A dust concentration intelligent detection method and system

Abstract

The present invention relates to the field of dust concentration detection, and more specifically, the present invention relates to an intelligent dust concentration detection method and system. The method comprises: obtaining a grayscale image based on a preset sampling interval, obtaining no less than one area in the grayscale image by edge detection, and calculating the boundary clarity of the target area; dividing the target area into two sub-areas, and determining the regional consistency of the target area according to the regional consistency of the two sub-areas; taking the inverse of the product of the boundary clarity and the regional consistency as the dust score, and when the dust score is greater than the preset threshold, the target area is a dust area; traversing to obtain the total area of ​​the dust area in the grayscale image, taking the ratio of the total area of ​​the dust area to the area of ​​the grayscale image as the particle density, calculating the final concentration value according to the particle density, and completing the dust concentration detection. Through the technical solution of the present invention, the accuracy of dust concentration detection can be improved.

Description

Technical Field

[0001] The present invention relates to the field of dust concentration detection, and more specifically, to an intelligent dust concentration detection method and system. Background Art

[0002] Dust is one of the common pollutants in industrial production processes, and its concentration is directly related to air quality, the health of workers, and the safe operation of equipment. In industries such as mining, construction, and manufacturing, excessive dust concentration can not only cause respiratory diseases, but can also lead to serious safety accidents such as fires and explosions. With the improvement of global awareness of environmental protection and the strict implementation of environmental protection laws and regulations, modern dust detectors also have higher requirements. They not only need to have higher accuracy and stability, but also need to achieve remote monitoring and data analysis through wireless communication and cloud platform integration to achieve real-time monitoring and intelligent alarm, and improve response speed and early warning capabilities.

[0003] The existing Chinese patent application document with publication number CN114387273A discloses an environmental dust concentration detection method and system based on computer image recognition. The method divides the noise image into multiple areas, calculates the color component ratio, contrast, number of edges and number of edge corner points of the working area image and the denoised image respectively, and then calculates the first background complexity corresponding to the working area image and the second background complexity corresponding to the denoised image respectively; calculates the difference between the first background complexity and the second background complexity, and obtains the dust concentration according to the difference.

[0004] However, the above application document uses noise intensity to characterize the dust concentration in the working area. The noise intensity is directly proportional to the dust concentration. The greater the noise intensity, the higher the dust concentration. However, it takes a certain amount of time for dust to diffuse in the working environment. Directly judging the dust concentration based on the noise intensity will lead to inaccurate dust detection results. Summary of the invention

[0005] In order to solve the problem of inaccurate prediction of dust detection results, the present invention proposes an intelligent dust concentration detection method and system.

[0006] In a first aspect, the present invention discloses an intelligent dust concentration detection method, comprising: obtaining a grayscale image based on a preset sampling interval, obtaining no less than one area in the grayscale image by edge detection, taking any area as a target area, and calculating the boundary clarity of the target area; dividing the target area into two sub-areas, calculating the grayscale value variance of any pixel point in the sub-area within its 8-neighborhood, and taking the mean of the grayscale value variance of all pixels in the sub-area after normalization as the regional consistency, and determining the regional consistency of the target area according to the regional consistency of the two sub-areas; taking the inverse of the product of the boundary clarity and the regional consistency as a dust score, and when the dust score is greater than a preset threshold, the target area is a dust area; traversing to obtain the total area of ​​the dust area in the grayscale image, taking the ratio of the total area of ​​the dust area to the area of ​​the grayscale image as the particle density, and calculating the final concentration value according to the particle density to complete the dust concentration detection; wherein the final concentration value satisfies the relationship: , represents the final concentration value, and Respectively represent the concentration values ​​detected by the detector at two adjacent sampling moments, and Respectively represent the particle density of the grayscale image at two adjacent sampling moments.

[0007] According to the characteristics of dust diffusion, the target area is divided into two inner and outer sub-areas of the same area, and the regional consistency of the two sub-areas is calculated, which effectively quantifies the image characteristics of the dust area. Based on the regional consistency difference between the two sub-areas, it can be known whether the low dust content in the target area is due to the low dust content itself or due to the presence of a dust source in the first sub-area, and the dust has not spread out, thereby preventing misjudgment or misjudgment of the dust area. The grayscale image cannot directly obtain the dust concentration, but the trend of particle density changes in the grayscale image at adjacent sampling moments is consistent with the trend of dust concentration changes in the actual environment. The calculation of the final concentration value can be combined with the particle density information of the grayscale image, making the dust concentration result obtained by the detector more accurate.

