A municipal construction dust reduction system and equipment based on image processing

Abstract

The present invention relates to the field of image processing technology, and in particular to a municipal construction dust reduction system and equipment based on image processing. A time-series grayscale image of the construction site is acquired, the grayscale changes and movement of the pixels are analyzed, and the particulate matter area is accurately divided in the grayscale image. Taking into account the differences in the movement performance of dust and water mist particles in the image, the grayscale distribution, quantity distribution and movement pattern of the pixels are comprehensively considered, and the dust reduction characteristic value of the particulate matter area is analyzed to reflect the dust reduction effect in the current image. On this basis, the system adaptively adjusts the initial water mist concentration based on the dust reduction characteristic value to obtain the final water mist concentration, and performs dust reduction on the construction site based on the final water mist concentration. This comprehensive evaluation system can comprehensively reflect the dust conditions of the construction site by analyzing the images of the construction site, provide a scientific basis for formulating targeted dust reduction strategies, and significantly enhance the dust reduction effect.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and in particular to a municipal construction dust reduction system and equipment based on image processing. Background Art

[0002] During municipal construction, large amounts of dust are often generated on-site. This not only affects the air quality of the construction environment, but can also adversely impact the living environment and health of surrounding residents. Therefore, dust reduction measures are crucial during municipal construction. High-pressure spray systems are typically used to produce a fine mist. When the mist comes into contact with dust particles in the air, the fine mist particles adhere to the dust, making the dust particles heavier and reducing their ability to remain suspended in the air, thereby reducing dust concentration in the air.

[0003] In the dust reduction process, water mist concentration is an important indicator, which is closely related to the dust reduction efficiency. In order to improve the dust reduction effect and efficiency, the existing technology usually uses grayscale information to identify the dust in the current construction scene, so as to determine the final required water mist concentration; however, since the construction site usually has a complex background, these backgrounds may be similar to the grayscale and texture of the dust, which increases the difficulty of dust identification, and the water mist generated by the fog cannon may cause the dust to be dispersed. These problems will increase the difficulty of identifying dust using only grayscale information, thereby affecting the subsequent dust reduction efficiency and effect. Summary of the Invention

[0004] In order to solve the technical problems that construction sites usually have complex backgrounds, which may be similar to the grayscale and texture of dust, making dust identification more difficult, and the water mist generated by the fog cannon may cause dust to scatter, all of which increase the difficulty of identifying dust using only grayscale information, thereby affecting the subsequent dust reduction efficiency and effect, the purpose of the present invention is to provide a municipal construction dust reduction system and equipment based on image processing. The technical solutions adopted are as follows:

[0005] A municipal construction dust reduction system based on image processing, comprising:

[0006] A data acquisition module is used to acquire two grayscale images of the construction site that are adjacent in time sequence, and use the previous grayscale image in time sequence as the target image;

[0007] The particle area segmentation module is used to determine the fog coefficient of each pixel in the target image based on the grayscale fluctuation characteristics of the pixels in the local area. In two adjacent grayscale images, the module analyzes the movement and grayscale changes of the pixels and, based on the fog coefficient of each pixel, divides the target image into regions to obtain the particle area.

[0008] A dustfall analysis module is used to obtain a dustfall characteristic value of the particle area based on the number distribution of pixels and the atomization coefficient of the pixels and the movement pattern of the pixels in the particle area;

[0009] The water mist concentration adjustment and dust reduction module is used to adjust the initial water mist concentration according to the dust reduction characteristic value of the particulate matter area to obtain the final water mist concentration; and to reduce dust at the construction site based on the final water mist concentration.

[0010] Furthermore, the method for obtaining the atomization coefficient includes:

[0011] In the target image, based on the grayscale values ​​of the pixels, a maximum value point is selected;

[0012] Construct a preset second neighborhood with each pixel as the center;

[0013] In the preset second neighborhood corresponding to each pixel point, a maximum point is randomly selected as the test point, and the maximum point with the smallest Euclidean distance to the test point among the remaining maximum points is selected as the comparison point. The grayscale value mean between the test point and the comparison point is calculated as the first grayscale mean, and the grayscale value mean of the pixel points whose grayscale values ​​exist between the test point and the comparison point is calculated as the second grayscale mean. The Euclidean distance between the test point and the comparison point is used as the distance factor;

[0014] The absolute value of the difference between the first grayscale mean and the second grayscale mean between the test point and the comparison point is used as the grayscale fluctuation factor, and the value after normalizing the ratio of the grayscale fluctuation factor to the distance factor is used as the fog factor corresponding to the test point;

[0015] In the preset second neighborhood corresponding to each pixel point, the sum of the fog factors corresponding to all maximum value points is normalized and used as the fog coefficient of each pixel point.

[0016] Furthermore, the method for obtaining the maximum point includes:

[0017] In the target image, a preset first neighborhood of each pixel is determined, and within the preset first neighborhood corresponding to each pixel, the pixel with the largest grayscale value is taken as a maximum value point;

[0018] The preset first neighborhood is smaller than the preset second neighborhood.

[0019] Furthermore, the method for obtaining the particulate matter area includes:

[0020] Use two grayscale images as input to the optical flow method, and output the optical flow vector corresponding to the pixel point in the target image;

[0021] In the target image, the granularity factor of each pixel is determined based on the fog coefficient and optical flow vector corresponding to each pixel, as well as the grayscale disorder of the pixel in the corresponding preset second neighborhood.

