An image sequence-based method for extracting blast shock wave pressure

Abstract

The application discloses a kind of based on video image's shock wave wave front automatic detection and pressure extraction method: step 1: input explosion shock wave propagation video frame sequence, explosion center coordinates, survey line angle and pixel size;Step 2: extract the gray scale of each frame image;Step 3: adjacent frame gray scale is carried out interframe difference processing, and binary difference image is generated;Step 4: with explosion center as center, detect initial wave front point;Step 5: initial wave front point is filtered by level;Step 6: wave front point after filtering is fitted;The intersection of specified survey line and fitted wave front curve and its distance from explosion center are calculated;The instantaneous velocity of shock wave propagation is calculated, and wave front propagation velocity is fitted;Step 7: the shock wave pressure of each frame on test angle is calculated.The application realizes the full-automatic detection, tracking and pressure extraction of shock wave wave front, improves analysis efficiency and accuracy, and is suitable for damage assessment, explosion power analysis field.

Description

Technical Field

[0001] This invention belongs to the field of image processing and explosion power testing technology, specifically relating to a method for extracting explosion shock wave pressure based on image sequences. Background Technology

[0002] In weapon damage effect assessment, the peak overpressure of the shock wave is a key indicator for evaluating the explosive power. Traditional methods mainly rely on pressure sensors for single-point measurements, and the obtained pressure-time curves can be used to calculate the peak overpressure, barometric time, and specific impulse. However, sensor methods can only obtain discrete information from a limited number of locations and cannot intuitively reflect the continuous propagation process of the shock wave in space.

[0003] In recent years, with the development of high-speed camera technology, it has become possible to directly observe the propagation of shock wave fronts through video images. In existing technologies, manually annotating the wavefront position frame by frame is a common method, but it suffers from drawbacks such as low efficiency, high subjectivity, and difficulty in standardization. Furthermore, using inter-frame difference extraction to extract moving regions still has shortcomings in noise suppression and maintaining wavefront continuity, severely affecting the accuracy of pressure calculations.

[0004] Therefore, there is an urgent need for a complete technical solution that can automatically and accurately detect the shock wave front, track its propagation trajectory, and extract pressure from explosion videos. Summary of the Invention

[0005] To address the problems of low computational accuracy and low efficiency in existing technologies, this invention provides an automatic detection and pressure extraction method for shock wave fronts based on video images. The aim is to achieve fully automatic, end-to-end processing from video sequences to wave front geometry, propagation speed, and relative pressure, thereby improving accuracy and efficiency.

[0006] To achieve the above objectives, the present invention adopts the following technical solution:

[0007] A method for extracting the pressure of an explosion shock wave based on an image sequence includes the following steps: Step 1: The user inputs the video frame sequence of the explosion shock wave propagation, the coordinates of the explosion center, the survey line angle, and the pixel size; Step 2: Read the video frame sequence of the explosion shock wave propagation and extract the grayscale image of each frame in the video frame sequence; Step 3: Perform inter-frame difference processing on the grayscale images of adjacent frames to obtain a difference image, and binarize the difference image to generate a binary difference image; Step 4: Detection of wavefront points using ray tracing: With the explosion center as the center, emit rays along multiple angular directions, search for foreground points in the binary difference image on each ray, and detect the initial wavefront points; Step 5: Perform hierarchical filtering on the detected initial wavefront points, including sequential radius percentage filtering, radius anomaly filtering, and local anomaly point filtering, to obtain the filtered wavefront points. Step 6: Fit the filtered wavefront points in the spatial dimension to obtain the fitted continuous wavefront curve; calculate the coordinates of the intersection point of the specified survey line and the fitted continuous wavefront curve, and calculate the distance between the intersection point and the explosion center; use the central difference method to calculate the instantaneous velocity of the shock wave propagation; fit the instantaneous velocity in the time dimension to obtain a stable wavefront propagation velocity. Step 7: Calculate the shock wave pressure at each frame at the test angle based on the wavefront propagation velocity.

[0008] Compared with the prior art, the present invention has the following advantages: (1) Fully automatic processing: From raw video input to wavefront detection, wavefront anomaly filtering, velocity calculation, and pressure calculation, no manual intervention is required, which greatly improves analysis efficiency.

[0009] (2) Strong robustness: Through methods such as ray neighborhood search, multi-level point filtering (percentage filtering, radius anomaly filtering, local anomaly filtering), and wavefront fitting, noise and fragment interference are effectively suppressed.