[0008] Preferably, the boundary definition includes: constructing a window with the edge pixel point of the target area as the center and a preset length as the size, and taking the pixel points in the window other than the edge pixel points as reference points; the boundary definition of the target area satisfies the relationship:

[0009] , Indicates the boundary clarity, Represents the total number of edge pixels. Indicates the window size, and Represent edge pixels Left side reference point and The gray value of the reference point, and Represent edge pixels Right side reference point and The gray value of the reference point, Represents the normalization function.

[0010] By comparing the grayscale value difference of the reference points on both sides of the edge pixel, the degree of edge blur can be identified, and the gradient characteristics of the area boundary can be accurately captured, thereby better performing image analysis and processing and improving the algorithm's ability to recognize the target area boundary.

[0011] Preferably, the direction of the window is consistent with the gradient direction of the edge pixel points.

[0012] Preferably, the boundary clarity also includes: calculating the difference cumulative value between each neighborhood point and the edge pixel point, the neighborhood point is the 8 neighborhood pixel points of the edge pixel point; calculating the product of the difference cumulative value and the gradient value of the edge pixel point, and taking the average product of all edge pixels in the target area after normalization as the boundary clarity.

[0013] It not only enhances the edge contrast, but also improves the accuracy of edge definition assessment, making the boundary clarity between different target areas comparable and enhancing the ability to recognize image edge features.

[0014] Preferably, dividing the target area into two sub-areas includes: iteratively corroding the target area inward along the edge, taking the area remaining from the corrosion as the first sub-area, and taking the pixels in the target area except the first sub-area as the second sub-area; in response to the areas of the first sub-area and the second sub-area being equal, stopping the iteration.

[0015] By iteratively corroding from the edge of the target area inward, the core area where dust is concentrated is identified as the first sub-area, and the rest constitutes the second sub-area. This division method is consistent with the dust diffusion process and determines the dust diffusion situation in the local range of the target area.

[0016] Preferably, the regional consistency of the target area satisfies the relationship:

[0017] , Indicates the regional consistency of the target area, represents the regional consistency of the first sub-region, represents the regional consistency of the second sub-region, Indicates the boundary clarity, Represents the sub-region difference parameter.

[0018] If the regional consistency of the first sub-region and the second sub-region is significantly different, simply using the average of the regional consistency of the two sub-regions to represent the regional consistency of the target region may lead to misjudgment. At this time, correcting the regional consistency of the first sub-region by edge clarity can more accurately judge the regional consistency of the target region.

[0019] Preferably, the regional consistency of the target area also satisfies the relationship:

[0020] , Indicates the regional consistency of the target area, represents the regional consistency of the first sub-region, represents the regional consistency of the second sub-region, Indicates the degree of diffusion, Represents the sub-region difference parameter.

[0021] Preferably, the degree of diffusion includes: obtaining the centroid of the target area, obtaining the matching area of ​​the target area, wherein the grayscale image of the matching area and the grayscale image of the target area are grayscale images at two adjacent sampling moments, and the distance between the centroid of the matching area and the centroid of the target area is the smallest; respectively calculating the average distance of the target area and the matching area, the average distance being the mean of the distance between each edge pixel and the centroid, and calculating the difference between the average distance of the target area and the average distance of the matching area; calculating the ratio of the area of ​​the target area to the area of ​​the matching area, and normalizing the product of the ratio and the difference as the degree of diffusion.

[0022] In a second aspect, the present invention discloses an intelligent dust concentration detection system, comprising: a processor; and a memory, wherein the memory stores computer instructions, and when the computer instructions are executed by the processor, the system executes the above-mentioned intelligent dust concentration detection method.

[0023] Beneficial effects of the present invention:

[0024] 1. Image analysis technology based on dust diffusion characteristics has made significant progress in accurately identifying dust areas. By carefully dividing the target area, the image characteristics of the dust area can be quantified and the interference of non-dust moving objects can be effectively eliminated, thus ensuring the accuracy of the image data.

[0025] 2. The dust score is used to calculate the particle density of the grayscale image, and combined with the dust density change trend detected by the detector, the dust density can be weighted and corrected to achieve more accurate dust area identification, thereby improving the accuracy of dust detection results. This is of great significance to ensuring the safety of workers' own work. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] By reading the following detailed description with reference to the accompanying drawings, the above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-limiting manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein:

[0027] Figure 1 It is a flow chart of an intelligent dust concentration detection method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0028] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0029] It should be understood that when the terms "first", "second", etc. are used in the claims, descriptions, and drawings of the present invention, they are only used to distinguish different objects, rather than to describe a specific order. The terms "include" and "comprise" used in the description and claims of the present invention indicate the presence of the described features, wholes, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components, and / or their collections.

