Real-time tracking and early warning system for rockfall trajectory based on radar monitoring

Abstract

The application discloses a real-time tracking and early warning system for rockfall trajectory based on radar monitoring, comprising: an acquisition module, which acquires radar data and image data of rockfall in a monitoring area, and acquires terrain data and road center line data of the monitoring area through a terrain database; an analysis module, which analyzes the radar data and image data, generates shape data of rockfall, and outputs a rolling index; performs kinematic analysis on the radar data, the rolling index and the terrain data, generates predicted trajectory data, and acquires predicted position information; generates a threat value based on the predicted trajectory data and the road center line data; and an early warning module, which sets different real-time sampling frequencies according to different threat values; when the radar data exceeds an early warning threshold, a rockfall warning is generated; the application classifies tracking according to threat levels, saves computing resources, and thus improves the accuracy and reliability of rockfall disaster warning.

Description

Technical Field

[0001] This invention relates to the field of radar monitoring and early warning technology, and in particular to a real-time tracking and early warning system for rolling stone trajectories based on radar monitoring. Background Technology

[0002] Rockfall hazards are a common geological disaster in mountainous areas, characterized by their suddenness and destructive power, posing a serious threat to the safety of roadsides, railway lines, and residential areas. Traditional rockfall monitoring methods mostly rely on manual patrols or fixed sensor monitoring, which suffer from limited monitoring range, poor real-time performance, and untimely early warnings.

[0003] In recent years, with the development of radar technology, it has become possible to monitor rolling stones using radar. Chinese Patent Publication No. CN119044960A discloses an adaptive constant false alarm rate (CFAR) rolling stone detection method. This method divides the reference window of the target unit into regions and then evaluates the echo type within the reference window using skewness and uniformity judgments. Skewness judgment determines whether the selected region's reference unit contains interfering targets; uniformity judgment assesses whether different regions' reference units contain clutter edges. Based on the evaluation results of the echoes within the reference window, this invention adaptively employs different threshold strategies to generate a CFAR detection threshold. This invention can effectively detect targets such as rolling stones even when clutter edges or interfering targets are present.

[0004] However, since multiple rocks may fall simultaneously within the monitoring area, monitoring multiple rockfall targets at the same time places high demands on the radar system and requires significant costs. This has hindered the widespread adoption of radar systems for rockfall monitoring on slopes.

[0005] Therefore, how to achieve real-time and accurate tracking and early warning of multiple rockfall trajectories under limited computing power is a technical problem that urgently needs to be solved in the field of rockfall disaster monitoring. Summary of the Invention

[0006] The purpose of this invention is to address the problems of insufficient accuracy and timeliness of early warning in existing rockfall monitoring systems, as well as poor data incompleteness handling capabilities. This invention provides a real-time tracking and early warning system for rockfall trajectories based on radar monitoring. This system can acquire rockfall data in real time, accurately predict rockfall trajectories, and provide graded early warnings according to the degree of threat. It can also effectively handle situations where rockfall image data is incomplete, thereby improving the accuracy and reliability of rockfall disaster early warning.

[0007] To address the aforementioned technical problems, this invention provides the following technical solution: a real-time tracking and early warning system for rolling stone trajectories based on radar monitoring, comprising: an acquisition module for acquiring radar data and image data of rolling stones within a monitoring area, and acquiring and storing terrain data and road centerline data of the monitoring area through a terrain database; an analysis module for analyzing the image data to generate the shape data of the rolling stones, and generating a rolling index of the rolling stones through a rolling analysis strategy; performing kinematic analysis on the radar data, rolling index, and terrain data to generate predicted trajectory data and acquire predicted location information; generating a threat value based on the predicted trajectory data and road centerline data; and an early warning module for setting different real-time sampling frequencies for the radar according to different threat values; generating a rolling stone alarm when the threat value exceeds an early warning threshold.

[0008] Preferably, the acquisition module includes: a millimeter-wave radar unit for transmitting and receiving frequency-modulated continuous wave signals and converting the frequency-modulated continuous wave signals into radar data; the radar data includes the real-time position and speed of the rolling stones; an image acquisition unit for acquiring image data of the rolling stones; and an environmental acquisition unit for retrieving terrain data and road centerline data of the monitoring area from the terrain database; the terrain data includes digital elevation model data, slope friction coefficient data, and obstacle vector data; the obstacle vector data includes the position, height, and outline information of the obstacles.

