A method and system for scanning and measuring the particle size distribution of sand and gravel particles

Abstract

The embodiment of the application discloses a method and system for scanning and measuring the particle size grading of sand and gravel particles. The method comprises: performing binaryzation processing on each frame of image of a high-speed image sequence of falling sand and gravel particles; extracting each connected domain in the binaryzation image; calculating the texture energy reduction degree of the connected domain relative to the nearest neighbor connected domain based on the image high-frequency energy attenuation principle to obtain a texture energy reduction value of the connected domain; calculating the discrete degree of the distance from the edge sampling point in the connected domain to the geometric center based on the morphological skeleton extraction principle to obtain a skeleton distance reduction amount of the connected domain; fusing the texture energy reduction value and the skeleton distance reduction amount to obtain a fuzzy domain score of the connected domain; determining a credibility quantization value of the connected domain according to the motion feature and the fuzzy domain score of the connected domain in adjacent two frames of binaryzation images; when the credibility quantization value is less than a preset credibility threshold, the connected domain is removed; otherwise, the connected domain is reserved as an effective sand and gravel particle region.

Description

Technical Field

[0001] This application relates to the field of image data processing technology, and relates to, but is not limited to, a method and system for scanning and measuring the particle size distribution of sand and gravel particles. Background Technology

[0002] Particle size distribution refers to the percentage distribution of different particle sizes by mass in sand and gravel. In engineering, particle size distribution affects the workability, strength, durability, and economy of concrete. Well-graded sand and gravel can reduce porosity, achieving ideal flowability with less cement paste. Simultaneously, the dense skeletal structure improves the strength, impermeability, and frost resistance of concrete, saving 10%-15% on cement usage. Traditional particle size distribution measurement mainly uses sieving, which involves layer-by-layer sieving of sand and gravel through a series of sieves with different apertures. However, this method is time-consuming and labor-intensive, making it difficult to meet the real-time online monitoring requirements of industrial production.

[0003] The related technology is a particle size analysis method based on the principle of laser diffraction. This method uses a laser beam to irradiate a dispersed group of sand and gravel particles, and inverts the particle size distribution by measuring the angular distribution of the scattered light. Laser diffraction has advantages such as fast measurement speed, good repeatability, and no need to contact the sample. It can perform statistical averaging on a large number of particles and has been applied to a certain extent for the rapid detection of sand and gravel particle size.

[0004] However, laser diffraction relies on the assumption of spherical particles to invert particle size distribution, while actual sand and gravel particles are highly irregular in shape, often angular or flaky, leading to a systematic deviation between the measured results and the true particle size. Furthermore, laser diffraction cannot obtain the shape parameters of individual particles, such as roundness and aspect ratio, which are crucial for assessing aggregate quality and concrete performance. Therefore, how to simultaneously achieve rapid measurement and obtain particle morphology parameters has become a key technical challenge for improving the accuracy of sand and gravel particle size distribution detection. Summary of the Invention

[0005] In view of this, the embodiments of this application provide a method and system for scanning and measuring the particle size distribution of sand and gravel, which at least solves the technical problem of not being able to improve the accuracy of sand and gravel particle size distribution detection while simultaneously performing rapid measurement.

[0006] The technical solution of this application embodiment is implemented as follows: In a first aspect, embodiments of this application provide a method for scanning and measuring the particle size distribution of sand and gravel particles, the method comprising: A high-speed image sequence containing falling sand and gravel particles is acquired; each frame of the high-speed image sequence is binarized to obtain a binarized image. Extract each connected component from the binarized image; the connected component is a set of pixels in the binarized image that have the same pixel value and are adjacent in position; For each connected component, based on the principle of high-frequency energy attenuation in images, the degree of texture energy reduction of the connected component relative to its nearest neighbor connected component is calculated to obtain the texture energy reduction value of the connected component. For each connected component, based on the principle of morphological skeleton extraction, the dispersion of the distance from the edge sampling point to the geometric center within the connected component is calculated to obtain the skeleton distance reduction of the connected component. The texture energy reduction value and the skeleton distance reduction are fused according to a preset weight to obtain the fuzzy domain score of the connected component. For each connected component, based on the principle of inter-frame motion tracking, the confidence quantization value of the connected component is determined according to the motion features of the connected component in the binarized images of two adjacent frames and the fuzzy domain score. The credibility quantification value is compared with a preset credibility threshold; when the credibility quantification value is less than the preset credibility threshold, the connected component is determined to be a ghosting region and removed from the binarized image; otherwise, the connected component is retained as a valid sand and gravel particle region.

[0007] Secondly, embodiments of this application provide a scanning measurement system for particle size distribution of sand and gravel particles, the system comprising: an image acquisition and processing module, a connected component extraction module, and a machine vision analysis module; The image acquisition and processing module is used to acquire a high-speed image sequence containing falling sand and gravel particles; and to perform binarization processing on each frame of the high-speed image sequence to obtain a binarized image. The connected component extraction module is used to extract each connected component in the binarized image; the connected component is a set of pixels in the binarized image that have the same pixel value and are adjacent in position; The machine vision analysis module is used to calculate the degree of texture energy reduction of each connected component relative to its nearest neighbor connected component based on the principle of high-frequency energy attenuation in images, and to obtain the texture energy reduction value of the connected component. The machine vision analysis module is used to calculate the dispersion of the distance from the edge sampling point to the geometric center within each connected component based on the principle of morphological skeleton extraction, and to obtain the skeleton distance reduction of the connected component. The machine vision analysis module is used to fuse the texture energy reduction value and the skeleton distance reduction amount according to a preset weight to obtain the fuzzy domain score of the connected component. The machine vision analysis module is used to determine the confidence quantization value of each connected component based on the principle of inter-frame motion tracking, according to the motion features of the connected component in two adjacent binarized images and the fuzzy domain score. The machine vision analysis module is further configured to compare the credibility quantification value with a preset credibility threshold; when the credibility quantification value is less than the preset credibility threshold, the connected component is determined to be a ghosting region and removed from the binarized image; otherwise, the connected component is retained as a valid sand and gravel particle region.