[0022] Segmenting the target image based on the Otsu threshold segmentation method to obtain two regions, wherein the regions are divided into a foreground region and a background region;

[0023] In each area, the mean particle factor of all pixels is calculated as the particle index corresponding to each area;

[0024] The area with the largest particle matter index is regarded as the particle matter area.

[0025] Furthermore, the method for obtaining the particle factor includes:

[0026] In the preset second neighborhood corresponding to each pixel, the mean value of the grayscale gradients of all pixels is used as the grayscale change factor of each pixel;

[0027] Calculate the modulus of the optical flow vector corresponding to each pixel as the motion intensity corresponding to each pixel;

[0028] For any pixel point, the product of the fog coefficient of the pixel point and the grayscale change factor is used as the grayscale change coefficient. The ratio of the grayscale change coefficient of the pixel point and the motion intensity is normalized with the absolute value of the difference between the preset first constant to obtain the particle factor of the pixel point.

[0029] Furthermore, the method for obtaining the dustfall characteristic value includes:

[0030] Dividing the particle area based on a superpixel segmentation method to obtain a plurality of superpixel blocks;

[0031] Based on the number distribution, motion, and fog coefficient of all superpixel blocks in the particle area, the water fog factor of each superpixel block and the water fog combination index of the particle area are determined;

[0032] In the particle area, analyzing the movement and diffusion of pixels, and combining the mist factors of all superpixel blocks in the particle area to determine the dust diffusion coefficient of the particle area;

[0033] The normalized value of the ratio of the water mist combination index to the dust diffusion coefficient is used as the dustfall characteristic value of the particulate matter area.

[0034] Furthermore, the method of determining the mist factor of each superpixel block and the mist combination index of the particle area based on the number distribution, motion and fog coefficient of pixels in all superpixel blocks in the particle area includes:

[0035] In each superpixel block, the average of the fog coefficients of all pixels is taken as the fog mean;

[0036] The number of iterations is preset. In each iteration, a number of pixels are randomly selected in each superpixel block as water mist pixels.

[0037] The ratio of the number of water fog pixels in each superpixel block to the number of all pixels is calculated as the water fog density, and the variance of the modulus of the optical flow vectors of all water fog pixels in each superpixel block is used as the confusion factor of each superpixel block.

[0038] The absolute value of the difference between the ratio of the fog mean and the water fog density corresponding to each superpixel block and the preset second constant is used as the true factor. The ratio of the true factor to the confusion factor is negatively correlated and normalized to the value obtained as the water fog loss factor corresponding to each superpixel block in each iteration;

[0039] The sum of the water mist loss factors corresponding to all superpixel blocks is used as the water mist loss value corresponding to the particle area in each iteration;

[0040] In all iterations, under the minimum water mist loss value, the water mist density corresponding to each superpixel block is used as the water mist factor of each superpixel block; the normalized value of the difference between the maximum water mist density and the minimum water mist density in all superpixel blocks is used as the water mist combination index of the particle matter area.

[0041] Furthermore, the method for obtaining the dust diffusion coefficient includes:

[0042] In each superpixel block, the mean of the optical flow vectors of all pixels is calculated as the comprehensive vector;

[0043] Decomposing the integrated vector, removing the vertical downward component and synthesizing the vector to obtain a diffusion vector;

[0044] The ratio of the modulus of the diffusion vector corresponding to each superpixel block to the water fog factor is used as the diffusion factor corresponding to each superpixel block;

[0045] The normalized value of the mean diffusion factor of all superpixel blocks is used as the dust diffusion coefficient of the particulate matter area.

[0046] Furthermore, the method for obtaining the final water mist concentration includes:

[0047] Using the difference between the preset third constant and the dustfall characteristic value as an adjustment coefficient;

[0048] The product of the adjustment coefficient and the initial water mist concentration is taken as the final water mist concentration.

[0049] A municipal construction dust reduction device based on image processing includes a processor and a memory, wherein the memory stores at least one instruction, at least one program, code set or instruction set. When the at least one instruction, at least one program, code set or instruction set is loaded and executed by the processor, the steps of the municipal construction dust reduction system based on image processing are implemented.

[0050] The present invention has the following beneficial effects:

[0051] First, a time-series grayscale image of the construction site is acquired to provide immediate and accurate environmental data. Since dust typically appears as smoke, it can appear as a mixture of light and dark in the image. Therefore, in the particle region segmentation module, the grayscale variation of each pixel is analyzed in detail to determine the atomization coefficient. This coefficient reflects the grayscale variation around the pixel, facilitating subsequent determination of whether it is a dust particle pixel. Dust particles of varying sizes exhibit distinct motion patterns at a construction site. Therefore, the atomization coefficient is combined with the pixel's motion and grayscale distribution to more accurately segment the particle region within the image. The particle region acquired in the aforementioned operation contains water mist particles generated by a fog cannon. Therefore, to more accurately assess the dust reduction situation at the current construction site, it is necessary to analyze the particle diffusion and the combination of the water mist. Given that the water mist particles produced by industrial fog cannons are small, evenly distributed, and have a consistent motion pattern, the number distribution of pixels, their motion pattern, and the atomization coefficient within the particle region are analyzed to obtain a dust reduction characteristic value for the particle region, which reflects the dust reduction effect within the current image. Finally, in the water mist concentration adjustment and dust suppression module, the initial water mist concentration is adaptively adjusted based on the dust suppression characteristic value to obtain the final water mist concentration. This final water mist concentration is then used to suppress dust at the construction site. This comprehensive assessment system can more comprehensively reflect the dust conditions at the construction site, providing a scientific basis for developing targeted dust suppression strategies and significantly enhancing dust suppression effectiveness. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0053] Figure 1 A system block diagram of a municipal construction dust reduction system based on image processing provided by one embodiment of the present invention;

