A method, electronic device and storage medium for identifying vehicle collisions and scrapes using elastic waves.

Abstract

This application discloses a method, electronic device, and storage medium for identifying vehicle collision and scrape incidents using elastic waves. The method includes: continuously acquiring raw elastic wave signals from a vehicle body and environmental noise signals; filtering and smoothing the raw elastic wave signals to obtain a primary elastic wave signal; dynamically constructing a threshold baseline based on the environmental noise signal; performing a difference operation between the primary elastic wave signal and the threshold baseline to obtain a signal residual, and performing a secondary filtering operation on the primary elastic wave signal based on the signal residual to obtain a secondary elastic wave signal; extracting features from the secondary elastic wave signal to obtain multidimensional time-domain features; determining the scrape type based on the multidimensional time-domain features, and generating and outputting the corresponding data recording results. This application can improve the reliability and environmental adaptability of vehicle collision and scrape incident identification.

Description

Technical Field

[0001] This application relates to the field of vehicle assisted driving technology, and in particular to an elastic wave vehicle collision and scrape recognition method, electronic device and storage medium. Background Technology

[0002] With the rapid iteration and development of intelligent connected vehicle technology, the demand for vehicle body safety and intelligent anti-theft protection continues to increase. Identifying abnormal events such as vehicle collisions and scrapes based on vibration sensing has become a mainstream technology direction in the field of vehicle security. This type of technology mainly collects vehicle vibration signals through sensing devices to identify external impact behavior, thereby realizing the monitoring and early warning of abnormal vehicle events, and is widely used in the safety protection systems of various passenger vehicles.

[0003] Currently, traditional vehicle vibration event recognition solutions in the industry generally employ fixed threshold judgment modes or complex frequency domain transformation processing methods. Fixed threshold judgment methods use a uniform standard to identify vibration signals, which cannot adapt to complex and changing external environmental noise such as driving bumps and wind and rain interference. They are prone to misjudging invalid environmental interference as collision events and easily miss weak scrape signals in quiet environments, resulting in poor environmental adaptability and significant false alarm and missed alarm problems. On the other hand, processing solutions using complex algorithms such as frequency domain transformation rely on a large number of floating-point operations, placing stringent demands on hardware computing power. This leads to high hardware costs and high system power consumption, and makes real-time deployment on resource-constrained low-cost embedded devices difficult. Furthermore, it is challenging to balance recognition accuracy, anti-interference capabilities, and the low-cost, low-power requirements for engineering implementation. Summary of the Invention

[0004] This application aims to address at least one of the technical problems existing in the prior art. To this end, this application proposes an elastic wave vehicle collision and scrape identification method, electronic device, and storage medium, which can improve the reliability and environmental adaptability of vehicle collision and scrape identification.

[0005] In a first aspect, this application provides a method for identifying elastic wave vehicle collisions and scrapes, including: Continuously collect raw vibration elastic wave signals of the vehicle body and environmental noise signals; The original vibration elastic wave signal is filtered and smoothed to obtain the first-order vibration elastic wave signal. A threshold baseline is dynamically constructed based on the environmental noise signal; The difference between the primary vibration elastic wave signal and the threshold baseline is calculated to obtain the signal residual. The primary vibration elastic wave signal is then subjected to secondary filtering based on the signal residual to obtain the secondary vibration elastic wave signal. Feature extraction is performed on the second-order vibration elastic wave signal to obtain multidimensional time-domain features; Based on the multidimensional time-domain features, the scraping type is determined, and the corresponding data record results are generated and output.

[0006] The elastic wave vehicle collision and scrape recognition method according to the first aspect of this application has at least the following beneficial effects: First, the original vibration elastic wave signal of the vehicle body and the environmental noise signal are continuously collected. The original vibration elastic wave signal is sequentially filtered and smoothed to obtain the first-level vibration elastic wave signal. Then, a threshold baseline is dynamically constructed based on the environmental noise signal. The difference between the first-level vibration elastic wave signal and the threshold baseline is calculated to obtain the signal residual. Based on the signal residual, the first-level vibration elastic wave signal is subjected to a second-level filtering process to obtain the second-level vibration elastic wave signal. Subsequently, feature extraction is performed on the second-level vibration elastic wave signal to obtain multi-dimensional time-domain features. Finally, the scrape type is determined based on the multi-dimensional time-domain features, and the corresponding data record result is generated and output. This scheme no longer uses a fixed judgment benchmark. It can combine real-time environmental noise to dynamically adapt the threshold baseline and combine multi-layer signal processing and multi-dimensional time-domain feature comprehensive judgment to overcome the defects of traditional recognition methods, such as weak environmental adaptability and easy misjudgment and missed judgment. It can accurately distinguish between real collision and scrape behavior and external environmental interference, and can improve the reliability and environmental adaptability of vehicle collision and scrape recognition.

[0007] According to some embodiments of the first aspect of this application, the step of filtering and smoothing the original vibration elastic wave signal to obtain a first-order vibration elastic wave signal includes: Obtain the set of valid physical addresses and valid sensor channels that have been pre-entered and stored on the local machine; The device physical address field and the sensor channel number field are extracted from the original vibration elastic wave signal; The device physical address field is compared one by one with the set of legal physical addresses, and the original vibration elastic wave signals that do not match the device physical address field are eliminated. The sensor channel number field of the original vibration elastic wave signal that matches the physical address field of the device is compared one by one with the set of valid sensor channels, and the original vibration elastic wave signals that do not match the sensor channel number field are eliminated. The original vibration elastic wave signals that have been screened and retained are sorted and organized according to the sampling timestamp, and isolated abnormal data points with timestamp jumps or abnormal sampling intervals are removed. The original vibration elastic wave signal after screening is smoothed and filtered to obtain the first-order vibration elastic wave signal.