[0008] The beneficial effects of the technical solutions provided in this application include at least the following: This application acquires a high-speed image sequence containing falling sand and gravel particles; each frame of the high-speed image sequence is binarized to obtain a binarized image. Binarization separates the sand and gravel particle region from the background region, making the sand and gravel particles appear as a set of pixels with the same pixel value in the binarized image. Each connected component in the binarized image is extracted; a connected component is a set of pixels with the same pixel value and adjacent positions in the binarized image. Identifying and separating each connected component from the binarized image allows for the identification of the separated sand and gravel particle regions one by one, with each connected component corresponding to a sand and gravel particle and a trailing region. For each connected component, based on the principle of high-frequency energy attenuation in images, the degree of texture energy reduction of the connected component relative to its nearest neighbor connected component is calculated, obtaining the texture energy reduction value of the connected component. Motion blur can cause a decrease in the amplitude of high-frequency components in connected components. By comparing the difference in high-frequency component amplitudes between the current connected component and its nearest neighbor connected components, the texture energy reduction value of the connected component is obtained. The larger the texture energy reduction value, the more severe the high-frequency energy attenuation of the current connected component, and the higher the probability that the connected component is a blur region. For each connected component, based on the principle of morphological skeleton extraction, the dispersion of the distance from the edge sampling points to the geometric center is calculated to obtain the skeleton distance reduction amount of the connected component. Real sand and gravel particles are usually compact in shape, and the distance from the edge sampling points to the geometric center of the connected component fluctuates little, while blur regions often present a slender shape extending from the edge of the body, and the distance from the edge sampling points to the geometric center fluctuates greatly. By calculating the skeleton distance reduction amount, real particles and blur regions can be distinguished, avoiding misjudging small particles with low texture energy as blur. The texture energy reduction value and the skeleton distance reduction amount are fused according to preset weights to obtain the blur region score of the connected component. Texture energy reduction values ​​focus on detecting high-frequency energy loss caused by motion blur, while skeleton distance reduction focuses on detecting the elongated shape characteristics unique to motion blur. Single features may lead to misclassification when identifying fine particles. By weighted fusion of texture energy reduction values ​​and skeleton distance reduction, both texture and shape information can be utilized simultaneously, allowing the fuzzy domain score to more accurately reflect motion blur in connected components, thus reducing misclassification. For each connected component, based on the principle of inter-frame motion tracking, a reliability quantification value is determined according to the motion characteristics and fuzzy domain score of the connected component in two adjacent binarized images. The direction of motion blur-induced motion blur should be consistent with the actual movement direction of the sand particles, and the faster the movement speed, the more severe the motion blur. By calculating the reliability quantification value, it is possible to verify whether the motion blur conforms to the motion law, thereby further improving the accuracy of the judgment and avoiding misclassification of static textured blurred areas as motion blur. The lower the reliability quantification value, the more likely the area is to be a motion blur area that needs to be removed.The reliability quantification value is compared with a preset reliability threshold. When the reliability quantification value is less than the preset reliability threshold, the connected component is identified as a motion blur region and removed from the binarized image; otherwise, the connected component is retained as a valid sand and gravel particle region. By setting a preset reliability threshold, connected components with low reliability are filtered out and removed, thereby eliminating the influence of motion blur on sand and gravel particle size calculation, while retaining the real sand and gravel particle regions with high reliability. Therefore, this application improves the accuracy of sand and gravel particle size distribution detection while performing rapid measurement. Attached Figure Description

[0009] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort, wherein: Figure 1 A schematic flowchart illustrating a method for scanning and measuring the particle size distribution of sand and gravel particles, provided in an embodiment of this application; Figure 2 This is a schematic diagram of a scanning measurement system for particle size distribution of sand and gravel provided in an embodiment of this application. Detailed Implementation

[0010] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. The following embodiments are used to illustrate this application, but are not intended to limit the scope of this application. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0011] In the following description, references are made to “some embodiments,” which describe a subset of all possible embodiments. However, it is understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0012] It should be noted that the terms "first, second, and third" used in the embodiments of this application are merely to distinguish similar objects and do not represent a specific ordering of objects. It is understood that "first, second, and third" can be interchanged in a specific order or sequence where permitted, so that the embodiments of this application described herein can be implemented in an order other than that illustrated or described herein.

[0013] It will be understood by those skilled in the art that, unless otherwise defined, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art to which the embodiments of this application pertain. It should also be understood that terms such as those defined in general dictionaries should be understood to have a meaning consistent with their meaning in the context of the prior art, and should not be interpreted in an idealized or overly formal sense unless specifically defined as herein.

[0014] This application provides a method for scanning and measuring the particle size distribution of sand and gravel particles. Figure 1 A schematic flowchart illustrating a method for scanning and measuring the particle size distribution of sand and gravel particles, as provided in this application embodiment, is shown below. Figure 1 As shown, the method includes at least the following steps: Step S110: Obtain a high-speed image sequence containing falling sand and gravel particles; perform binarization processing on each frame of the high-speed image sequence to obtain a binarized image.

[0015] A high-speed image sequence is a series of high-speed images containing falling sand and gravel particles, captured continuously by a high-speed industrial camera. Each frame of the high-speed image is a single image from the high-speed image sequence.

[0016] Binarization is the process of converting each frame of a high-speed image into an image with only two pixel values. A binarized image is the image obtained after binarization.

[0017] Image data of sand and gravel particles during their movement is acquired and converted into a binarized image format suitable for subsequent processing. Binarization separates the sand and gravel particle area from the background area, making the sand and gravel particles appear as a set of pixels with the same pixel value in the binarized image.

[0018] Step S120: Extract each connected component in the binarized image; the connected component is a set of pixels in the binarized image that have the same pixel value and are adjacent in position.

