A fast estimation method of shield slag gradation based on monitoring video

Abstract

The present invention provides a method for quickly estimating shield slag gradation based on monitoring video, which is characterized by comprising the following steps: installing a video sensor at a shield slag outlet to realize real-time collection of slag video recordings, and then using a frame difference method to obtain white pixel areas in each video, i.e., slag blocks, after averaging the slag video recordings, thereby solving the problem that a large number of slag blocks cannot be captured by the traditional frame difference method; at the same time, a calibration coefficient is obtained based on the ratio of the image size of a chessboard calibration plate to the size of a chessboard calibration plate; the size of the white pixel area, i.e., the block, in the video frame image is calculated based on the calibration coefficient, so that the calculated block size is closer to its actual size; finally, because the final slag gradation result is obtained by calculating the slag gradation of each gradation section and then based on the average calculation of the image sequence, the estimation of the final slag gradation is more accurate.

Description

Technical Field

[0001] The invention relates to a method for quickly estimating shield slag gradation based on monitoring video. Background Art

[0002] With the acceleration of urbanization, the demand for shield tunnel construction in my country is increasing, among which slurry shield is one of the important construction methods. During the construction of slurry shield engineering, a large amount of slag and rock waste will be generated as the excavation face advances. After screening, the slag has the opportunity to be recycled and become a building material resource again. Due to the characteristics of the shield tunnel construction method, there are few opportunities to directly contact the geological excavation face. The gradation parameters of the unscreened slag depend on the slurry ingredients and surrounding rock characteristics, which can indirectly reflect the geological parameters to a certain extent.

[0003] At present, the calculation and control of shield slag discharge parameters mainly rely on the experience of the shield driver, which is subjective. As for the gradation of shield slag discharge, there is currently no real-time calculation method in the slag discharge stage. At present, it is mainly estimated by screening after slag discharge. Because this method has a lag, the distribution of block stone particle size can only be determined after the secondary screening of slag, and the slag discharge loses the effect of dynamically reflecting geological parameters in real time. Moreover, the amount of slag in each ring of shield excavation is usually large, so the particle size determination level of the slag is also relatively coarse, which is not conducive to understanding the geological conditions of the excavation process through slag discharge. Therefore, a fast calculation method for shield slag discharge parameters is urgently needed in shield tunnel engineering.

[0004] Existing video moving target detection methods mainly include frame difference method, optical flow method and background subtraction method. Among them, the traditional frame difference method is only suitable for capturing a small number of moving objects, while in the capture of shield slag, there are many moving targets, and the focus is not only on the capture of moving targets, but also on the distribution of block stone particle gradation within a video range, which is difficult to deal with by the traditional frame difference method. Therefore, it is worth studying how to make full use of the results of motion capture by the frame difference method and construct an appropriate shield slag gradation calculation method. Summary of the invention

[0005] In order to solve the above problems, a method for accurately and real-time estimating slag gradation is provided. The present invention adopts the following technical solutions:

[0006] The present invention provides a method for quickly estimating shield slag gradation based on monitoring video, which is characterized by comprising the following steps: step S1, installing a video sensor at the shield slag outlet to obtain monitoring video, and marking the slag outlet position based on the monitoring video; step S2, selecting a time period when the shield does not discharge slag, placing chessboard calibration plates of different sizes at the slag outlet position, obtaining chessboard calibration plate images of different sizes corresponding to the chessboard calibration plates of different sizes based on the video sensor, and obtaining a calibration coefficient according to the ratio of the size of the chessboard calibration plate image to the actual size of the chessboard calibration plate; step S3, setting a video recording time interval t for each slag discharge for the video sensor, obtaining a slag discharge video in real time, and dividing the slag discharge video into n parts, each video They are respectively denoted as fi(x), wherein i=1,2…n, and x represents the xth frame of the image in the i-th video; step S4, obtaining m consecutive color image frames in each video fi(x); step S5, processing the m color image frames to obtain a predetermined number of black-and-white images, calculating the actual sizes of the white pixel areas in the predetermined number of black-and-white images based on the calibration coefficients and counting their distribution, to obtain the number s of white pixel areas in each gradation section; step S6, based on the number s of white pixel areas in each gradation section, counting the final slag gradation values ​​of the n videos, wherein the white pixel areas are the slag blocks, and each gradation section is obtained by averaging the preset interval of the block stone particle size.

