CN120722359A - Active detection method and system for mountain road traps based on sound wave reflection - Google Patents
Active detection method and system for mountain road traps based on sound wave reflectionInfo
- Publication number
- CN120722359A CN120722359A CN202510989169.2A CN202510989169A CN120722359A CN 120722359 A CN120722359 A CN 120722359A CN 202510989169 A CN202510989169 A CN 202510989169A CN 120722359 A CN120722359 A CN 120722359A
- Authority
- CN
- China
- Prior art keywords
- data
- trap
- mountain road
- echo
- propagation path
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S15/00—Systems using the reflection or reradiation of acoustic waves, e.g. sonar systems
- G01S15/02—Systems using the reflection or reradiation of acoustic waves, e.g. sonar systems using reflection of acoustic waves
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S15/00—Systems using the reflection or reradiation of acoustic waves, e.g. sonar systems
- G01S15/02—Systems using the reflection or reradiation of acoustic waves, e.g. sonar systems using reflection of acoustic waves
- G01S15/04—Systems determining presence of a target
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
- G01S7/52—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S15/00
- G01S7/539—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S15/00 using analysis of echo signal for target characterisation; Target signature; Target cross-section
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Remote Sensing (AREA)
- Radar, Positioning & Navigation (AREA)
- General Physics & Mathematics (AREA)
- Computer Networks & Wireless Communication (AREA)
- Data Mining & Analysis (AREA)
- Acoustics & Sound (AREA)
- Theoretical Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Measurement Of Velocity Or Position Using Acoustic Or Ultrasonic Waves (AREA)
Abstract
The embodiment of the invention discloses a mountain road trap active detection method and system based on sound wave reflection, wherein the method comprises the following steps of collecting echo signals of multi-band detection sound waves reflected after multipath propagation in a mountain road area to be detected through a microphone array, and evaluating and scoring propagation paths according to a path reliability scoring mechanism; and if the trigger condition is not met, inputting the second characteristic data and the first characteristic data into a pre-trained machine learning model, and judging to obtain the trap type of the mountain area to be tested according to the machine learning model. The system has second-level response capability, greatly improves safety, realizes full-flow automation, reduces manual intervention, and greatly improves mountain road inspection efficiency.
Description
Technical Field
The embodiment of the invention relates to the technical field of road detection, in particular to a mountain road trap active detection method and system based on sound wave reflection, electronic equipment and a storage medium.
Background
At present, in the field of mountain road safety monitoring, a visual image recognition method is often adopted, a camera or an unmanned aerial vehicle is used for aerial image recognition, ground cracks and collapse areas are recognized, and then the images are automatically judged through image segmentation, edge detection and deep learning. However, visual image recognition methods have certain limitations, such as exposure to light, large influence of weather conditions, and difficulty in recognizing potential traps below the surface.
At present, acoustic wave detection is used for detecting highway traps such as pits, collapse, holes and the like, but is mainly focused on structural health monitoring, concrete defect detection, pavement disease identification and the like, but mountain road traps have some remarkable differences from highway traps in detection, such as great differences in terrain conditions, environmental noise, surface materials, trap types, signal attenuation modes and the like. Therefore, the existing sound wave highway detection method does not have trap recognition capability under complex terrains, cannot be suitable for unstructured complex mountain road terrains, and still belongs to the technical blank of active intelligent detection under mountain road environments.
Disclosure of Invention
In order to overcome the defects of the prior art, the embodiment of the invention aims to provide a mountain road trap active detection method, a mountain road trap active detection system and electronic equipment based on acoustic wave reflection, so that all-weather and all-terrain detection applicability is improved, trap detection coverage rate can be improved by using acoustic wave reflection in different angle directions, and more accurate, real-time and intelligent mountain road trap detection capability is realized.
In order to solve the above problems, a first aspect of the embodiments of the present invention discloses an active detection method for mountain traps based on acoustic wave reflection, which includes the following steps:
Collecting echo signals reflected back after multi-band detection sound waves are subjected to multi-path propagation in a mountain road area to be detected through a microphone array, and obtaining propagation paths of the echo signals;
Evaluating and scoring the propagation paths according to a path credibility scoring mechanism, and judging whether echo signals corresponding to the propagation paths when the evaluation scores are lower than a credibility threshold meet triggering conditions or not;
if the trigger condition is judged to be met, judging that a trap exists in the mountain road area to be detected, and carrying out risk rapid early warning;
If the triggering condition is judged not to be met, acquiring a propagation path when the evaluation score is greater than or equal to a credible threshold value, acquiring a credible propagation path, and extracting first characteristic data of echo signals corresponding to the credible propagation path;
identifying material data of a reflection point corresponding to the trusted propagation path, and encoding the material data to serve as second characteristic data;
inputting the second characteristic data and the first characteristic data into a pre-trained machine learning model, and judging and obtaining the trap category of the mountain road area to be tested according to the machine learning model, wherein the trap category comprises an instant trap, a potential trap and no trap.
Preferably, the extracting the first characteristic data of the echo signal corresponding to the trusted propagation path includes:
And acquiring topographic map data of the mountain road area to be detected, correcting echo signals corresponding to the trusted propagation path according to the topographic map data, and extracting characteristic data of the corrected echo signals to obtain first characteristic data.
Preferably, the identifying the material data of the reflection point corresponding to the trusted propagation path includes:
obtaining geological data of a mountain road area to be detected, and establishing an acoustic wave reflection database mapped with the geological data;
acquiring echo data corresponding to the reflection points;
And obtaining the material data of the reflection points according to the echo data and the acoustic wave reflection database.