[0019] In some embodiments, in the binarized image, foreground pixels correspond to sand and gravel particle regions, and background pixels correspond to non-particle regions. Specifically, sand and gravel particle regions are displayed as white in the image, while background regions without sand and gravel particles are displayed as black. Therefore, white regions are foreground regions, corresponding to sand and gravel particles; black regions are background regions, corresponding to non-particle parts. In this case, each independent white region in the binarized image is a connected component.

[0020] A connected component is a set of pixels in a binarized image that have the same pixel value and are adjacent in position. Specifically, when a grain of sand falls at high speed and experiences motion blur (i.e., a trailing image), the white pixels of the sand grain itself and the trailing image will connect together, forming a connected component larger than the actual sand grain. This connected component includes both the actual projection of the sand grain and the trailing image caused by motion blur.

[0021] Identifying and separating each connected component from a binarized image allows for the identification of individual sand and gravel particle regions within the image, with each connected component corresponding to a sand and gravel particle and a trailing region.

[0022] Step S130: For each connected component, based on the principle of high-frequency energy attenuation in images, calculate the degree of texture energy reduction of the connected component relative to its nearest neighbor connected component, and obtain the texture energy reduction value of the connected component.

[0023] The nearest neighbor connected component is the connected component that is spatially closest to the current connected component. The principle of high-frequency energy attenuation in images refers to the principle that the amplitude of high-frequency components in an image region decreases when motion blur occurs. Texture energy reduction is the degree of difference between the high-frequency component amplitudes of the current connected component and those of the nearest neighbor connected component.

[0024] Motion blur can cause a decrease in the amplitude of high-frequency components in connected regions. By comparing the difference in the amplitude of high-frequency components between the current connected region and the nearest neighbor connected region, the texture energy reduction value of the connected region can be obtained. The larger the texture energy reduction value, the more severe the high-frequency energy attenuation of the current connected region, and the higher the probability that the connected region is a motion blur region.

[0025] Step S140: For each connected component, based on the principle of morphological skeleton extraction, calculate the dispersion of the distance from the edge sampling point to the geometric center within the connected component, and obtain the skeleton distance reduction of the connected component.

[0026] Morphological skeleton extraction works by extracting the central axis of a connected component through image morphological operations. Edge sampling points are pixels selected at regular intervals along the edges of the connected component. The geometric center is the average position of the coordinates of all pixels within the connected component. The dispersion of the distances from the edge sampling points to the geometric center represents the degree of fluctuation of these distances relative to their average value.

[0027] For each connected component, based on the principle of morphological skeleton extraction, the discreteness of the distance from the edge sampling point to the geometric center within the connected component is calculated to obtain the skeleton distance reduction, which is used to characterize the elongation of the shape of the connected component.

[0028] Real sand and gravel particles are typically compact in shape, with minimal fluctuations in the distance from the edge sampling points of connected domains to the geometric center. In contrast, ghosting regions often exhibit elongated shapes extending from the body's edge, with greater fluctuations in the distance from their edge sampling points to the geometric center. By calculating the skeleton distance reduction, we can distinguish between real particles and ghosting regions, avoiding misclassifying small particles with low texture energy as ghosting.

[0029] Step S150: The texture energy reduction value and the skeleton distance reduction amount are fused according to a preset weight to obtain the fuzzy domain score of the connected domain.

[0030] The preset weights are pre-defined coefficients used to adjust the proportions of texture energy reduction and skeleton distance reduction in the fusion process. The fuzzy domain score of the connected component is a value obtained through the fusion operation, used to characterize the probability that the connected component is a blurred or trailing region.

[0031] Texture energy reduction values ​​focus on detecting high-frequency energy loss caused by motion blur, while skeleton distance reduction focuses on detecting the elongated shape characteristics unique to motion blur. Single features may lead to misclassification when identifying fine particles. By weighted fusion of texture energy reduction values ​​and skeleton distance reduction, both texture and shape information can be utilized simultaneously, allowing the fuzzy domain score to more accurately reflect motion blur in the connected domain, thereby reducing misclassification.

[0032] Step S160: For each connected component, based on the principle of inter-frame motion tracking, determine the quantized value of the connected component's credibility according to the motion features of the connected component in the binarized images of two adjacent frames and the fuzzy domain score.

[0033] The principle of inter-frame motion tracking is to estimate the motion state of sand and gravel particles by comparing the position changes of the same connected region in two adjacent frames.

[0034] Motion characteristics are information about positional changes, which can include displacement direction and magnitude. The reliability metric is used to characterize the reliability of the connected domain for particle size distribution calculations.

[0035] The direction of the motion blur should be consistent with the actual movement direction of the sand and gravel particles, and the faster the movement speed, the more severe the motion blur. By calculating the confidence quantization value, it can be verified whether the motion blur conforms to the motion law, thereby further improving the accuracy of the judgment and avoiding misjudging static textured blurred areas as motion blur. The lower the confidence quantization value, the more likely the area is to be a motion blur area that needs to be removed.

[0036] Step S170: Compare the confidence quantification value with a preset confidence threshold; when the confidence quantification value is less than the preset confidence threshold, the connected component is determined to be a ghosting region and removed from the binarized image; otherwise, the connected component is retained as a valid sand and gravel particle region.

[0037] The preset confidence threshold is a pre-defined numerical limit used to determine whether a confidence quantification is qualified. Motion blur regions are false pixel regions caused by motion blur, not real sand particles. Valid sand particle regions are connected components of real sand particles.

[0038] By setting a preset confidence threshold, connected components with low confidence are filtered out and removed, thereby eliminating the influence of motion blur on the calculation of sand and gravel particle size, while retaining the real sand and gravel particle regions with high confidence. Therefore, while achieving rapid measurement, the accuracy of sand and gravel particle size distribution detection is also improved.