[0030] The present invention provides an intelligent dust concentration detection method. Figure 1 As shown, an intelligent dust concentration detection method includes steps S1 to S4, which are described in detail below.

[0031] S1, obtaining a grayscale image based on a preset sampling interval, obtaining at least one region in the grayscale image by edge detection, taking any region as a target region, and calculating the boundary clarity of the target region.

[0032] In one embodiment, an industrial camera is used to capture an RGB image (Red Green Blue) of the environment at a preset sampling interval, and electrical signal data of a dust concentration detector is acquired at the same preset sampling interval. For example, the preset sampling interval is set to once every 2 seconds.

[0033] The environment RGB image is grayed to obtain a grayscale image, and a mixed Gaussian model is used to model the background. After the modeling is completed, the background and foreground of the grayscale image are detected.

[0034] The canny edge detection algorithm is used on the foreground area of ​​each grayscale image to obtain at least one area in the grayscale image, and any area is taken as the target area.

[0035] The edge must meet the following conditions: it is inside the foreground and both end points of the edge are connected to the boundary of the foreground, and the foreground area can be divided into two parts by the edge. No operation is performed on edges that do not meet the conditions.

[0036] It should be noted that due to factors such as the diffusion characteristics of dust itself, the dynamic changes in environmental conditions, and the gradual distribution of dust concentration, the concentration of dust gradually decreases as it diffuses outward from the source. Therefore, the boundary of the dust area will appear fuzzy and unclear, while people or other objects will have clear edge boundaries. Therefore, the possibility of the target area being a dust area can be measured based on the boundary clarity of the target area.

[0037] Calculate the gradient direction of the edge pixel point, take the gradient direction as the reference direction, set the window, the center of the window is the edge pixel point of the target area, the window size is the preset length, and the pixels in the window except the edge pixel point are used as reference points.

[0038] The boundary clarity of the target area satisfies the relationship:

[0039] , Indicates the boundary clarity, Represents the total number of edge pixels. Indicates the window size, and Represent edge pixels Left side reference point and The gray value of the reference point, and Represent edge pixels Right side reference point and The gray value of the reference point, Represents the normalization function.

[0040] For example, the window size is set to 11, that is, .

[0041] Indicates edge pixels The grayscale value difference of pixels at the same position on the left and right sides is used to indicate the difference in dust concentration near the edge.

[0042] and The difference between two adjacent reference points is calculated respectively, which can characterize the change of concentration value near the edge.

[0043] In one embodiment, the boundary clarity also includes: calculating the difference cumulative value between each neighborhood point and the edge pixel point, the neighborhood point is the 8 neighborhood pixel points of the edge pixel point; calculating the product of the difference cumulative value and the gradient value of the edge pixel point, and taking the average product of all edge pixel points in the target area after normalization as the boundary clarity.

[0044] It should be noted that dust generally reduces the clarity of the image, and the area blocked by dust will become blurred, but the occlusion of dust will not completely eliminate the features of the original area, that is, the grayscale value changes slightly in the dust-blocked area. The occlusion caused by people and other influencing objects during movement is complete occlusion, and the original features of the blocked area are completely changed, that is, the grayscale value changes drastically.

[0045] S2, divide the target area into two sub-areas, calculate the gray value variance of any pixel in the sub-area within its 8-neighborhood, and take the mean of the gray value variance of all pixels in the sub-area after normalization as the regional consistency, and determine the regional consistency of the target area based on the regional consistency of the two sub-areas.

[0046] In one embodiment, the target area is iteratively eroded inward along the edge, the area remaining from the erosion is used as the first sub-area, and the pixels in the target area except the first sub-area are used as the second sub-area; in response to the areas of the first sub-area and the second sub-area being equal, the iteration is stopped.

[0047] It should be noted that if there is a dust source inside the target area and dust is continuously being generated, the newly generated dust will not diffuse over a large enough area in a short period of time, resulting in uneven dust concentration in the target area, which is manifested as the closer to the edge of the target area, the lower the dust concentration and the smaller the regional consistency; the closer to the center of the target area, the greater the dust concentration and the greater the regional consistency. In this case, if the sub-areas are not divided and calculated separately, the regional consistency of the target area will be inaccurate.

[0048] The regional consistency of the sub-region satisfies the relationship:

[0049] ,in, Indicates Regional consistency of sub-regions, Indicates The total number of pixels in the sub-region, Indicates In the sub-area The gray value variance of a pixel in its 8-neighborhood is Represents the exponential function, used for normalization.

[0050] when The regional consistency of the first sub-region is obtained when When the gray value variance is smaller, it means that the gray value difference between the pixel point and its neighboring pixels is small and the consistency is high.