[0054] Figure 2A flow chart of a method for obtaining a particulate matter region provided by one embodiment of the present invention;

[0055] Figure 3 A flow chart of a method for obtaining dustfall characteristic values ​​provided by one embodiment of the present invention;

[0056] Figure 4 A schematic structural diagram of a municipal construction dust reduction device based on image processing provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0057] To further illustrate the technical means and effectiveness of the present invention in achieving its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, describes in detail the specific implementation, structure, features, and effectiveness of an image processing-based municipal construction dust reduction system and equipment proposed by the present invention. In the following description, references to different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.

[0058] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0059] The following describes in detail a specific solution of a municipal construction dust reduction system and equipment based on image processing provided by the present invention with reference to the accompanying drawings.

[0060] See also Figure 1 , which shows a system block diagram of a municipal construction dust reduction system based on image processing provided by an embodiment of the present invention. The system includes a data acquisition module 101, a particulate matter area division module 102, a dust reduction situation analysis module 103, and a water mist concentration adjustment and dust reduction module 104.

[0061] The data acquisition module 101 is used to acquire two grayscale images of the construction site that are adjacent in time sequence, and use the previous grayscale image in time sequence as the target image.

[0062] In order to reduce dust at municipal construction sites, industrial fog cannons are used to spray water mist. The industrial fog cannon consists of a high-pressure pump, a nozzle, and a control unit. The high-pressure pump draws water from the water tank and pressurizes it. The high-pressure water is released through the nozzle to form a fine water mist. The type and design of the nozzle determine the particle size and spraying range of the water mist. The control unit can adjust the water mist concentration, etc., where the control unit can be a microprocessor or a microcontroller.

[0063] To effectively reduce dust at construction sites, it's necessary to analyze the dust density in the air. Therefore, in this embodiment of the present invention, a high-resolution industrial camera is installed at the bottom of the industrial fog cannon to capture video images of the construction site. Professional video processing software is then used to extract video images from the recorded video at set time intervals. Because dust analysis doesn't require a color space, the captured video images of the construction site are grayscaled to produce grayscale images. Then, within the video image sequence, two temporally adjacent grayscale images are selected—for example, the nth frame and the n+1th frame. Of these two adjacent grayscale images, the temporally preceding grayscale image (the nth frame) is used as the target image.

[0064] It should be noted that the time interval for extracting video images can be set to 1 second, and the specific time interval can be adjusted according to the implementation scenario, which is not limited here; it should be noted that the method for obtaining grayscale images can adopt average grayscale, maximum value method or weighted grayscale method, etc., and they are all technical means well known to technical personnel in this field, and are not limited or elaborated here.

[0065] The particle area segmentation module 102 is used to determine the haze coefficient of each pixel in the target image based on the grayscale fluctuation characteristics of the pixels in the local range; in two adjacent grayscale images, the pixel movement and grayscale changes are analyzed, and the target image is segmented into regions based on the haze coefficient of each pixel to obtain the particle area.

[0066] Since the dust in the construction site will float and move in the air, the dust part in the image will appear smoke-like in the image and may show grayscale fluctuation characteristics of alternating light and dark. Therefore, in the target image, based on the grayscale fluctuation of pixels in the local range, the fog coefficient of each pixel can be obtained to reflect the grayscale changes around the pixel. Furthermore, due to the differences in the movement patterns of dust particles of different sizes at the construction site, for example, small dust particles are less likely to settle when they are lifted up due to their smaller mass, so the movement changes are weaker, and due to their smaller volume, the grayscale changes around them are more obvious; on the contrary, large dust particles are more likely to settle after being lifted up due to their larger mass, so the movement changes are more obvious, but due to their larger volume, the grayscale changes around them are weaker; therefore, based on this feature, the present invention analyzes the movement and grayscale changes of pixels between two adjacent grayscale images on the basis of the aforementioned calculated atomization coefficient, so as to more accurately quantify the possibility that each pixel in the target image is particulate dust, and then divides the target image into regions to obtain more accurate particulate matter regions, providing support for subsequent analysis processes.

[0067] First, in the target image, the fog coefficient of each pixel is determined based on the grayscale fluctuation characteristics of the pixels in the local range.

[0068] Preferably, in one embodiment of the present invention, the method for obtaining the atomization coefficient includes:

[0069] The maximum points are usually the peaks of local grayscale changes. These points may correspond to the edges, corners and other features of the particles in the target image. Therefore, in the target image, the maximum points are screened based on the grayscale values ​​of the pixels: the preset first neighborhood of each pixel is determined, and within the preset first neighborhood corresponding to each pixel, the pixel with the largest grayscale value is taken as the maximum point.

[0070] At this point, all the maximum points in the target image, that is, the pixels where the grayscale changes suddenly, can be screened out.