[0039] This application acquires a high-speed image sequence containing falling sand and gravel particles; each frame of the high-speed image sequence is binarized to obtain a binarized image. Binarization separates the sand and gravel particle region from the background region, making the sand and gravel particles appear as a set of pixels with the same pixel value in the binarized image. Each connected component in the binarized image is extracted; a connected component is a set of pixels with the same pixel value and adjacent positions in the binarized image. Identifying and separating each connected component from the binarized image allows for the identification of the separated sand and gravel particle regions one by one, with each connected component corresponding to a sand and gravel particle and a trailing region. For each connected component, based on the principle of high-frequency energy attenuation in images, the degree of texture energy reduction of the connected component relative to its nearest neighbor connected component is calculated, obtaining the texture energy reduction value of the connected component. Motion blur can cause a decrease in the amplitude of high-frequency components in connected components. By comparing the difference in high-frequency component amplitudes between the current connected component and its nearest neighbor connected components, the texture energy reduction value of the connected component is obtained. The larger the texture energy reduction value, the more severe the high-frequency energy attenuation of the current connected component, and the higher the probability that the connected component is a blur region. For each connected component, based on the principle of morphological skeleton extraction, the dispersion of the distance from the edge sampling points to the geometric center is calculated to obtain the skeleton distance reduction amount of the connected component. Real sand and gravel particles are usually compact in shape, and the distance from the edge sampling points to the geometric center of the connected component fluctuates little, while blur regions often present a slender shape extending from the edge of the body, and the distance from the edge sampling points to the geometric center fluctuates greatly. By calculating the skeleton distance reduction amount, real particles and blur regions can be distinguished, avoiding misjudging small particles with low texture energy as blur. The texture energy reduction value and the skeleton distance reduction amount are fused according to preset weights to obtain the blur region score of the connected component. Texture energy reduction values ​​focus on detecting high-frequency energy loss caused by motion blur, while skeleton distance reduction focuses on detecting the elongated shape characteristics unique to motion blur. Single features may lead to misclassification when identifying fine particles. By weighted fusion of texture energy reduction values ​​and skeleton distance reduction, both texture and shape information can be utilized simultaneously, allowing the fuzzy domain score to more accurately reflect motion blur in connected components, thus reducing misclassification. For each connected component, based on the principle of inter-frame motion tracking, a reliability quantification value is determined according to the motion characteristics and fuzzy domain score of the connected component in two adjacent binarized images. The direction of motion blur-induced motion blur should be consistent with the actual movement direction of the sand particles, and the faster the movement speed, the more severe the motion blur. By calculating the reliability quantification value, it is possible to verify whether the motion blur conforms to the motion law, thereby further improving the accuracy of the judgment and avoiding misclassification of static textured blurred areas as motion blur. The lower the reliability quantification value, the more likely the area is to be a motion blur area that needs to be removed.The reliability quantification value is compared with a preset reliability threshold. When the reliability quantification value is less than the preset reliability threshold, the connected component is identified as a motion blur region and removed from the binarized image; otherwise, the connected component is retained as a valid sand and gravel particle region. By setting a preset reliability threshold, connected components with low reliability are filtered out and removed, thereby eliminating the influence of motion blur on sand and gravel particle size calculation, while retaining the real sand and gravel particle regions with high reliability. Therefore, this application improves the accuracy of sand and gravel particle size distribution detection while performing rapid measurement.

[0040] In some embodiments, a high-speed image sequence containing falling sand and gravel particles is acquired as follows: A high-speed industrial camera captures real-time video frames of sand and gravel dispersion. The high-speed industrial camera acquires light signals through dual telecentric lenses. A line light source is synchronously triggered with the high-speed industrial camera, driven by a stroboscopic controller outputting pulses with a fixed pulse width of 500 nanoseconds to freeze the movement of the falling sand and gravel particles. The line frequency of the high-speed industrial camera is locked at 48 kHz and hard-synchronized with the conveyor belt encoder signal, with a constant spatial sampling interval of 0.061 mm per pixel. Every 2048 lines accumulate to form a complete image frame, with a physical field of view of 124.9 mm by 124.9 mm for each frame. With the above configuration, a high-speed image sequence containing falling sand and gravel particles is acquired.

[0041] Optionally, before binarizing each frame of the high-speed image in the high-speed image sequence, each frame of the high-speed image is preprocessed, including: performing flat field correction on each frame of the high-speed image to eliminate light source non-uniformity and lens vignetting; performing grayscale processing on the flat field corrected high-speed image to obtain an 8-bit grayscale image; and performing one-dimensional median filtering with an 8×1 pixel window on the 8-bit grayscale image to suppress inter-line random noise.

[0042] Optionally, binarizing each frame of the high-speed image in the high-speed image sequence to obtain a binarized image includes: dividing each frame of the high-speed image into multiple image sub-blocks according to a grid of preset size, wherein each image sub-block is a sub-region of the high-speed image; calculating a binarization threshold for each image sub-block using the maximum inter-class variance method; and performing binarization processing on each image sub-block using the binarization threshold to obtain the binarized image.

[0043] Each frame of the high-speed image is converted to a grayscale image. The grayscale image is then divided into multiple image sub-blocks using a 64-pixel multi-64-pixel grid, each sub-region of the grayscale image. For each sub-block, a binarization threshold is independently calculated using the Otsu's method (maximum inter-class variance). Otsu's method is an adaptive thresholding method that calculates a threshold based on the grayscale distribution of pixels within the sub-block, maximizing the inter-class variance between the foreground and background regions. For each sub-block, the calculated binarization threshold is used to binarize it, setting pixels with values ​​greater than or equal to the threshold as foreground pixels and pixels with values ​​less than the threshold as background pixels. In the binarized image, foreground pixels correspond to sand and gravel particle regions, and background pixels correspond to non-particle regions.

[0044] In some embodiments, each high-speed image frame, after preprocessing, is packaged together with the corresponding connected component structure array to form a single-frame data unit. This single-frame data unit is transmitted to the machine vision analysis module at a fixed rate of 6.8 gigabits per second (Gbps) via the Camera Link Full interface. The frame sampling interval is locked at 42.7 milliseconds, corresponding to a frame rate of 23.4 frames per second.