[0007] The present invention provides a method for quickly estimating shield slag gradation based on monitoring video, which may also have the following technical features, wherein step S5 includes the following sub-steps: step S5-1, graying m frames of color images to obtain m frames of grayscale images; step S5-2, eliminating part of the noise in the m frames of grayscale images; step S5-3, processing the denoised m frames of grayscale images based on a frame difference method to obtain m-1 black-and-white images; step S5-4, recording the white pixel areas in the m-1 black-and-white images as slag blocks, and obtaining the maximum size h(x) of each white pixel area according to the calibration coefficient, and counting the distribution of h(x) in the m-1 black-and-white images to obtain the number s of white pixel areas in each gradation section.

[0008] The present invention provides a method for quickly estimating shield slag gradation based on monitoring video, which may also have such a technical feature, wherein, in step S5-2, a Gaussian filter is used for smoothing to eliminate part of the noise.

[0009] The present invention provides a method for quickly estimating shield slag gradation based on monitoring video, which may also have such a technical feature, wherein step S5-3 includes the following steps:

[0010] Subtract the m frames of grayscale images in order, that is, subtract the m-1 frame of grayscale image from the m-th frame of grayscale image, to obtain m-1 grayscale images, satisfying the expression:

[0011] K i (j) = fi(m) - fi(m-1)

[0012] In the formula, K i (j) is the grayscale image sequence of the i-th video, where j = 1, 2...m-1; for the grayscale image sequence K of the i-th video i (j) Binarization is performed to obtain the corresponding black and white image sequence G i (j), the sequence G i (j) contains m-1 black and white images, where the binarization processing formula is:

[0013]

[0014] Wherein, L is a preset threshold value, which is set as: L = (I max -(I max -I min )) / 2, where I max and I min Respectively represent the maximum and minimum grayscale values ​​of the image.

[0015] The present invention provides a method for quickly estimating shield slag gradation based on monitoring video, which may also have such a technical feature, wherein the calculation formula of the final slag gradation value is:

[0016]

[0017] Where W z is the final slag gradation value of n videos, s ij (z) is the number of white pixel areas in the zth grading segment of the jth black-and-white image of the i-th video.

[0018] Function and Effect of the Invention

[0019] According to a method for quickly estimating shield slag gradation based on monitoring video of the present invention, a video sensor is installed at the shield slag outlet to realize real-time collection of slag video. Since the frame difference method is used to obtain the white pixel area in each video, i.e., the slag block, after the slag video is evenly divided, the problem that the traditional frame difference method cannot capture a large number of slag blocks is solved. At the same time, the calibration coefficient in the chessboard calibration method is used to calculate the size of the block in the video frame image, so that the calculated block size is closer to its actual size. Finally, since the final slag gradation result is obtained by calculating the slag gradation of each grading section and then based on the average calculation of the image sequence, the estimation of the final slag gradation is more accurate.

[0020] The method for quickly estimating shield slag gradation based on monitoring video of the present invention realizes dynamic capture of pixel areas of slag blocks by combining video segmentation with the frame difference method, and obtains the slag gradation result of the entire video based on statistics of the pixel areas of the blocks in the segmented video, thereby establishing a method for quickly estimating shield slag gradation based on monitoring video (or image). BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 It is a flow chart of a method for quickly estimating shield slag gradation based on monitoring video in an embodiment of the present invention;

[0022] Figure 2 is a diagram showing an example of installing a video sensor in an embodiment of the present invention;

[0023] Figure 3 is a schematic diagram of a chessboard calibration plate image in an embodiment of the present invention;

[0024] Figure 4 It is a flow chart of obtaining the number of white pixel areas in each grading section in an embodiment of the present invention. DETAILED DESCRIPTION

[0025] In order to make the technical means, creative features, objectives and effects achieved by the present invention easy to understand, the method for rapid estimation of shield slag gradation based on monitoring video of the present invention is specifically described below in combination with embodiments and drawings.

[0026] <Example>

[0027] Figure 1 It is a flow chart of a method for quickly estimating shield slag gradation based on monitoring video in an embodiment of the present invention.

[0028] like Figure 1 As shown in FIG. 1 , the shield slag gradation rapid estimation method based on monitoring video includes the following steps:

[0029] Step S1, installing a video sensor at the slag outlet of the shield machine to obtain a monitoring video, and marking the slag outlet position based on the monitoring video.