[0010] In summary, compared with the prior art, the method of the present invention improves efficiency, reduces reliance on manual labor, and effectively suppresses noise, thereby improving the accuracy of pressure calculation. Attached Figure Description

[0011] Figure 1 This is an overall flowchart of the method of the present invention.

[0012] Figure 2 This is a schematic diagram of wavefront point detection using the ray method.

[0013] Figure 3 This is a flowchart for filtering local outliers.

[0014] Figure 4 This is a screenshot showing the effect of filtering out local anomalies.

[0015] Figure 5 This is a diagram showing the fitting effect of the wavefront.

[0016] Figure 6 This is a schematic diagram illustrating the propagation distance of the shock wave front on the measuring line.

[0017] Figure 7 This is a graph showing the fitting effect of the shock wave velocity. Detailed Implementation

[0018] The present invention will be further described in detail below with reference to the embodiments. It should be noted that the present invention is not limited to the following specific embodiments. Those skilled in the art can make various other specific modifications and combinations based on the technical teachings disclosed in the present invention without departing from the essence of the present invention. These modifications and combinations are still within the protection scope of the present invention.

[0019] like Figure 1 As shown, the method for extracting explosion shock wave pressure based on image sequences provided by this invention includes the following steps: Step 1: The user inputs the video frame sequence of the explosion shock wave propagation, the coordinates of the blast center, the survey line angle, and the pixel size. The specific steps are as follows: The user inputs the original explosion shockwave video image, the coordinates of the explosion center (blast center) on the image, the survey line angle, and the pixel size. The survey line refers to a ray extending from the blast center in a certain direction, and the survey line angle refers to the angle between this ray and the horizontal direction to the right clockwise; the pixel size refers to the length of the actual object represented by each pixel, which is calibrated using a calibration object at the test site before capturing the shockwave image.

[0020] Step 2: Read the video frame sequence of the explosion shock wave propagation and extract the grayscale image of each frame in the video frame sequence. The specific operations are as follows: Use a video reading tool (such as OpenCV) to read the video frame by frame, and then... The color image of a frame is converted to a grayscale image, denoted as . ,in For pixel coordinates, , , W and H represent the width and height of the image, respectively, and N represents the number of video frames.

[0021] Step 3: Perform inter-frame difference processing on adjacent grayscale images to obtain a difference image, and then binarize the difference image to generate a binary difference image. The specific operations are as follows: Calculate the absolute difference between two adjacent grayscale images:

[0022] In the formula, N represents the number of video frames.

[0023] Then, the difference image is denoised (e.g., Gaussian denoising, median denoising, etc.), and then a given threshold is applied. Binarization:

[0024] After binarization, morphological opening operations and connected component analysis are used to remove isolated points, resulting in a clean binary difference image. , Represents the pixels in a binary difference image The grayscale value.

[0025] Step 4: Ray tracing detection of wavefront points: Using the explosion center as the focal point, rays are emitted along multiple angular directions. The foreground points in the binary difference image are searched along each ray to detect the initial wavefront points. This includes the following operations: With a given core Launch with center A single ray, with a uniform emission angle covering the entire area. : , .

[0026] For the A ray, from Start with step size Pixel outward search, maximum radius Check its neighborhood at each step. (size is) , For square area The side length (in pixels, default value is 3, users can modify this parameter as needed) contains a foreground point ( If it exists, record the point in the neighborhood that is farthest from the blast center. As candidate points:

[0027] The point furthest from the explosion center among all candidate points corresponding to the current ray is taken as the wavefront point of the ray:

[0028] By traversing all rays, the initial wavefront point set is obtained. .

[0029] by Figure 2 For example, in a typical binary image, white pixels have a grayscale value of 1, representing foreground points, while black pixels have a grayscale value of 0, representing background points. In this embodiment, a black square is used to represent the foreground point. Figure 2 In the image, numerous black squares cluster near the epicenter, representing the foreground of the fireball and smoke generated by the explosion. The outward-extending semi-circular band of black squares indicates the foreground of the shock wave. Figure 2 In the middle, three rays were drawn out from the explosion center, and the foreground points farthest from the explosion center on each ray were detected as follows: , and Obviously, and These are not actual wavefront points and need to be further eliminated.

[0030] Step 5: Perform hierarchical filtering on the detected initial wavefront points, including sequential radius percentage filtering, radius anomaly filtering, and local anomaly point filtering, to obtain the filtered wavefront points.