Preferably, the identifying the material data of the reflection point corresponding to the trusted propagation path includes:
According to the time delay, frequency response and reflection intensity data of echo data corresponding to the obtained reflection point, inputting the time delay, frequency response and reflection intensity data into a pre-trained acoustic wave material classification model to obtain a first discrimination material of the reflection point;
acquiring laser point cloud echo data corresponding to the reflection point, and acquiring a second discrimination material of the reflection point according to the laser point cloud echo data;
And carrying out confidence fusion on the first discrimination material and the second discrimination material to obtain comprehensive discrimination materials, and taking the comprehensive discrimination materials as material data of reflection points.
Preferably, the inputting the second feature data and the first feature data into a pre-trained machine learning model, and determining the trap class of the mountain road area to be detected according to the machine learning model includes:
The machine learning model adopts a multi-layer perceptron MLP model, training data of the multi-layer perceptron MLP model adopts data marked with the combination of second characteristic data and first characteristic data and trap categories to be mapped as sample data, the second characteristic data and the first characteristic data are input into the pre-trained multi-layer perceptron MLP model, and the trap categories of the mountain road area to be tested are judged according to the multi-layer perceptron MLP model.
Preferably, the scoring the propagation path according to the path credibility scoring mechanism includes:
And constructing a scoring model, wherein the evaluation indexes of the scoring model comprise a time delay consistency score, an integrity score, a smoothness score, a reflection point stability score, a material consistency score and an echo repeatability score, and summing the scores of all the evaluation indexes to obtain an evaluation score.
Preferably, when the trap type is judged to be a potential trap, judging whether echo data corresponding to the trap area meets a risk threshold condition or not;
when judging that the risk threshold condition is met, acquiring each frame of echo data corresponding to the trap area in a set time period and the occurrence time corresponding to each frame of echo data;
And inputting the time feature sequence formed by the echo data of each frame and the occurrence time thereof into a trained risk prediction model, obtaining a risk probability sequence of the trap region in a preset future time period according to the risk prediction model, and carrying out risk early warning according to the risk probability sequence.
The second aspect of the embodiment of the invention discloses a mountain road trap active detection system based on acoustic wave reflection, which comprises the following components:
The acquisition unit is used for acquiring echo signals reflected by the multi-band detection sound waves after multipath propagation in the mountain area to be detected through the microphone array and acquiring propagation paths of the echo signals;
the evaluation unit is used for evaluating and scoring the propagation paths according to a path credibility scoring mechanism and judging whether echo signals corresponding to the propagation paths when the evaluation score is lower than a credibility threshold value meet triggering conditions or not;
the triggering unit is used for judging that the mountain road area to be detected has a trap if the triggering condition is judged to be met, and carrying out risk rapid early warning;
The characteristic unit is used for acquiring a propagation path when the evaluation score is larger than or equal to a credible threshold value if the triggering condition is judged not to be met, obtaining a credible propagation path and extracting first characteristic data of echo signals corresponding to the credible propagation path;
The identification unit is used for identifying the material data of the reflection point corresponding to the trusted propagation path, and encoding the material data to be used as second characteristic data;
The judging unit is used for inputting the second characteristic data and the first characteristic data into a pre-trained machine learning model, judging and obtaining the trap category of the mountain road area to be tested according to the machine learning model, wherein the trap category comprises an instant trap, a potential trap and no trap.
The third aspect of the embodiment of the invention discloses electronic equipment, which comprises a memory storing executable program codes, a processor coupled with the memory, and a processor calling the executable program codes stored in the memory and used for executing the mountain road trap active detection method based on acoustic wave reflection disclosed in the first aspect of the embodiment of the invention.
A fourth aspect of the embodiment of the present invention discloses a computer-readable storage medium storing a computer program, where the computer program causes a computer to execute the mountain road trap active detection method based on acoustic wave reflection disclosed in the first aspect of the embodiment of the present invention.
Compared with the prior art, the embodiment of the invention has the beneficial effects that:
According to the invention, the microphone array is used for collecting echo signals of the multi-band detection sound waves reflected after multipath propagation in the mountain road area to be detected, and obtaining the propagation paths of the echo signals, so that the multipath sound wave collecting capability of the complex mountain road environment can be realized, the multipath echo signals can be obtained under the environments of steep fluctuation of the mountain road and complex structure of the easy-to-shelter space, the detection coverage area is remarkably improved, meanwhile, the path reliability scoring mechanism is used for improving the identification accuracy, the most reasonable propagation paths are reserved for analysis, the misjudgment and missed judgment probability is remarkably reduced, and when the low path reliability is detected and the specific triggering condition is met, the high-risk trap can be immediately judged and early warning is sent out, the second-level response capability of the detection system is ensured, the safety is greatly improved, the system is judged by a machine learning model, the adaptability is high, the later maintenance cost is low, the full-flow automation is realized, the manual intervention is reduced, and the mountain road inspection efficiency is greatly improved.
Furthermore, after the trap is judged to be the potential trap, the risk is dynamically modeled by utilizing the change trend of multi-frame echo data in the time dimension through a time sequence analysis means, so that whether the potential trap has the trend of deteriorating, expanding and converting into the immediate trap can be effectively identified, the trend prejudgment is realized, the dynamic development track of the trap is reflected, the static detection is converted into the trend prediction, the early warning can be realized, the actual demand is more met, and the method is applicable to the early warning of geological disaster hidden dangers.