[0009] Preferably, the rolling analysis strategy includes the following steps: performing edge detection on the image data to obtain a binary edge image; marking the connected components of the binary edge image and extracting the connected components as the contour data of the rolling stones; generating the contour area and contour perimeter using the Gaussian area formula based on the contour data, and extracting the equivalent diameter and roundness of each rolling stone; constructing a two-dimensional coordinate system based on the contour data, obtaining the axis-aligned bounding box of the contour by calculating the difference between the minimum and maximum values ​​on the horizontal and vertical axes, and using the ratio of the length to the width of the axis-aligned bounding box as the aspect ratio of the rolling stone; and generating the rolling index of the rolling stone using a preset rolling algorithm based on the equivalent diameter, roundness, and aspect ratio.

[0010] Preferably, the image acquisition unit further includes a contour completion subunit, used to complete the binary region split by the rolling stone due to occlusion into a complete connected region; specifically, it includes the following steps: Step a, matching all connected regions in the image data of two adjacent frames one by one, and calculating the motion vector of each connected region; if the difference in motion vectors between two connected regions is less than a preset vector threshold, and the centroid distance between the two connected regions is less than a preset centroid distance threshold, then the two connected regions are determined to be fragments of the same rolling stone; Step b, taking the major axis direction of the overall minimum bounding rectangle of the fragment group as the main direction, expanding the edge points of each connected region in the fragment group to both sides of the main direction by one pixel width to generate a closed contour; Step c, filling the closed contour to obtain a single complete connected region after completion, and replacing the original fragment group.

[0011] Preferably, the following sub-step is included between step b and step c: calling the slope of the monitored area in the terrain database, and interpolating to generate a virtual edge consistent with the slope in the missing area of ​​the fragment group.

[0012] Preferably, the analysis module includes: an initial setting unit, which sets the initial state of the predicted trajectory based on the real-time position and velocity data of the rolling stone; a physical model unit, which generates the future position sequence of the rolling stone using a kinematic algorithm based on the initial state, rolling index, digital elevation model data, and friction coefficient data; and a trajectory output unit, which generates predicted trajectory data based on the future position sequence; the predicted trajectory data includes a position sequence and a corresponding timestamp; wherein the position sequence includes the predicted position.

[0013] Preferably, the analysis module includes: a distance calculation unit, which calculates the minimum distance value using the Euclidean distance algorithm based on the predicted trajectory data and the preset road centerline data; a speed acquisition unit, which acquires the speed value of the rolling stone based on the position sequence and corresponding timestamp in the predicted trajectory data; and a threat generation unit, which extracts the maximum value after normalization of the minimum distance value and the speed value as the threat value.

[0014] Preferably, the early warning module further includes: a frequency setting unit, which sets the sampling frequency of the millimeter-wave radar based on the threat value using a linear generation strategy; wherein the linear generation strategy includes: the sampling frequency is positively correlated with the threat value, the maximum threat value corresponds to the maximum sampling frequency, the minimum threat value corresponds to the minimum sampling frequency, and intermediate threat values ​​are calculated to generate corresponding sampling frequencies through linear interpolation; and a frequency adjustment unit, which outputs a frequency adjustment signal to the millimeter-wave radar based on the sampling frequency to adjust the data acquisition frequency.

[0015] Preferably, the early warning module further includes: a threshold storage unit for storing preset attention thresholds, early warning thresholds, and emergency thresholds; wherein the attention thresholds, early warning thresholds, and emergency thresholds are arranged in ascending order; a comparison unit for comparing the threat value sequentially with the attention thresholds, early warning thresholds, and emergency thresholds; if the threat value is greater than the attention threshold but less than or equal to the early warning threshold, an attention signal is generated; if the threat value is greater than the early warning threshold but less than or equal to the emergency threshold, an early warning signal is generated; if the threat value is greater than the emergency threshold, an emergency signal is generated; and an early warning output unit for outputting the attention signal, early warning signal, or emergency signal.