[0008] According to some embodiments of the first aspect of this application, the step of smoothing and filtering the screened original vibration elastic wave signal to obtain a first-order vibration elastic wave signal includes: A sliding window is constructed according to a preset fixed time domain length, and each sampling point in the original vibration elastic wave signal after filtering is sequentially traversed according to the sliding window. Extract the amplitude data of all sampling points within the current sliding window, and remove outlier sampling points whose amplitudes exceed the preset amplitude range; Calculate the amplitude difference between the current sampling point and the adjacent previous sampling point and the adjacent next sampling point; If all the amplitude differences are less than the preset amplitude difference threshold, the interval where the current sampling point is located is determined to be a steady-state interval; If any of the amplitude differences is greater than or equal to the amplitude difference threshold, the interval where the current sampling point is located is determined to be a waveform abrupt change interval; Apply the complete sliding window to the steady-state interval and calculate the average amplitude of the remaining sampling points within the sliding window; For the waveform abrupt change interval, reduce the value length of the sliding window, and calculate the average amplitude of the sampling points within the reduced sliding window; Based on the average amplitude of the corresponding interval, the sampling point amplitude of the original vibration elastic wave signal is replaced point by point; The replaced sampling data are rearranged according to the sampling time sequence to obtain the first-order vibration elastic wave signal.

[0009] According to some embodiments of the first aspect of this application, the step of dynamically constructing a threshold baseline based on the environmental noise signal includes: Get the preset segmented time length; According to the segmented time length obtained by matching, the environmental noise signal is continuously segmented and sampled to obtain multiple sets of environmental noise sampling data arranged in sequence. The amplitude distribution information of each group of environmental noise sampling data is statistically analyzed, and the current external environmental noise is classified into levels based on the amplitude distribution information of each group to obtain the environmental noise level; The corresponding baseline iteration step size is matched according to the environmental noise level, and the initial threshold baseline is dynamically adjusted by iterating point by point according to the time sequence based on the baseline iteration step size. When multiple consecutive sets of environmental noise sampling data correspond to the same environmental noise level, the iterative update of the threshold baseline is stopped. Furthermore, the upper limit and lower limit of the amplitude of the threshold baseline are preset to control the amplitude of the threshold baseline not to exceed the upper limit and lower limit during the iteration process. The current vehicle operating conditions are obtained, and the segmented time length is dynamically adjusted based on the current vehicle operating conditions and the ambient noise level.

[0010] According to some embodiments of the first aspect of this application, the step of obtaining the current vehicle operating condition and dynamically adjusting the segmented time length based on the current vehicle operating condition and the ambient noise level includes: Obtain the current vehicle operating conditions; wherein, the vehicle operating conditions include driving conditions and parking conditions; Based on the preset weight allocation rules, corresponding operating condition weights and noise weights are allocated according to the current vehicle operating conditions and the ambient noise level, respectively. Based on the operating condition weight and the noise weight, a comprehensive weight value is obtained; The adjusted segmented time length is obtained by multiplying the comprehensive weight value with the basic segmented time length.

[0011] According to some embodiments of the first aspect of this application, the step of performing secondary filtering processing on the primary vibration elastic wave signal based on the signal residual to obtain a secondary vibration elastic wave signal includes: Preset a fixed residual threshold value; The signal residuals at each time step are compared one by one with the residual threshold values; When the signal residual is less than the residual threshold, the corresponding signal is determined to be invalid background noise, and the signal value of the sampling point corresponding to the first-order vibration elastic wave signal is set to 0. If the signal residual is greater than or equal to the residual threshold value, then the original signal value of the sampling point corresponding to the first-order vibration elastic wave signal is retained; The sampling points in the processed primary vibration elastic wave signal are rearranged sequentially according to the time sequence to obtain the secondary vibration elastic wave signal.

[0012] According to some embodiments of the first aspect of this application, the step of extracting features from the second-order vibration elastic wave signal to obtain multidimensional time-domain features includes: The effective signal segments in the second-order vibration elastic wave signal are traversed segment by segment. Traverse the amplitude data of all sampling points within the current effective signal segment, compare the amplitude values ​​one by one, filter out the maximum value of the amplitude data, and extract the signal peak value; The amplitude data of each sampling point of the current effective signal segment is compared with the threshold baseline one by one, and the duration of the amplitude data continuously exceeding the threshold baseline is counted to extract the signal width. Locate the start and end times of the current valid signal segment, count the time intervals before and after the valid signal segment where there are no valid signal segments, and extract the silent window. The extracted signal peak value, signal width, and silent window are integrated to form the multidimensional time-domain feature.

[0013] According to some embodiments of the first aspect of this application, the step of determining the scraping type based on the multidimensional time-domain features and generating and outputting the corresponding data record result includes: Preset peak value determination threshold, width determination threshold, and silent window determination threshold; If the signal peak value is greater than or equal to the peak value determination threshold, the signal width is greater than or equal to the width determination threshold, and the silence window is greater than or equal to the silence window determination threshold, then it is determined as the first event result. If the signal peak value is less than the peak value determination threshold, or the signal width is less than the width determination threshold, or the silence window is less than the silence window determination threshold, then it is determined to be the second event result; The first event result or the second event result is associated and matched with the corresponding multidimensional time-domain features and sampling timestamps to generate data record results; The data recording results are displayed on a preset electronic display screen.

[0014] Secondly, this application also provides an electronic device, comprising: At least one memory; At least one processor; At least one program; The program is stored in the memory, and the processor executes at least one of the programs to implement the elastic wave vehicle collision and scrape recognition method as described in any embodiment of the first aspect.

[0015] Thirdly, this application also provides a computer-readable storage medium storing a computer-executable program for performing the elastic wave vehicle collision and scrape recognition method as described in any embodiment of the first aspect.