[0071] Then, in order to more comprehensively analyze the grayscale changes around each pixel, a local range larger than the preset first neighborhood is constructed, that is, a preset second neighborhood is constructed with each pixel as the center.

[0072] In the preset second neighborhood corresponding to each pixel point, a maximum point is selected as the test point, and the maximum point with the smallest Euclidean distance to the test point among the remaining maximum points is selected as the comparison point.

[0073] The grayscale value mean between the test point and the comparison point is calculated as the first grayscale mean, which reflects the average grayscale level between the two points; the grayscale value mean of the pixel points whose grayscale values ​​exist between the test point and the comparison point is calculated as the second grayscale mean, which reflects the average grayscale level of the pixel points whose grayscale values ​​exist between the two points; the Euclidean distance between the test point and the comparison point is used as the distance factor.

[0074] The absolute value of the difference between the first grayscale mean and the second grayscale mean between the test point and the comparison point is used as the grayscale fluctuation factor. The larger the grayscale fluctuation factor, the more obvious the grayscale value fluctuation of the pixel point, and the smaller the distance factor, the faster the grayscale value change of the pixel point, which also means that the grayscale change is more obvious. Therefore, the ratio of the grayscale fluctuation factor to the distance factor is normalized and used as the atomization factor corresponding to the test point. At this time, the larger the atomization factor, the more significant the grayscale change around the pixel point. Normalization is a technical means well known to those skilled in the art. The normalization function can be linear normalization or standard normalization, etc. The specific normalization method is not limited here.

[0075] At this point, the fog factors of all maximum points in the preset second neighborhood corresponding to each pixel point can be obtained.

[0076] Finally, within the preset second neighborhood corresponding to each pixel, the sum of the fog factors corresponding to all maximum points is normalized and used as the fog coefficient for each pixel. The larger the fog coefficient, the more obvious the grayscale variation characteristics within the local area where the pixel is located. Normalization is a technical means well known to those skilled in the art. The normalization function can be linear normalization or standard normalization, etc. The specific normalization method is not limited here.

[0077] It should be noted that, in this embodiment of the present invention, when calculating the Euclidean distance, a two-dimensional coordinate system can be constructed in the target image to obtain the position coordinates of each pixel point for calculating the Euclidean distance; the first neighborhood is preset to 3×3, and the second neighborhood is preset to 5×5.

[0078] After obtaining the haze coefficient, since particles of different sizes have different motion patterns, in order to obtain the particle area in the target image, the movement and grayscale changes of the pixels in two adjacent grayscale images are further analyzed. Combined with the haze coefficient of each pixel, the target image is divided into regions to obtain the accurate particle area.

[0079] Preferably, in one embodiment of the present invention, the method for obtaining the particulate matter area includes:

[0080] See also Figure 2 , which shows a flow chart of a method for obtaining a particulate matter region in one embodiment of the present invention, the method comprising the following steps:

[0081] Step S201: using two grayscale images as inputs of the optical flow method, thereby outputting optical flow vectors corresponding to pixels in the target image.

[0082] The optical flow method is a technology used to detect and analyze the motion of objects in video or image sequences. It can estimate the motion of objects based on the changes in the grayscale values ​​of pixels in the image, and the obtained optical flow vector can reflect the movement direction and distance of the object in the image between adjacent frames. This is important for the subsequent analysis of the motion pattern of particulate matter and thus the determination of the particulate matter area. Therefore, in this embodiment of the present invention, two grayscale images are used as input to the optical flow method, and the optical flow vector corresponding to the pixel point in the target image is output.

[0083] It should be noted that the optical flow method is a well-known technology and the specific process will not be described here in detail.

[0084] Step S202: In the target image, the granularity factor of each pixel is determined according to the haze coefficient and optical flow vector corresponding to each pixel, and the grayscale disorder of the pixel in the corresponding preset second neighborhood.

[0085] Since large dust particles have a larger volume, the grayscale disorder of pixels in a local range may be smaller; conversely, small dust particles have a smaller volume, so the grayscale disorder of pixels in a local range may be larger.

[0086] Therefore, within the preset second neighborhood corresponding to each pixel, the mean of the grayscale gradients of all pixels is used as the grayscale variation factor for each pixel. A larger grayscale variation factor indicates a more dramatic grayscale variation within the preset second neighborhood for that pixel, making it more likely to be a small dust particle. A smaller grayscale variation factor indicates a smaller grayscale variation within the preset second neighborhood for that pixel, making it more likely to be a large dust particle. Large dust particles, due to their larger mass, are more likely to settle quickly at a construction site, resulting in more noticeable movement. Conversely, small dust particles, due to their smaller mass, are more likely to remain suspended in the air at a construction site, making their movement less noticeable.

[0087] Therefore, the modulus of the optical flow vector corresponding to each pixel is calculated as the motion intensity corresponding to each pixel. At this time, the greater the motion intensity, the more obvious the motion feature, and the more likely it is to represent large dust particles. The smaller the motion intensity, the less obvious the motion feature, and the more likely it is to represent small dust particles.