[0045] Optionally, the step of extracting each connected component in the binarized image includes: performing a single-scan connected component labeling algorithm on the binarized image to extract each connected component in each binarized image, and recording the geometric center coordinates, circumscribed rectangle boundary, pixel area, and second moment of each connected component.

[0046] The single-scan connected component labeling algorithm is an algorithm that can identify connected components in a single scan of an image. Starting from the top left corner of the binarized image, the algorithm scans each pixel row by row. When a foreground pixel is encountered, the algorithm checks whether the pixels above and to the left of that pixel have already been labeled as a connected component. If either the pixel above or to the left has been labeled, the current pixel is assigned to the same connected component; otherwise, a new connected component label is assigned to the current pixel. In this way, the algorithm can identify and label all connected components in the binarized image in a single scan.

[0047] Here, the geometric center coordinates are the average of the coordinates of all pixels within the connected component. The circumscribed rectangle boundary is the coordinates of the top-left and bottom-right corners of the smallest rectangle that can enclose the connected component. The pixel area is the total number of foreground pixels within the connected component. The second moment is the second-order statistical moment of the pixels within the connected component relative to the geometric center, used to describe the shape distribution characteristics of the connected component.

[0048] Optionally, for each connected component, calculating the degree of texture energy reduction of the connected component relative to its nearest neighbor connected component based on the principle of high-frequency energy attenuation in images, and obtaining the texture energy reduction value of the connected component, includes: performing a two-dimensional Fourier transform on the pixel region corresponding to the connected component to obtain a two-dimensional spatial energy distribution structure map; obtaining the sum of high-frequency component amplitudes in the two-dimensional spatial energy distribution structure map; obtaining the sum of high-frequency component amplitudes of the nearest neighbor connected component of the connected component; and obtaining the texture energy reduction value of the connected component based on the reciprocal of the sum of high-frequency component amplitudes of the connected component, the sum of high-frequency component amplitudes of the connected component, and the sum of high-frequency component amplitudes of the nearest neighbor connected component.

[0049] No. Within the frame, the first The calculation of the texture energy reduction value of each connected component is shown in formula (1): Formula (1); in, For the first The texture energy reduction value of each connected domain; Let the sum of the high-frequency component amplitudes of the nth connected component be denoted as . It is the sum of the high-frequency component amplitudes of the nearest neighboring connected component to the nth connected component. This is the product coefficient.

[0050] A larger product coefficient indicates a lower sum of high-frequency component amplitudes in the connected region, meaning a lower energy level in the connected region. This is more likely to correspond to a blurred region caused by motion blur. Conversely, a lower product coefficient indicates a larger proportion of high-frequency components in the connected region, which is consistent with the high-frequency characteristics of normal sand and gravel particle surface texture distribution.

[0051] Meanwhile, the trailing shadows created by sand and gravel particles are extracted by the algorithm as independent connected components, and the sand and gravel particles themselves also have corresponding connected components. Therefore, when a connected component of a sand and gravel particle contains a trailing shadow, the connected component closest to it is the trailing portion of the particle. In this case, there must be a significant difference between the high-frequency energy amplitude of the sand and gravel particle itself and the high-frequency energy amplitude of the trailing portion, because the pixel energy of the sand and gravel particle itself is greater than the pixel energy of the trailing portion. Therefore, the greater the difference between the sum of the high-frequency energy amplitudes of the two connected components, the greater the energy difference between them, and the more likely this situation is due to uneven energy distribution caused by the trailing shadow.

[0052] The larger the texture energy reduction value, the higher the texture energy reduction of the connected region, and the higher the probability of the connected region having ghosting, that is, the closer it is to the blur region.

[0053] Optionally, for each connected component, based on the principle of morphological skeleton extraction, calculating the dispersion of the distance from the edge sampling points within the connected component to the geometric center to obtain the skeleton distance reduction of the connected component includes: obtaining the geometric center of the connected component; setting multiple edge sampling points at preset pixel intervals on the edges of the connected component; calculating the distance value from each edge sampling point to the geometric center to obtain multiple distance values; obtaining the maximum and minimum distance values ​​of the multiple distance values, and calculating the average value of the multiple distance values; and obtaining the skeleton distance reduction based on the maximum distance value, the minimum distance value, each of the distance values, and the average value.

[0054] Relying solely on the energy characteristics of connected components to determine motion blur caused by blur has the problem of insufficient feature recognition capability within the scene. For example, for relatively fine sand particles, the surface texture distribution is not obvious enough, resulting in low energy of the corresponding connected components, but these particles themselves do not belong to the motion blur region.

[0055] Therefore, based on energy distribution characteristics, morphological features need to be introduced for supplementary judgment. For real sand and gravel particles, their binarized morphological skeleton is usually continuous and smooth. However, for the blurred trailing shadow region, its morphology presents as a slender and gradually narrowing protruding structure extending from the edge of the particle body. The distance transformation skeleton of this structure extends into the trailing shadow region, and the skeleton width monotonically decreases along the extension direction. The rate of change of distance values ​​at various points on this skeleton is the morphological feature factor of the blurred trailing shadow. The skeleton distance reduction amount is a numerical indicator used to quantify the morphological feature factor; the larger the value of the skeleton distance reduction amount, the more significant the slender and gradual transition feature of the trailing shadow region.

[0056] The calculation of the decrease in skeleton distance is shown in formula (2): Formula (2); in, For the first The skeleton distance decrease of each connected component. This represents the distance between the m-th edge sampling point within the n-th connected component and its geometric center. This is the average distance from all edge sampling points to the geometric center within the nth connected component. This represents the maximum distance from all edge sampling points within the nth connected component to the geometric center. M represents the minimum distance from all edge sampling points to the geometric center within the nth connected component, where M is the total number of edge sampling points. It reflects the degree of dispersion of the distance between edge sampling points.