[0016] Preferably, the early warning module includes: a deviation calculation unit, which obtains the real-time position of the current rolling stone, extracts the predicted position of the corresponding time from the predicted trajectory data, calculates the difference between the real-time position and the predicted position of the current rolling stone, and generates a deviation value; a deviation comparison unit, which generates a correction trigger signal if the deviation value exceeds a preset deviation threshold; and a trajectory update unit, which drives the trajectory output unit to update and replace the predicted trajectory data based on the correction trigger signal.

[0017] The beneficial effects of this invention are as follows: First, the real-time tracking and early warning system for rolling stones based on radar monitoring provided by this invention quantifies and evaluates the shape, movement state, and relative position of the rolling stones to the road. It dynamically adjusts the radar sampling frequency according to the threat level, increasing the sampling frequency when the threat level is high to ensure the real-time performance and accuracy of data at critical moments, and decreasing the sampling frequency when the threat level is low to save system resources and improve system operating efficiency. Second, the real-time tracking and early warning system for rolling stones based on radar monitoring provided by this invention effectively addresses the problem of incomplete image data caused by occlusion of rolling stones through the rolling analysis strategy and contour completion subunit in the image acquisition unit. By completing the fragmented areas, it ensures the integrity and accuracy of the rolling stone shape data, thereby improving the reliability of the rolling index and trajectory prediction. Third, the real-time tracking and early warning system for rolling stones based on radar monitoring provided by this invention, through deviation calculation and trajectory update mechanisms, can monitor the deviation between the actual and predicted positions of the rolling stones in real time, and promptly correct the predicted trajectory when the deviation exceeds a threshold, ensuring the real-time performance and accuracy of the prediction and effectively addressing the uncertainties in the movement of rolling stones. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the basic structure of the real-time tracking and early warning system for rolling stone trajectories based on radar monitoring provided by the present invention. Detailed Implementation

[0019] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0020] Please see Figure 1 The present invention provides a real-time tracking and early warning system for rolling stone trajectories based on radar monitoring, characterized in that it includes: an acquisition module, an analysis module, and an early warning module.

[0021] The acquisition module is used to acquire radar and image data of rolling stones within the monitoring area, and to acquire and store terrain data and road centerline data of the monitoring area through a terrain database. The radar data provides real-time location and speed information of the rolling stones, the image data is used to analyze the shape characteristics of the rolling stones, and the terrain and road centerline data provide environmental background information for rolling stone trajectory prediction and threat assessment.

[0022] The analysis module analyzes image data to generate the shape data of the rolling stones and, through a rolling analysis strategy, generates the rolling index. It then performs kinematic analysis on radar data, the rolling index, and terrain data to generate predicted trajectory data and obtain predicted location information. Based on the predicted trajectory data and road centerline data, it generates a threat value. This module is the core of the system, responsible for transforming raw data into meaningful prediction and evaluation results.

[0023] The early warning module is used to set different real-time sampling frequencies for the radar based on different threat values; when the threat value exceeds the early warning threshold, a rolling stone alarm is generated. This module dynamically adjusts the system's operating mode based on the threat values ​​generated by the analysis module and promptly issues early warning information.

[0024] In a preferred embodiment, the acquisition module includes: a millimeter-wave radar unit, an image acquisition unit, and an environment acquisition unit.

[0025] The millimeter-wave radar unit transmits and receives frequency-modulated continuous wave (FMCH) signals and converts them into radar data. This radar data includes the real-time position and velocity of the rolling stone. Millimeter-wave radar is capable of penetrating fog, rain, snow, and other adverse weather conditions, enabling all-weather, high-precision acquisition of the rolling stone's motion information.

[0026] The image acquisition unit is used to acquire image data of the rolling stones. The image data can provide visual information about the rolling stones, which is used to identify features such as the shape and size of the stones. These features are crucial for assessing the rolling behavior of the stones.