[0088] Finally, for any pixel, the product of the pixel's haze coefficient and the grayscale variation factor is used as the grayscale variation coefficient. A larger grayscale variation coefficient indicates a more pronounced grayscale variation characteristic within the pixel's local area, and a higher likelihood of representing a small dust particle. A smaller grayscale variation coefficient indicates a weaker grayscale variation characteristic within the pixel's local area, and a higher likelihood of representing a large dust particle. Based on the previous analysis, a greater motion intensity indicates a higher likelihood of a large dust particle, while a smaller motion intensity indicates a smaller dust particle. Therefore, the ratio of the pixel's grayscale variation coefficient to the motion intensity is calculated. The greater the difference between the denominator and the numerator in this fraction, the more likely the pixel is a dust pixel. The absolute value of the difference between this ratio and a preset first constant is then normalized to obtain the pixel's grain factor. The larger the absolute value of this difference, the more likely the pixel represents particulate matter in the target image. Therefore, a larger grain factor indicates a more likely particulate matter pixel. Normalization is a technical means well known to those skilled in the art. The normalization function may be linear normalization or standard normalization, etc. The specific normalization method is not limited here.

[0089] It should be noted that the preset first constant is set to 1 for the purpose of facilitating comparison and calculation; the grayscale gradient can be obtained based on the Canny operator, and the specific process is not described here.

[0090] Step S203: performing image segmentation on the target image and determining the particulate matter area based on the particle factors of the pixels in the target image.

[0091] Otsu threshold segmentation is a method that automatically determines the optimal segmentation threshold. It can divide an image into two parts based on the grayscale histogram of the image. First, the target image is segmented based on the Otsu threshold segmentation method to obtain two regions, which are divided into foreground area and background area.

[0092] Then, in each area, the mean particle factor of all pixels is calculated as the particle matter index corresponding to each area. At this time, the larger the particle matter index, the more concentrated the pixels representing particle matter in the area are, and the higher the possibility that the pixels in the area are particle matter pixels. Therefore, in the end, of the two areas, the area with the largest particle matter index is taken as the particle matter area.

[0093] It should be noted that the Otsu threshold segmentation method is a well-known technology, and the specific process will not be described here in detail.

[0094] The dustfall analysis module 103 is used to obtain the dustfall characteristic value of the particle area based on the number distribution of pixels and the atomization coefficient of the pixels and the movement pattern of the pixels.

[0095] The particulate matter region obtained through the above operation requires an analysis of dust fall conditions based on the particulate matter portion. However, since the industrial fog cannon is operating and performing dust removal during image acquisition, water mist particles may be present in the obtained particulate matter region. Therefore, directly analyzing the dust fall conditions based on the particulate matter portion will result in errors. In this embodiment of the present invention, to analyze and determine the dust fall conditions of the dust in the current image, it is necessary to analyze the diffusion of the particulate matter and integrate the water mist binding. Since the water mist emitted by the industrial fog cannon is intended to contact and adhere to dust particles in the air, increasing the dust mass and achieving the purpose of dust reduction, the water mist particles emitted by the industrial fog cannon are smaller and have good air suspension. Furthermore, since the water mist emitted by the industrial fog cannon is relatively uniform, the distribution of the water mist in the air is relatively uniform and its movement pattern is relatively consistent. However, due to the inconsistent particle size and different causes of dust emission, the distribution and movement pattern of the dust in the air are relatively disordered. Therefore, the dust fall characteristic value of the particulate matter region can be obtained based on the number distribution, movement, and atomization coefficient of the pixels in the particulate matter region.

[0096] Preferably, in one embodiment of the present invention, the method for obtaining the dustfall characteristic value includes:

[0097] See also Figure 3, which shows a method flow chart of a method for obtaining dustfall characteristic values ​​in one embodiment of the present invention, the method comprising the following steps:

[0098] Step S301: Divide the particle area based on a superpixel segmentation method to obtain multiple superpixel blocks.

[0099] The superpixel segmentation method can divide pixels with similar texture features or attributes into a superpixel block, reducing complexity while retaining key features, thereby simplifying the subsequent analysis process. Therefore, the particulate matter area is divided based on the superpixel segmentation method to obtain multiple superpixel blocks.

[0100] It should be noted that in the embodiment of the present invention, the number of superpixel blocks is set to 32. The specific number can be adjusted according to the implementation scenario and is not limited here. The superpixel segmentation method is a well-known technology and the specific process will not be described here.

[0101] Step S302: Based on the number distribution, motion and fog coefficient of pixels in all superpixel blocks in the particle area, determine the fog factor of each superpixel block and the fog combination index of the particle area.

[0102] In each superpixel block, the average of the fog coefficients of all pixels is taken as the fog mean, which reflects the average grayscale change of all pixels in the superpixel block.

[0103] In order to more accurately evaluate the water mist combination, in this embodiment of the present invention, the number of iterations is preset. In each iteration, a number of pixels are randomly selected as water mist pixels in each superpixel block. This can simulate the distribution of water mist under different conditions, thereby screening out water mist pixels that can better represent real water mist, thereby improving the accuracy of subsequent dust fall analysis.

[0104] The ratio of the number of all water fog pixels to the number of all pixels in each superpixel block is calculated as the water fog density.

[0105] Based on the above analysis, it can be seen that the water mist emitted by the industrial fog cannon is relatively uniform, which leads to a more uniform distribution of water mist in the air and a more consistent movement pattern. Therefore, in order to evaluate whether the water mist pixel points in each superpixel are appropriately selected, the variance of the modulus of the optical flow vector of the water mist pixel points in each superpixel block is used as the confusion factor of each superpixel block. At this time, the smaller the confusion factor, the more appropriate the selection of the water mist pixel points is, and the more it can represent the real water mist.