[0057] for By analyzing the differences between the sampling distance and the mean distance of several equally spaced sampling points within a connected domain, the slender and gradually narrowing protruding morphological features of the edge of the connected domain are quantified.

[0058] As the difference gradually increases and the distance value changes step by step, the edge morphology of the connected region tends to be non-circular, and the edges are gradually elongated. This is because the deviation between the sampling distance and the mean becomes larger and larger. Furthermore, the larger the sum of all deviation values, the more obvious the irregularity of the connected region, the more irregular the edge skeleton shape of the sand particles, and the more significant the blurring and elongation effect.

[0059] This represents the difference between the maximum and minimum distance values. A larger difference indicates a greater discrepancy between the longest extension caused by the ghosting and the normal width of the sand and gravel particles. The greater this difference, the more pronounced the ghosting effect, and the higher the corresponding reduction in skeleton distance.

[0060] Optionally, the step of fusing the texture energy reduction value and the skeleton distance reduction amount according to a preset weight to obtain the fuzzy domain score of the connected component includes: weighting the texture energy reduction value using a first preset weight to obtain a first weighted value; weighting the skeleton distance reduction amount using a second preset weight to obtain a second weighted value; wherein the sum of the first preset weight and the second preset weight is 1; and adding the first weighted value and the second weighted value to obtain the fuzzy domain score of the connected component.

[0061] Based on the numerical analysis of texture energy reduction, skeleton distance reduction analysis can be performed to avoid misclassifying small connected components with low texture energy as ghosting regions. Furthermore, the fuzzy domain score of the connected component is determined. The fuzzy domain score characterizes the probability that the connected component belongs to a ghosting region. A higher fuzzy domain score indicates a more pronounced ghosting effect, a more severe interference with particle size distribution calculations caused by the ghosting effect, and therefore a lower reliability quantification value for the connected component. The severity of the ghosting effect is also related to inter-frame motion characteristics; that is, the more consistent the mean gradient direction of the connected component is with the direction of the geometric center displacement vector in the binary images of two adjacent frames, and the larger the magnitude of the geometric center displacement vector, the more likely the connected component is ghosting caused by high-speed motion, and the more pronounced the ghosting effect.

[0062] The fuzzy domain score is calculated as shown in formula (3): Formula (3); in, To score the fuzzy domain, As the first preset weight, This is the second preset weight.

[0063] The fuzzy domain score ranges from 0 to 1. When the fuzzy domain score exceeds 0.8, the connected component is considered a fuzzy domain.

[0064] Optionally, the first preset weight is equal to the second preset weight.

[0065] In other words, the values ​​of the first preset weight and the second preset weight are both 0.5.

[0066] Optionally, for each connected component, based on the principle of inter-frame motion tracking, determining the reliability quantification value of the connected component according to the motion features of the connected component in two adjacent binarized images and the fuzzy domain score includes: determining the first geometric center coordinates and the second geometric center coordinates of the connected component in two adjacent binarized images; calculating the geometric center displacement vector of the connected component in two adjacent binarized images based on the first geometric center coordinates and the second geometric center coordinates; obtaining the grayscale image of the corresponding region of the connected component in the high-speed image, performing gradient calculation on each pixel in the grayscale image belonging to the connected component, and obtaining the gradient direction of each pixel; performing vector averaging on all gradient directions to obtain the mean gradient direction of the connected component; calculating the angle difference between the mean gradient direction and the direction of the geometric center displacement vector; and determining the reliability quantification value of the connected component based on the fuzzy domain score, the angle difference, and the magnitude of the geometric center displacement vector.

[0067] Inter-frame motion feature verification of connected components is a process of verifying the spatiotemporal consistency of the morphological features of connected components based on the laws of physical motion. In real physical scenarios, rigid body motion follows strict spatiotemporal continuity, that is, the direction of the displacement trajectory of an object between adjacent frames should be highly consistent with its body extension direction, and the faster the motion speed, the larger the magnitude of the geometric center displacement vector generated on the image plane.

[0068] However, pseudo-connected regions generated by blurring often exhibit randomness in direction, or their internal gradient directions deviate significantly from the overall displacement direction. Therefore, this application constructs a geometric center displacement vector and compares it with the mean gradient direction calculated based on the grayscale image. The angle difference between the two is used to quantify the degree of anomaly that violates physical laws. Simultaneously, the magnitude of the geometric center displacement vector is introduced as a priori constraint for dynamic blurring, reflecting the negative impact of instantaneous motion speed on image matching accuracy. By using both the angle difference and the magnitude as penalty terms, interference regions that do not conform to true kinematic characteristics can be effectively removed, thereby achieving high-precision and robust removal of blur trails.

[0069] The credibility quantification value of the connected component is determined based on the fuzzy domain score, the angle difference, and the magnitude of the geometric center displacement vector, as shown in formula (4): Formula (4); in, The credibility quantification value, Score the fuzzy domain. The angle difference, Let be the magnitude of the displacement vector at the geometric center.

[0070] The fuzzy domain score reflects the degree of irregularity of the connected domain. Since fuzzy trails usually exhibit a slender structure, their fuzzy domain score is significantly higher than that of normal particles, thus occupying the denominator in the formula and playing a role in suppressing their score.

[0071] Angular difference is used to verify the consistency between the internal texture orientation of a connected component and its actual motion trajectory. In real physical motion, the motion directions of different parts of an object should remain highly consistent; however, the local gradient directions of false connected components or severely ghosted regions are often chaotic or deviate significantly from the displacement direction. Therefore, a larger angular difference implies lower realism.

[0072] The magnitude of the geometric center displacement vector incorporates a velocity dimension. In particle size measurement, excessively high motion speeds can lead to edge blurring and geometric centroid shift. This application, by using the magnitude of the geometric center displacement vector as a penalty term, can effectively filter out connected regions that suffer from morphological distortion due to high-speed motion and are unusable for accurate dimensional measurement.