[0016] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0017] Additional aspects and advantages of this application will become apparent and readily understood in conjunction with the following description of the embodiments, in which: Figure 1 A flowchart of the elastic wave vehicle collision and scrape recognition method provided in this application; Figure 2 For this application Figure 1 The flowchart for step S120 is shown below; Figure 3 For this application Figure 2 The flowchart for step S260 is shown below; Figure 4 For this application Figure 1 The flowchart for step S130 is shown below; Figure 5 For this application Figure 4 The flowchart for step S460 is shown below; Figure 6 For this application Figure 1 The flowchart for step S140 is shown below; Figure 7 For this application Figure 1 The flowchart for step S150 is shown below; Figure 8 For this application Figure 1 The flowchart for step S160. Detailed Implementation

[0018] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain this application, and should not be construed as limiting this application.

[0019] In the description of this application, it should be understood that the orientation descriptions, such as up, down, front, back, left, right, etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this application.

[0020] In the description of this application, the use of "first" and "second" is for the purpose of distinguishing technical features only, and should not be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated or the order of the technical features indicated.

[0021] In the description of this application, unless otherwise expressly defined, terms such as "setup," "installation," and "connection" should be interpreted broadly, and those skilled in the art can reasonably determine the specific meaning of the above terms in this application in conjunction with the specific content of the technical solution.

[0022] With the rapid development of intelligent connected vehicle technology, the requirements for vehicles' perception capabilities of the external environment are increasing. In addition to traditional vision (cameras) and distance perception (radar), the vehicle's ability to perceive its own "tactile" senses—that is, the ability to identify vibrations and impacts occurring on the vehicle's surface—has become a crucial element in improving vehicle safety and intelligence. This perception capability primarily relies on vibration sensors (such as accelerometers) and the signal processing algorithms behind them. Its core task is to convert the raw analog signals collected by the sensors into digital signals, accurately extract useful information (such as collisions and scrapes), and filter out useless interference (such as road bumps and wind noise).

[0023] Prior to this invention, the mainstream vibration signal processing solutions in the industry mainly relied on the technical route of "analog front-end + fixed threshold + floating-point operation".

[0024] Analog Front-End and Fixed Threshold Judgment: Traditional solutions typically use analog circuits (such as comparators) to directly process sensor signals. The system sets a fixed voltage threshold; when the sensor signal momentarily exceeds this threshold, a collision is determined. This approach is simple in structure and low in cost. However, its fatal weakness lies in its "one-size-fits-all" judgment logic. When a vehicle is traveling on bumpy roads or encountering strong winds, the vibration signals caused by environmental noise are often large and easily exceed the fixed threshold, leading to false alarms. Conversely, in very quiet environments, weak malicious knocking or scraping signals may be missed because they are below the threshold. This lack of environmental adaptability severely restricts the reliability of vehicle security systems.

[0025] Digital signal processing based on floating-point arithmetic: To improve the accuracy of judgments, some high-end solutions have begun to introduce microprocessors (such as DSPs or high-performance MCUs) for digital signal processing. These solutions can run complex algorithms, such as Fast Fourier Transform (FFT) or wavelet transform, attempting to distinguish between collisions and noise by analyzing the frequency domain characteristics of the signal. Although theoretically, this improves accuracy, it relies heavily on floating-point operations. While floating-point operations offer high precision and ease of programming, they place extremely high demands on the processor's computing power, resulting in high system power consumption, high heat generation, and high hardware costs. More importantly, these complex algorithms are often difficult to implement in real-time on resource-constrained, low-cost embedded platforms (such as general-purpose MCUs), often requiring external high-performance computing units, increasing the complexity and latency of system integration.

[0026] In summary, existing technical solutions face a dilemma: low-cost analog solutions suffer from poor accuracy and high false alarm rates, while high-precision digital solutions are costly, power-hungry, and difficult to implement. Neither approach effectively addresses the issues of poor environmental adaptability, high false alarm and false negative rates, and difficulties in engineering implementation, thus failing to meet the comprehensive requirements of modern intelligent vehicles for high reliability, low cost, and low power consumption.

[0027] Based on this, this application provides an elastic wave vehicle collision and scrape identification method, electronic device and storage medium to solve the above-mentioned technical problems. The technical solutions provided by this application will be described in detail below.

[0028] Firstly, referring to Figure 1 This application provides a method for identifying elastic wave vehicle collisions and scrapes, including but not limited to the following steps: Step S110: Continuously collect the original vibration elastic wave signal of the vehicle body and the environmental noise signal.

[0029] Step S120: The original vibration elastic wave signal is screened and smoothed to obtain the first-order vibration elastic wave signal.

[0030] Step S130: Dynamically construct a threshold baseline based on the environmental noise signal.

[0031] Step S140: Perform difference calculation processing on the primary signal of the vibration elastic wave and the threshold baseline to obtain the signal residual, and perform secondary filtering processing on the primary signal of the vibration elastic wave based on the signal residual to obtain the secondary signal of the vibration elastic wave.

[0032] Step S150: Extract features from the second-order vibration elastic wave signal to obtain multidimensional time-domain features.

[0033] Step S160: Determine the type of scraping based on multidimensional time-domain features, and generate and output the corresponding data record results.

[0034] In steps S110 to S160, the original vibration elastic wave signal and environmental noise signal of the vehicle body are continuously collected. The original vibration elastic wave signal is sequentially filtered and smoothed to obtain the primary vibration elastic wave signal. Then, a threshold baseline is dynamically constructed based on the environmental noise signal. The difference between the primary vibration elastic wave signal and the threshold baseline is calculated to obtain the signal residual. Based on the signal residual, the primary vibration elastic wave signal is subjected to secondary filtering to obtain the secondary vibration elastic wave signal. Subsequently, feature extraction is performed on the secondary vibration elastic wave signal to obtain multi-dimensional time-domain features. Finally, the scratch type is determined based on the multi-dimensional time-domain features, and the corresponding data record result is generated and output. This scheme no longer uses a fixed judgment benchmark. It can dynamically adapt the threshold baseline by combining real-time environmental noise and the comprehensive judgment method of multi-layer signal processing and multi-dimensional time-domain features. It overcomes the defects of traditional recognition methods, such as weak environmental adaptability and easy misjudgment and missed judgment. It can accurately distinguish between real collision and scratch behavior and external environmental interference, and can improve the reliability and environmental adaptability of vehicle collision and scratch recognition.

