an 8-bit precision.

[0031] Compared with the prior art, the beneficial effects of the present invention are:

[0032] This application preprocesses the raw data collected by the vibration sensor to obtain the absolute value after removing redundant jitter, low-frequency data, and AC data. Then, it uses a data rolling processing method to obtain the rising edge length data and falling edge length data, which are used to determine whether the valid waveform meets the conditions. Based on the waveform data that meets the conditions, an alarm judgment is made. By processing and judging the rising edge length data and falling edge length data, this application eliminates invalid waveforms, reduces or even eliminates false judgments, improves accuracy, and saves time and manpower costs for dispatching police. Attached Figure Description

[0033] Figure 1 This is a schematic diagram of the method flow of the present invention;

[0034] Figure 2 This is a schematic diagram of ADC data acquisition.

[0035] Figure 3 This is a schematic diagram of the raw data collected.

[0036] Figure 4 A schematic diagram for calculating the rolling average value based on frequency;

[0037] Figure 5 A diagram illustrating the process of calculating the difference between rolling average data.

[0038] Figure 6 This is a diagram illustrating the absolute value of the difference data according to frequency F;

[0039] Figure 7 A diagram illustrating the process of finding the maximum absolute value within the Ft2 order;

[0040] Figure 8 A schematic diagram illustrating the average value of the maximum value within the Ft3 order;

[0041] Figure 9 This is a schematic diagram for effective waveform data analysis. Detailed Implementation

[0042] To facilitate understanding of the technical content of this invention by those skilled in the art, the invention will be further described in detail below with reference to the accompanying drawings and specific examples. It should be understood that the specific examples described herein are merely illustrative and not intended to limit the scope of the invention.

[0043] The term cover_cnt mentioned in the text refers to the number of data collections.

[0044] Cycle_cnt represents the number of times the loop is entered before the next data collection;

[0045] F indicates the frequency of current data processing.

[0046] F t1This indicates the number of processing steps at the current processing frequency: 1.

[0047] F t2 This indicates the number of processing steps at the current processing frequency: 2.

[0048] F t3 This indicates the number of processing steps at the current processing frequency: 3.

[0049] Data_Sum, Data_Sum1 indicates that F t3 The sum of the maximum absolute values ​​of the preprocessed counts;

[0050] Rising_Flg indicates a rising edge flag;

[0051] Fall_Flg indicates a falling edge flag;

[0052] `cover_cntrising` indicates the number of acquisitions corresponding to the current rising edge, used to calculate the rising edge width; `cover_cnt` fall This indicates the number of acquisitions corresponding to the current falling edge, used to calculate the falling edge width; Fall_Length indicates the width of the falling edge.

[0053] Rising_Length represents the width of the rising edge;

[0054] Wave_cnt represents the number of currently valid waveforms.

[0055] Example 1

[0056] like Figure 1 , Figure 2 , Figure 4 , Figure 5 , Figure 6 , Figure 7 , Figure 8 and Figure 9 As shown, a method for detecting vibration analysis characteristics and generating an alarm includes the following steps:

[0057] S1: Enable ADC acquisition to collect and extract the raw data of the current half-cycle waveform of the vibration sensor;

[0058] S2: Preprocess data. Collect raw data until the number of collections is sufficient to enable the judgment of valid waveforms. Each time raw data is collected, the collected raw data is preprocessed once. The number of loops is used to ensure that the preprocessing is only performed once before the next collection, so as to obtain the absolute value of removing excess jitter, low frequency data and AC data.

[0059] The steps for preprocessing data in S2 include:

[0060] S2.1: Calculate the average value of the raw data by rolling to smooth the collected data and remove unnecessary jitter;

[0061] S2.2: Perform a rolling calculation of the average value based on the frequency F. t1 Secondary difference, filtering low-frequency data;

[0062] S2.3: Take the absolute value of the difference data from the previous step according to the frequency F, filter the AC data, and calculate F in a rolling manner when the corresponding number of times is collected in cover_cnt. t3 The sum of the second largest absolute values, Data_Sum, Data_Sum1;

[0063] Where F represents the frequency of current data processing;

[0064] F t1 This indicates the number of processing steps at the current processing frequency: 1.

[0065] F t3 This indicates the number of processing steps at the current processing frequency: 3.