Drawings
FIG. 1 is a schematic flow chart of a mountain road trap active detection method based on acoustic wave reflection according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a mountain trap active detection system based on acoustic wave reflection according to an embodiment of the present invention;
Fig. 3 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
This detailed description is merely illustrative of the embodiments of the invention and is not intended to limit the embodiments of the invention, since modifications of the embodiments can be made by those skilled in the art without creative contribution as required after reading the specification, but are protected by the patent laws within the scope of the claims of the embodiments of the invention.
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by one of ordinary skill in the art without undue burden from the present disclosure, are intended to be within the scope of the embodiments of the present invention.
The terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
In embodiments of the invention, words such as "exemplary" or "such as" are used to mean serving as an example, instance, or illustration. Any embodiment or design described herein as "exemplary" or "e.g." in an embodiment should not be taken as preferred or advantageous over other embodiments or designs. Rather, the use of words such as "exemplary" or "such as" is intended to present related concepts in a concrete fashion.
Example 1
Referring to fig. 1-3, a control method of a mountain road trap active detection system based on acoustic wave reflection, as shown in fig. 1, includes the following steps:
s110, collecting echo signals reflected back after multi-band detection sound waves are subjected to multi-path propagation in a mountain area to be detected through a microphone array, and obtaining propagation paths of the echo signals;
in the step, the multi-band detection sound wave can use broadband sound waves of 20Hz-20kHz, the penetrability and detail resolution capability of signals are considered, the low-frequency sound wave can penetrate the road surface well, and the high-frequency sound wave provides finer reflection information, so that the accurate identification of different types of traps is facilitated.
In this step, since in an actual mountain road or irregular terrain, the reflection of the sound wave is not always direct, but may undergo multiple reflections or refractions. Therefore, the adoption of multipath propagation can help the system acquire echo information from different paths, and the accuracy and reliability of trap detection are further improved.
The microphone array is adopted, that is, a plurality of microphones simultaneously receive signals in different directions, multipath signals are distinguished, echo characteristics of different propagation paths are accurately analyzed, environmental noise and irrelevant sound interference, such as wind noise, traffic noise and the like, can be effectively restrained through a signal processing technology, and the definition and accuracy of the received echo signals can be improved.
In this embodiment, the propagation path of the echo signal includes a path from the transmitting point to the reflecting point and a path from the reflecting point to the receiving point, where the sound wave is transmitted from the transmitting source, reflected by rocks, soil, cavities, and the like, and the reflected sound wave continues to propagate, and is finally captured by the microphone array.
Step S120, evaluating and scoring the propagation paths according to a path credibility scoring mechanism, and judging whether echo signals corresponding to the propagation paths when the evaluation score is lower than a credibility threshold meet triggering conditions or not;
specifically, the scoring the propagation path according to the path credibility scoring mechanism includes:
And constructing a scoring model, wherein the evaluation indexes of the scoring model comprise a time delay consistency score, an integrity score, a smoothness score, a reflection point stability score, a material consistency score and an echo repeatability score, and summing the scores of all the evaluation indexes to obtain an evaluation score.
Specific evaluation indexes and score ranges and scoring criteria are as follows:
① And (3) scoring the consistency of the time delay, namely 0-25 minutes, and determining whether the measured time delay accords with the geometric propagation distance and the medium propagation speed, wherein the smaller the deviation is, the higher the score is.
② Signal integrity grading, 0-20 minutes, whether echo intensity and frequency spectrum characteristics are continuous and stable, high-frequency jitter and serious attenuation, and withholding can be judged by signal-to-noise ratio, frequency spectrum width and other metrics.
③ And (3) grading the patency of the path, namely grading from 0 to 20 minutes, simulating whether the path passes through a shielding area or a terrain mutation area, specifically modeling the terrain and simulating the propagation of sound waves, and if the path is shielded in the simulation, such as being blocked by rocks/large-angle slopes, reducing the grading.
④ And (5) grading the stability of the reflection points, and judging whether the multi-angle echoes of the same target area are concentrated to the same reflection point or not according to 0-15 points. If multiple paths in the same area reflect at a certain point, the reflection point is trusted, and the score is high.
⑤ And (3) scoring the consistency of the materials, namely scoring 0-10 points, and judging whether the inferred materials in the propagation path are matched with the reflection types or not, if the attenuation is stronger due to the cavity, buckling.
⑥ Echo repeatability scores, 0-10 points, whether the path appears stably in multiple detection, whether the path is an accidental path, high reliability of high-frequency repeaters and low accidental path score.
In specific implementation, the trusted threshold can be set empirically, for example, 70 minutes, the trusted path is that the total score is equal to or more than 70 minutes, and only echo signals on the trusted path participate in trap judgment.
And for the echo signals corresponding to the propagation paths when the evaluation score is lower than the credible threshold, judging whether the triggering condition is met or not.
Step S130, if the trigger condition is judged to be met, judging that a trap exists in the mountain area to be detected, and carrying out risk rapid early warning;
In this embodiment, although the path is not trusted for echo signals corresponding to propagation paths when the evaluation score is below the trusted threshold, the corresponding echo signals may still have dangerous signs, such as sudden strong attenuation, extreme frequency reflection, and multipath convergence.
And whether the triggering condition is met or not is judged by utilizing the echo signals of the unreliable propagation paths, so that the risk of error trap leakage caused by imperfect scoring mechanism can be reduced, the response speed can be improved, and the dangerous signals can trigger warning at the first time without completely depending on the reliable paths and the machine learning model judgment.