[0035] It should be noted that in step S110, the original vibration elastic wave signal is obtained by attaching a dedicated vibration sensor to the vehicle body sheet metal, door, or chassis, etc., to sense the mechanical vibration of the vehicle body caused by collisions, scrapes, and road bumps in real time. The mechanical vibration is converted into a continuous electrical signal, which is then sampled to obtain the original vibration elastic wave signal. The ambient noise signal is collected by independently deployed acoustic sensors or sound pickup sensors to specifically pick up external background acoustic signals such as wind noise, traffic noise, environmental noise, and wind and rain interference around the vehicle. These are sampled independently and acquired as ambient noise signals, collected separately from the original vibration elastic wave signal to avoid confusion.

[0036] Reference Figure 2 It is understood that step S120 may include, but is not limited to, the following steps: Step S210: Obtain the set of legal physical addresses and the set of valid sensor channels of the local machine that have been pre-entered and stored.

[0037] Step S220: Extract the device physical address field and the sensor channel number field from the original vibration elastic wave signal.

[0038] Step S230: Compare the device physical address field with the set of valid physical addresses one by one, and eliminate the original vibration elastic wave signals that do not match the device physical address field.

[0039] Step S240: Compare the sensor channel number field of the original vibration elastic wave signal that matches the device physical address field with the set of valid sensor channels one by one, and eliminate the original vibration elastic wave signals that do not match the sensor channel number field.

[0040] Step S250: Sort and organize the selected and retained original vibration elastic wave signals according to the sampling timestamp, and remove isolated abnormal data points with timestamp jumps or abnormal sampling intervals.

[0041] Step S260: Perform smoothing filtering on the screened original vibration elastic wave signal to obtain the first-order vibration elastic wave signal.

[0042] In steps S210 to S260, by comparing and filtering the set of legitimate physical addresses and the set of valid sensor channels, invalid original vibration and elastic wave signals with mismatched device physical addresses or sensor channel numbers can be eliminated, avoiding interference from unauthorized devices and invalid channel signals in subsequent processing. The selected original vibration and elastic wave signals are then sorted and regularized according to the sampling timestamp, eliminating isolated abnormal data points with timestamp jumps or abnormal sampling intervals, ensuring the continuity and regularity of the signal timing arrangement. Simultaneously, the selected original vibration and elastic wave signals undergo smoothing filtering to remove random interference spikes, improving the purity of the original vibration and elastic wave signals and resulting in higher quality primary vibration and elastic wave signals. This provides stable and reliable basic signal support for subsequent threshold baseline construction, signal residual calculation, and multi-dimensional time-domain feature extraction processes, reducing identification bias caused by abnormal and invalid data.

[0043] Reference Figure 3 It is understood that step S260 may include, but is not limited to, the following steps: Step S310: Construct a sliding window according to a preset fixed time domain length, and according to the sliding window, sequentially traverse each sampling point in the original vibration elastic wave signal after screening.

[0044] Step S320: Extract the amplitude data of all sampling points within the current sliding window, and remove outlier sampling points whose amplitudes exceed the preset amplitude range.

[0045] Step S330: Calculate the amplitude difference between the current sampling point and the adjacent previous sampling point and the adjacent next sampling point.

[0046] Step S340: If the amplitude difference is less than the preset amplitude difference threshold, the interval where the current sampling point is located is determined to be a steady state interval.

[0047] Step S350: If any of the amplitude difference values ​​is greater than or equal to the amplitude difference threshold, the interval where the current sampling point is located is determined to be a waveform change interval.

[0048] Step S360: Apply a complete sliding window to the steady-state interval and calculate the average amplitude of the remaining sampling points within the sliding window.

[0049] Step S370: Reduce the length of the sliding window for the waveform abrupt change interval, and calculate the average amplitude of the sampling points within the reduced sliding window.

[0050] Step S380: Replace the sampling point amplitude of the original vibration elastic wave signal point by point according to the average amplitude of the corresponding interval.

[0051] Step S390: Rearrange the replaced sampling data according to the sampling time sequence to obtain the first-order signal of the vibration elastic wave.

[0052] In steps S310 to S320, by constructing a sliding window according to a preset fixed time domain length to traverse the filtered original vibration elastic wave signal sampling points, outlier sampling points with amplitudes exceeding the preset amplitude range within the window are first removed. This can effectively remove abrupt abnormal amplitude interference points in the signal and avoid outlier abnormal data from causing adverse effects on subsequent filtering processing.

[0053] In steps S330 to S390, the steady-state interval and waveform abrupt change interval are accurately divided by calculating the amplitude difference between the current sampling point and the adjacent sampling points. For the steady-state interval, a complete sliding window is used to calculate the average amplitude for smoothing, effectively filtering out minor random noise in the stable signal segment. For the waveform abrupt change interval, the sliding window length is reduced before calculating the average amplitude, which not only smooths and reduces noise around the abrupt change but also fully preserves the waveform details corresponding to collisions and scrapes, avoiding excessive smoothing of effective waveform features due to an excessively large window. This differentiated window smoothing method balances noise suppression and effective feature preservation, improving the waveform quality and fidelity of the processed first-order vibration elastic wave signal.

[0054] Reference Figure 4 It is understood that step S130 may include, but is not limited to, the following steps: Step S410: Obtain the preset segmented time length.

[0055] Step S420: According to the segmented time length obtained by matching, the environmental noise signal is continuously sampled in segments to obtain multiple sets of environmental noise sampling data arranged in sequence.