[0106] Under normal circumstances, a larger fog coefficient (that is, a greater grayscale variation) indicates better integration with the mist, and the true mist density will decrease. Conversely, a smaller fog coefficient indicates poorer integration with the mist, and the true mist density will increase. Therefore, the ratio of the fog mean value to the mist density for each superpixel block is calculated. The greater the difference between the numerator and denominator in this fraction, the more representative the selected mist pixel is of the true mist. Therefore, the absolute value of the difference between this ratio and a preset second constant is calculated as the true factor. The larger the true factor, the more representative the mist pixel is of the true mist.

[0107] Then the ratio of the true factor to the confusion factor is negatively correlated and normalized, and the value obtained is used as the water mist loss factor corresponding to each super pixel block in each iteration. Based on the above analysis, the smaller the confusion factor, the higher the accuracy of the water mist pixel selection, and the larger the true factor, the higher the accuracy of the water mist pixel selection. Therefore, the smaller the water mist loss factor obtained, the more the water mist pixel selected in the super pixel block in this iteration can represent the real water mist. The negative correlation mapping and normalization processing here can be used using the formula ,in, It represents the exponential function with the natural constant e as the base, and x represents the independent variable.

[0108] The sum of the water mist loss factors corresponding to all superpixel blocks is used as the water mist loss value corresponding to the particle matter area in each iteration. Similarly, the smaller the water mist loss value, the higher the accuracy of the water mist pixel points selected in the particle matter area in this iteration.

[0109] Finally, among all the iterations, the minimum water mist loss value is selected. At this time, the more correct the selection of water mist pixels is, the more accurate the actual water mist situation in the particle area can be measured on this basis.

[0110] Under the minimum water mist loss value, the water mist density corresponding to each superpixel block is used as the water mist factor of each superpixel block to reflect the water mist distribution in each superpixel block; because the better the water mist binding efficiency, the greater the binding difference between different concentrations will be, so the normalized value of the difference between the maximum water mist density and the minimum water mist density in all superpixel blocks is used as the water mist binding index of the particle matter area. This index reflects the uniformity or difference of the water mist distribution in the particle matter area. When the difference is large, it indicates that the water mist distribution in the particle matter area is uneven, which means that the water mist binding efficiency is stronger.

[0111] It should be noted that the number of iterations is set to 20 times, and the specific value can be adjusted according to the implementation scenario and is not limited here; the preset second constant here is set to 1 for the purpose of facilitating comparison and calculation.

[0112] Step S303: In the particle area, the motion diffusion of the pixels is analyzed, and the dust diffusion coefficient of the particle area is determined by combining the mist factors of all superpixel blocks in the particle area.

[0113] In each superpixel block, the mean of the optical flow vectors of all pixels is calculated as a comprehensive vector, which can represent the overall motion of the pixels in each superpixel block.

[0114] Since the principle of dust reduction is that water mist contacts and adheres to dust particles in the air, causing the dust mass to increase, thereby achieving the purpose of dust reduction, the comprehensive vector is decomposed, and the vertical downward component is removed before being synthesized to obtain the diffusion vector. At this time, the diffusion vector can characterize the diffusion of dust and reduce the impact of the dust reduction process on its calculation.

[0115] Since the dust diffusion situation of the superpixel block with a smaller water fog factor will be more serious, the ratio of the modulus of the diffusion vector corresponding to each superpixel block to the water fog factor is used as the diffusion factor corresponding to each superpixel block. At this time, the larger the diffusion factor, the more serious the diffusion of dust particles in the superpixel block.

[0116] Finally, the average of the diffusion factors across all superpixels is normalized to form the dust diffusion coefficient for the particle area. Similarly, a larger diffusion coefficient indicates more severe dust particle diffusion within the particle area. Normalization is a well-known technique for those skilled in the art, and the normalization function can be linear normalization or standard normalization. The specific normalization method is not limited here.

[0117] Step S304: The water mist combination index and the dust diffusion coefficient corresponding to the particle area are integrated to obtain a dustfall characteristic value corresponding to the particle area.

[0118] The dustfall characteristic value can be used to reflect the dustfall situation at the current construction site, which helps to adjust the water mist concentration in the subsequent process. The dustfall characteristic value can be quantified by the water mist binding index corresponding to the particulate matter area and the dust diffusion coefficient.

[0119] When the water mist combination index is larger, it means that the combination of water mist and dust in the particulate matter area is better, which can be regarded as the better dust reduction effect. When the dust diffusion coefficient is smaller, it means that the diffusion of dust particles in the particulate matter area is weaker, which can also be regarded as the better dust reduction effect. Therefore, the ratio of the water mist combination index to the dust diffusion coefficient is calculated, and the value after normalization of the ratio is used as the dust reduction characteristic value of the particulate matter area. At this time, the larger the dust reduction characteristic value, the better the dust reduction effect corresponding to the particulate matter area, that is, the better the dust reduction effect at the current construction site. Normalization is a technical means well known to those skilled in the art. The choice of normalization function can be linear normalization or standard normalization, etc. The specific normalization method is not limited here.

[0120] The water mist concentration adjustment and dust reduction module 104 is used to adjust the initial water mist concentration according to the dust reduction characteristic value of the particulate matter area to obtain a final water mist concentration; and to reduce dust at the construction site based on the final water mist concentration.