As an embodiment, the trigger condition may be determined by setting a quick trigger determination rule.
For example, if it is determined that the trigger condition is satisfied, determining that a mountain area to be detected has a trap, and performing risk fast warning includes:
And when the evaluation score is lower than the credible threshold value, the echo signals corresponding to the propagation paths are judged to meet any one or more of the following conditions, namely triggering risk rapid early warning.
Condition 1 echo frequency concentration low frequency (< 300 Hz) and the material is "cavity", which may be a collapsed or suspended area.
Condition 2, abrupt delay extension and reduced reflection intensity, which indicates that the acoustic wave propagation path is abnormally elongated and may traverse an abnormal structure.
Condition 3. Multiple untrusted paths reflect in the same area, with the multipath concentrated pointing to an outlier.
And 4, abnormal echo time delay change, wherein the time delay of the echo suddenly changes greatly and exceeds the normal terrain change range, and the abnormal cavity, crack or underground water flow and the like can be indicated to be penetrated by the signal. At this time, a threshold value of the delay change rate is set, and when the delay change rate exceeds the set threshold value, an abnormality alarm is triggered.
The above trigger conditions are taken as examples, and the trigger conditions are not listed here.
As another embodiment, if it is determined that the trigger condition is satisfied, determining that a mountain area to be detected has a trap, and performing risk fast warning may specifically include:
Step S1301, dividing the regional risk level according to the topography of the mountain road region to be detected, the history of occurrence risk or accident and the known high-risk region, wherein the regional risk level comprises high risk, medium risk and low risk;
step S1302, obtaining a region where a reflection point corresponding to a propagation path is located when the evaluation score is lower than a credible threshold value, and judging the risk level of the region;
Step S1303, when the area is a high risk area, counting the number of propagation paths when the evaluation score is lower than a credible threshold, and when the evaluation score is greater than or equal to the number threshold, directly triggering risk rapid early warning;
When the area is in the middle risk area, waveform analysis is carried out on the echo signal, and when obvious asymmetry of the reflected waveform is detected, the risk early warning is triggered;
When the area is in the low risk area, the risk rapid early warning is not triggered.
In this department, when the system triggers the quick early warning of risk, need not to wait for complete analysis flow, can export early warning signal immediately to realize higher response speed and stronger emergent prevention and control ability.
Step 140, if the triggering condition is not met, acquiring a propagation path when the evaluation score is greater than or equal to a credible threshold value, obtaining a credible propagation path, and extracting first characteristic data of echo signals corresponding to the credible propagation path;
In this embodiment, echo signal characteristics corresponding to the trusted propagation path are analyzed, and characteristic data is extracted from the trusted propagation path for trap judgment.
Specifically, the first feature data may specifically include the following:
Echo delay, which is the time reflecting the propagation of sound waves, is commonly used to determine distance and depth.
Echo intensity is the amplitude of the reflected wave, and can reflect the characteristics of the reflection surface.
Echo frequency-the main frequency characteristic of the echo, can help to distinguish material types, such as cavities, rocks, etc.
Attenuation, the loss of energy of an acoustic wave during propagation, is generally related to the nature of the propagation medium.
Echo duration, the length of time of the echo, reflects the complexity of the signal reflection.
In specific implementation, when the first characteristic data of the echo signal corresponding to the trusted propagation path is extracted, the method can include the steps of extracting information such as time delay and intensity by carrying out envelope detection on the echo signal, extracting spectrum characteristics by using fast Fourier transform, and calculating main frequency, bandwidth, spectrum energy and the like of the echo signal so as to obtain the corresponding first characteristic data.
In this embodiment, the method is used in complicated environments such as mountain roads, where the reflection characteristics of different materials directly affect the identification of potential traps, for example, echo signals show longer delay and attenuation characteristics, and after the identification of materials, the echo signals are determined to be cavities or groundwater layers, so that potential traps can be identified, and meanwhile, the trap identification model can use the physical characteristics to more accurately classify and determine the characteristics of the materials of the reflection points as input, so as to improve the precision and accuracy of trap determination.
As another embodiment, the extracting the first characteristic data of the echo signal corresponding to the trusted propagation path includes:
step S1401, obtaining topographic map data of a mountain area to be detected, and correcting echo signals corresponding to the trusted propagation path according to the topographic map data;
In this step, the topographic map data may be obtained from digital topographic elevation map DEM data, satellite topographic model data/laser radar point cloud data, and may specifically include data such as altitude, elevation value, slope information, a topographic relief map, and the like. Because the sound wave propagates in the mountain region and is not an ideal straight line, refraction, shielding, path bending and other phenomena can occur due to relief of the topography, the topography data can be used for adjusting the path propagation distance and propagation time estimation, and the geometric accuracy of the propagation path is improved.
In this step, the correction of the echo signal may specifically include correction of propagation time, correction of reflection point position, correction of reflection angle, incidence angle, and shading compensation.
For example, according to elevation data, it is identified which areas are likely to have an occlusion, when there is an obstacle occlusion, a possible diffraction path is calculated, and the time delay and intensity of the echo signal are adjusted. Multipath propagation compensation, namely, when an obstacle is blocked, calculating possible diffraction paths and adjusting the time delay and the intensity of echo signals. For example, the reflection angle of the acoustic wave is calculated from the gradient and slope data displayed by the topographic map data, and the intensity and form of the actual echo are corrected.