[0073] In this application, the connected components are distinguished into trailing regions and effective sand and gravel particle regions by comparing the quantified value of the confidence metric with the preset confidence threshold. Based on this, the trailing regions need to be processed to eliminate their influence on particle size distribution calculation. Trailing removal involves removing connected components identified as trailing regions from the binarized image and filling the removed regions with the background color, thereby avoiding errors caused by trailing regions in particle size distribution calculation. Through trailing removal processing, the interference of trailing caused by the high-speed falling of sand and gravel particles on particle size distribution scanning is eliminated, improving the accuracy of particle size distribution scanning.

[0074] The following describes a method for scanning and measuring the particle size distribution of sand and gravel particles according to this application, with reference to a specific embodiment. This specific embodiment is only for better illustrating the invention and does not constitute an undue limitation of the invention.

[0075] Figure 2 This application provides a scanning measurement system for the particle size distribution of sand and gravel particles, such as... Figure 2As shown, this application proposes a scanning measurement system for particle size distribution of sand and gravel. The system 200 includes: an image acquisition and processing module 210, a connected component extraction module 220, and a machine vision analysis module 230. The image acquisition and processing module 210 is used to acquire a high-speed image sequence containing falling sand and gravel particles; and to perform binarization processing on each frame of the high-speed image sequence to obtain a binarized image. The connected component extraction module 220 is used to extract each connected component in the binarized image; the connected component is a set of pixels in the binarized image that have the same pixel value and are adjacent in position; The machine vision analysis module 230 is used to calculate the degree of texture energy reduction of each connected component relative to its nearest neighbor connected component based on the principle of high-frequency energy attenuation in images, and to obtain the texture energy reduction value of the connected component. The machine vision analysis module 230 is used to calculate the dispersion of the distance from the edge sampling point to the geometric center within each connected component based on the principle of morphological skeleton extraction, and to obtain the skeleton distance reduction of the connected component. The machine vision analysis module 230 is used to fuse the texture energy reduction value and the skeleton distance reduction amount according to a preset weight to obtain the fuzzy domain score of the connected component. The machine vision analysis module 230 is used to determine the confidence quantization value of each connected component based on the principle of inter-frame motion tracking, according to the motion features of the connected component in two adjacent binarized images and the fuzzy domain score. The machine vision analysis module 230 is further configured to compare the credibility quantification value with a preset credibility threshold; when the credibility quantification value is less than the preset credibility threshold, the connected component is determined to be a ghosting region and removed from the binarized image; otherwise, the connected component is retained as a valid sand and gravel particle region.

[0076] It should be noted that the description of the above system embodiments is similar to the description of the above method embodiments, and has similar beneficial effects. For technical details not disclosed in the system embodiments of this application, please refer to the description of the method embodiments of this application for understanding.

[0077] It should be noted that, in the embodiments of this application, if the above-described method for scanning and measuring the particle size distribution of sand and gravel is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiments of this application, or the part that contributes to the related technology, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause an electronic device to execute all or part of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), magnetic disks, or optical disks. Thus, the embodiments of this application are not limited to any specific hardware and software combination.

[0078] Correspondingly, embodiments of this application provide a computer-readable storage medium storing a computer program thereon. When executed by a processor, this computer program implements the steps in any of the above-described methods for scanning and measuring the particle size distribution of sand and gravel. Correspondingly, embodiments of this application also provide a computer program product, which, when executed by a processor of an electronic device, is used to implement the steps in any of the above-described methods for scanning and measuring the particle size distribution of sand and gravel.

[0079] It should be understood that the phrase "one embodiment" or "an embodiment" throughout the specification means that a specific feature, structure, or characteristic related to the embodiment is included in at least one embodiment of this application. Therefore, "in one embodiment" or "in an embodiment" appearing throughout the specification does not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. It should be understood that in the various embodiments of this application, the sequence numbers of the above-described processes do not imply a sequential order of execution; the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application. The sequence numbers of the above-described embodiments are merely descriptive and do not represent the superiority or inferiority of the embodiments.

[0080] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0081] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods, such as: multiple units or components can be combined, or integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the various components shown or discussed can be through some interfaces, and the indirect coupling or communication connection between devices or units can be electrical, mechanical, or other forms.

[0082] The units described above as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units; some or all of the units may be selected to achieve the purpose of the embodiments of this application according to actual needs. In addition, each functional unit in the embodiments of this application may be fully integrated into one processing unit, or each unit may be a separate unit, or two or more units may be integrated into one unit; the integrated unit may be implemented in hardware or in the form of hardware plus software functional units.

[0083] Alternatively, if the integrated units described above are implemented as software functional modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of this application, or the parts that contribute to related technologies, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause the device automatic test line to execute all or part of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, ROMs, magnetic disks, or optical disks.

[0084] The methods disclosed in the several method embodiments provided in this application can be arbitrarily combined to obtain new method embodiments without conflict. The features disclosed in the several method or device embodiments provided in this application can be arbitrarily combined to obtain new method embodiments or device embodiments without conflict.

[0085] The above description is merely an embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for scanning and measuring the particle size distribution of sand and gravel particles, characterized in that, The method includes: A high-speed image sequence containing falling sand and gravel particles is acquired; each frame of the high-speed image sequence is binarized to obtain a binarized image. Extract each connected component from the binarized image; the connected component is a set of pixels in the binarized image that have the same pixel value and are adjacent in position; For each connected component, based on the principle of high-frequency energy attenuation in images, the degree of texture energy reduction of the connected component relative to its nearest neighbor connected component is calculated to obtain the texture energy reduction value of the connected component. For each connected component, based on the principle of morphological skeleton extraction, the dispersion of the distance from the edge sampling point to the geometric center within the connected component is calculated to obtain the skeleton distance reduction of the connected component. The texture energy reduction value and the skeleton distance reduction are fused according to a preset weight to obtain the fuzzy domain score of the connected component. For each connected component, based on the principle of inter-frame motion tracking, the confidence quantization value of the connected component is determined according to the motion features of the connected component in the binarized images of two adjacent frames and the fuzzy domain score. The credibility quantification value is compared with a preset credibility threshold; when the credibility quantification value is less than the preset credibility threshold, the connected component is determined to be a ghosting region and removed from the binarized image; otherwise, the connected component is retained as a valid sand and gravel particle region.

2. The method according to claim 1, characterized in that, Extracting each connected component from the binarized image includes: A single-scan connected component labeling algorithm is performed on the binarized image to extract each connected component in each binarized image, and the geometric center coordinates, circumscribed rectangle boundary, pixel area, and second moment of each connected component are recorded.

3. The method according to claim 1, characterized in that, For each connected component, based on the principle of high-frequency energy attenuation in images, the degree of texture energy reduction of the connected component relative to its nearest neighbor connected component is calculated to obtain the texture energy reduction value of the connected component, including: A two-dimensional Fourier transform is performed on the pixel region corresponding to the connected component to obtain a two-dimensional spatial energy distribution structure map; Obtain the amplitude of the high-frequency components in the two-dimensional spatial energy distribution structure diagram; Obtain the sum of the high-frequency component amplitudes of the nearest neighbor connected components of the connected component; The texture energy reduction value of the connected region is obtained based on the reciprocal of the sum of the high-frequency component amplitudes of the connected region, the sum of the high-frequency component amplitudes of the connected region, and the sum of the high-frequency component amplitudes of the nearest neighbor connected region.

4. The method according to claim 1, characterized in that, For each connected component, based on the principle of morphological skeleton extraction, the dispersion of the distance from the edge sampling points within the connected component to the geometric center is calculated to obtain the skeleton distance reduction of the connected component, including: Obtain the geometric center of the connected domain; Multiple edge sampling points are set at preset pixel intervals on the edge of the connected domain; Calculate the distance from each edge sampling point to the geometric center to obtain multiple distance values; Obtain the maximum and minimum distance values ​​of the plurality of distance values, and calculate the average value of the plurality of distance values; The skeleton distance reduction amount is obtained based on the maximum distance value, the minimum distance value, each of the distance values, and the average value.

5. The method according to claim 1, characterized in that, The step of fusing the texture energy reduction value and the skeleton distance reduction amount according to a preset weight to obtain the fuzzy domain score of the connected component includes: The texture energy reduction value is weighted using a first preset weight to obtain a first weighted value; The decrease in skeleton distance is weighted using a second preset weight to obtain a second weighted value; wherein the sum of the first preset weight and the second preset weight is 1; The first weighted value is added to the second weighted value to obtain the fuzzy domain score of the connected component.

6. The method according to claim 5, characterized in that, The first preset weight is equal to the second preset weight.

7. The method according to claim 1, characterized in that, For each connected component, based on the principle of inter-frame motion tracking, the credibility quantification value of the connected component is determined according to the motion features of the connected component in two adjacent binarized images and the fuzzy domain score, including: Determine the first and second geometric center coordinates of the connected component in two adjacent frames of the binarized image; Based on the first geometric center coordinates and the second geometric center coordinates, calculate the geometric center displacement vector of the connected component in two adjacent frames of the binarized image; Obtain the grayscale image of the corresponding region of the connected component in the high-speed image; perform gradient calculation on each pixel in the grayscale image belonging to the connected component to obtain the gradient direction of each pixel; and perform vector averaging on all the gradient directions to obtain the average gradient direction of the connected component. Calculate the angle difference between the mean gradient direction and the direction of the geometric center displacement vector; Based on the fuzzy domain score, the angle difference, and the magnitude of the geometric center displacement vector, the quantified value of the connected component is determined.

8. The method according to claim 7, characterized in that, The confidence quantification value of the connected component is determined based on the fuzzy domain score, the angle difference, and the magnitude of the geometric center displacement vector, as shown in the following formula: ； in, The credibility quantification value, Score the fuzzy domain. The angle difference, Let be the magnitude of the displacement vector at the geometric center.

9. The method according to claim 1, characterized in that, The step of binarizing each frame of the high-speed image in the high-speed image sequence to obtain a binarized image includes: Each frame of the high-speed image is divided into multiple image sub-blocks according to a grid of preset size, and the image sub-blocks are sub-regions of the high-speed image; For each of the image sub-blocks, a binarization threshold is calculated using the maximum inter-class variance method; For each of the image sub-blocks, binarization processing is performed using the binarization threshold to obtain the binarized image.

10. A scanning measurement system for particle size distribution of sand and gravel, characterized in that, The system includes: an image acquisition and processing module, a connected component extraction module, and a machine vision analysis module; The image acquisition and processing module is used to acquire a high-speed image sequence containing falling sand and gravel particles; and to perform binarization processing on each frame of the high-speed image sequence to obtain a binarized image. The connected component extraction module is used to extract each connected component in the binarized image; the connected component is a set of pixels in the binarized image that have the same pixel value and are adjacent in position; The machine vision analysis module is used to calculate the degree of texture energy reduction of each connected component relative to its nearest neighbor connected component based on the principle of high-frequency energy attenuation in images, and to obtain the texture energy reduction value of the connected component. The machine vision analysis module is used to calculate the dispersion of the distance from the edge sampling point to the geometric center within each connected component based on the principle of morphological skeleton extraction, and to obtain the skeleton distance reduction of the connected component. The machine vision analysis module is used to fuse the texture energy reduction value and the skeleton distance reduction amount according to a preset weight to obtain the fuzzy domain score of the connected component. The machine vision analysis module is used to determine the confidence quantization value of each connected component based on the principle of inter-frame motion tracking, according to the motion features of the connected component in two adjacent binarized images and the fuzzy domain score. The machine vision analysis module is further configured to compare the credibility quantification value with a preset credibility threshold; when the credibility quantification value is less than the preset credibility threshold, the connected component is determined to be a ghosting region and removed from the binarized image; otherwise, the connected component is retained as a valid sand and gravel particle region.

Images