[0121] Module 103 calculates the dust reduction characteristic value of the current construction site, which can be used to reflect the dust reduction effect of the construction site. In this module, the initial water mist concentration can be adaptively adjusted according to the dust reduction characteristic value of the particulate matter area, so that the final water mist concentration is more in line with the actual dust reduction requirements in the current construction site, thereby improving the dust reduction effect.

[0122] Preferably, in one embodiment of the present invention, the method for obtaining the final water mist concentration includes:

[0123] When the dust reduction characteristic value is smaller, it means that the dust reduction effect at the current construction site is poor. Then it is necessary to increase the water mist concentration so that the water mist can better adhere to the dust particles to improve the dust reduction effect. Therefore, the difference between the preset third constant and the dust reduction characteristic value is used as the adjustment coefficient. At this time, the larger the adjustment coefficient, the greater the degree to which the water mist concentration needs to be increased. Finally, the product of the adjustment coefficient and the initial water mist concentration is used as the final water mist concentration.

[0124] After obtaining the final water mist concentration, the control unit can be used to adjust the water mist concentration of the industrial fog cannon to reduce dust at the construction site.

[0125] It should be noted that the preset third constant here is set to 2, the purpose of which is to ensure that the final water mist concentration is greater than the initial water mist concentration but not too large, thereby ensuring the dust reduction effect.

[0126] To summarize, we first acquire a time-series grayscale image of the construction site to provide immediate and accurate environmental data. Since dust typically manifests as smoke, it appears as a mixture of light and dark in the image. Therefore, in the particle region segmentation module, the grayscale variation of each pixel is analyzed in detail to determine the atomization coefficient. This coefficient reflects the grayscale variation around the pixel, facilitating subsequent determination of whether it is a dust particle. Dust particles of varying sizes exhibit distinct motion patterns at the construction site, so the atomization coefficient is combined with the pixel's motion and grayscale distribution to more accurately delineate the particle region within the image. The particle region captured in the aforementioned operation contains water mist particles generated by a fog cannon. Therefore, to more accurately assess the dust reduction situation at the construction site, it is necessary to analyze the diffusion of the particles and the combination of the water mist. Given that the water mist particles produced by industrial fog cannons are small, evenly distributed, and have a consistent motion pattern, the number distribution of pixels, their motion pattern, and the atomization coefficient within the particle region are analyzed to obtain a dust reduction characteristic value for the particle region, which reflects the dust reduction effect within the current image. Finally, in the water mist concentration adjustment and dust suppression module, the initial water mist concentration is adaptively adjusted based on the dust suppression characteristic value to obtain the final water mist concentration. This final water mist concentration is then used to suppress dust at the construction site. This comprehensive assessment system can more comprehensively reflect the dust conditions at the construction site, providing a scientific basis for developing targeted dust suppression strategies and significantly enhancing dust suppression effectiveness.

[0127] The embodiment of the present invention also provides a municipal construction dust reduction device based on image processing, please refer to Figure 4 , which shows a structural schematic diagram of a municipal construction dust reduction equipment based on image processing provided by an embodiment of the present invention, including a processor 400, a memory 401, a bus 402 and a communication interface 403, wherein the processor 400, the communication interface 403 and the memory 401 are connected via the bus 402; wherein the memory 401 may include a high-speed random access memory, the bus 402 may be an ISA bus, a PCI bus or an EISA bus, etc., and the processor 400 may be an integrated circuit chip with signal processing capabilities; the memory 401 stores at least one instruction, at least one program, a code set or an instruction set, and when the at least one instruction, at least one program, a code set or an instruction set is loaded and executed by the processor, a step in a municipal construction dust reduction system based on image processing is implemented.

[0128] It should be noted that the order in which the embodiments of the present invention are described above is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or sequential order shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0129] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments.

Claims

1. A municipal construction dust reduction system based on image processing, characterized in that: The system comprises: A data acquisition module is used to acquire two grayscale images of the construction site that are adjacent in time sequence, and use the previous grayscale image in time sequence as the target image; The particle area segmentation module is used to determine the fog coefficient of each pixel in the target image based on the grayscale fluctuation characteristics of the pixels in the local area. In two adjacent grayscale images, the module analyzes the movement and grayscale changes of the pixels and, based on the fog coefficient of each pixel, divides the target image into regions to obtain the particle area. A dustfall analysis module is used to obtain a dustfall characteristic value of the particle area based on the number distribution of pixels and the atomization coefficient of the pixels and the movement pattern of the pixels in the particle area; A water mist concentration adjustment and dust reduction module is used to adjust the initial water mist concentration according to the dust reduction characteristic value of the particulate matter area to obtain a final water mist concentration; and to reduce dust at the construction site based on the final water mist concentration; The method for obtaining the atomization coefficient includes: In the target image, based on the grayscale values ​​of the pixels, a maximum value point is selected; Construct a preset second neighborhood with each pixel as the center; In the preset second neighborhood corresponding to each pixel point, a maximum point is randomly selected as the test point, and the maximum point with the smallest Euclidean distance to the test point among the remaining maximum points is selected as the comparison point. The grayscale value mean between the test point and the comparison point is calculated as the first grayscale mean, and the grayscale value mean of the pixel points whose grayscale values ​​exist between the test point and the comparison point is calculated as the second grayscale mean. The Euclidean distance between the test point and the comparison point is used as the distance factor; The absolute value of the difference between the first grayscale mean and the second grayscale mean between the test point and the comparison point is used as the grayscale fluctuation factor, and the value after normalizing the ratio of the grayscale fluctuation factor to the distance factor is used as the fog factor corresponding to the test point; In the preset second neighborhood corresponding to each pixel point, the sum of the fog factors corresponding to all maximum value points is normalized and used as the fog coefficient of each pixel point; The method for obtaining the final water mist concentration includes: Using the difference between the preset third constant and the dustfall characteristic value as an adjustment coefficient; The product of the adjustment coefficient and the initial water mist concentration is taken as the final water mist concentration.