Step S1402, extracting the characteristic data of the modified echo signal to obtain first characteristic data.
In this step, the extracted feature data may include time features, amplitude features, frequency domain features, shape features, and spatial consistency data.
In this embodiment, after the echo data corresponding to the trusted propagation path is fused with the topographic information, the echo signal of the trusted path is corrected, for example, after the propagation time is corrected, the echo signal is extracted, so as to obtain a standardized feature vector set, that is, the first feature data.
In the implementation process, by correcting the echo signal of the trusted path, misjudgment caused before correction can be avoided, for example, when correction is not performed on an uphill reflection point, the echo delay can be underestimated, and thus the echo delay is misjudged as a 'closer reflection' or a 'shallow structure'. For example, when there is a height difference or a shade, data which is not corrected is input into a machine learning model for judgment, and false trap alarm is easily caused.
Step S150, identifying material data of a reflection point corresponding to the trusted propagation path, and encoding the material data as second characteristic data;
in this step, the material type of the identified reflection points, such as rock, loose soil, holes, vegetation, etc., is converted into a numerical form or a vector form processed by the machine learning model so as to be input as a model together with the first characteristic data.
As an embodiment, the identifying the material data of the reflection point corresponding to the trusted propagation path includes:
Step S1501, obtaining geological data of a mountain road area to be detected, and establishing an acoustic wave reflection database mapped with the geological data;
In this step, the geological data may be obtained by satellite remote sensing image analysis, or may be obtained from a public geological database, such as geological bureau data, geological map, remote sensing DEM data, etc. The geological data comprises data of earth surface type, lithology distribution, stratum structure, hydrologic condition and the like. And establishing a one-to-one mapping relation between the geological types and the acoustic wave propagation and reflection characteristics to form an acoustic wave reflection database for comparison.
Specifically, the acoustic wave reflection database is formed by establishing a mapping relation between each common material such as rock, sand, cavity, water body, vegetation layer and typical acoustic wave reflection parameters such as reflection coefficient, spectral characteristics, attenuation rate and the like.
Step S1502, acquiring echo data corresponding to the reflection points;
In this step, for each of the trusted propagation paths, an endpoint, i.e., a reflection point, is determined, echo signals at the path endpoint are extracted, echo raw data associated with the reflection point is recorded, and data including echo delay, echo intensity, dominant frequency and frequency bandwidth, echo energy attenuation ratio, waveform symmetry, duration, and the like are extracted as echo data corresponding to the reflection point.
And step S1503, obtaining the material data of the reflection points according to the echo data and the acoustic wave reflection database.
In the step, the characteristics of the actually collected echo data of the reflecting points are compared with the characteristics of different materials in the acoustic wave reflecting database, the most similar material types are matched, and the identification of the materials of the reflecting points is realized.
Specifically, a material type with the smallest euclidean distance can be selected as the recognition result of the reflection point through similarity calculation.
As another embodiment, the identifying the material data of the reflection point corresponding to the trusted propagation path may include:
Step S15011, according to the time delay, frequency response and reflection intensity data of echo data corresponding to the obtained reflection point, inputting the time delay, frequency response and reflection intensity data into a pre-trained acoustic wave material classification model to obtain a first discrimination material of the reflection point;
In this step, the acoustic material classification model may use a conventional classifier, such as an SVM, a decision tree, a random forest, or a lightweight neural network, to output a first discrimination material and a first confidence value thereof by performing sample training on a data set in which acoustic reflection feature vectors and material types form a map.
Step S15012, acquiring laser point cloud echo data corresponding to the reflection point, and acquiring a second discrimination material of the reflection point according to the laser point cloud echo data;
In this step, the laser point cloud echo data includes reflection intensity, morphological characteristics and laser multi-echo structure data, the laser radar system is used to obtain the spatial structure and reflection intensity information of the reflection point, and the laser material model is used to infer the material type, so as to obtain the second discrimination material and the second confidence coefficient thereof.
And step S15013, performing confidence fusion on the first discrimination material and the second discrimination material to obtain a comprehensive discrimination material, and taking the comprehensive discrimination material as material data of the reflection point.
In this step, the confidence levels of the first and second discrimination materials are compared, and a discrimination material corresponding to the discrimination material having a high confidence level is selected as the material data of the reflection point. And the recognition results of the acoustic wave channel and the laser channel are fused, and the final material judgment is output, so that the accuracy of the system is improved. In this embodiment, through dual-channel identification of the acoustic wave and the laser point cloud, accuracy and robustness of material judgment can be improved.
Step S160, inputting the second characteristic data and the first characteristic data into a pre-trained machine learning model, and judging and obtaining the trap category of the mountain road area to be tested according to the machine learning model, wherein the trap category comprises an instant trap, a potential trap and no trap.
Specifically, step S160 may include:
The machine learning model adopts a multi-layer perceptron MLP model, training data of the multi-layer perceptron MLP model adopts data marked with the combination of second characteristic data and first characteristic data and trap categories to be mapped as sample data, the second characteristic data and the first characteristic data are input into the pre-trained multi-layer perceptron MLP model, and the trap categories of the mountain road area to be tested are judged according to the multi-layer perceptron MLP model.
In specific implementation, the input layer of the multi-layer perceptron MLP model includes data after the first feature and the second feature are spliced. The first layer of the hidden layer adopts 64 neurons, the second layer adopts 32 neuron output layers and comprises 3 neurons, which respectively represent 'instant trap', 'potential trap', 'no trap'. The activation function uses the ReLU activation function to introduce a nonlinear transformation, softmax for the probability distribution of the output layer.