[0051] The regional consistency of the target area is negatively correlated with the possibility that the target area is a dust area, because the dust itself is composed of discrete dust particles, and it cannot be guaranteed that the dust is distributed in a uniform state, especially when the dust is relatively thin, the regional consistency is low.

[0052] The regional consistency of the target area satisfies the relationship:

[0053] , Indicates the regional consistency of the target area, represents the regional consistency of the first sub-region, represents the regional consistency of the second sub-region, Indicates the boundary clarity, Represents the sub-region difference parameter.

[0054] Exemplarily, the regional difference parameter may be set by a person skilled in the art, and the present invention sets the regional difference parameter to 0.5.

[0055] when When , the difference in regional consistency between the first sub-region and the second sub-region is small, indicating that the dust distribution inside the target area is uniform. At this time, the regional consistency of the target area can be expressed by the mean of the regional consistency of the two sub-regions.

[0056] when When the regional consistency of the first sub-region and the second sub-region is different, due to the continuous generation of dust in the first sub-region of the target area, when the dust in the first sub-region is not dispersed in time, using the mean of the regional consistency of the two sub-regions to represent the regional consistency of the target area will be biased and cannot reflect the low regional consistency characteristics that are common in dust areas, thus affecting the identification of dust areas. At this time, it is necessary to correct the regional consistency of the first sub-region according to the boundary clarity of the target area to avoid excessive influence of the regional consistency of the first sub-region on the regional consistency of the target area.

[0057] In one embodiment, the centroid of the target area is obtained, and the matching area of ​​the target area is obtained, wherein the grayscale image of the matching area and the grayscale image of the target area are grayscale images at two adjacent sampling moments, and the distance between the centroid of the matching area and the centroid of the target area is the smallest.

[0058] The average distances of the target area and the matching area are calculated respectively. The average distance is the mean of the distances between each edge pixel and the centroid. The difference between the average distance of the target area and the average distance of the matching area is calculated. The ratio of the area of ​​the target area to the area of ​​the matching area is calculated, and the product of the ratio and the difference is normalized as the diffusion degree.

[0059] The regional consistency of the first sub-region is corrected by the diffusion degree, and the regional consistency of the target region satisfies the relationship:

[0060] , Indicates the regional consistency of the target area, represents the regional consistency of the first sub-region, represents the regional consistency of the second sub-region, Indicates the degree of diffusion, Represents the sub-region difference parameter.

[0061] S3, the inverse of the product of boundary clarity and regional consistency is used as the dust score. When the dust score is greater than a preset threshold, the target area is a dust area.

[0062] Exemplarily, the preset threshold is set to 0.3, that is, when the dust score of the target area is greater than 0.3, it is a dust area.

[0063] S4, traverse to obtain the total area of ​​the dust region in the grayscale image, take the ratio of the total area of ​​the dust region to the area of ​​the grayscale image as the particle density, calculate the final concentration value according to the particle density, and complete the dust concentration detection.

[0064] In one embodiment, all regions in a grayscale image are divided into dust regions and non-dust regions according to the method for determining dust regions in step S3, and the particle density corresponding to each grayscale image is calculated.

[0065] The final concentration value is calculated based on the particle density, where the final concentration value satisfies the relationship:

[0066] , represents the final concentration value, and Respectively represent the concentration values ​​detected by the detector at two adjacent sampling moments, and Respectively represent the particle density of the grayscale image at two adjacent sampling moments.

[0067] and The difference is positively correlated with the correction strength of the concentration value correction process detected by the detector. A large difference indicates that the change trend of the particle density in the grayscale image is very different from the change trend of the concentration value of the detector, and it is necessary to increase the correction strength of the concentration value correction process detected by the detector; conversely, a small difference indicates that the change trend of the particle density in the grayscale image is small compared with the change trend of the concentration value of the detector, and the correction strength of the concentration value correction process detected by the detector can be reduced. Ensure the accuracy of dust concentration detection.

[0068] An embodiment of the present invention further discloses an intelligent dust concentration detection system, including a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, an intelligent dust concentration detection method according to the present invention is implemented.

[0069] The above system also includes other components well known to those skilled in the art, such as a communication bus and a communication interface, and their configuration and functions are known in the art, so they will not be described in detail here.

[0070] In the present invention, the aforementioned memory may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, apparatus or device. For example, a computer-readable storage medium may be any appropriate magnetic storage medium or magneto-optical storage medium, such as a resistive random access memory RRAM (Resistive Random Access Memory), a dynamic random access memory DRAM (Dynamic Random Access Memory), a static random access memory SRAM (Static Random-Access Memory), an enhanced dynamic random access memory EDRAM (Enhanced Dynamic Random Access Memory), a high-bandwidth memory HBM (High-Bandwidth Memory), a hybrid memory cube HMC (Hybrid Memory Cube), etc., or any other medium that can be used to store the required information and can be accessed by an application, a module or both. Any such computer storage medium may be part of a device or accessible or connectable to a device.