[0056] Step S430: Calculate the amplitude distribution information of each group of environmental noise sampling data, and classify the current external environmental noise into levels based on the amplitude distribution information of each group to obtain the environmental noise level.

[0057] Step S440: Match the corresponding baseline iteration step size according to the environmental noise level, and iterate point by point according to the time sequence based on the baseline iteration step size to dynamically adjust the initial threshold baseline.

[0058] Step S450: When multiple consecutive sets of environmental noise sampling data correspond to the same environmental noise level, stop iteratively updating the threshold baseline. Furthermore, preset the upper and lower limits of the threshold baseline amplitude to control the amplitude of the threshold baseline not to exceed the upper and lower limits during the iteration process.

[0059] Step S460: Obtain the current vehicle operating conditions and dynamically adjust the segment time length based on the current vehicle operating conditions and environmental noise level.

[0060] In step S440, the required speed of baseline adjustment varies depending on the level of external noise. In high-noise environments, the baseline needs to rise rapidly to quickly adapt to the background noise level; in low-noise, quiet environments, the baseline needs to fall slowly and slightly to maintain sensitivity to detect weak collisions and scrapes. If a uniform fixed iteration step size is used, the baseline in noisy environments may not keep up with noise changes, leading to false alarms; conversely, in quiet environments, the baseline falling too quickly can amplify minor interferences and cause misjudgments. Point-by-point time-series iteration ensures smooth, abrupt changes in the threshold baseline, conforming to the gradual changes in real noise and preventing abrupt baseline changes from incorrectly filtering out valid vibration signals.

[0061] In step S450, when the environmental noise is high and remains constant over a long period, if the iteration continues without stopping, the baseline will rise indefinitely, eventually exceeding the amplitude of the actual collision / scratching signal, drowning out all valid signals and causing missed detections. Conversely, in extremely quiet environments, the baseline will also drop indefinitely, triggering false alarms even with slight environmental disturbances. Adding a steady-state stopping iteration feature ensures that the baseline remains at a reasonable position in stable noise scenarios, preventing ineffective increases or decreases. Adding upper and lower amplitude limits as a fallback constraint prevents extreme baseline conditions of excessively high or low values, always locking the threshold baseline within a reasonable working range. This mechanism simultaneously avoids the defects of unlimited increases or decreases, balancing anti-interference capability and detection sensitivity.

[0062] In step S460, during vehicle operation, the road surface is bumpy, there are many external interferences, and the noise fluctuates frequently and complexly; when the vehicle is parked and stationary, the environmental noise is stable and fluctuates little. If a fixed segment time length is used throughout the process, if the segment time length is set too short, it is easy to be skewed by instantaneous sudden noise, and the environmental noise level will be inaccurate; if the segment time length is set too long, there will be more redundant calculations in the quiet parking scenario, which will consume more processing resources and reduce real-time performance.

[0063] In steps S410 to S460, the segmented time length is dynamically adjusted according to the current vehicle operating conditions and environmental noise level. Then, the environmental noise signal is continuously sampled in segments according to the matched segmented time length. The environmental noise level is divided by statistically analyzing the amplitude distribution information of each group of environmental noise sampling data. The initial threshold baseline is dynamically adjusted point by point according to the baseline iteration step size matched with the environmental noise level. At the same time, the iteration update of the threshold baseline stops when the environmental noise level corresponding to multiple consecutive groups of environmental noise sampling data is the same. Combined with the preset upper and lower amplitude limits, the range of change of the threshold baseline is constrained to avoid the threshold baseline from rising or falling indefinitely. This makes the adjustment process of the threshold baseline more stable and reasonable, and closer to the real-time changes in the external environment.

[0064] It should be noted that the threshold baseline is the core reference benchmark for distinguishing between the effective collision and scraping components and the invalid background noise components in the vibration elastic wave signal. It is constructed and adaptively adjusted in real time based on the environmental noise signal, which can adapt to different vehicle operating conditions and noisy external environments, and can improve the environmental adaptability and discrimination accuracy of the overall recognition process.

[0065] Reference Figure 5 It is understood that step S460 may include, but is not limited to, the following steps: Step S510: Obtain the current vehicle operating conditions; wherein, the vehicle operating conditions include driving conditions and parking conditions.

[0066] Step S520: Based on the preset weight allocation rules, assign corresponding operating condition weights and noise weights according to the current vehicle operating conditions and environmental noise levels.

[0067] Step S530: Obtain the comprehensive weight value based on the operating condition weight and the noise weight.

[0068] Step S540: Multiply the comprehensive weight value with the basic segmented time length to obtain the adjusted segmented time length.

[0069] In steps S510 to S540, the current vehicle operating conditions are distinguished into two categories: driving conditions and parking conditions. Based on preset weighting rules, corresponding operating condition weights and noise weights are assigned to the current vehicle operating condition and environmental noise level, respectively. A comprehensive weight value is then calculated from the operating condition weight and noise weight. This comprehensive weight value is multiplied by the base segmented time length to obtain the adjusted segmented time length, breaking the limitation of a fixed segmented time length. This method can simultaneously consider two influencing factors: vehicle operating conditions and environmental noise levels. Through weight fusion, it achieves adaptive adjustment of the segmented time length, allowing the segmented sampling duration of the environmental noise signal to adapt to the current actual operation and noise environment. This makes the subsequent amplitude distribution information statistics and environmental noise level classification results of the environmental noise sampling data more closely match the real external conditions, providing a reliable foundation for the accurate iterative adjustment of the threshold baseline.