When the MLP model of the multilayer perceptron trains data, the first characteristic and the second characteristic are spliced and then input into the model, the model optimizes weight through a gradient descent algorithm, the model is adjusted according to the training data, the generalization capability of the model is estimated by using a cross verification technology, and the effect of the model is estimated by using the accuracy.
The mapping relationship between the combination of the second feature data and the first feature in the sample data and the trap class can be referred to as the following table 1:
TABLE 1 relation table of combinations of second feature data and first feature in sample data and trap class mapping
Optionally, the method of the present invention further includes step 170, and step 170 may specifically include:
Step 1701, when the trap type is judged to be a potential trap, judging whether echo data corresponding to the trap area meets a risk threshold condition or not;
In this embodiment, the risk level is further evaluated for the area determined to be "potential trap", and the object to be continuously monitored and predicted is screened out.
In this step, the risk threshold condition may be a condition that satisfies a certain policy model, for example, classification by a classifier, or may be a simple setting condition, for example, when the echo delay change rate of the continuous frame increases by more than 18% of the set value, which indicates that the underground structure may continue to evolve, and continuous monitoring and prediction are required. For example, a variance of the energy attenuation fluctuation amplitude greater than a certain set value, such as greater than 0.05, indicates that the material changes or cracks are enlarged, and continuous monitoring and prediction are required.
Step 1702, when it is determined that the risk threshold condition is met, acquiring echo data of each frame corresponding to the trap area in a set time period and occurrence time corresponding to the echo data of each frame;
In this step, a frame sequence generated when echo detection is performed on the mountain road to be detected is acquired, for example, once every 5 seconds, and each frame of echo signal generated by the trap area in a set period of time, for example, the last 10 minutes, including characteristic data such as time delay, frequency response, reflection intensity, etc., is recorded, and a time stamp is acquired, where the unit may be accurate to ms.
Step 1703, inputting a time feature sequence formed by each frame of echo data and the occurrence time thereof into a trained risk prediction model, and obtaining a risk probability sequence of the trap region in a preset future time period according to the risk prediction model;
in this step, the risk prediction model may be an LSTM long-term memory network model, and the input of the model may include echo data of each frame corresponding to the trap in the set time period, that is, feature data and time feature data of a multi-frame historical echo signal.
For example, the multi-frame historical echo signal may be 20 frames of data, each frame including characteristic data such as time delay, frequency response, reflection intensity, etc., and the time characteristic data is characteristic data such as a relative time interval or time frequency converted from a corresponding time stamp.
The output of the model is a sequence of risk probabilities for the trap region for a preset future time period, which may be set to the next 5 minutes or 10 minutes in an implementation.
For example, when the preset future time period is 10 minutes, a risk probability sequence with the length of 10 is output, wherein each value epsilon [0,1] represents the probability of occurrence of a trap in the minute.
And step 1704, performing risk early warning according to the risk probability sequence.
In this step, the time feature sequence formed by each frame of echo data and the occurrence time thereof is input into a trained risk prediction model, modeling is performed on the time sequence, and the static detection is converted into trend prediction, so that early warning can be realized, the actual requirements are met, and quantitative management and progressive warning are supported.
Specifically, performing risk early warning according to the risk probability sequence may include:
and 17041, setting an instant trap threshold, comparing the risk probability value in the risk probability sequence with the instant trap threshold, and carrying out risk early warning of different levels according to the comparison result.
And 17042, carrying out red early warning when the risk probability value in the occurrence risk probability sequence is greater than or equal to the instant trap threshold value, and carrying out orange early warning when the difference value between the risk probability value in the occurrence risk probability sequence and the instant trap threshold value exceeds a preset difference value.
For example, the immediate trap threshold is set to 0.8.
If the risk probability sequence in the next 5 minutes of the trap area is output as [0.62,0.68,0.74,0.81,0.85], the risk gradually rises and gradually approaches the threshold value of 'instant trap' in the next 5 minutes. And the risk probability values are 0.81 and 0.85, which are larger than 0.8, and red early warning is carried out.
If the risk probability sequence within the next 5 minutes of the trap area is output to be [0.62,0.68,0.74,0.76,0.79], the risk probability rate is 0.79, the difference value between the risk probability value in the risk probability sequence and the immediate trap threshold exceeds a preset difference value (for example, 0.02), the serious risk is approached, and orange warning is entered.
In actual implementation, the trap position can be fed back, for example, by combining inertial navigation or GPS data, the detected trap position is subjected to space coordinate calibration, and the trap position, the trap type and the early warning level are fed back in real time through a visual interface or an early warning system, so that the active identification and early warning of the mountain road potential trap are realized.
Example two
The embodiment of the invention discloses an active detection system of a mountain road trap based on acoustic wave reflection, as shown in fig. 2, fig. 2 is an active detection system of a mountain road trap based on acoustic wave reflection, comprising:
The acquisition unit 210 is configured to acquire echo signals reflected back after multipath propagation of the multi-band detection sound wave in the mountain area to be detected through the microphone array, and acquire propagation paths of the echo signals;
an evaluation unit 220, configured to evaluate and score the propagation paths according to a path reliability scoring mechanism, and determine whether an echo signal corresponding to the propagation path when the evaluation score is lower than a reliability threshold meets a triggering condition;
The triggering unit 230 is configured to determine that a trap exists in the mountain area to be detected and perform risk fast early warning if the triggering condition is determined to be satisfied;
a feature unit 240, configured to obtain a propagation path when the evaluation score is greater than or equal to the trusted threshold if the trigger condition is determined not to be satisfied, obtain a trusted propagation path, and extract first feature data of an echo signal corresponding to the trusted propagation path;
an identifying unit 250, configured to identify material data of a reflection point corresponding to the trusted propagation path, and encode the material data as second feature data;
The judging unit 260 is configured to input the second feature data and the first feature data into a pre-trained machine learning model, and judge, according to the machine learning model, a trap class of the mountain area to be tested, where the trap class includes an immediate trap, a potential trap and a no trap.