[0066] cover_cnt represents the number of data collections;

[0067] S3: For each collection of raw data, the absolute value obtained is processed twice by data rolling to obtain the rising edge length data and the falling edge length data;

[0068] The data rolling processing steps in S3 include:

[0069] S3.1: Find F t2 The maximum absolute value within a given number of iterations;

[0070] S3.2: Rolling calculation of F t3 The average of the maximum values ​​within a range is used to smooth the waveform data. The preprocessed data Data_Sum and Data_Sum1 are then incremented by the corresponding number of acquisitions plus F. t2 The maximum number of times, minus the maximum number of times the previous collection was collected, and then divided by the average number of times;

[0071] S3.3: Find the F of the average value t3 The difference between the current sampling count and the baseline is checked. If the difference is greater than the baseline, the rising edge flag Rising_Flg is incremented by one, and the falling edge flag Fall_Flg is cleared. If Rising_Flg equals 1, the current sampling count cover_cnt is assigned to cover_cntrising, where cover_cntrising represents the sampling count corresponding to the current rising edge, and Fall_Length = cover_cnt - cover_cnt fall Get the width of the falling edge;

[0072] S3.4: If the number is less than the baseline, the falling edge flag Fall_Flg is incremented by one, and the rising edge flag Rising_Flg is cleared to zero. If Fall_Flg equals 1, the current sampling count cover_cnt is assigned to cover_cnt. fall cover_cnt fall This indicates the number of acquisitions corresponding to the current falling edge, and Rising_Length = cover_cnt - cover_cntrising; This gives the rising edge width.

[0073] Among them, F t2 This indicates the number of processing steps at the current processing frequency: 2.

[0074] S4: Based on the rising edge length data and falling edge length data, determine whether the valid waveform meets the conditions. If it does, save the waveform data.

[0075] The specific judgment method in S4 is as follows: if the waveform rising edge length Rising_Length is greater than the predetermined length, the falling edge length Fall_Length is greater than the predetermined length, and the total waveform length Rising_Length + Fall_Length is greater than the predetermined length, then the current waveform is determined to be a valid waveform, and its waveform start position is recorded as cover_cnt - Rising_Length - Fall_Length, waveform length Rising_Length + Fall_Length, and waveform Wave_cnt count is incremented by 1;

[0076] S5: Determine whether the number and period of valid waveforms meet the alarm requirements based on the saved waveform data. If they do, perform alarm processing; otherwise, process the valid waveforms and return to S1.

[0077] In this embodiment, the present application preprocesses the raw data collected by the vibration sensor to obtain the absolute values ​​of excess jitter, low-frequency data and AC data. Then, the absolute values ​​are processed using a data rolling method to obtain the rising edge length data and falling edge length data, which are used to determine whether the valid waveform meets the conditions. Then, an alarm judgment is made based on the waveform data that meets the conditions. By processing and judging the rising edge length data and falling edge length data, the present application eliminates invalid waveforms, reduces or even eliminates false judgments, improves accuracy and saves time and manpower costs for dispatching police.

[0078] Example 2

[0079] The difference between this embodiment and Embodiment 1 is that: Figure 3As shown, the waveform output in S1 fluctuates around 125, which is the middle value between 0 and 255. If no vibration is detected, the data fluctuates slightly around 125.

[0080] Example 3

[0081] The difference between this embodiment and embodiment 2 is that the judgment method in S5 is as follows: it is determined whether the number of valid waveform data in the waveform data reaches 5, and the maximum value of the last five valid waveform periods max_data is found. If the difference between the five adjacent waveform periods is less than the average period of the five waveforms, and the maximum period is less than the predetermined value, the alarm condition is met. If one of the conditions is not met, the first valid waveform data recorded at the moment is removed and the process returns to S1 to continue collecting data.

[0082] Example 4

[0083] The difference between this embodiment and embodiment 3 is that: after the alarm is triggered, the source of the vibration is determined to be a person, vehicle or animal based on the frequency of the five most recent valid waveforms, thereby generating a corresponding alarm. This solves the problem of control in unmanned patrol environments, border defense and some important military sites, and effectively reduces labor costs.

[0084] Example 5

[0085] The difference between this embodiment and embodiment 1 is that the original data in S1 is amplified and after removing excess precision, the output data has a precision of 8 bits.

[0086] In this embodiment, the unit of data processing by the computer system is often a byte, which is 8 bits. Using 10 bits may waste some hardware data transmission and processing capabilities, resulting in processing performance lower than 8-bit encoding.