[0070] In one embodiment, a fixed basic segment time length is preset, along with preset weight allocation rules. These rules specify a driving condition weight of 1.2, a parking condition weight of 0.8, a high noise level weight of 1.2, and a low noise level weight of 0.8. When the current vehicle operating condition is detected as driving and the ambient noise level is high, the corresponding driving condition weight of 1.2 and noise weight of 1.2 are assigned respectively. These two weights are multiplied to obtain a comprehensive weight value, which is then multiplied by the basic segment time length to obtain the amplified segment time length. When the current vehicle operating condition is detected as driving and the ambient noise level is low, or the current vehicle operating condition is parking and the ambient noise level is high, the corresponding driving condition weight and noise weight are matched, and the comprehensive weight value is calculated to ultimately obtain an appropriate segment time length. When the current vehicle operating condition is detected as parking and the ambient noise level is low, a smaller operating condition weight and noise weight are applied, resulting in a smaller overall weight value, which in turn reduces the basic segmented time length. Through the above steps, the segmented time length is adaptively adjusted according to the vehicle operating condition and ambient noise level.

[0071] Reference Figure 6 It is understood that step S140 may include, but is not limited to, the following steps: Step S610: Preset a fixed residual threshold value.

[0072] Step S620: Compare the signal residuals under each time series with the residual threshold values ​​one by one.

[0073] Step S630: When the signal residual is less than the residual threshold, the corresponding signal is determined to be invalid background noise, and the signal value of the corresponding sampling point of the first-order vibration elastic wave signal is set to 0.

[0074] Step S640: If the signal residual is greater than or equal to the residual threshold, then retain the original signal value of the sampling point corresponding to the first-order signal of the vibration elastic wave.

[0075] Step S650: Arrange the sampling points in the processed primary vibration elastic wave signal in sequence according to the time sequence to obtain the secondary vibration elastic wave signal.

[0076] In steps S610 to S650, by setting a fixed residual threshold, the signal residuals at each time step are compared with the residual threshold. This accurately identifies invalid background noise and valid signal components in the primary vibration elastic wave signal. Sampling points with signal residuals less than the residual threshold are identified as invalid background noise and set to zero, effectively filtering out background interference components attached to the valid signal. Simultaneously, this step fully preserves the original signal values ​​of sampling points with signal residuals greater than or equal to the residual threshold, preventing the erroneous removal of valid vibration information corresponding to collisions or scrapes. The processed sampling points are then rearranged according to the time sequence to obtain the secondary vibration elastic wave signal, simplifying the data volume of subsequent signal processing, further purifying the signal waveform, and eliminating the interference caused by invalid background noise.

[0077] Reference Figure 7 It is understood that step S150 may include, but is not limited to, the following steps: Step S710: Traverse the effective signal segments in the second-order vibration elastic wave signal segment by segment.

[0078] Step S720: Traverse the amplitude data of all sampling points within the current valid signal segment, compare the amplitude values ​​one by one, filter out the maximum amplitude data, and extract the signal peak value.

[0079] Step S730: Compare the amplitude data of each sampling point of the current effective signal segment with the threshold baseline one by one, count the duration of amplitude data that continuously exceeds the threshold baseline, and extract the signal width.

[0080] Step S740: Locate the start and end times of the current valid signal segment, count the time intervals before and after the valid signal segment where there are no valid signal segments, and extract the silent window.

[0081] Step S750: Integrate the extracted signal peak value, signal width, and silent window to form a multidimensional time domain feature.

[0082] In steps S710 to S750, by traversing the effective signal segments in the secondary vibration elastic wave signal segment by segment, the signal peak value, signal width, and silence window are extracted sequentially and integrated to form a multi-dimensional time-domain feature. This comprehensively depicts the waveform characteristics and temporal distribution of the effective signal segments from multiple different time-domain dimensions. The various features complement and support each other, which can comprehensively characterize the inherent characteristics of vehicle body vibration events and provide a rich and highly discriminative feature basis for subsequent scratch type determination. Among them, the signal peak value is used to characterize the amplitude of the vibration elastic wave signal, intuitively reflecting the severity of the external force impact on the vehicle body; the signal width is used to characterize the duration of the effective signal, which can distinguish between instantaneous impact and continuous scratch-type vibration behavior; the silence window is used to characterize the time interval between adjacent effective signal segments, which can help determine the independence and integrity of vibration events.

[0083] In step S710, all time-series sampling points of the vibration elastic wave secondary signal are traversed, and whether the signal value of the sampling point is zero is used as the dividing criterion; a group of adjacent sampling points that continuously present non-zero values ​​are defined as an independent effective signal segment, and the interval where the signal value is continuously zero is defined as the silent interval. In this way, all effective signal segments can be completely segmented and screened from the signal.

[0084] Reference Figure 8 It is understood that step S160 may include, but is not limited to, the following steps: Step S810: Preset peak value determination threshold, width determination threshold, and silent window determination threshold.

[0085] Step S820: When the signal peak value is greater than or equal to the peak value determination threshold, the signal width is greater than or equal to the width determination threshold, and the silence window is greater than or equal to the silence window determination threshold, it is determined as the first event result.

[0086] Step S830: When the signal peak value is less than the peak value determination threshold, or the signal width is less than the width determination threshold, or the silence window is less than the silence window determination threshold, it is determined as the second event result.

[0087] Step S840: Associate and match the first event result or the second event result with the corresponding multidimensional time-domain features and sampling timestamps to generate data record results.

[0088] Step S850: Display the data recording results on a preset electronic display screen.

[0089] In steps S810 to S850, by using preset peak value judgment thresholds, width judgment thresholds, and silent window judgment thresholds, and combining the three time-domain features of signal peak value, signal width, and silent window for joint logical judgment, the results of the first event and the second event can be standardized and distinguished, overcoming the limitations of single feature judgment and improving the rigor and discriminability of scratch type judgment. The event results, multi-dimensional time-domain features, and sampling timestamps are then correlated and matched to generate data record results, which are then visually displayed on a preset electronic display screen. This not only completely preserves the event-related feature data and time information, facilitating subsequent traceability analysis, but also provides real-time visualization of the recognition results.