Optionally, the mountain path trap active detection system based on acoustic wave reflection of the present invention further includes:
The prediction unit 270 is configured to determine whether echo data corresponding to the trap area meets a risk threshold condition when it is determined that the trap type is a potential trap, obtain each frame of echo data corresponding to the trap area in a set time period and an occurrence time corresponding to each frame of echo data when it is determined that the risk threshold condition is met, input a time feature sequence formed by each frame of echo data and the occurrence time thereof into a trained risk prediction model, obtain a risk probability sequence of the trap area in a preset future time period according to the risk prediction model, and perform risk early warning according to the risk probability sequence.
Example III
Referring to fig. 3, fig. 3 is a schematic structural diagram of an electronic device according to an embodiment of the invention. As shown in fig. 3, the electronic device may include:
a memory 310 in which executable program code is stored;
A processor 320 coupled to the memory 310;
The processor 320 invokes executable program codes stored in the memory 310 to perform some or all of the steps in the active detection method for mountain road traps based on acoustic wave reflection in the first embodiment.
The embodiment of the invention discloses a computer readable storage medium storing a computer program, wherein the computer program enables a computer to execute part or all of the steps in a mountain road trap active detection method based on acoustic wave reflection in the first embodiment.
The embodiment of the invention also discloses a computer program product, wherein when the computer program product runs on a computer, the computer is caused to execute part or all of the steps in the mountain road trap active detection method based on acoustic wave reflection in the first embodiment.
The embodiment of the invention also discloses an application release platform, wherein the application release platform is used for releasing a computer program product, and when the computer program product runs on a computer, the computer is caused to execute part or all of the steps in the mountain road trap active detection method based on acoustic wave reflection in the first embodiment.
In various embodiments of the present invention, it should be understood that the size of the sequence numbers of the processes does not mean that the execution sequence of the processes is necessarily sequential, and the execution sequence of the processes should be determined by the functions and internal logic thereof, and should not constitute any limitation on the implementation process of the embodiments of the present invention.
The units described as separate units may or may not be physically separate, and units shown as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the embodiment.
In addition, each functional unit in the embodiments of the present invention may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units may be implemented in hardware or in software functional units.
The integrated units, if implemented in the form of software functional units and sold or used as stand-alone products, may be stored in a computer-accessible memory. Based on this understanding, the technical solution of the present invention, or a part contributing to the prior art or all or part of the technical solution, may be embodied in the form of a software product stored in a memory, comprising several requests for a computer device (which may be a personal computer, a server or a network device, etc., in particular may be a processor in a computer device) to execute some or all of the steps of the method according to the embodiments of the present invention.
In the embodiments provided herein, it should be understood that "B corresponding to a" means that B is associated with a, from which B can be determined. It should also be understood that determining B from a does not mean determining B from a alone, but may also determine B from a and/or other information.
Those of ordinary skill in the art will appreciate that some or all of the steps of the various methods of the described embodiments may be implemented by hardware associated with a program that may be stored in a computer-readable storage medium, including Read-Only Memory (ROM), random-access Memory (Random Access Memory, RAM), programmable Read-Only Memory (Programmable Read-Only Memory, PROM), erasable programmable Read-Only Memory (Erasable Programmable Read-Only Memory, EPROM), one-time programmable Read-Only Memory (One-time Programmable Read-Only Memory, OTPROM), electrically erasable programmable Read-Only Memory (EEPROM), compact disc Read-Only Memory (Compact Disc Read-Only Memory, CD-ROM), or other optical disc Memory, magnetic disc Memory, tape Memory, or any other medium capable of being used to carry or store data that is readable by a computer.
The above describes the mountain road trap active detection method, device, electronic equipment and storage medium based on acoustic wave reflection in detail, and specific examples are applied to illustrate the principle and implementation of the present invention, and the above description of the embodiment is only used to help understand the method and core idea of the present invention, and meanwhile, to those skilled in the art, according to the idea of the present invention, the changes in the specific implementation and application range will be apparent, so that the content of the present invention should not be interpreted as limiting the present invention.