2. The municipal construction dust reduction system based on image processing according to claim 1 is characterized in that: The method for obtaining the maximum value point includes: In the target image, a preset first neighborhood of each pixel is determined, and within the preset first neighborhood corresponding to each pixel, the pixel with the largest grayscale value is taken as a maximum value point; The preset first neighborhood is smaller than the preset second neighborhood.

3. The municipal construction dust reduction system based on image processing according to claim 1 is characterized in that: The method for obtaining the particulate matter area includes: Use two grayscale images as input to the optical flow method, and output the optical flow vector corresponding to the pixel point in the target image; In the target image, the granularity factor of each pixel is determined based on the fog coefficient and optical flow vector corresponding to each pixel, as well as the grayscale disorder of the pixel in the corresponding preset second neighborhood. Segmenting the target image based on the Otsu threshold segmentation method to obtain two regions, wherein the regions are divided into a foreground region and a background region; In each area, the mean particle factor of all pixels is calculated as the particle index corresponding to each area; The area with the largest particle matter index is regarded as the particle matter area.

4. The municipal construction dust reduction system based on image processing according to claim 3 is characterized in that: The method for obtaining the particle factor includes: In the preset second neighborhood corresponding to each pixel, the mean value of the grayscale gradients of all pixels is used as the grayscale change factor of each pixel; Calculate the modulus of the optical flow vector corresponding to each pixel as the motion intensity corresponding to each pixel; For any pixel point, the product of the fog coefficient of the pixel point and the grayscale change factor is used as the grayscale change coefficient. The ratio of the grayscale change coefficient of the pixel point and the motion intensity is normalized with the absolute value of the difference between the preset first constant to obtain the particle factor of the pixel point.

5. The municipal construction dust reduction system based on image processing according to claim 3 is characterized in that: The method for obtaining the dustfall characteristic value includes: Dividing the particle area based on a superpixel segmentation method to obtain a plurality of superpixel blocks; Based on the number distribution, motion, and fog coefficient of all superpixel blocks in the particle area, the water fog factor of each superpixel block and the water fog combination index of the particle area are determined; In the particle area, analyzing the movement and diffusion of pixels, and combining the mist factors of all superpixel blocks in the particle area to determine the dust diffusion coefficient of the particle area; The normalized value of the ratio of the water mist combination index to the dust diffusion coefficient is used as the dustfall characteristic value of the particulate matter area.

6. The municipal construction dust reduction system based on image processing according to claim 5, characterized in that: The method of determining the water mist factor of each superpixel block and the water mist combination index of the particle area based on the number distribution, motion and fog coefficient of pixels in all superpixel blocks in the particle area includes: In each superpixel block, the average of the fog coefficients of all pixels is taken as the fog mean; The number of iterations is preset. In each iteration, a number of pixels are randomly selected in each superpixel block as water mist pixels. The ratio of the number of water fog pixels in each superpixel block to the number of all pixels is calculated as the water fog density, and the variance of the modulus of the optical flow vectors of all water fog pixels in each superpixel block is used as the confusion factor of each superpixel block. The absolute value of the difference between the ratio of the fog mean and the water fog density corresponding to each superpixel block and the preset second constant is used as the true factor. The ratio of the true factor to the confusion factor is negatively correlated and normalized to the value obtained as the water fog loss factor corresponding to each superpixel block in each iteration; The sum of the water mist loss factors corresponding to all superpixel blocks is used as the water mist loss value corresponding to the particle area in each iteration; In all iterations, under the minimum water mist loss value, the water mist density corresponding to each superpixel block is used as the water mist factor of each superpixel block; the normalized value of the difference between the maximum water mist density and the minimum water mist density in all superpixel blocks is used as the water mist combination index of the particle matter area.

7. The municipal construction dust reduction system based on image processing according to claim 5, characterized in that: The method for obtaining the dust diffusion coefficient includes: In each superpixel block, the mean of the optical flow vectors of all pixels is calculated as the comprehensive vector; Decomposing the integrated vector, removing the vertical downward component and synthesizing the vector to obtain a diffusion vector; The ratio of the modulus of the diffusion vector corresponding to each superpixel block to the water fog factor is used as the diffusion factor corresponding to each superpixel block; The normalized value of the mean diffusion factor of all superpixel blocks is used as the dust diffusion coefficient of the particulate matter area.

8. A municipal construction dust reduction device based on image processing, characterized in that: It includes a processor and a memory, wherein the memory stores at least one instruction, at least one program, code set or instruction set, and when the at least one instruction, at least one program, code set or instruction set is loaded and executed by the processor, a municipal construction dust reduction system based on image processing as described in any one of claims 1 to 7 is implemented.

Images