[0027] The environmental acquisition unit retrieves terrain data and road centerline data for the monitored area from the terrain database. The terrain data includes digital elevation model (DEM) data, slope friction coefficient data, and obstacle vector data. The obstacle vector data includes the location, height, and outline information of obstacles. The terrain database can be pre-built and stores detailed geographic information of the monitored area, providing fundamental data for boulder kinematic analysis. The DEM data provides surface elevation information, the slope friction coefficient data reflects the frictional characteristics between the boulder and the ground surface, and the obstacle vector data is used to identify potential collision points along the boulder's path.

[0028] In a preferred embodiment, the rolling analysis strategy includes the following steps: performing edge detection on the image data to obtain a binary edge image; labeling connected components in the binary edge image; and extracting the connected components as the contour data of the rolling stones. Edge detection can employ operators such as the Canny operator or the Sobel operator to separate the rolling stones from the background in the image. Connected component labeling algorithms (such as the Two-pass algorithm) are used to identify individual rolling stones in the image.

[0029] Based on the contour data, the contour area and perimeter are generated using the Gaussian area formula, and the equivalent diameter and roundness of each stone are extracted. The Gaussian area formula (also known as the shoelace formula) can calculate the area of ​​the contour based on the vertex coordinates. Roundness measures how close an object's shape is to a circle, and is usually defined as 4π × area / perimeter squared. The equivalent diameter can be calculated from the area, for example, (4 × area / π)¹ / ².

[0030] Based on the contour data, a two-dimensional coordinate system is constructed. By calculating the difference between the minimum and maximum values ​​on the horizontal and vertical axes, the axis-aligned bounding box of the contour is obtained. The ratio of the length to the width of the axis-aligned bounding box is used as the aspect ratio of the rolling stone. The aspect ratio is an indicator of how slender an object is.

[0031] Based on the equivalent diameter, roundness, and aspect ratio, a rolling index is generated using a pre-defined rolling algorithm. The rolling index is a comprehensive indicator used to quantify the rolling tendency of the stone. The pre-defined rolling algorithm can be an empirical formula or a machine learning model, for example:

[0032] Rolling index = w1 × roundness + w2 × (1 - aspect ratio) + w3 × equivalent diameter; where w1, w2, and w3 are weighting coefficients determined based on practical experience or training data. The higher the roundness, the closer the aspect ratio is to 1 (i.e., the closer to a circle), and the larger the equivalent diameter, the stronger the rolling tendency of the stone, and the higher the rolling index.

[0033] Specifically, the rolling algorithm is constructed in the following ways: (1) Collect no less than 200 sets of historical rolling stone sample data, each set of data including roundness, aspect ratio, equivalent diameter and measured rolling distance; (2) Train the model using a support vector machine regression model, with RBF kernel function, penalty parameter C=100 and kernel parameter γ=0.1; (3) Normalize the output rolling index to the [0,1] interval, the larger the value, the stronger the rolling tendency.

[0034] For example, for granite boulders, typical model parameters are: w1=0.45, w2=0.35, w3=0.20; for sandstone boulders, w1=0.30, w2=0.50, w3=0.20. These parameters were calibrated through field experiments under the following conditions: slope 30°, friction coefficient 0.5, and boulder diameter 0.2~1.2 m.

[0035] In a preferred embodiment, the image acquisition unit further includes a contour completion subunit, used to complete the binary regions split by occlusion of the rolling stone into complete connected regions. Specifically, this includes the following steps: Step a, matching all connected regions within the image data of two adjacent frames one by one, and calculating the motion vector of each connected region; if the difference in motion vectors between two connected regions is less than a preset vector threshold, and the centroid distance between the two connected regions is less than a preset centroid distance threshold, then the two connected regions are determined to be fragments of the same rolling stone. Through inter-frame matching, multiple image regions belonging to the same rolling stone but split due to occlusion can be identified. The motion vector difference and centroid distance are the basis for determining whether they are fragments of the same rolling stone.

[0036] Step b: Using the major axis of the smallest bounding rectangle of the fragment group as the main direction, extend the edge points of each connected region within the fragment group by one pixel on each side of the main direction to generate a closed contour. By extending the edges outward, contour loss caused by occlusion can be compensated.