Claims

1. A method of detecting seismic analysis features and generating an alert, the method comprising: The method comprises the following steps: S1: starting ADC collection, collecting and extracting original data of a half-cycle waveform of a current vibration sensor; S2: preprocessing data, collecting original data until the collection times are sufficient to start judging effective waveforms, and preprocessing the collected original data once each time the original data is collected, and using the cycle times to ensure that the preprocessing is entered only once before the next collection, to obtain absolute values with redundant jitter, low-frequency data and alternating current data removed; The preprocessing data step in S2 comprises: S2.1: rolling average of the original data is performed to smooth the collected data and remove redundant jitter; S2.2: Data rolling average value of rolling is calculated according to frequency F t1 Sub-difference value, filter low frequency data; S2.3: take absolute value of the difference data in the last step, filter the alternating current data, when the corresponding time segment is collected in cover_cnt according to frequency F, roll calculation F t3 the sum of the maximum values of the absolute values of the differences Data_Sum, Data_Sum1; Wherein, F represents the frequency used in current data processing; F t1 represents the number of treatments 1 at the current treatment frequency; F t3 represents the number of treatments 3 at the current treatment frequency; cover_cnt represents the collection times; S3: each time the original data is collected, the obtained absolute values are subjected to rolling secondary processing to obtain rising edge length data and falling edge length data; The data rolling processing step in S3 comprises: S3.1: Find F t2 the maximum of the absolute values over the number of times S3.2: Rolling F t3 The average of the maximum value of the number of times, smooth the waveform data, add the preprocessed data Data_Sum, Data_Sum1 to the corresponding acquisition number plus F t2 The maximum value of the number of times, subtract the corresponding maximum value of the last acquisition number and divide by the average number of times; S3.3: F of average value t3 the difference between the two times is obtained and it is determined whether the difference is greater than a reference number. If the difference is greater than the reference number, the rising edge flag Rising_Flg is incremented by one, and the falling edge flag Fall_Flg is cleared. If Rising_Flg is equal to 1, the current acquisition time cover_cnt is assigned to cover_cntrising, which represents the acquisition time corresponding to the current rising edge. Furthermore, Fall_Length = cover_cnt - cover_cntrising fall the falling edge width is obtained; S3.4: If the reference number is less than the current number, the falling edge flag Fall_Flg is incremented by one, the rising edge flag Rising_Flg is cleared, and if Fall_Flg equals 1, the current acquisition number cover_cnt is assigned to cover_cnt fall , cover_cnt fall represents the acquisition number corresponding to the current falling edge, and Rising_Length = cover_cnt - cover_cntrising; the rising edge width is obtained. where F t2 represents the number of treatments 2 at the current treatment frequency; S4: judging whether the effective waveforms meet the conditions according to the rising edge length data and the falling edge length data, and saving the waveform data if they meet the conditions; The judgment method in S4 is specifically as follows: if the waveform rising edge length Rising_Length is greater than a predetermined length, the falling edge length Fall_Length is greater than a predetermined length, and the entire waveform length Rising_Length+Fall_Length is greater than a predetermined length, the current waveform is judged to be an effective waveform, and the waveform starting position, i.e. cover_cnt-Rising_Length-Fall_Length, the waveform length Rising_Length+Fall_Length, and the waveform Wave_cnt count are recorded; S5: judging whether the number and period of the effective waveforms meet the alarm requirements according to the saved waveform data, and performing alarm processing if they meet the requirements, or processing the effective waveforms and returning to S1 if they do not meet the requirements; The judgment method in S5 is specifically as follows: judging whether the number of effective waveform data in the waveform data reaches 5, finding the maximum value max_data of the five latest effective waveform periods, and meeting the alarm condition if the difference between the adjacent five waveform periods is less than the average period of 5 waveforms, and the maximum period is less than a predetermined value, or removing the first recorded effective waveform data and returning to S1 to continue collecting data if one of the conditions is not met, and judging the vibration source to be a person, a vehicle or an animal based on the frequency of the five latest effective waveforms to generate a corresponding alarm alarm.

2. A method of detecting and analyzing characteristics of vibrations and generating an alarm according to claim 1, characterized in that, The waveform output in S1 fluctuates up and down at a position of 125 in the middle value of 0-255, and the data fluctuates up and down at 125 with a small amplitude if no vibration is detected.

3. A method of detecting and analyzing characteristics of vibrations and generating an alarm as claimed in claim 1, wherein, The original data in S1 is collected and amplified, and the output data precision is 8-bit data after removing redundant precision.

Images