[0031] The initial wavefront points detected by the ray-mapping method contain many non-real wavefront points, such as foreground points generated by fireballs or smoke. Therefore, it is necessary to filter out different types of non-real wavefront points. Step 5 includes the following sub-steps: Step 5.1: Filter by radius percentage Filtering by radius percentage mainly removes foreground points generated by fireballs and smoke within the wavefront. The operation is as follows: Calculate each initial wavefront point radius to the center of the explosion Click on all initial wavefronts. Sort by largest to smallest, and retain the ones with the largest distance. The point that is kept is the farthest point. The points:

[0032] in, For the sorted number Distance values.

[0033] Step 5.2: Radius Anomaly Filtering Radius anomaly filtering primarily removes anomaly points far from the wavefront. The operation is as follows: [The text abruptly ends here, likely due to an incomplete sentence or a formatting error.] For points within the radius, calculate the mean and standard deviation of the radius:

[0034]

[0035] Remove radius exceeding The points within the range are used to obtain the wavefront point set. .

[0036] Step 5.3: Local Anomaly Filtering The radius percentage filtering method and the radius anomaly filtering method are based on the global statistical characteristics of the initial point of the wavefront detected by the ray-mapping method. They cannot effectively remove local anomalies on the wavefront, so further removal of local anomalies is required. For details, please refer to [link to relevant documentation]. Figure 3 .

[0037] Computation of wavefront point set Each point in the diagram is relative to the center of the explosion. Polar angle:

[0038] Convert the polar angle to ;Will The points in the array are sorted in ascending order of polar angle to obtain an ordered point sequence. ; Assumption There are n points The corresponding radius is The corresponding polar angle satisfies ; exist Randomly select m seed points For each selected seed point Initialize the point set ; Using the dynamic threshold method, at the seed point left-hand point set and seed point The point set on the right Looking for points to join point; Seed point The point set on the right To illustrate the dynamic threshold method, let's take an example. The specific steps include: (a) Initialize the current point as , radius is The next point is , radius is ; (b) Given a threshold If the radius of the current point and the radius of the next point satisfy the following relationship:

[0039] but Update the current point and the next point using step (c); otherwise, update the next point using step (d). (c) Update the current point and the next point: , , , Repeat step (b); (d) Update the next point as , Repeat step (b); Repeat steps (a)-(d) above until point 1. Up to this point, the seed point The point set on the right has been traversed; Use a similar method at the seed point left-hand point set Find points that meet the criteria and add them to the point set. ; For m seed points Let the final set of points be... ;right Points in Calculate the probability of it appearing in the point set:

[0040] in

[0041] For a given threshold ,like Then remove Otherwise, keep All retained point sets are denoted as the filtered wavefront points. .

[0042] Figure 4 This is a schematic diagram showing two local outliers that were filtered out after extending the point set.

[0043] Step 6: Fit the filtered wavefront points in the spatial dimension to obtain the fitted continuous wavefront curve; calculate the coordinates of the intersection point of the specified survey line and the fitted continuous wavefront curve, and calculate the distance between the intersection point and the explosion center; use the central difference method to calculate the instantaneous velocity of the shock wave propagation; fit the instantaneous velocity in the time dimension to obtain a stable wavefront propagation velocity.

[0044] Step 6 includes the following sub-steps: 6.1: Wavefront Spatial Dimension Fitting The filtered wavefront point set obtained in step 5 The process involves fitting points within a given area to obtain a smoother, continuous wavefront curve, facilitating subsequent wavefront velocity calculations. Fitting methods include, but are not limited to, circular, elliptical, and polynomial fitting. Wavefront fitting can be global or specific to a designated local region. Figure 5 This is a diagram showing the effect after fitting the wavefront.

[0045] 6.2: Fitting the time dimension of velocity data According to the survey angle specified by the user Calculate the intersection point of the wavefront point fitting curve and the survey line for each frame at this survey line angle. The calculation method is to solve the following system of equations:

[0046] In the formula, The equation for the fitted curve is... For the angle is The equation for the survey line.

[0047] Calculate the distance between the intersection of the survey lines and the blast center. (Distance values ​​must include cell size), forming a distance-time sequence. ,in .like Figure 6 As shown.

[0048] Instantaneous velocity is calculated using central difference on the distance-time series:

[0049] in, Instantaneous velocity The distance between the intersection of the survey lines and the blast center. Frame number Half the width of the window For frame rate.

[0050] For points near the beginning or end, use forward or backward difference.

[0051] The instantaneous velocity of the shock wave calculated based on distance-time series exhibits significant fluctuations, affecting the calculation of shock wave pressure. Further fitting of the instantaneous velocity along the time dimension is needed to obtain the wavefront propagation velocity. For example... Figure 7 As shown.

[0052] Step 7: Pressure Calculation Based on wavefront propagation speed The test angle was calculated based on the Rankine-Hugoniot equation in shock wave theory. The shock wave pressure in each frame is given by the Rankine-Hugoniot equation:

[0053] In the formula, Shock wave pressure The ambient pressure is approximately 0.101 MPa at standard sea level. For specific heat ratio, take 1.4 for air; It is the ratio of the shock wave velocity to the local speed of sound. The unit is Mach number, where For the shock wave velocity, The speed of sound.

[0054] The above fully discloses a preferred embodiment of the present invention. Those skilled in the art can make reasonable adjustments to the parameters (such as the difference threshold, polynomial order, filtering percentage, etc.) without departing from the principle of the present invention, and these adjustments still fall within the protection scope of the present invention.

[0055] This invention extracts moving regions from video image sequences by converting them to grayscale and performing inter-frame difference operations on the grayscale images; it obtains an initial wavefront by deriving several rays from the explosion center and searching for the farthest foreground point on each ray using a sliding neighborhood search; it obtains a clean shock wavefront by removing different types of outliers through a hierarchical filtering method; and it improves the accuracy of shock wave pressure calculation by performing spatiotemporal fitting of the data (global or local fitting of the wavefront and fitting of the wavefront propagation velocity).

[0056] This invention automates the processing of high-speed video sequences from an explosion test, acquiring peak shock wave overpressure data at distances of 1.5 meters, 2 meters, and 2.5 meters from the explosion center, and comparing the data with sensor measurements. The experimental results are shown in Table 1. The maximum relative error between the image-measured pressure values ​​and the sensor measurements is 5.9%, indicating that the method of this invention calculates the peak shock wave overpressure with high accuracy. Compared to traditional manual processing methods, this invention automates data processing, significantly improving data processing efficiency.

[0057] Table 1. Estimation and error of image-based overpressure peak values.

Claims

1. A method for extracting the pressure of an explosion shock wave based on an image sequence, characterized in that, Includes the following steps: Step 1: The user inputs the video frame sequence of the explosion shock wave propagation, the coordinates of the explosion center, the survey line angle, and the pixel size; Step 2: Read the video frame sequence of the explosion shock wave propagation and extract the grayscale image of each frame in the video frame sequence; Step 3: Perform inter-frame difference processing on the grayscale images of adjacent frames to obtain a difference image, and binarize the difference image to generate a binary difference image; Step 4: Detection of wavefront points using ray tracing: With the explosion center as the center, emit rays along multiple angular directions, search for foreground points in the binary difference image on each ray, and detect the initial wavefront points; Step 5: Perform hierarchical filtering on the detected initial wavefront points, including sequential radius percentage filtering, radius anomaly filtering, and local anomaly point filtering, to obtain the filtered wavefront points. Step 6: Fit the filtered wavefront points in the spatial dimension to obtain the fitted continuous wavefront curve; calculate the coordinates of the intersection point of the specified survey line and the fitted continuous wavefront curve, and calculate the distance between the intersection point and the explosion center; use the central difference method to calculate the instantaneous velocity of the shock wave propagation; fit the instantaneous velocity in the time dimension to obtain a stable wavefront propagation velocity. Step 7: Calculate the shock wave pressure at each frame at the test angle based on the wavefront propagation velocity.

2. The method for extracting explosion shock wave pressure based on image sequences as described in claim 1, characterized in that, The specific steps for step 1 are as follows: The user inputs the original explosion shock wave video image, the coordinates of the explosion center on the image, the survey line angle, and the pixel size; the survey line refers to a ray extending from the explosion center in a certain direction, the survey line angle refers to the angle between the ray and the horizontal rightward clockwise direction; the pixel size refers to the length of the actual object represented by each pixel, which is calibrated using a calibration object at the test site before capturing the shock wave image.

3. The method for extracting explosion shock wave pressure based on image sequences as described in claim 1, characterized in that, The specific steps for step 2 are as follows: Read the video frame by frame, and then... The color image of a frame is converted to a grayscale image, denoted as . ,in For pixel coordinates, , , W and H represent the width and height of the image, respectively, and N represents the number of video frames.

4. The method for extracting explosion shock wave pressure based on image sequences as described in claim 1, characterized in that, The specific steps for step 3 are as follows: Calculate the absolute difference between two adjacent grayscale images: In the formula, N represents the number of video frames; Then, the difference image is denoised, and a given threshold is applied. Binarization: After binarization, morphological opening operations and connected component analysis are used to remove isolated points, resulting in a clean binary difference image. , Represents the pixels in a binary difference image The grayscale value.

5. The method for extracting explosion shock wave pressure based on image sequences as described in claim 1, characterized in that, Step 4 specifically includes the following operations: With a given core Launch with center A single ray, with a uniform emission angle covering the entire area. : ， ； For the A ray, from Start with step size Pixel outward search, maximum radius Check its neighborhood at each step size. Does the foreground area exist inside? ; If it exists, record the point in the neighborhood that is farthest from the blast center. As candidate points: The point furthest from the explosion center among all candidate points corresponding to the current ray is taken as the wavefront point of the ray: By traversing all rays, the initial wavefront point set is obtained. .

6. The method for extracting explosion shock wave pressure based on image sequences as described in claim 1, characterized in that, Step 5 includes the following sub-steps: Step 5.1: Filter by radius percentage Filtering by radius percentage mainly removes foreground points generated by fireballs and smoke within the wavefront. The operation is as follows: Calculate each initial wavefront point radius to the center of the explosion ; Click on all initial wavefronts Sort by largest to smallest, and retain the ones with the largest distance. The point that is kept is the farthest point. The points: in, For the sorted number One distance value; Step 5.2: Radius Anomaly Filtering Radius anomaly filtering primarily removes anomaly points far from the wavefront. The operation is as follows: [The text abruptly ends here, likely due to an incomplete sentence or a formatting error.] For points within the radius, calculate the mean and standard deviation of the radius: Remove radius exceeding The points within the range are used to obtain the wavefront point set. ; Step 5.3: Local Anomaly Filtering Computation of wavefront point set Each point in the diagram is relative to the center of the explosion. Polar angle: Convert the polar angle to ;Will The points in the array are sorted in ascending order of polar angle to obtain an ordered point sequence. ; Assumption There are n points The corresponding radius is The corresponding polar angle satisfies ; exist Randomly select m seed points For each selected seed point Initialize the point set ; Using the dynamic threshold method, at the seed point left-hand point set and seed point The point set on the right Find points that can be added to the set point; For m seed points Let the final set of points be... ;right Points in Calculate the probability of it appearing in the point set: in For a given threshold ,like Then remove Otherwise, keep ; denote all retained point sets as the filtered wavefront points. .

7. The method for extracting explosion shock wave pressure based on image sequences as described in claim 1, characterized in that, Step 6 includes the following sub-steps: 6.1: Wavefront Spatial Dimension Fitting The filtered wavefront point set obtained in step 5 The points in the curve are fitted to obtain the fitted continuous wavefront curve. 6.2: Fitting the time dimension of velocity data According to the survey angle specified by the user Calculate the intersection point of the wavefront point fitting curve and the survey line for each frame at this survey line angle. The calculation method is to solve the following system of equations: In the formula, The equation for the fitted curve is... For the angle is The equation of the survey line; Calculate the distance between the intersection of the survey lines and the blast center. This forms a distance-time sequence. ,in ; Instantaneous velocity is calculated using central difference on the distance-time series: in, Instantaneous velocity The distance between the intersection of the survey lines and the blast center. Frame number Half the width of the window For frame rate; For points near the beginning or end, use forward or backward difference; By fitting the instantaneous velocity along the time dimension, the propagation velocity of the wavefront is obtained.

8. The method for extracting explosion shock wave pressure based on image sequences as described in claim 1, characterized in that, The specific operation of step 7 is as follows: Based on the wavefront propagation speed The test angle was calculated based on the Rankine-Hugoniot equation in shock wave theory. The shock wave pressure in each frame is given by the Rankine-Hugoniot equation: In the formula, Shock wave pressure The ambient pressure is approximately 0.101 MPa at standard sea level. For specific heat ratio, take 1.4 for air; It is the ratio of the shock wave velocity to the local speed of sound. The unit is Mach number, where For the shock wave velocity, The speed of sound.

Images