[0037] Step c: Fill the closed contour to obtain a complete, single connected region, and replace the original fragment group. The filling operation can use a flooding fill algorithm, marking all regions inside the closed contour as rolling stone regions, thus forming a complete connected region.

[0038] In a preferred embodiment, the following sub-step is included between step b and step c: The slope of the monitored area is retrieved from the terrain database, and a virtual edge consistent with the slope is generated by interpolation within the missing area of ​​the fragment group. The virtual edge uses linear interpolation: elevation is calculated at 0.1 m intervals per pixel along the slope direction. This sub-step utilizes terrain information to assist in contour completion, making the completed contour more consistent with the shape of the rolled stone on the actual terrain, thus improving the accuracy of the completion. For example, during the rolling of a stone, its contact surface with the ground may exhibit a specific shape due to the slope; by introducing slope information, the contour of the missing part can be more reasonably inferred.

[0039] In a preferred embodiment, the analysis module includes: an initial setting unit, a physical model unit, a trajectory output unit, a distance calculation unit, a velocity acquisition unit, a threat generation unit, a frequency setting unit, a frequency adjustment unit, a threshold storage unit, a comparison unit, and a warning output unit. The initial setting unit sets the initial state of the predicted trajectory based on the real-time position and velocity data of the rolling stone. This real-time position and velocity data is provided by the millimeter-wave radar unit.

[0040] The physical model unit, based on initial state, rolling index, digital elevation model (DEM) data, and friction coefficient data, generates the future position sequence of the rolling stones using a kinematic algorithm (the core algorithm of which is based on the kinematic calculation formulas in the thesis "Zhou You, Experimental and Numerical Simulation Study on the Motion Characteristics of Rolling Stones on High and Steep Slopes," [Sichuan University] Master's Thesis, 2023), which models the sliding, rolling, and collision stages of the rolling stones respectively). The kinematic algorithm can employ the Discrete Element Method (DEM), Finite Element Method (FEM), or a rigid body dynamics model based on Newtonian mechanics. This model considers factors such as gravity, friction, and collision forces, combined with the shape of the rolling stones (represented by the rolling index), terrain undulations (DEM), and surface friction characteristics (friction coefficient), to simulate the movement path of the rolling stones over a future period.

[0041] Specifically, the steps to generate the future position sequence of the rolling stone are as follows: 1. Initial state input: initial position Centroid coordinates taken from the real-time output of the millimeter-wave radar unit; initial velocity Obtained by differential analysis of the same radar data; slope inclination angle From Digital Elevation Model (DEM) First-order difference calculation at the point; coefficient of sliding friction Rolling friction coefficient Normal recovery coefficient Tangential recovery coefficient Directly refer to the median value of the corresponding interval for "granite slab" in Tables 3.4 to 3.6 ( =0.50, =0.40, =0.50, =0.73); equivalent radius of the rolling stone With quality Obtained from the "equivalent diameter" and "volume × density" output by the image acquisition unit; moment of inertia. Approximation of a sphere calculate.

[0042] Tables 3.4, 3.5, and 3.6 are shown below:

[0043] Slope surface types coefficient of sliding friction Granite slab surface, concrete surface 0.40 - 0.60 gravel accumulation slope 0.55 - 0.70 dense gravel slope, hard soil slope 0.55 - 0.85 Loose gravel slopes and soft soil slopes 0.50 - 0.85

[0044] Table 3.4: Recommended values ​​for sliding friction coefficient

[0045] Slope features rolling friction coefficient Granite slab surfaces, paved surfaces, shotcrete surfaces, and other work surfaces 0.3~0.6 Soft rock surface, strongly weathered hard rock surface 0.4~0.6 Rock pile slope 0.55~0.70 The slope is characterized by dense gravel accumulation, hard soil, and well-developed vegetation, primarily shrubs. 0.55~0.85 Slopes with dense gravel or hard soil, devoid of vegetation or with only a few weeds. 0.50~0.75 On loose gravel slopes and soft soil slopes, vegetation is mainly shrubs. 0.50~0.85 Soft soil slope, with no vegetation or only a few weeds 0.50~0.85

[0046] Table 3.5: Recommended values ​​for rolling friction coefficient

[0047] Slope features Normal recovery coefficient Tangential recovery coefficient Granite slab surfaces, paved surfaces, shotcrete surfaces 0.25~0.75 0.55~0.91 Soft rock surface, strongly weathered hard rock surface 0.15~0.37 0.53~0.90 Rock piles covered with grass and shrubs 0.12~0.33 0.30~0.85 dense gravel slope, hard soil slope 0.12~0.32 0.65~0.95 Loose gravel slopes and soft soil slopes 0.10~0.25 0.30~0.80 Soft or loose soil layers 0.10~0.30 0.50~0.80

[0048] Table 3.6: Recommended Values ​​for Collision Recovery Coefficient

[0049] II. Determining the Sliding Phase and Calculating Acceleration: If directional velocity components Less than 0.1 m / s and slope angle If the angle is less than 45°, it is determined to be a sliding state, and formula 3-4 in the paper is used: ; For sliding acceleration, For gravity; update speed along the slope direction: ,in, Current time, time step =0.05s.

[0050] III. Rolling Phase Judgment and Acceleration Calculation: When the roundness of the rolling stone is greater than 0.8, and greater than the critical roll angle ( When the time is reached, it enters the rolling state, using formula 3-14 in the paper: ; It is rolling acceleration; The equivalent rolling friction arm is set to 0.05R (the experimental calibration value in the paper), where R is the equivalent radius of the rolling stone; the speed update method is the same as that in the sliding stage.

[0051] IV. Collision Phase Speed ​​Correction: When Displacement increment in direction If the value is less than 0 and the contact discrimination function C(x,y) (C(x,y)=DEM elevation−y) is less than or equal to 0, a collision is triggered, and the velocity is immediately corrected. ; The negative sign indicates that the normal velocity is reversed.

[0052] V. Position Integration: For each time step Δt, use explicit Euler integration: ; The process iterates 100 times to obtain the future position sequence within 5 seconds, which is then output as the predicted trajectory data to the trajectory output unit. The trajectory output unit generates the predicted trajectory data based on the future position sequence. The predicted trajectory data includes the position sequence and its corresponding timestamp. The position sequence includes the predicted position. The distance calculation unit calculates the minimum distance value using the Euclidean distance algorithm based on the predicted trajectory data and the preset road centerline data. Euclidean distance is the straight-line distance between two points; here, it calculates the minimum distance from a point on the predicted trajectory to the road centerline. The speed acquisition unit obtains the speed value of the rolling stone based on the position sequence and corresponding timestamp in the predicted trajectory data. The speed can be calculated by dividing the distance between adjacent predicted position points by the time interval. The threat generation unit extracts the maximum value after normalizing the minimum distance value and the speed value as the threat value. Normalization converts data of different dimensions to a unified scale, such as the [0, 1] interval.

[0053] For example, the normalization method can use the Min-Max approach: Distance factor = (d_max - d) / (d_max - d_min), where the maximum distance d_max = 50 m and the minimum distance d_min = 1 m; Speed ​​factor = (v - v_min) / (v_max - v_min), where the maximum speed v_max = 20 m / s and the minimum speed v_min = 0.5 m / s; Threat value = max(distance factor, speed factor), with a threshold range of [0,1]. When the threat value is greater than 0.7, an emergency signal is triggered. The smaller the minimum distance value, the greater the threat; the larger the speed value, the greater the threat. By taking the maximum value, it ensures that if any one factor reaches a high threat level, the overall threat value is high.