[0030] Figure 2 2 is a diagram showing an installation example of a video sensor in an embodiment of the present invention.

[0031] In this embodiment, the video sensor is installed on the slag discharge scaffolding of the slurry shield, and the installation position and distance are based on the principle of taking the most ideal picture and not interfering with the construction.

[0032] The video sensor is a solar camera with wireless network card transmission function (such as Figure 2 The battery capacity of the solar camera is enough to last for more than 25 consecutive rainy days. To ensure the video quality, the solar camera configuration should be at least higher than 2 million pixels.

[0033] Step S2, selecting a time period when the shield does not discharge slag, placing chessboard calibration plates of different sizes at the slag discharge position, and obtaining chessboard calibration plate images of different sizes corresponding to the chessboard calibration plates of different sizes based on the video sensor (such as Figure 3 As shown), the calibration coefficient is obtained according to the ratio of the size of the chessboard calibration plate image to the actual size of the chessboard calibration plate.

[0034] In this embodiment, the size of the chessboard calibration plate image is set to be m, and the actual size of the chessboard calibration plate is n, then the calibration coefficient k=n / m.

[0035] Step S3, setting the video sensor with a recording time interval t for each slag discharge, obtaining the slag discharge video in real time, and dividing the slag discharge video into n parts, each of which is recorded as fi(x), where i=1,2…n, and x represents the xth frame of the image in the i-th video.

[0036] Step S4, obtaining m consecutive color image frames in each video fi(x).

[0037] Step S5, processing m frames of color images to obtain a predetermined number of black-and-white images, calculating the actual sizes of white pixel areas in the predetermined number of black-and-white images based on the calibration coefficients and counting their distribution, to obtain the number s of white pixel areas in each grading section.

[0038] In this embodiment, the white pixel area is recorded as the slag block, and the grading is the interval range of the particle size of the block. The interval range can be preset in advance according to actual needs. In this implementation, the minimum particle size of the block is set to Q, and the maximum particle size is set to Q1. Then the particle size interval of the block is [Q, Q1]. The particle size interval is evenly divided to obtain each grading section, and the proportion of each grading section is counted separately.

[0039] Figure 4 It is a flow chart of obtaining the number of white pixel areas in each grading section in an embodiment of the present invention.

[0040] like Figure 4 As shown, step S5 includes the following sub-steps:

[0041] Step S5-1, grayscale processing is performed on m frames of color images to obtain m frames of grayscale images.

[0042] Step S5-2, using a Gaussian filter to perform smoothing processing to eliminate part of the noise of the m frames of grayscale images.

[0043] Step S5-3, using the frame difference method to process the m frames of grayscale images after noise removal to obtain m-1 black-and-white images.

[0044] The above step S5-3 is specifically as follows:

[0045] First, subtract the m-frame grayscale images in order, that is, subtract the m-1-frame grayscale image from the m-frame grayscale image to obtain m-1 grayscale images, satisfying the expression:

[0046] K i (j) = fi(m) - fi(m-1)

[0047] In the formula, K i (j) is the grayscale image sequence of the i-th video, where j=1, 2...m-1.

[0048] Then, the grayscale image sequence Ki(j) of the above i-th video is binarized to obtain the corresponding black-and-white image sequence G i (j), the black and white image sequence G i (j) Contains m-1 black and white images.

[0049] Among them, the binarization processing formula is:

[0050]

[0051] Wherein, L is a preset threshold value, which is set as: L = (I max -(I max -I min )) / 2, where I max and I min Respectively represent the maximum and minimum grayscale values ​​of the image.

[0052] Step S5-4, record the white pixel areas in the m-1 black-and-white images as slag blocks, obtain the maximum size h(x) of each white pixel area according to the calibration coefficient, and count the distribution of h(x) in the m-1 black-and-white images to obtain the number s of white pixel areas in each grading section.

[0053] Step S6: based on the number s of white pixel areas in each gradation section, the final slag gradation values ​​of the n videos are counted.

[0054] In this embodiment, the calculation formula of the slag gradation is:

[0055]

[0056] Where W z is the final slag gradation value of n videos, s ij (z) is the number of white pixel areas in the zth grading segment of the jth black-and-white image of the i-th video.