[0071] Although this specification has shown and described a number of embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and alternatives without departing from the ideas and spirit of the present invention. It should be understood that in the practice of the present invention, various alternatives to the embodiments of the present invention described herein may be employed.

[0072] The above are all preferred embodiments of the present invention, and are not intended to limit the protection scope of the present invention. Therefore, any equivalent changes made based on the structure, shape, and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. An intelligent dust concentration detection method, characterized in that: include: A grayscale image is obtained based on a preset sampling interval, and at least one region in the grayscale image is obtained by edge detection. Any region is taken as a target region, and the boundary clarity of the target region is calculated. The boundary clarity includes: A window is constructed with the edge pixel of the target area as the center and the preset length as the size, and the pixels in the window other than the edge pixel are used as reference points; The boundary clarity of the target area satisfies the relationship: , Indicates the boundary clarity, Represents the total number of edge pixels. Indicates the window size, and Represent edge pixels Left side reference point and The gray value of the reference point, and Represent edge pixels Right side reference point and The gray value of the reference point, represents the normalization function; The target area is divided into two sub-areas, the gray value variance of any pixel in the sub-area within its 8-neighborhood is calculated, and the mean of the gray value variance of all pixels in the sub-area after normalization is taken as the regional consistency. The regional consistency of the target area is determined based on the regional consistency of the two sub-areas. The regional consistency satisfies the relationship: , Indicates the regional consistency of the target area, represents the regional consistency of the first sub-region, represents the regional consistency of the second sub-region, Indicates the boundary clarity, represents the sub-region difference parameter; The inverse of the product of boundary clarity and regional consistency is taken as the dust score. When the dust score is greater than the preset threshold, the target area is a dust area. The total area of ​​the dust region in the grayscale image is obtained by traversing, and the ratio of the total area of ​​the dust region to the area of ​​the grayscale image is used as the particle density. The final concentration value is calculated according to the particle density to complete the dust concentration detection; Among them, the final concentration value satisfies the relationship: , represents the final concentration value, and Respectively represent the concentration values ​​detected by the detector at two adjacent sampling moments, and Respectively represent the particle density of the grayscale image at two adjacent sampling moments.

2. The intelligent dust concentration detection method according to claim 1 is characterized in that: The direction of the window is consistent with the gradient direction of the edge pixel points.

3. The intelligent dust concentration detection method according to claim 1 is characterized in that: The boundary definition also includes: Calculate the difference between each neighborhood point and the edge pixel point, where the neighborhood point is the 8-neighborhood pixel point of the edge pixel point; The product of the difference accumulation value and the gradient value of the edge pixel point is calculated, and the average product of all edge pixels in the target area after normalization is taken as the boundary clarity.

4. The intelligent dust concentration detection method according to claim 1 is characterized in that: The dividing the target area into two sub-areas comprises: Iteratively corrode the target area inward along the edge, take the remaining area as the first sub-area, and take the pixels in the target area except the first sub-area as the second sub-area; In response to the areas of the first sub-region and the second sub-region being equal, the iteration is stopped.

5. The intelligent dust concentration detection method according to claim 1 is characterized in that: The regional consistency of the target area also satisfies the relationship: , Indicates the regional consistency of the target area, represents the regional consistency of the first sub-region, represents the regional consistency of the second sub-region, Indicates the degree of diffusion, Represents the sub-region difference parameter.

6. The intelligent dust concentration detection method according to claim 5 is characterized in that: The extent of spread includes: Obtaining the centroid of the target area and obtaining the matching area of ​​the target area, wherein the grayscale image where the matching area is located and the grayscale image where the target area is located are grayscale images at two adjacent sampling moments, and the distance between the centroid of the matching area and the centroid of the target area is the smallest; Calculate the average distance of the target area and the matching area respectively. The average distance is the mean of the distance between each edge pixel and the centroid. Calculate the difference between the average distance of the target area and the average distance of the matching area. The ratio of the area of ​​the target area to the area of ​​the matching area is calculated, and the product of the ratio and the difference is normalized as the diffusion degree.

7. An intelligent dust concentration detection system, characterized in that: include: Processor; and A memory storing computer instructions, wherein when the computer instructions are executed by a processor, the system executes an intelligent dust concentration detection method according to any one of claims 1 to 6.

Images