Claims (10)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202510989169.2A CN120722359A (en) | 2025-07-17 | 2025-07-17 | Active detection method and system for mountain road traps based on sound wave reflection |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202510989169.2A CN120722359A (en) | 2025-07-17 | 2025-07-17 | Active detection method and system for mountain road traps based on sound wave reflection |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN120722359A true CN120722359A (en) | 2025-09-30 |
Family
ID=97166524
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202510989169.2A Pending CN120722359A (en) | 2025-07-17 | 2025-07-17 | Active detection method and system for mountain road traps based on sound wave reflection |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN120722359A (en) |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107334205A (en) * | 2017-07-19 | 2017-11-10 | 防城港市港口区晶通科技有限公司 | One kind is based on the anti-lost hole chemical industry Labor protection shoes of sonic detection |
| CN110991507A (en) * | 2019-11-22 | 2020-04-10 | 北京中科蓝图科技有限公司 | Road underground cavity identification method, device and system based on classifier |
| CN115220115A (en) * | 2022-09-21 | 2022-10-21 | 江苏筑升土木工程科技有限公司 | Road void online detection system and detection method |
| US20220397666A1 (en) * | 2021-06-11 | 2022-12-15 | Robert Bosch Gmbh | Ultrasonic system and method for classifying obstacles using a machine learning algorithm |
| CN118038686A (en) * | 2024-02-24 | 2024-05-14 | 张南雨 | A road condition detection method based on ultrasonic array and deep learning |
| CN118269997A (en) * | 2022-12-30 | 2024-07-02 | 华为技术有限公司 | A road surface detection method and related device |
| CN119131514A (en) * | 2024-11-13 | 2024-12-13 | 中国矿业大学(北京) | A method for detecting open-pit mine disasters based on drones |
| CN119672914A (en) * | 2024-12-19 | 2025-03-21 | 辽宁工程技术大学 | A road collapse monitoring and early warning system and early warning method |
| CN119915352A (en) * | 2025-04-02 | 2025-05-02 | 贵州中南锦天科技有限责任公司 | Inspection data monitoring method and server for early warning of roads and bridges in mountainous areas |
-
2025
- 2025-07-17 CN CN202510989169.2A patent/CN120722359A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107334205A (en) * | 2017-07-19 | 2017-11-10 | 防城港市港口区晶通科技有限公司 | One kind is based on the anti-lost hole chemical industry Labor protection shoes of sonic detection |
| CN110991507A (en) * | 2019-11-22 | 2020-04-10 | 北京中科蓝图科技有限公司 | Road underground cavity identification method, device and system based on classifier |
| US20220397666A1 (en) * | 2021-06-11 | 2022-12-15 | Robert Bosch Gmbh | Ultrasonic system and method for classifying obstacles using a machine learning algorithm |
| CN115220115A (en) * | 2022-09-21 | 2022-10-21 | 江苏筑升土木工程科技有限公司 | Road void online detection system and detection method |
| CN118269997A (en) * | 2022-12-30 | 2024-07-02 | 华为技术有限公司 | A road surface detection method and related device |
| CN118038686A (en) * | 2024-02-24 | 2024-05-14 | 张南雨 | A road condition detection method based on ultrasonic array and deep learning |
| CN119131514A (en) * | 2024-11-13 | 2024-12-13 | 中国矿业大学(北京) | A method for detecting open-pit mine disasters based on drones |
| CN119672914A (en) * | 2024-12-19 | 2025-03-21 | 辽宁工程技术大学 | A road collapse monitoring and early warning system and early warning method |
| CN119915352A (en) * | 2025-04-02 | 2025-05-02 | 贵州中南锦天科技有限责任公司 | Inspection data monitoring method and server for early warning of roads and bridges in mountainous areas |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Bianco et al. | Convective boundary layer depth: Improved measurement by Doppler radar wind profiler using fuzzy logic methods | |
| CN116028852B (en) | A method for detecting underground road fire based on BP neural network and D-S evidence theory | |
| CN120196909B (en) | A high-cut slope stability assessment method based on multi-source data fusion | |
| CN119902230B (en) | Sea fog detection method and device based on satellite-borne laser radar | |
| CN119148129A (en) | Pipeline periphery soil body detection method and system based on ground penetrating radar | |
| CN111222743A (en) | A method for judging vertical offset distance and threat level of optical fiber sensing events | |
| CN120491207A (en) | Remote sensing-based investigation and mapping method and surveying instrument thereof | |
| CN121053747B (en) | A Big Data-Based Early Warning System for Marine Algal Blooms | |
| Heck et al. | Automatic detection of avalanches combining array classification and localization | |
| Guérin et al. | Tsunami detection by high-frequency radar in British Columbia: performance assessment of the time-correlation algorithm for synthetic and real events | |
| CN120065229A (en) | Vehicle detection method and system of millimeter wave radar | |
| CN120781017B (en) | System and method for dynamically identifying isolated mature convection cloud cluster by combining satellite and radar | |
| Rachmadan et al. | Developing ground motion prediction models for West Java: A machine learning approach to support Indonesia's earthquake early warning system | |
| CN120932414A (en) | Road side slope shallow level displacement monitoring method, system, device, equipment and medium | |
| CN120722359A (en) | Active detection method and system for mountain road traps based on sound wave reflection | |
| CN120831665A (en) | A radar-sensing highway monitoring data processing system and method | |
| Zhang et al. | Automatic identification of diverse tunnel threats with machine learning–based distributed acoustic sensing | |
| CN120229279B (en) | Distributed fiber optic acoustic sensing track safety monitoring system | |
| CN118671714B (en) | Radar data acquisition method and system combined with scene error analysis and electronic equipment | |
| CN118707633A (en) | A method for observation and identification of dynamic signals triggered by strong convection | |
| Wang et al. | Exploring magnitude estimation for earthquake early warning using available P-wave time windows based on Chinese strong-motion records | |
| Loftis et al. | A deep learning algorithmic approach to develop a video inundation monitoring system | |
| CN118837907B (en) | Thunderstorm warning generation method and device based on quantum radar | |
| CN120278611B (en) | A method for constructing a multi-dimensional social group activity monitoring model | |
| CN120561828B (en) | Method and device for classifying vibration events of ship-to-bridge collision |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |