Method for digitizing observation chart of short-period seismograph and related device

Abstract

This invention belongs to the field of seismic image signal processing, and provides a method and related apparatus for digitizing short-period seismograph observation maps. The method includes: removing noise from scanned short-period seismograph observation map images to obtain denoised images; using a correction formula based on the length of the recording pen, correcting for arc distortion in the denoised images to obtain arc-distortion-corrected images; since the images also have tilt distortion, based on the working principle of short-period seismographs, deriving an inherent displacement formula for the recording pen, and using this formula to correct for tilt distortion in the arc-distortion-corrected images to obtain distortion-free seismic waveform images; and digitizing the distortion-free seismic waveform images. This application solves the problem mentioned in the background art that previous solutions only completed the electronic scanning and storage of simulated seismic observation maps, without realizing the digitization of the electronic image waveforms, by performing image preprocessing and digitization processing on simulated seismic observation maps.

Description

Technical Field

[0001] This invention belongs to the field of earthquake image processing, and specifically relates to a method and related apparatus for digitizing observation maps from short-period seismographs. Background Technology

[0002] my country experiences frequent earthquakes, and since large-scale earthquake observation began in the last century, a large number of simulated earthquake observation maps have been gradually accumulated, containing valuable records of several strong earthquakes, which have significant strategic value. The earthquake system has completed the electronic scanning of over 4 million maps, saving them in PNG format, and is attempting to digitally extract data from the DD-1 short-period simulated observation maps.

[0003] The initial plan only completed the digital scanning and storage of simulated earthquake observation maps, without effectively extracting waveform information. This resulted in a large number of valuable electronic images being left idle and wasting storage resources. They could not be transformed into effective data that could be used for earthquake scientific research, earthquake prevention and disaster reduction, etc., and the rescue and revitalization of simulated earthquake data was not truly realized, which restricted the in-depth utilization of historical earthquake data. Summary of the Invention

[0004] The purpose of this invention is to provide a method and related apparatus for digitizing short-period seismograph observation maps, in order to solve the problem that the current lack of effective extraction of waveform information leads to a large number of valuable electronic images being left idle and wasting storage resources.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: In a first aspect, the present invention provides a method for digitizing short-period seismograph observation maps, comprising: Noise removal is performed on the short-period seismograph observation images obtained from the scan to obtain denoised images; Using a correction formula based on the length of the recording pen, the denoised image is corrected for arc distortion to obtain an image with arc distortion correction. Using a displacement formula based on the inherent properties of the recording pen, tilt distortion correction is performed on the image after arc distortion correction to obtain a distortion-free seismic waveform image; The waveform pixels are extracted from the distortion-free seismic waveform image, and the median of the ordinate corresponding to the same horizontal coordinate is taken. The pixel coordinates are mapped to the time domain and amplitude domain, the waveform is smoothed by applying a filter, and baseline correction is performed to determine the zero point, so as to achieve accurate digitization of the waveform of the observation map.

[0006] Furthermore, the noise removal process performed on the short-period seismograph observation images obtained from the scan to obtain denoised images includes: For noise that is different from the waveform color, each pixel of the image is read, the color value is analyzed, a color threshold range is set based on the color of the waveform lines, and the color of each pixel is compared with the threshold range. If the pixel color is not within the range, it is determined to be noise, and the color of the pixel is directly replaced with the set background color. This process traverses all pixels of the entire image.

[0007] For noise that is the same as or close to the waveform color, manual processing is performed by setting the background color to white. After processing, the image is saved, resulting in a denoised image with a uniform white background and a clear black curve waveform.

[0008] Furthermore, the step of using a correction formula based on the length of the recording pen to correct arc distortion in the denoised image, resulting in an image with arc distortion correction, includes: Based on the width and height of the denoised image, a brand new blank image is created as the output. A regular grid is generated on the pixel coordinates of the output image, and its coordinate values ​​are normalized to the range of [-1, 1]. Based on the curvature radius parameter obtained by converting the physical length R of the recording pen, the inverse mapping of the arc correction formula is solved. For each target coordinate point in the grid of the output image, its corresponding source coordinate point in the original denoised image is calculated using the inverse mapping formula. The calculated source coordinates are restricted to the effective coordinate range of the original image. Using the bilinear interpolation algorithm, the pixel value of the target point is calculated based on the pixel values ​​around the source coordinate point in the original image, and filled into the newly created blank image, finally obtaining the image after the arc distortion is corrected.

[0009] Furthermore, the reverse mapping is the process of finding the original coordinates (x, y) from the corrected coordinates (x1, y1), which is the inverse process of solving the forward mapping formula. Solve the equation by finding x and y from x1 and y1 respectively: .

[0010] Furthermore, the step of using a displacement formula based on the recording pen's inherent properties to correct the arc distortion of the image, thereby obtaining a distortion-free seismic waveform image, includes: Obtain the dimensions of the image after arc correction. Calculate the total displacement parameter D based on the instrument pen length and scanning resolution. Combined with the image height h, calculate a fixed slope value. Create a blank image with the same dimensions as the input image and a white background. Iterate through each pixel row of the input image. For the y-th row, calculate the required horizontal compensation displacement based on the slope. Iterate through each pixel in the row. For the x-th column, calculate the corresponding original column position in the input image. x Check the corresponding source column position original xIf the value is within the image width range, then the input image will be displayed in (original) format. x The pixel value at position (x, y) is assigned to the blank image at position (x, y); if it is outside the range, the position remains white with background; after traversing all pixels, the corrected image is output.

[0011] Furthermore, the slope value is calculated using the following formula: slope = -D / h, where D is the total pixel displacement; The formula for calculating the horizontal compensation displacement is as follows: compensation = -slope×y The formula for calculating the corresponding source column position is as follows: original x =x + compensation.

[0012] Furthermore, the process of extracting waveform pixels from the distortion-free seismic waveform image, taking the median of the ordinates corresponding to the same abscissa, mapping the pixel coordinates to the time and amplitude domains, applying filters to smooth the waveform, and performing baseline correction to determine the zero point, thereby achieving accurate digitization of the waveform in the observation drawing, includes: Load the corrected waveform image and convert it into a grayscale image. Apply Gaussian blur filter to the grayscale image and perform binarization processing on the image using an adaptive threshold algorithm. Perform morphological closing operation on the binary image, scan the entire image, identify and record the coordinate set of all white pixels, and the coordinates constitute the original pixel-level expression of the waveform lines. From all the acquired white pixels, find all unique x-coordinate values ​​(X values) and sort them in ascending order, with each X value representing a time sampling point. For each unique X value, find all corresponding y-coordinate values ​​for that column. Using the median algorithm, calculate the most representative y-coordinate value from these Y values. This method effectively eliminates interference from discrete noise points that may exist in the column and robustly determines the center position of the waveform. Map the pixel coordinates to physical quantities: based on the total duration T of the seismic signal recorded on the drawing, map the pixel x-coordinates to time values ​​proportionally. Simultaneously, linearly map the y-coordinates to the normalized amplitude range of [-1, 1]. Obtain a preliminary set of time-amplitude data pairs. Apply a Savitzky-Golay filter to this amplitude sequence for smoothing. The filter smooths the data by performing polynomial fitting within a sliding window, effectively suppressing high-frequency noise while better preserving the original morphological characteristics of the waveform. Baseline correction and amplitude normalization are performed by selecting a segment of signal data considered as background ground vibrations before the earthquake event and calculating the median amplitude value, which is then used as the baseline offset. The amplitude value of each data point in the entire waveform data sequence is subtracted from the baseline offset to make the signal fluctuate around the zero line before the event. After baseline correction, the maximum absolute value of the entire waveform amplitude is calculated, and all amplitude values ​​are divided by this maximum value to normalize the amplitude range of the entire waveform to the interval [-1, 1]. The obtained data points are resampled, a standard target sampling rate is set, and an equally spaced time axis is generated based on the original time range. The cubic spline interpolation method is used to calculate the amplitude value on the new equally spaced time axis based on the existing non-equally spaced time-amplitude data points. A high-quality digital seismic waveform time series with a fixed and regular sampling rate is output, completing the full conversion from analog drawing images to computable digital signals.

[0013] Secondly, the present invention provides a digitization system for short-period seismograph observation maps, comprising: The data preprocessing module is used to remove noise from the short-period seismograph observation images obtained by scanning, so as to obtain denoised images; The arc distortion correction module is used to correct the arc distortion of the denoised image using a correction formula based on the length of the recording pen, so as to obtain the image after arc distortion correction. The tilt distortion correction module is used to correct the tilt distortion of images after arc distortion correction using a correction formula based on a fixed speed and time of the recording pen, so as to obtain distortion-free seismic waveform images. The digitization module is used to extract waveform pixels from distortion-free seismic waveform images, take the median of the ordinate corresponding to the same horizontal coordinate, map the pixel coordinates to the time and amplitude domains, apply filters to smooth the waveform, and perform baseline correction to determine the zero point, thereby achieving accurate digitization of the waveform in the observation drawing.

[0014] Thirdly, the present invention provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the method for digitizing short-period seismograph observation maps.

[0015] Fourthly, the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the method for digitizing short-period seismograph observation maps.

[0016] Compared with the prior art, the present invention has the following technical effects: This application solves the problem mentioned in the background art that the previous scheme only completed the electronic scanning and storage of the simulated earthquake observation drawings, but did not realize the digitization of the waveform of the electronic images, by performing image preprocessing and digitization processing on the DD-1 type simulated earthquake observation drawings.

[0017] First, from the preprocessed waveform image, by extracting all pixels corresponding to the vertical amplitude axis for each horizontal time axis and taking the median, the discrete pixel sequence with background noise is accurately transformed into a continuous trajectory line representing the seismic waveform. Then, the pixel coordinates of the image are mapped to the actual time and physical amplitude values ​​according to the time information recorded on the drawing, thus transforming the image information into structured time-series data. A filter is applied to smooth the data sequence, effectively suppressing high-frequency noise introduced during drawing scanning, denoising, and correction, improving data quality. Finally, through a baseline correction step, using the average value of stable ground pulsation signals before the earthquake event as a benchmark, the DC component of the entire waveform is zeroed, ensuring that the physical meaning of the waveform amplitude values ​​is clear and the zero-crossing points are accurate. Electronic images stored in formats such as PNG, which cannot be directly used for calculation, are transformed into standard digital waveform signals that can be directly used for earthquake analysis, scientific research, and engineering calculations. This transforms a massive amount of idle scanned image resources into effective data that can be deeply utilized, achieving a true rescue and revitalization of historical simulated earthquake observation data. Attached Figure Description

[0018] Figure 1 This is a flowchart of the present invention.

[0019] Figure 2 This is a simulation observation drawing for the present invention.

[0020] Figure 3 This is a schematic diagram of the annotation on the simulated observation map of the present invention.

[0021] Figure 4 This is a denoised image of the actual earthquake waveform after the present invention.

[0022] Figure 5 This is a schematic diagram of the DD-1 type short-period seismograph of the present invention.

[0023] Figure 6 For the present invention to Figure 4 Create a magnified view of the area where the arc correction was performed.

[0024] Figure 7 For the present invention Figure 6 Create a magnified view of the tilt distortion correction area.

[0025] Figure 8 This is a digitized schematic diagram illustrating the process of restoring a distorted image according to the present invention.

[0026] Figure 9 This is a comparison of the results of digital processing of the synthesized signal in this invention.

[0027] Figure 10 This is a schematic diagram illustrating the digital processing and analysis of upper and lower components in an embodiment of the present invention.

[0028] Figure 11 This is a waveform record after preprocessing in an embodiment of the present invention.

[0029] Figure 12 This is a diagram of the correct seismic waveform after arc correction and tilt correction, as shown in the embodiment of the present invention.

[0030] Figure 13 This is the waveform digitally extracted by the present invention.

[0031] Figure 14 In this embodiment of the invention, the wideband signal is converted into DD based on the instrument response. 1. Short-period signal, and extract waveform segments of the Z component within the time period.

[0032] Figure 15 The simulation results for an embodiment of the present invention are shown in the figure. Detailed Implementation

[0033] The present invention will be further described below with reference to the accompanying drawings: Regarding the DD-1 type simulation observation drawings: The drawings produced by the DD-1 type short-period seismograph are DD-1 type simulation observation drawings.

[0034] The dimensions of the drawing, after measurement, are 79 cm x 45 cm. Because the complete DD-1 seismograph system includes three DS-1 seismometers, such as... Figure 2 The drawing shows three components, from top to bottom: N (North-South component, the first of the three records in the simulation drawing), marked as North-South; E (East-West component, the second of the three records in the simulation drawing), marked as East-West; and Z (Vertical component, the third of the three records in the simulation drawing), marked as Up-Down. A red stamp is printed in the upper left corner of the drawing, containing the following complete information: 1. Station Name: Xiangfan Seismic Station; 2. Date: [Year] [Month] [Day]; 3. Clock Difference Information: t=0; 4. Magnification VO (the three components are 3.4×10⁻⁶) 4 3.1×10 4 2.4×10 45. Duty Officer: Zhang; 6. Analyst: Zhang; 7. Proofreader: None; 8. Remarks: None. Besides the seismic waveform, the drawings also include manually annotated upper and lower time values: [Date and Time of Updating] (update and end date are both dates and times). The earthquake was massive, and we can clearly see the arc distortion in the waveform recorded by the DD-1 short-period recorder. The seismic waveform cannot be directly digitized; preprocessing of the images is necessary to obtain the correct seismic waveform before digitization. To correctly preprocess the images, we need to understand the working principle of the DD-1 short-period seismograph and the characteristics of the production drawings to obtain the correct seismic waveform.

[0035] Example 1: This invention provides a method for digitizing short-period seismograph observation maps, including: S1, noise removal is performed on the short-period seismograph observation images obtained from the scan to obtain the denoised images; Remove noise from scanned images of drawings, including noise that is significantly different in color from the waveform and noise that is similar in color to the waveform.

[0036] First, the color image is converted to grayscale for processing. The program iterates through each pixel of the image, obtaining its grayscale value. A threshold range based on typical grayscale values ​​of the waveform lines is preset. For each pixel, if its grayscale value is outside this threshold range, it is determined to be background or heterochromatic noise, and its grayscale value is directly set to 255 (pure white). This process automatically removes most noise with significant color differences. After automatic processing, the resulting image is opened in image editing software. The operator manually inspects it and carefully erases any residual traces whose grayscale values ​​are very close to the waveform and were not removed by the program, while ensuring that the continuity of the waveform curve is not disrupted. Combining the efficiency of automated batch processing with the precision of manual intervention, a clean background and clear waveform image can be quickly obtained, eliminating interference for subsequent geometric correction and ensuring the clarity and integrity of the waveform outline.

[0037] S2, using a correction formula based on the length of the recording pen, the denoised image is corrected for arc distortion to obtain the image after arc distortion correction; Correct the waveform arc distortion caused by the limited length of the recording pen of the seismograph. Obtain the width w and height h of the denoised image. According to the physical pen length of the DD-1 instrument and the image scanning resolution, calculate the radius of curvature R in pixels. Create a blank output image with the same size (w×h) as the input image. Construct the coordinate grid (x1, y1) of each pixel point in the output image and normalize it to the coordinate system of [-1, 1]. Using the inverse operation of the correction formula based on the pen length R (i.e., the known mathematical model of arc distortion), for each target point (x1, y1) in the output image, solve for its corresponding source point coordinates (x, y) in the original input image. Check the calculated source coordinates to ensure that they fall within the valid range of the original image (0≤x<w, 0≤y<h). Since the calculated source coordinates are usually floating-point numbers, use the bilinear interpolation algorithm to calculate the gray value of the target point by weighting the gray values of the four actual pixel points around the source point and fill it into the corresponding position of the blank output image. After traversing all pixels, the corrected image is obtained. Through geometric inverse transformation, the circular arc trajectory recorded due to the pen swing is restored to the real straight trajectory of the seismic wave, solving the first-layer waveform distortion caused by the mechanical structure of the instrument.

[0038] S3. Use the correction formula based on the fixed speed and time of the recording pen to correct the tilt distortion of the image after arc distortion correction, and obtain a distortion-free seismic waveform image; Correct the tilt distortion remaining after the processing in step S2, which is caused by the continuous uniform rotation of the recording drum and this factor was not considered in the previous correction. Obtain the width w and height h of the image after arc correction. According to the instrument parameters (pen length) and scanning parameters, calculate the total pixel displacement D in the vertical direction of the image caused by the rotation of the drum. Calculate the fixed slope slope = -D / h. Create a blank white image with the same size as the input image as the output. Traverse the input image row by row with the y coordinate from 0 to h - 1. For each row y, calculate the required horizontal compensation displacement compensation = -slope×y. For each column x in this row, the corresponding source column position in the input image is original x = x + compensation. Judge whether original x is within the range of [0, w). If it is, then the pixel in the input image at (original xThe pixel value of (x, y) is copied to the (x, y) position of the output image; if it is not found, it means that the position does not exist in the original data, and the (x, y) position of the output image remains white. Through a linear translation operation proportional to the row number, the overall waveform tilt caused by the uniform motion of the roller is compensated. Combined with step S2, the composite distortion (arc + tilt) in the original image is finally completely corrected, resulting in an image that accurately reflects the real ground motion waveform.

[0039] S4 extracts waveform pixels from the distortion-free seismic waveform image and takes the median of the ordinate corresponding to the same horizontal coordinate. The pixel coordinates are mapped to the time and amplitude domains, a filter is applied to smooth the waveform, and baseline correction is performed to determine the zero point, thus achieving accurate digitization of the waveform in the observation drawing.

[0040] The process involves extracting and generating a usable digital signal sequence from the corrected waveform image. First, the image is binarized to make the waveform white and the background black, and morphological closing operations are performed to connect minor breaks. Each column of the image's x-coordinate is traversed, and the row coordinate y-values ​​of all white pixels in that column are collected. The median of all collected y-values ​​in that column is taken as the waveform amplitude position corresponding to that time point x. This effectively resists interference from sporadic noise points that may exist in a single column. Next, coordinate mapping is performed: the pixel's x-coordinate is divided by the total width of the image, then multiplied by the actual total time length of the waveform segment as indicated on the drawing (in seconds) to obtain the timestamp of each data point; the median y-coordinate is linearly transformed to map to amplitude values ​​in the range [-1, 1], where the top of the image corresponds to +1 and the bottom to -1. Then, the obtained amplitude sequence is smoothed using a Savitzky-Golay filter to remove high-frequency glitches while preserving the main waveform features. Finally, baseline correction is performed: the amplitude data of a stable waveform preceding the arrival of the seismic signal is selected, and its average value is calculated as the baseline value. This baseline value is then subtracted from the amplitude sequence of the entire waveform, causing the signal to fluctuate around the zero line during seismic-free periods. Image pixels are converted into high-fidelity time-amplitude series data. Median extraction ensures the robustness of the amplitude values, coordinate mapping gives the data physical meaning, filtering improves data quality, and baseline correction ensures data accuracy. The final result is a digital waveform file that can be directly used for seismic analysis.

[0041] Example 2: This invention provides a method for digitizing short-period seismograph observation maps, including: Waveform preprocessing of drawings: Image preprocessing includes image noise removal and arc correction to obtain the correct seismic waveform.

[0042] Image noise removal: from Figure 2 and Figure 3We can see that the waveform records on the original drawings have manual annotations, such as red text, "top / bottom," "5 / 12," "1441," or pencil writing. Furthermore, due to the drawings being decades old and fading, the colors on the drawings are uneven. Our previous digitization of scanned images resulted in a lot of noise during waveform extraction, significantly impacting the extraction process and causing waveform distortion. Through practice, we found that waveform denoising is essential. After denoising, setting the waveform background to white and the curves to black ensures accurate waveform extraction. While denoising isn't the primary focus, it's crucial. We'll use the most direct and efficient method for denoising, handling it in two ways: 1. For noise with a different color than the waveform, we'll use code to set the noise pixels to white. 2. For noise with a similar or identical color to the waveform, we'll remove it manually. Since each image has three channels—east / west, north / south, and top / bottom—we'll only study the same channel; other channels are handled similarly. For ease of study, we'll select the complete third channel, the top / bottom channel, for our research. First, we'll use Python... Figure 2 We perform horizontal image segmentation to obtain the third component waveform. Figure 2 After removing noise from the segmented image, we get Figure 4 As shown, the noise in the image has been completely removed, with excellent results.

[0043] Correcting arc distortion: After removing noise from the image, we obtained a clean image. Next, we performed arc correction on the image. This work has been done before. Based on the arc correction formula in the literature "Waveform Correction Method in DD-1 Type Pen Drawing Record and 65 Type Smoke Record", we simplified the formula in the literature as shown in formula (1). Figure 5 The illustration is from the document "Digitalization Method and Preliminary Study of Historical Earthquake Simulation Waveform Drawings in Hebei". Figure 2 That is, to correct point B to point A.

[0044]

[0045] In formula (1), R is the length of the recording pen of the DD-1 instrument.

[0046] The time resolution Δt of the drawing is calculated as scan resolution / drum speed. We know the scan resolution is 600 DPI, and through simple calculation, Δt = 0.127 seconds / pixel. Based on the "Introduction to Seismic Instruments," the parameters of the DD-1 instrument, DD-1 drawings, and its digitized images are summarized in Table 1 below. Table 1: Summary of Parameters Required for Image Preprocessing

[0047] Through the Figure 4 Curve correction does not result in a straight line, but rather a slanted line. For example... Figure 6 The image shown is a magnified view of a section of the waveform. This result reveals that the curve, visible to the naked eye, has become a slanted line, indicating that the correct seismic waveform has not been corrected. We define the distortion resulting from the above-mentioned arc distortion correction as tilt distortion.

[0048] Correcting tilt distortion: The analysis of the tilt distortion is due to the failure to consider that the actual arc distortion is caused by two factors. The arc correction formula (1) only considers the limited length of the DD-1 oscillator pen, without taking into account that the drum recorder is constantly rotating. It does not consider that the center of the arc changes because the pen moves at a fixed speed v0. The shape of the arc is thus deformed and is no longer a very symmetrical arc, that is... Figure 5 The angle α is not a fixed value. They used a fixed instantaneous α to calculate the coordinates of points A and B. In fact, this coordinate not only changes with R, but also with v0. They did not consider that v0 would cause the arc correction to be a slanted line instead of a straight line. Only by considering that the center of the arc changes can the arc correction be completely correct. To make more changes, we should make less changes. Based on the results of the literature "Waveform Correction Method in DD-1 Type Pen Drawing Record and 65 Type Smoke Record", this patent further corrects the tilt distortion after the arc correction to obtain the correct seismic waveform. When there is no earthquake, the recording trace formula is as shown in expression (2):

[0049] Where x0 is the x-coordinate of the initial position of the recording pen on the drawing, v0 is the fixed speed of the recording pen, t is the recording time, and y0 is the y-coordinate of the initial position of the recording pen on the drawing. The above coordinates are in a rectangular coordinate system.

[0050] The tilt distortion correction formula is shown in expression (3):

[0051] Use expression (3) Figure 6 The results were tilt-corrected to obtain Figure 7 As you can see, the diagonal line has become a straight line. We have obtained the correct seismic waveform. The pseudocode for the image preprocessing algorithm is shown in Algorithm 1. In Algorithm 1, the instrument parameters are: radius of curvature R = 3561 pixels, total displacement D = 500 pixels (both are pen length parameters of 15 cm), and the image size and scanning resolution of 600 dpi are obtained through simple calculations (the process is omitted).

[0052] The process involves using a correction formula based on the length of the recording pen to correct arc distortion in the denoised image, resulting in an image with arc distortion correction, including: Based on the width and height of the denoised image, a brand new blank image is created as the output. A regular grid is generated on the pixel coordinates of the output image, and its coordinate values ​​are normalized to the range of [-1, 1]. Based on the curvature radius parameter obtained by converting the physical length R of the recording pen, the inverse mapping of the arc correction formula is solved. For each target coordinate point in the grid of the output image, its corresponding source coordinate point in the original denoised image is calculated using the inverse mapping formula. The calculated source coordinates are restricted to the effective coordinate range of the original image. Using the bilinear interpolation algorithm, the pixel value of the target point is calculated based on the pixel values ​​around the source coordinate point in the original image, and filled into the newly created blank image, finally obtaining the image after the arc distortion is corrected.

[0053] The reverse mapping is the process of finding the original coordinates (x, y) from the corrected coordinates (x1, y1), which is the inverse process of solving the forward mapping formula. Solve the equation by finding x and y from x1 and y1 respectively: .

[0054] The method uses a correction formula based on a fixed speed and time of the recording pen to correct the tilt distortion of the image after arc distortion correction, resulting in a distortion-free seismic waveform image, including: Obtain the dimensions of the image after arc correction. Calculate the total displacement parameter D based on the instrument pen length and scanning resolution. Combined with the image height h, calculate a fixed slope value. Create a blank image with the same dimensions as the input image and a white background. Iterate through each pixel row of the input image. For the y-th row, calculate the required horizontal compensation displacement based on the slope. Iterate through each pixel in the row. For the x-th column, calculate the corresponding source column position `originalx` in the input image. Check if the value of `originalx` is within the image width range. If it is, assign the pixel value at position (`originalx`, y`) in the input image to the blank image at position (`x`, y`). If it is outside the range, leave the position as a white background. After iterating through all pixels, output the corrected image.

[0055] The formula for calculating the slope value is as follows: slope = -D / h, where D is the total pixel displacement; The formula for calculating the horizontal compensation displacement is as follows: compensation = -slope×y The formula for calculating the corresponding source column position is as follows: originalx = x + compensation.

[0056] Image waveform digitization: First, image preprocessing and waveform contour extraction are performed. The corrected waveform image is loaded and converted to grayscale. A Gaussian blur filter is applied to the grayscale image to suppress random noise. Next, an adaptive thresholding algorithm is used to binarize the image. This algorithm dynamically calculates the threshold based on the grayscale distribution of each pixel's neighborhood, thus robustly separating the waveform curve (usually set to white) from the background (set to black). Subsequently, morphological closing operations are performed on the binary image to connect any subtle breaks in the waveform lines that may have been caused by the original recording or preprocessing, ensuring the continuity of the contour. Finally, the entire image is scanned, and the coordinates of all white pixels are identified and recorded. These coordinates constitute the original pixel-level representation of the waveform lines.

[0057] Next, key waveform points are extracted and a preliminary sequence is generated. From all the acquired white pixels, all unique x-coordinate values ​​(X-values) are identified and sorted in ascending order, with each X-value representing a time sampling point. For each unique X-value, all corresponding y-coordinate values ​​(Y-values) for that column (i.e., the same time point) are identified. Using the median algorithm, the most representative y-coordinate value is calculated from these Y-values. This method effectively eliminates potential discrete noise interference in the column, robustly determining the center position of the waveform. Then, the pixel coordinates are mapped to physical quantities: based on the total duration T (seconds) of the seismic signal recorded on the drawing, the pixel x-coordinate is mapped proportionally to a time value; simultaneously, the y-coordinate is linearly mapped to a normalized amplitude range of [-1, 1], typically set to +1 for the top of the image and -1 for the bottom. This yields a preliminary set of "time-amplitude" data pairs. To further improve data quality, the Savitzky-Golay filter was applied to smooth the amplitude sequence. This filter smooths the data by performing polynomial fitting within a sliding window, effectively suppressing high-frequency noise while preserving the original waveform characteristics.

[0058] Next, baseline correction and amplitude normalization are performed. Due to potential zero-point drift in the recording equipment, the extracted signal baseline may deviate from zero. Signal data from a period preceding the seismic event (e.g., the first N seconds) is selected as background ground fluctuations, and the median amplitude value is calculated. This value is used as the baseline offset. This baseline offset is subtracted from the amplitude value of each data point in the entire waveform data sequence, thus making the signal appear to fluctuate around the zero line before the event. After baseline correction, the maximum absolute value of the entire waveform amplitude is calculated, and all amplitude values ​​are divided by this maximum value, thereby normalizing the amplitude range of the entire waveform to the interval [-1, 1]. This step eliminates DC offset and scale differences in the amplitude values, standardizing the data.

[0059] Finally, equal-interval resampling is performed. The time intervals of the data points obtained in the preceding steps are determined by the pixel distribution of the original image and may be non-uniform. To meet the requirements of most seismic analysis algorithms for equal-interval time series, resampling is necessary. A standard target sampling rate is set (e.g., 100Hz, i.e., 0.01-second intervals), and an equal-interval time axis is generated based on the original time range. Using cubic spline interpolation, the amplitude values ​​on the new equal-interval time axis are calculated based on the existing, non-equal-interval time-amplitude data points. Through this step, a high-quality digital seismic waveform time series with a fixed, regular sampling rate is finally output, completing the full conversion from analog drawing images to computable digital signals.

[0060] To evaluate the reliability of the drawing digitization method, the digitized signal needs to be compared with the reference signal, which is achieved through the following process.

[0061] First, load the two signal files to be compared. Read the CSV data file, automatically identifying the time and amplitude columns. Calculate the average time interval for each signal; its reciprocal is the sampling frequency of that signal. This completes the signal loading and basic parameter extraction.

[0062] The core similarity calculation is then performed. To ensure a fair comparison, the two signals are first standardized to the same sampling rate, typically using the higher of the two. The sinc interpolation method is then used to resample the two signals. Next, similarity metrics are calculated from three dimensions: First, the maximum cross-correlation coefficient between the two signals is calculated. This value reflects the overall shape similarity of the signals in the time domain and is obtained by normalizing the maximum value of the cross-correlation function. Second, spectral correlation is calculated. This involves estimating the power spectral density of the two signals using the Welch method and then calculating the correlation coefficient between the two power spectral sequences to assess the consistency of the signals' energy distribution in the frequency domain. Third, phase similarity is evaluated. Specifically, the positions of the zero-crossing points of the amplitude values ​​of the two signals (i.e., crossing the zero line from positive to negative or from negative to positive) are detected. Bidirectional matching is then performed between the zero-crossing points of the two signals to find the closest pair. The average time difference between all paired zero-crossing points is calculated and compared with a preset tolerance threshold to finally calculate a quantified phase similarity score. Finally, the three indicators (cross-correlation coefficient, spectral correlation, and phase similarity) are assigned different weights (e.g., 25%, 45%, and 30% respectively) and weighted averaged to obtain a comprehensive similarity score, which serves as a quantitative basis for overall reliability.

[0063] To visually demonstrate the comparison results, a comprehensive comparison chart is generated. This chart contains four sub-charts: The first sub-chart is a time-domain waveform comparison chart, plotting the two signals on the same time axis, distinguishing them with different colors or line styles, and marking their respective zero-crossing points, connecting matching zero-crossing pairs with line segments. The second sub-chart is a spectrum comparison chart, displaying the power spectral density curves of the two signals and annotating the calculated spectral correlation coefficient. The third sub-chart is a magnified view, typically focusing on the waveform details of the first 20 seconds to observe the alignment and matching of the signals during key periods. The fourth sub-chart is an indicator bar chart, displaying the values ​​of the cross-correlation coefficient, spectral correlation, and phase similarity side-by-side in a bar chart format, with the calculated comprehensive similarity indicated above the chart. Finally, this comprehensive comparison chart is saved as an image file.

[0064] The entire evaluation process is as follows: load the original signal file as the benchmark and the signal file digitized by this method, call the above similarity calculation function for processing, and finally output various quantitative indicators and a visual comparison chart, so as to evaluate the accuracy of the digitization results in a multi-dimensional and quantitative manner.

[0065] Experimental comparative analysis was conducted based on synthetic signal models constructed using drawing features. To verify the effectiveness and reliability of the digitization method, this study constructed a synthetic seismic signal based on the recording characteristics of actual seismic maps for comparative experiments. The expression of the synthetic seismic signal model is shown in expression (4): (4) in, In the experiment, the generated original waveforms were saved in two formats: one was an image format (PNG format) to simulate the actual drawing output. Considering that our DD-1 drawings have arc distortion and tilt distortion, we simulated arc distortion and tilt distortion in the image format of the original waveforms, and then restored these distortions to obtain the restored image, which was then digitized as follows. Figure 8 As shown. Another method is a digital sampling point sequence (saved as a CSV file), used to establish a benchmark for evaluating the effectiveness of digitalization. Figure 9 The results of digital processing of synthetic signals are shown in four sub-figures: (a) A time-domain comparison diagram of synthetic seismic signals and digitally synthesized seismic signals, including phase difference statistics; (b) Comparison of power spectral density between synthetic seismic signals and digitally synthesized seismic signals; (c) Comparison of local magnified time domain images of synthetic seismic signals and digitally synthesized seismic signals.

[0066] (d) Bar chart of similarity between synthetic seismic signal and digital synthetic seismic signal based on maximum cross-correlation, spectral correlation and phase correlation.

[0067] As can be seen from the figure, the overall similarity between the digital extraction result and the original waveform is 0.9908, which is very similar, indicating that the digital method used in this study has a very good effect on the recovery of the synthesized signal.

[0068] Digital processing and comparative analysis of actual earthquake drawings Based on the aforementioned synthetic signal processing method, this study further applies it to the digital processing and analysis of actual historical seismic maps. (Selecting examples such as...) Figure 3 The simulated earthquake record map of Wenchuan shown is used as the research object. This map records the magnitude 7.8 earthquake event that occurred at 14:28:04 on May 12, xxxx, and includes three components: east-west, north-south, and vertical. This study selects the vertical component (corresponding to...) Figure 10 Digital processing and analysis were carried out. To focus on verifying core digital technologies, a waveform segment from 14:28:00 to 14:29:50 (a total of 110 seconds) on [date] was selected as the digitization object. Before digitization, ground motion interference from non-target time periods was first removed through manual intervention to obtain the pre-processed waveform record. Figure 11 This method not only achieves effective digitization of strong earthquake waveforms, but also helps clarify the key technical aspects and principles in the digitization process, laying a methodological foundation for subsequent processing of more complex recorded drawings (such as drawings with severe signal aliasing or exceeding recording limits).

[0069] Using Algorithm 2, the image waveform digitization algorithm is used to... Figure 12 The seismic waveform is digitized to obtain a digitized waveform as follows: Figure 13 As shown.

[0070] To assess the accuracy of the digitized results, this study used a single digital broadband seismograph (BBVS) simultaneously operating at the Xiangfan station in Hubei Province. Data recorded by the 60 speedometer was used as a reference. Earthquake waveform records from 14:19:01 to 15:19:01 on [date] (a total of one hour) were obtained from the Hubei Earthquake Bureau. This broadband signal was converted to DD based on the instrument response. 1. A short-period signal was used, and a waveform segment from 14:28:00 to 14:29:50 (a total of 110 seconds) of the Z component was extracted. The result is as follows. Figure 14 As shown. Initial visual comparison reveals some differences between the digital signal and the waveform of the analog drawing. After applying a bandpass filter of 0.1–20Hz to the digital signal, the waveforms of the two become more consistent.

[0071] Further comprehensive comparative analysis of the two signals yielded the following simulation results: Figure 15 As shown, its evaluation indicators are similar to Figure 9 The results are completely consistent, and related illustrations will not be repeated. Analysis shows that the overall similarity between the digitized extraction result and the original waveform reaches 0.9104, indicating high consistency. This demonstrates that the digitization method used in this study performs well in actual seismic map processing. In summary, the method proposed in this study possesses good feasibility and accuracy in the digitization of actual seismic maps.

[0072] The algorithm described above for digitizing DD-1 short-period analog observation maps is studied using the actual Xiangfan seismograph station in Hubei Province as an example. Through the preprocessing of seismic images, the digitization of seismic waveforms, and the comprehensive comparison between the digitized signals and the actual digital signals, the entire process demonstrates that our digitization method has high accuracy and can be practically applied. This application is applicable to any maps generated by the DD-1 short-period seismograph.

[0073] In another embodiment of the present invention, a short-period seismograph observation map digitization system is provided, which can be used to implement the above-described short-period seismograph observation map digitization method. Specifically, the system includes: The data preprocessing module is used to remove noise from the short-period seismograph observation images obtained by scanning, so as to obtain denoised images; The arc distortion correction module is used to correct the arc distortion of the denoised image using a correction formula based on the length of the recording pen, so as to obtain the image after arc distortion correction. The tilt distortion correction module is used to correct the tilt distortion of images after arc distortion correction using a correction formula based on a fixed speed and time of the recording pen, so as to obtain distortion-free seismic waveform images. The digitization module is used to extract waveform pixels from distortion-free seismic waveform images, take the median of the ordinate corresponding to the same horizontal coordinate, map the pixel coordinates to the time and amplitude domains, apply filters to smooth the waveform, and perform baseline correction to determine the zero point, thereby achieving accurate digitization of the waveform in the observation drawing.

[0074] The module division in this embodiment of the invention is illustrative and represents only one logical functional division. In actual implementation, other division methods may be used. Furthermore, the functional modules in the various embodiments of the invention can be integrated into a single processor, exist as separate physical entities, or be integrated into a single module. The integrated modules described above can be implemented in hardware or as software functional modules.

[0075] In another embodiment of the present invention, a computer device is provided, comprising a processor and a memory. The memory stores a computer program, which includes program instructions. The processor executes the program instructions stored in the computer storage medium. The processor may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing and control core of the terminal, suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions in the computer storage medium to achieve a corresponding method flow or corresponding function. The processor described in this embodiment of the present invention can be used for the operation of a method for digitizing short-period seismograph observation maps.

[0076] In another embodiment of the present invention, a storage medium is provided, specifically a computer-readable storage medium (Memory), which is a memory device in a computer device used to store programs and data. It is understood that the computer-readable storage medium here can include both the built-in storage medium in the computer device and extended storage media supported by the computer device. The computer-readable storage medium provides storage space that stores the terminal's operating system. Furthermore, the storage space also stores one or more instructions suitable for loading and execution by a processor. These instructions can be one or more computer programs (including program code). It should be noted that the computer-readable storage medium here can be high-speed RAM or non-volatile memory, such as at least one disk storage device. The processor can load and execute one or more instructions stored in the computer-readable storage medium to implement the corresponding steps of the short-period seismograph observation map digitization method in the above embodiments.

[0077] 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 embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0078] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0079] 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.

[0080] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0081] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for digitizing short-period seismograph observation maps, characterized in that, include: Noise removal is performed on the short-period seismograph observation images obtained from the scan to obtain denoised images; Using a correction formula based on the length of the recording pen, the denoised image is corrected for arc distortion to obtain an image with arc distortion correction. Using a displacement formula based on the inherent properties of the recording pen, tilt distortion correction is performed on the image after arc distortion correction to obtain a distortion-free seismic waveform image; The waveform pixels are extracted from the distortion-free seismic waveform image, and the median of the ordinate corresponding to the same horizontal coordinate is taken. The pixel coordinates are mapped to the time domain and amplitude domain, the waveform is smoothed by applying a filter, and baseline correction is performed to determine the zero point, so as to achieve accurate digitization of the waveform of the observation map.

2. The method for digitizing short-period seismograph observation maps according to claim 1, characterized in that, The process of removing noise from the short-period seismograph observation images obtained by scanning to obtain denoised images includes: For noise that is different from the waveform color, read every pixel of the image, analyze the color value, set a color threshold range based on the color of the waveform lines, compare the color of each pixel with the threshold range, if the pixel color is not in the range, it is determined to be noise, and the color of the pixel is directly replaced with the set background color. This process traverses all pixels of the entire image. For noise that is the same as or close to the waveform color, manual processing is performed by setting the background color to white. After processing, the image is saved, resulting in a denoised image with a uniform white background and a clear black curve waveform.

3. The method for digitizing short-period seismograph observation maps according to claim 1, characterized in that, The process involves using a correction formula based on the length of the recording pen to correct arc distortion in the denoised image, resulting in an image with arc distortion correction, including: Based on the width and height of the denoised image, a brand new blank image is created as the output. A regular grid is generated on the pixel coordinates of the output image, and its coordinate values ​​are normalized to the range of [-1, 1]. Based on the curvature radius parameter obtained by converting the physical length R of the recording pen, the inverse mapping of the arc correction formula is solved. For each target coordinate point in the grid of the output image, its corresponding source coordinate point in the original denoised image is calculated using the inverse mapping formula. The calculated source coordinates are restricted to the effective coordinate range of the original image. Using the bilinear interpolation algorithm, the pixel value of the target point is calculated based on the pixel values ​​around the source coordinate point in the original image, and filled into the newly created blank image, finally obtaining the image after the arc distortion is corrected.

4. The method for digitizing short-period seismograph observation maps according to claim 3, characterized in that, The reverse mapping is the process of finding the original coordinates (x, y) from the corrected coordinates (x1, y1), which is the inverse process of solving the forward mapping formula. Solve the equation by finding x and y from x1 and y1 respectively: 。 5. The method for digitizing short-period seismograph observation maps according to claim 1, characterized in that, The method of using a displacement formula based on the recording pen's inherent properties to correct tilt distortion in the image after arc distortion correction, resulting in a distortion-free seismic waveform image, includes: Obtain the dimensions of the image after arc correction. Calculate the total displacement parameter D based on the instrument pen length and scanning resolution. Combined with the image height h, calculate a fixed slope value. Create a blank image with the same dimensions as the input image and a white background. Iterate through each pixel row of the input image. For the y-th row, calculate the required horizontal compensation displacement based on the slope. Iterate through each pixel in the row. For the x-th column, calculate the corresponding original column position in the input image. x Check the corresponding source column position original x If the value is within the image width range, then the input image will be displayed in (original) format. x The pixel value at position (x, y) is assigned to the blank image at position (x, y); if it is outside the range, the position remains white with background; after traversing all pixels, the corrected image is output.

6. The method for digitizing short-period seismograph observation maps according to claim 5, characterized in that, The formula for calculating the slope value is as follows: slope = -D / h, where D is the total pixel displacement; The formula for calculating the horizontal compensation displacement is as follows: compensation = -slope×y The formula for calculating the corresponding source column position is as follows: original x =x+compensation。 7. The method for digitizing short-period seismograph observation maps according to claim 1, characterized in that, The process of extracting waveform pixels from distortion-free seismic waveform images, taking the median of the ordinates corresponding to the same abscissa, mapping the pixel coordinates to the time and amplitude domains, applying filters to smooth the waveform, and performing baseline correction to determine the zero point, thereby achieving accurate digitization of the waveform in the observation map, includes: Load the corrected waveform image and convert it into a grayscale image. Apply Gaussian blur filter to the grayscale image and perform binarization processing on the image using an adaptive threshold algorithm. Perform morphological closing operation on the binary image, scan the entire image, identify and record the coordinate set of all white pixels, and the coordinates constitute the original pixel-level expression of the waveform lines. From all the acquired white pixels, find all unique x-coordinate values ​​(X values) and sort them in ascending order, with each X value representing a time sampling point. For each unique X value, find all corresponding y-coordinate values ​​for that column. Using the median algorithm, calculate the most representative y-coordinate value from these Y values. This method effectively eliminates interference from discrete noise points that may exist in the column and robustly determines the center position of the waveform. Map the pixel coordinates to physical quantities: based on the total duration T of the seismic signal recorded on the drawing, map the pixel x-coordinates to time values ​​proportionally. Simultaneously, linearly map the y-coordinates to the normalized amplitude range of [-1, 1]. Obtain a preliminary set of time-amplitude data pairs. Apply a Savitzky-Golay filter to this amplitude sequence for smoothing. The filter smooths the data by performing polynomial fitting within a sliding window, effectively suppressing high-frequency noise while better preserving the original morphological characteristics of the waveform. Baseline correction and amplitude normalization are performed by selecting a segment of signal data considered as background ground vibrations before the earthquake event and calculating the median amplitude value, which is then used as the baseline offset. The amplitude value of each data point in the entire waveform data sequence is subtracted from the baseline offset to make the signal fluctuate around the zero line before the event. After baseline correction, the maximum absolute value of the entire waveform amplitude is calculated, and all amplitude values ​​are divided by this maximum value to normalize the amplitude range of the entire waveform to the interval [-1, 1]. The obtained data points are resampled, a standard target sampling rate is set, and an equally spaced time axis is generated based on the original time range. The cubic spline interpolation method is used to calculate the amplitude value on the new equally spaced time axis based on the existing non-equally spaced time-amplitude data points. A high-quality digital seismic waveform time series with a fixed and regular sampling rate is output, completing the full conversion from analog drawing images to computable digital signals.

8. A short-period seismograph observation map digitization system, characterized in that, include: The data preprocessing module is used to remove noise from the short-period seismograph observation images obtained by scanning, so as to obtain denoised images; The arc distortion correction module is used to correct the arc distortion of the denoised image using a correction formula based on the length of the recording pen, so as to obtain the image after arc distortion correction. The tilt distortion correction module is used to correct the tilt distortion of images after arc distortion correction using a correction formula based on a fixed speed and time of the recording pen, so as to obtain distortion-free seismic waveform images. The digitization module is used to extract waveform pixels from distortion-free seismic waveform images, take the median of the ordinate corresponding to the same horizontal coordinate, map the pixel coordinates to the time and amplitude domains, apply filters to smooth the waveform, and perform baseline correction to determine the zero point, thereby achieving accurate digitization of the waveform in the observation drawing.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the short-period seismograph observation map digitization method as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the method for digitizing short-period seismograph observation maps as described in any one of claims 1 to 7.

Images