[0090] It should be noted that the first event outcome is characterized by a real collision, malicious knocking, or tailgate closure—events with valid external force action that have security monitoring value and are considered genuine abnormal vehicle events that need to be identified and recorded. The second event outcome is characterized by natural environmental interference vibration events that have no security warning significance, such as minor scratches from bushes or tree branches, road bumps, or wind vibrations—and are considered invalid interference events that do not need to be judged as valid anomalies.

[0091] In a second aspect, this application also provides an electronic device, comprising: at least one memory; at least one processor; at least one program; the program being stored in the memory, and the processor executing the at least one program to implement the elastic wave vehicle collision and scrape recognition method as described in any embodiment of the first aspect.

[0092] In this electronic device, the original vibration elastic wave signal of the vehicle body and the environmental noise signal are continuously collected first. The original vibration elastic wave signal is sequentially filtered and smoothed to obtain the primary vibration elastic wave signal. Then, a threshold baseline is dynamically constructed based on the environmental noise signal. The difference between the primary vibration elastic wave signal and the threshold baseline is calculated to obtain the signal residual. Based on the signal residual, the primary vibration elastic wave signal is subjected to secondary filtering to obtain the secondary vibration elastic wave signal. Subsequently, feature extraction is performed on the secondary vibration elastic wave signal to obtain multi-dimensional time-domain features. Finally, the scrape type is determined based on the multi-dimensional time-domain features, and the corresponding data record result is generated and output. This scheme no longer uses a fixed judgment benchmark. It can dynamically adapt the threshold baseline by combining real-time environmental noise, and combine multi-layer signal processing and multi-dimensional time-domain feature comprehensive judgment to overcome the defects of traditional recognition methods, such as weak environmental adaptability and easy misjudgment and missed judgment. It can accurately distinguish between real collision scrape behavior and external environmental interference, and can improve the reliability and environmental adaptability of vehicle collision scrape recognition.

[0093] Memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs, non-transitory computer-executable programs, and signals, such as the program instructions / signals corresponding to the processing module in the embodiments of this application. The processor executes various functional applications and data processing by running the non-transitory software programs, instructions, and signals stored in the memory, thereby implementing the touch signal extraction method of the above-described method embodiments.

[0094] The memory may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data related to the aforementioned touch signal extraction method. Furthermore, the memory may include high-speed random access memory and non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, the memory may optionally include memory remotely located relative to the processor, and these remote memories can be connected to the processing module via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0095] One or more signals are stored in a memory, and when executed by one or more processors, the touch signal extraction method in any of the above method embodiments is performed.

[0096] Thirdly, embodiments of this application provide a computer-readable storage medium storing a computer program that is executed by one or more processors, enabling the one or more processors to perform the elastic wave vehicle collision and scrape recognition method in the above method embodiments.

[0097] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0098] Based on the above description of the embodiments, those skilled in the art will understand that all or some of the steps and systems in the methods disclosed above can be implemented as software, firmware, hardware, and suitable combinations thereof. Some or all physical components can be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit. Such software can be distributed on a computer-readable medium, which may include computer storage media (or non-transitory media) and communication media (or transient media). As is known to those skilled in the art, the term computer storage media includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable signals, data structures, program modules, or other data). Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disc (DVD) or other optical disc storage, magnetic cartridges, magnetic tape, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and is accessible by a computer. Furthermore, as is known to those skilled in the art, communication media typically contain computer-readable signals, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium.

[0099] It should be understood that in this application, "at least one (item)" means one or more, and "more than" means two or more. "And / or" is used to describe the relationship between related objects, indicating that three relationships can exist. For example, "A and / or B" can represent three cases: only A exists, only B exists, and both A and B exist simultaneously, where A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one (item) of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one (item) of a, b, or c can represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0100] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of the units described above is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0101] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0102] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0103] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes multiple instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing programs, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0104] The embodiments of this application have been described in detail above with reference to the accompanying drawings. However, this application is not limited to the above embodiments. Within the scope of knowledge possessed by those skilled in the art, various changes can be made without departing from the spirit of this application.

Claims

1. A method for identifying vehicle collisions and scrapes using elastic waves, characterized in that, include: Continuously collect raw vibration elastic wave signals of the vehicle body and environmental noise signals; The original vibration elastic wave signal is filtered and smoothed to obtain the first-order vibration elastic wave signal. A threshold baseline is dynamically constructed based on the environmental noise signal; The difference between the primary vibration elastic wave signal and the threshold baseline is calculated to obtain the signal residual. The primary vibration elastic wave signal is then subjected to secondary filtering based on the signal residual to obtain the secondary vibration elastic wave signal. Feature extraction is performed on the second-order vibration elastic wave signal to obtain multidimensional time-domain features; Based on the multidimensional time-domain features, the scraping type is determined, and the corresponding data record results are generated and output.

2. The elastic wave vehicle collision and scrape identification method according to claim 1, characterized in that, The process of filtering and smoothing the original vibration elastic wave signal to obtain a first-order vibration elastic wave signal includes: Obtain the set of valid physical addresses and valid sensor channels that have been pre-entered and stored on the local machine; The device physical address field and the sensor channel number field are extracted from the original vibration elastic wave signal; The device physical address field is compared one by one with the set of legal physical addresses, and the original vibration elastic wave signals that do not match the device physical address field are eliminated. The sensor channel number field of the original vibration elastic wave signal that matches the physical address field of the device is compared one by one with the set of valid sensor channels, and the original vibration elastic wave signals that do not match the sensor channel number field are eliminated. The original vibration elastic wave signals that have been screened and retained are sorted and organized according to the sampling timestamp, and isolated abnormal data points with timestamp jumps or abnormal sampling intervals are removed. The original vibration elastic wave signal after screening is smoothed and filtered to obtain the first-order vibration elastic wave signal.