[0057] Example Function and Effect

[0058] According to the method for quickly estimating shield slag gradation based on monitoring video provided in this embodiment, a video sensor is installed at the shield slag outlet to realize real-time acquisition of slag video, and the slag video is averaged and smoothed by Gaussian filter to eliminate some noise in the video. Then, the white pixel area in each video, i.e., the slag block, is obtained by frame difference method, and the calibration coefficient in the chessboard calibration method is used to calculate the size of the block, so that the calculated block size is closer to its actual size. Finally, the number of blocks in each grading section is counted, and the final slag gradation result is obtained based on the average calculation of the image sequence, so that the final slag gradation estimation is not only fast but also more accurate.

[0059] In the embodiment, the slag discharge video is first divided into n parts, and then the frame difference method is used to capture the white pixel area, i.e., the rock blocks, in each video, thereby solving the problem that the traditional frame difference method cannot cope with the capture of rock blocks in the shield slag discharge. At the same time, since the final slag gradation result is obtained by calculating the slag gradation of each grading section and then averaging the image sequence, the estimation of the slag gradation is more accurate.

[0060] The above embodiments are only used to illustrate specific implementation modes of the present invention, and the present invention is not limited to the description scope of the above embodiments.

Claims

1. A method for quickly estimating shield slag gradation based on monitoring video, characterized in that: The following steps are involved: Step S1, installing a video sensor at the slag outlet of the shield machine to obtain a monitoring video, and marking the slag outlet position based on the monitoring video; Step S2, selecting a time period when the shield does not discharge slag, placing chessboard calibration plates of different sizes at the slag discharge position, acquiring chessboard calibration plate images of different sizes corresponding to the chessboard calibration plates of different sizes based on the video sensor, and acquiring a calibration coefficient according to the ratio of the size of the chessboard calibration plate image to the actual size of the chessboard calibration plate; Step S3, setting a video recording time interval t for each slag discharge for the video sensor, obtaining a slag discharge video in real time, and dividing the slag discharge video into n parts, each of which is recorded as fi(x), where i=1,2…n, and x represents the xth frame of the image in the i-th video; Step S4, obtaining m consecutive color image frames in each video fi(x); Step S5, processing the m color images by using a frame difference method to obtain a predetermined number of black-and-white images, calculating the actual sizes of white pixel areas in the predetermined number of black-and-white images based on the calibration coefficients and counting their distribution, and obtaining the number s of the white pixel areas in each grading section; Step S6, based on the number s of the white pixel areas in each grading section, the final slag grading values ​​of n videos are counted. The white pixel area is the slag block. Each grading section is obtained by evenly dividing the preset interval of the block stone particle size. The calculation formula of the final slag gradation value is: Where W z is the final slag gradation value of the n videos, s ij (z) is the number of white pixel areas in the zth grading segment of the jth black-and-white image of the ith video.

2. According to the method for rapid estimation of shield slag gradation based on monitoring video in claim 1, Features: in, The step S5 comprises the following sub-steps: Step S5-1, graying the m color image frames to obtain m gray image frames; Step S5-2, eliminating some noise of the m frames of grayscale images; Step S5-3, processing the m frames of grayscale images after denoising based on the frame difference method to obtain m-1 black-and-white images; Step S5-4, record the white pixel areas in the m-1 black-and-white images as slag blocks, and obtain the maximum size h(x) of each of the white pixel areas according to the calibration coefficient, and count the distribution of h(x) in the m-1 black-and-white images to obtain the number s of the white pixel areas in each grading section.

3. The method for quickly estimating shield slag gradation based on monitoring video according to claim 2 is characterized in that: in, In step S5-2, a Gaussian filter is used to perform smoothing processing to eliminate part of the noise.

4. The method for quickly estimating shield slag gradation based on monitoring video according to claim 2, Features: in, The step S5-3 comprises the following steps: Subtract the m frames of grayscale images in order, that is, subtract the m-1 frame of grayscale image from the m-th frame of grayscale image, to obtain m-1 grayscale images, satisfying the expression: K i (j)=fi(m)-fi(m-1) In the formula, K i (j) is the grayscale image sequence of the i-th video, where j = 1, 2…m-1; For the grayscale image sequence K of the i-th video i (j) Binarization is performed to obtain the corresponding black and white image sequence G i (j), the sequence G i (j) contains m-1 black and white images, Among them, the binarization processing formula is: Wherein, L is a preset threshold value, which is set as: L = (I max -(I max -I min )) / 2, where I max and I min Respectively represent the maximum and minimum grayscale values ​​of the image.

Images