[0054] The frequency setting unit sets the sampling frequency of the millimeter-wave radar based on the threat value using a linear generation strategy. This linear generation strategy includes: the sampling frequency is positively correlated with the threat value; the maximum threat value corresponds to the maximum sampling frequency, the minimum threat value corresponds to the minimum sampling frequency, and intermediate threat values ​​are calculated using linear interpolation to generate their corresponding sampling frequencies. The frequency adjustment unit outputs a frequency adjustment signal to the millimeter-wave radar based on the sampling frequency to adjust the data acquisition frequency. This invention can rationally utilize the system's computing resources, avoid insufficient computing power, and improve the system's tracking and early warning efficiency. The threshold storage unit stores preset attention thresholds, early warning thresholds, and emergency thresholds. These thresholds are arranged in ascending order. They are set according to the actual application scenario and security requirements. The comparison unit compares the threat value sequentially with the attention threshold, early warning threshold, and emergency threshold. If the threat value is greater than the attention threshold but less than or equal to the early warning threshold, an attention signal is generated; if the threat value is greater than the early warning threshold but less than or equal to the emergency threshold, an early warning signal is generated; if the threat value is greater than the emergency threshold, an emergency signal is generated. The early warning output unit outputs attention signals, warning signals, or emergency signals. These signals can trigger different early warning mechanisms, such as audible and visual alarms, SMS notifications, and traffic control.

[0055] In a preferred embodiment, the early warning module includes: a deviation calculation unit, a deviation comparison unit, and a trajectory update unit.

[0056] The deviation calculation unit acquires the real-time position of the rolling stone, extracts the predicted position for the corresponding time from the predicted trajectory data, and calculates the difference between the real-time position and the predicted position to generate a deviation value. The real-time position is provided by the millimeter-wave radar unit. The deviation comparison unit generates a correction trigger signal if the deviation value exceeds a preset deviation threshold. The preset deviation threshold is used to determine whether the predicted trajectory deviates significantly from the actual trajectory. The trajectory update unit, based on the correction trigger signal, drives the trajectory output unit to update and replace the predicted trajectory data. When there is a large deviation between the predicted trajectory and the actual trajectory, the system will promptly use the latest real-time position data to re-predict the trajectory, ensuring the accuracy of the prediction.

[0057] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media containing computer-usable program code. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0058] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A real-time tracking and early warning system for rolling stone trajectories based on radar monitoring, characterized in that, include: The acquisition module acquires radar and image data of the rolling stones within the monitoring area, and obtains and stores terrain data and road centerline data of the monitoring area through a terrain database; the analysis module analyzes the image data to generate the shape data of the rolling stones, and generates the rolling index of the rolling stones through a rolling analysis strategy; it performs kinematic analysis on the radar data, rolling index, and terrain data to generate predicted trajectory data and obtain predicted location information; based on the predicted trajectory data and road centerline data, it generates a threat value. The early warning module sets different real-time sampling frequencies for the radar based on different threat values; when the threat value exceeds the early warning threshold, a rolling stone alarm is generated.

2. The real-time tracking and early warning system for rolling stone trajectories based on radar monitoring according to claim 1, characterized in that, The acquisition module includes: a millimeter-wave radar unit for transmitting and receiving frequency-modulated continuous wave signals and converting the frequency-modulated continuous wave signals into radar data; the radar data includes the real-time position and speed of the rolling stones; an image acquisition unit for acquiring image data of the rolling stones; and an environmental acquisition unit for retrieving terrain data and road centerline data of the monitoring area from the terrain database; the terrain data includes digital elevation model data, slope friction coefficient data, and obstacle vector data; the obstacle vector data includes the position, height, and outline information of the obstacles.

3. The real-time tracking and early warning system for rolling stone trajectories based on radar monitoring according to claim 1, characterized in that, The rolling analysis strategy includes the following steps: edge detection is performed on the image data to obtain a binary edge image; connected component labeling is performed on the binary edge image, and the connected components are extracted as the contour data of the rolling stones; based on the contour data, the contour area and contour perimeter are generated using the Gaussian area formula, and the equivalent diameter and roundness of each rolling stone are extracted; based on the contour data, a two-dimensional coordinate system is constructed, and the axis-aligned bounding box of the contour is obtained by calculating the difference between the minimum and maximum values ​​on the horizontal and vertical axes, and the ratio of the length to the width of the axis-aligned bounding box is used as the aspect ratio of the rolling stone; based on the equivalent diameter, roundness, and aspect ratio, the rolling index of the rolling stone is generated using a preset rolling algorithm.

4. The real-time tracking and early warning system for rolling stone trajectories based on radar monitoring according to claim 2, characterized in that, The image acquisition unit also includes a contour completion subunit, used to complete the binary regions split by the occlusion of the rolling stone into complete connected regions; specifically, it includes the following steps: Step a, matching all connected regions in the image data of two adjacent frames one by one, and calculating the motion vector of each connected region; if the difference in motion vectors between two connected regions is less than a preset vector threshold, and the centroid distance between the two connected regions is less than a preset centroid distance threshold, then the two connected regions are determined to be fragments of the same rolling stone; Step b, taking the major axis direction of the overall minimum bounding rectangle of the fragment group as the main direction, expanding the edge points of each connected region in the fragment group by one pixel width to both sides of the main direction to generate a closed contour; Step c, filling the closed contour to obtain a single complete connected region after completion, and replacing the original fragment group.

5. The real-time tracking and early warning system for rolling stone trajectories based on radar monitoring according to claim 4, characterized in that, Between step b and step c, the following sub-step is also included: calling the slope of the monitored area in the terrain database, and interpolating to generate a virtual edge consistent with the slope in the missing area of ​​the fragment group.

6. The real-time tracking and early warning system for rolling stone trajectories based on radar monitoring according to claim 3, characterized in that, The analysis module includes: an initial setting unit, which sets the initial state of the predicted trajectory based on the real-time position and velocity data of the rolling stone; a physical model unit, which generates the future position sequence of the rolling stone through a kinematic algorithm based on the initial state, rolling index, digital elevation model data, and friction coefficient data; and a trajectory output unit, which generates predicted trajectory data based on the future position sequence. The predicted trajectory data includes a position sequence and a corresponding timestamp; wherein the position sequence includes the predicted position.

7. The real-time tracking and early warning system for rolling stone trajectories based on radar monitoring according to claim 6, characterized in that, The analysis module includes: a distance calculation unit, which calculates the minimum distance value using the Euclidean distance algorithm based on the predicted trajectory data and the preset road centerline data; a speed acquisition unit, which acquires the speed value of the rolling stone based on the position sequence and corresponding timestamp in the predicted trajectory data; and a threat generation unit, which extracts the maximum value after normalization of the minimum distance value and the speed value as the threat value.

8. The real-time tracking and early warning system for rolling stone trajectories based on radar monitoring according to claim 7, characterized in that, The early warning module further includes: a frequency setting unit, which sets the sampling frequency of the millimeter-wave radar based on the threat value using a linear generation strategy; wherein the linear generation strategy includes: the sampling frequency is positively correlated with the threat value, the maximum threat value corresponds to the maximum sampling frequency, the minimum threat value corresponds to the minimum sampling frequency, and intermediate threat values ​​are calculated to generate corresponding sampling frequencies through linear interpolation; and a frequency adjustment unit, which outputs a frequency adjustment signal to the millimeter-wave radar based on the sampling frequency to adjust the data acquisition frequency.

9. The real-time tracking and early warning system for rolling stone trajectories based on radar monitoring according to claim 8, characterized in that, The early warning module further includes: a threshold storage unit for storing preset attention thresholds, early warning thresholds, and emergency thresholds; wherein the attention thresholds, early warning thresholds, and emergency thresholds are arranged in ascending order; a comparison unit for comparing the threat value sequentially with the attention thresholds, early warning thresholds, and emergency thresholds; if the threat value is greater than the attention threshold but less than or equal to the early warning threshold, an attention signal is generated; if the threat value is greater than the early warning threshold but less than or equal to the emergency threshold, an early warning signal is generated; if the threat value is greater than the emergency threshold, an emergency signal is generated; and an early warning output unit for outputting an attention signal, an early warning signal, or an emergency signal.

10. The real-time tracking and early warning system for rolling stone trajectories based on radar monitoring according to claim 6, characterized in that, The early warning module includes: a deviation calculation unit, which obtains the real-time position of the rolling stone, extracts the predicted position of the corresponding time from the predicted trajectory data, calculates the difference between the real-time position and the predicted position of the rolling stone, and generates a deviation value; a deviation comparison unit, which generates a correction trigger signal if the deviation value exceeds a preset deviation threshold; and a trajectory update unit, which drives the trajectory output unit to update and replace the predicted trajectory data based on the correction trigger signal.

Images