3. The elastic wave vehicle collision and scrape identification method according to claim 2, characterized in that, The process of smoothing and filtering the filtered original vibration elastic wave signal to obtain the first-order vibration elastic wave signal includes: A sliding window is constructed according to a preset fixed time domain length, and each sampling point in the original vibration elastic wave signal after filtering is sequentially traversed according to the sliding window. Extract the amplitude data of all sampling points within the current sliding window, and remove outlier sampling points whose amplitudes exceed the preset amplitude range; Calculate the amplitude difference between the current sampling point and the adjacent previous sampling point and the adjacent next sampling point; If all the amplitude differences are less than the preset amplitude difference threshold, the interval where the current sampling point is located is determined to be a steady-state interval; If any of the amplitude differences is greater than or equal to the amplitude difference threshold, the interval where the current sampling point is located is determined to be a waveform abrupt change interval; Apply the complete sliding window to the steady-state interval and calculate the average amplitude of the remaining sampling points within the sliding window; For the waveform abrupt change interval, reduce the value length of the sliding window, and calculate the average amplitude of the sampling points within the reduced sliding window; Based on the average amplitude of the corresponding interval, the sampling point amplitude of the original vibration elastic wave signal is replaced point by point; The replaced sampling data are rearranged according to the sampling time sequence to obtain the first-order vibration elastic wave signal.

4. The elastic wave vehicle collision and scrape identification method according to claim 1, characterized in that, The step of dynamically constructing a threshold baseline based on the environmental noise signal includes: Get the preset segmented time length; According to the segmented time length obtained by matching, the environmental noise signal is continuously segmented and sampled to obtain multiple sets of environmental noise sampling data arranged in sequence. The amplitude distribution information of each group of environmental noise sampling data is statistically analyzed, and the current external environmental noise is classified into levels based on the amplitude distribution information of each group to obtain the environmental noise level; The corresponding baseline iteration step size is matched according to the environmental noise level, and the initial threshold baseline is dynamically adjusted by iterating point by point according to the time sequence based on the baseline iteration step size. When multiple consecutive sets of environmental noise sampling data correspond to the same environmental noise level, the iterative update of the threshold baseline is stopped. Furthermore, the upper limit and lower limit of the amplitude of the threshold baseline are preset to control the amplitude of the threshold baseline not to exceed the upper limit and lower limit during the iteration process. The current vehicle operating conditions are obtained, and the segmented time length is dynamically adjusted based on the current vehicle operating conditions and the ambient noise level.

5. The elastic wave vehicle collision and scrape identification method according to claim 4, characterized in that, The step of obtaining the current vehicle operating condition and dynamically adjusting the segmented time length based on the current vehicle operating condition and the ambient noise level includes: Obtain the current vehicle operating conditions; wherein, the vehicle operating conditions include driving conditions and parking conditions; Based on the preset weight allocation rules, corresponding operating condition weights and noise weights are allocated according to the current vehicle operating conditions and the ambient noise level, respectively. Based on the operating condition weight and the noise weight, a comprehensive weight value is obtained; The adjusted segmented time length is obtained by multiplying the comprehensive weight value with the basic segmented time length.

6. The elastic wave vehicle collision and scrape identification method according to claim 1, characterized in that, The step of performing a second-level filtering process on the first-level vibration elastic wave signal based on the signal residual to obtain a second-level vibration elastic wave signal includes: Preset a fixed residual threshold value; The signal residuals at each time step are compared one by one with the residual threshold values; When the signal residual is less than the residual threshold, the corresponding signal is determined to be invalid background noise, and the signal value of the sampling point corresponding to the first-order vibration elastic wave signal is set to 0. If the signal residual is greater than or equal to the residual threshold value, then the original signal value of the sampling point corresponding to the first-order vibration elastic wave signal is retained; The sampling points in the processed primary vibration elastic wave signal are rearranged sequentially according to the time sequence to obtain the secondary vibration elastic wave signal.

7. The elastic wave vehicle collision and scrape identification method according to claim 1, characterized in that, The feature extraction of the second-order vibration elastic wave signal yields multidimensional time-domain features, including: The effective signal segments in the second-order vibration elastic wave signal are traversed segment by segment. Traverse the amplitude data of all sampling points within the current effective signal segment, compare the amplitude values ​​one by one, filter out the maximum value of the amplitude data, and extract the signal peak value; The amplitude data of each sampling point of the current effective signal segment is compared with the threshold baseline one by one, and the duration of the amplitude data continuously exceeding the threshold baseline is counted to extract the signal width. Locate the start and end times of the current valid signal segment, count the time intervals before and after the valid signal segment where there are no valid signal segments, and extract the silent window. The extracted signal peak value, signal width, and silent window are integrated to form the multidimensional time-domain feature.

8. The elastic wave vehicle collision and scrape identification method according to claim 7, characterized in that, The step of determining the type of scraping based on the multidimensional time-domain features and generating and outputting the corresponding data record results includes: Preset peak value determination threshold, width determination threshold, and silent window determination threshold; If the signal peak value is greater than or equal to the peak value determination threshold, the signal width is greater than or equal to the width determination threshold, and the silence window is greater than or equal to the silence window determination threshold, then it is determined as the first event result. If the signal peak value is less than the peak value determination threshold, or the signal width is less than the width determination threshold, or the silence window is less than the silence window determination threshold, then it is determined to be the second event result; The first event result or the second event result is associated and matched with the corresponding multidimensional time-domain features and sampling timestamps to generate data record results; The data recording results are displayed on a preset electronic display screen.

9. An electronic device, characterized in that, include: At least one memory; At least one processor; At least one program; The program is stored in the memory, and the processor executes at least one of the programs to implement the elastic wave vehicle collision and scrape recognition method as described in any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer-executable program for performing the elastic wave vehicle collision and scrape identification method as described in any one of claims 1 to 8.

Images