Sludge dewatering machine risk early warning method and system based on image recognition

Abstract

The invention relates to the field of risk early warning, in particular to a sludge dewatering machine risk early warning method and system based on image recognition, and the method comprises the steps: obtaining a gray level image of a sludge dewatering machine filter screen at the current moment, screening out blocked pixel points in the gray level image, enhancing the blocked pixel points through employing an unsharp mask algorithm, and obtaining an enhanced image; performing anomaly detection on the enhanced image, and sending out an early warning signal in response to an abnormal area in the enhanced image; the filter screen detection device has the effect of improving the accuracy of filter screen detection.

Description

Technical Field

[0001] The present invention relates to the field of risk warning, and in particular to a sludge dewatering machine risk warning method and system based on image recognition. Background Art

[0002] Sludge dewatering machine can remove water from sludge, reduce the volume and humidity of sludge, and thus facilitate subsequent treatment. Sludge dewatering machine is used in sewage treatment processes in municipal, chemical, petroleum, printing and dyeing, food processing, papermaking and other industries. Dehydrated sludge usually contains less water, which is convenient for transportation, stacking or further use. Sludge dewatering machine often faces some potential failures and risks during operation, the most common of which is filter blockage. If filter blockage is not discovered and handled in time, it will lead to low equipment efficiency and even affect the environment. Traditional equipment inspection relies on manual labor, and the manual inspection method has a large workload, and there are problems of untimely and inaccurate detection, which is easy to miss potential risks. The risk warning system based on image recognition can monitor the working status of the sludge dewatering machine in real time, automatically monitor the equipment and promptly detect filter blockage. The potential blockage on the surface of the filter of the sludge dewatering machine is often manifested as blurred spots or unclear boundaries in the image.

[0003] A Chinese patent document with announcement number CN107301627B discloses a dynamic image unsharp mask de-artifact enhancement method, comprising: using an unsharp mask (USM) enhancement algorithm to decompose the dynamic image to be processed to obtain a high-frequency detail image and a low-frequency contour image; generating a detail enhancement de-artifact mask and a dynamic range compression de-artifact mask according to the high-frequency detail image; using the detail enhancement de-artifact mask and the dynamic range compression de-artifact mask to perform de-artifact detail enhancement processing on the high-frequency detail image to obtain a high-frequency detail image after de-artifact detail enhancement, and performing dynamic range compression on the low-frequency contour image to obtain a compressed low-frequency contour image; synthesizing the high-frequency detail image after de-artifact detail enhancement and the compressed low-frequency contour image to obtain a dynamic image after de-artifacting.

[0004] The unsharp mask algorithm can enhance the high-frequency details in the image, making the image clearer and the details more obvious. Since sewage droplets may be attached to the surface of the filter screen of the sludge dewatering machine during operation, the potential blockage of the sewage droplets and the filter screen surface will produce similar effects in the image. The unsharp mask algorithm enhances the clarity of the image by emphasizing the edges and details of the image. Due to the presence of sewage droplets, the unsharp algorithm will simultaneously sharpen and enhance the edges of the sewage droplets in the image and the edges of the filter screen blockage area, resulting in the wrong detection of sewage droplets as filter screen blockage, resulting in low accuracy of filter screen blockage detection. Summary of the invention

[0005] In order to solve the problem of low filter detection accuracy, the present invention provides a sludge dewatering machine risk warning method and system based on image recognition.

[0006] In the first aspect, the present invention provides a sludge dewatering machine risk warning method based on image recognition, which adopts the following technical solution: Obtaining a grayscale image of the filter screen of the sludge dewatering machine at the current moment, filtering out blocked pixels in the grayscale image, enhancing the blocked pixels using an unsharp masking algorithm to obtain an enhanced image, performing anomaly detection on the enhanced image, and issuing an early warning signal in response to the presence of an abnormal area in the enhanced image; Among them, the method for screening out blocked pixels in the grayscale image is as follows: calculating the anomaly index of each pixel, the anomaly index represents the degree of difference between the corresponding pixel and the remaining pixels; calculating the blocking index of each pixel, the blocking index represents the possibility of blocking in the corresponding pixel area, and the blocking index is positively correlated with the anomaly index; and taking the pixels whose blocking index is greater than the preset blocking threshold as blocked pixels.

[0007] By calculating the potential blockage index of each pixel, we can understand the blockage situation in the grayscale image, provide a theoretical basis for the enhancement process of the unsharp mask algorithm, and ensure that the unsharp mask algorithm only enhances the potential blockage areas in the image, reducing the interference of sewage drops or other irrelevant factors in image enhancement, and improving the accuracy of filter detection.

[0008] Preferably, the expression of the anomaly index is: ; In the formula, represents the anomaly index of the mth pixel, Indicates the maximum grayscale value of a pixel in a grayscale image. Represents the gray value of the mth pixel, Represents the grayscale mean value of the mth pixel in the grayscale image at the historical moment, is a hyperparameter, and norm represents the normalization function.

[0009] The anomaly index is calculated in multiple dimensions to improve the accuracy of the calculation results. The anomaly index can be used to preliminarily determine whether the corresponding pixel area is a potential blocking area.

[0010] Preferably, the method further comprises: A window is constructed with any pixel point as the center, and the absolute value of the difference between the anomaly index of other pixels in the window area and the central pixel point is calculated. In response to the absolute value of the difference being less than a preset threshold, the corresponding pixel point is taken as the first pixel point; a window is constructed with the first pixel point as the center, and the absolute value of the difference between the anomaly index of other pixels in the window area and the central pixel point is repeatedly calculated until the absolute value of the difference between the anomaly index of all pixels in the window area and the central pixel point is greater than or equal to the preset threshold; the area composed of all first pixels is taken as the area to which the corresponding pixel point belongs; edge detection is performed on the belonging area to obtain the edge of the belonging area, and the chain code of the edge is calculated.

[0011] Preferably, the expression of the blocking index is: ; In the formula, represents the potential blocking index of the m-th pixel, represents the anomaly index of the mth pixel, The information entropy representing the gradient amplitude of all edge pixels in the region to which the m-th pixel belongs, It represents the sum of the absolute values ​​of the differences between two adjacent chain code values ​​in the chain code in the area to which the m-th pixel belongs, and norm represents the normalization function.

[0012] The blocking index can ultimately determine whether the corresponding pixel area is a blocking area, so that the image can be enhanced in a targeted manner to improve the enhancement effect.

[0013] Preferably, the expression of the blocking index is: ; In the formula, represents the potential blocking index of the m-th pixel, represents the anomaly index of the mth pixel, The information entropy of the cosine value of the angle between the gradient direction and the horizontal direction of all edge pixels in the region to which the m-th pixel belongs, It represents the sum of the absolute values ​​of the differences between two adjacent chain code values ​​in the chain code in the area to which the m-th pixel belongs, and norm represents the normalization function.

[0014] The congestion index is obtained by calculating multiple dimensions, which improves the accuracy of the congestion index calculation result and facilitates the judgment of whether the corresponding pixel point is a congestion area.

[0015] Preferably, the Sobel operator is used to perform edge detection on the region to obtain the edge of the region.

[0016] Preferably, the method of enhancing blocked pixels using an unsharp mask algorithm is: The grayscale image is blurred to obtain a blurred image, and the grayscale image and the blurred image are differentiated to obtain a difference image; in the difference image, the pixels corresponding to the blocked pixels in the grayscale image are used as enhanced pixels, and then the enhanced pixels are multiplied by a preset enhancement factor; the difference image and the grayscale image are added to obtain an enhanced image.

[0017] By selectively enhancing the pixels in the grayscale image, the enhancement effect of the grayscale image is improved, and the interference of sewage drops and other substances in the image enhancement process is reduced.

[0018] In a second aspect, the present invention provides a sludge dewatering machine risk warning system based on image recognition, which adopts the following technical solution: A sludge dewatering machine risk warning system based on image recognition, a processor and a memory, the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned sludge dewatering machine risk warning method based on image recognition is implemented.

[0019] The beneficial effect is: the above-mentioned sludge dewatering machine risk warning method based on image recognition is generated into a computer program and stored in a memory so as to be loaded and executed by a processor, thereby making a system based on the memory and the processor for easy use.

[0020] The present invention has the following technical effects: By calculating the potential blockage index of each pixel, it is easy to understand the blockage situation in the grayscale image, thereby ensuring that the unsharp mask algorithm only performs targeted enhancement on the potential blockage areas in the image, reducing the interference of sewage drops or other irrelevant factors on image enhancement, and improving the accuracy of subsequent filter detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] By reading the detailed description below with reference to the accompanying drawings, the above and other purposes, 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-restrictive manner, and the same or corresponding reference numerals represent the same or corresponding parts.

[0022] Figure 1 The present invention is a flow chart of a sludge dewatering machine risk early warning method based on image recognition. DETAILED DESCRIPTION

[0023] 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.

[0024] 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.

[0025] The embodiment of the present invention discloses a risk warning method for a sludge dewatering machine based on image recognition, referring to Figure 1 , including the following steps, as follows: S1: Get the grayscale image of the sludge dewatering machine filter at the current moment.

[0026] The filter surface of the sludge dewatering machine is photographed with a high-definition camera to obtain the filter surface image of the sludge dewatering machine, and then the filter surface image of the sludge dewatering machine is grayed to obtain the grayscale image of the filter surface image of the sludge dewatering machine. The filter area in the grayscale image is extracted using the FCN semantic segmentation model to analyze the pixel points in the filter area.

[0027] S2: Filter out blocked pixels in the grayscale image.

[0028] The screening method is: S201: Calculate an anomaly index for each pixel, where the anomaly index represents the degree of difference between the corresponding pixel and the remaining pixels.

[0029] The expression of the anomaly index is: ; In the formula, represents the anomaly index of the mth pixel, Indicates the maximum grayscale value of a pixel in a grayscale image. Represents the gray value of the mth pixel, Represents the grayscale mean value of the mth pixel in the grayscale image at the historical moment, is a hyperparameter, , setting the hyperparameter is to prevent the corresponding factor from being zero, norm represents the normalization function. For example, It represents the grayscale mean of the mth pixel in the 10 grayscale images at the historical moment. It should be noted that It represents the anomaly index of the mth pixel in the grayscale image at the current moment.

[0030] in the formula The larger the value is, the smaller the gray value of the mth pixel is in the gray image of the filter area, indicating that the area where the pixel is located is more likely to be a potential blocking area, and the greater its anomaly index is. The larger it is, the more obvious the grayscale value changes of the m-th pixel at multiple moments are, which means that the possibility that the area where the m-th pixel is located is a potential blocking area is greater, and its anomaly index is larger.

[0031] The potential blockage phenomenon in the filter area usually refers to the blockage of the filter holes by impurities, dust or other substances. The blocked substances cause the grayscale value of the potential blockage area to decrease. However, there are also some normal areas with reduced grayscale values ​​on the filter surface. As time goes by, the dirt in the filter will gradually accumulate, and the degree of blockage will change, which will cause the grayscale value in the image to change, making the grayscale value of the same position different in different frames. Therefore, when analyzing the anomaly index of each pixel, the smaller the grayscale value of a pixel, the more obvious the change in its grayscale value at multiple times, indicating that the pixel is more likely to belong to a potential blockage area and its anomaly index is greater.

[0032] S202: Calculate the blocking index of each pixel point. The blocking index indicates the possibility of blocking in the corresponding pixel point area. The blocking index is positively correlated with the anomaly index. Pixel points with blocking indexes greater than a preset blocking threshold are regarded as blocked pixels.

[0033] In one embodiment, a window is constructed with any pixel as the center, and the absolute value of the difference between the anomaly index of other pixels in the window area and the central pixel is calculated. In response to the absolute value of the difference being less than a preset threshold, the corresponding pixel is taken as the first pixel; a window is constructed with the first pixel as the center, and the absolute value of the difference between the anomaly index of other pixels in the window area and the central pixel is repeatedly calculated until the absolute value of the difference between the anomaly index of all pixels in the window area and the central pixel is greater than or equal to the preset threshold; the area composed of all first pixels is taken as the area to which the corresponding pixel belongs; the Sobel operator is used to perform edge detection on the area to obtain the edge of the area to which the Sobel operator belongs, and the chain code of the edge is calculated. The chain code is a simplified method for representing the boundary or contour in an image. The chain code effectively represents the shape by describing the directional relationship between each continuous point on the image contour. The present invention adopts an 8-neighborhood chain code. The calculation of the chain code is a prior art, and the specific steps are not repeated here. The threshold is set artificially according to the actual situation. For example, the threshold is 0.01.

[0034] For example, in pixels Build a 3×3 window for the center, and the pixels in the window area With the center pixel If the absolute value of the difference of the anomaly index is less than the threshold value 0.01, the pixel point As the first pixel; Build a 3×3 window for the center and find the first pixel in the window area , then continue with pixel Build a 3×3 window as the center. If the first pixel is not found in the window area, stop searching. , , The area where the pixel is located The region to which it belongs.

[0035] In one embodiment, a circular area with a radius of r is constructed with any pixel point as the center, and the circular area is taken as the area to which the corresponding pixel point belongs; the Sobel operator is used to perform edge detection on the area to obtain the edge of the area, and the chain code of the edge is calculated.

[0036] In one embodiment, the expression of the congestion index is: ; In the formula, represents the potential blocking index of the m-th pixel, represents the anomaly index of the mth pixel, The information entropy representing the gradient amplitude of all edge pixels in the region to which the m-th pixel belongs is obtained by calculating the gradient amplitude using the Sobel operator. The calculation method of the information entropy is a prior art and will not be described in detail here. It represents the sum of the absolute values ​​of the differences between the two adjacent chain code values ​​in the chain code of the area to which the mth pixel belongs. If there are multiple chain codes, calculate the absolute value of each chain code. , take its maximum value, norm represents the normalization function.

[0037] In the formula, The larger the value is, the greater the anomaly index of the m-th pixel in terms of gray value expression and gray value change analysis in multiple gray images, and the greater the possibility that the pixel belongs to a potential blockage area.

[0038] The larger it is, the more chaotic the gradient amplitude change of the edge pixels in the area to which the m-th pixel belongs is, which means that the possibility that the area to which the m-th pixel belongs belongs to a potential blockage area is greater, and the possibility that it belongs to a sewage drop area is smaller, and the corresponding m-th pixel has a larger potential blockage index.

[0039] The larger it is, the more dramatic the change in the gradient direction between all adjacent edge pixels in the area to which the m-th pixel belongs is. This means that the edge of the area to which the m-th pixel belongs is more chaotic and the texture is more complex. The more complex the gradient change direction is, the greater the possibility that the area to which the m-th pixel belongs is a potential blockage area, and the smaller the possibility that it belongs to a sewage drop area. Therefore, the potential blockage index of the m-th pixel is also greater.

[0040] The blocking threshold is set manually according to the actual situation. For example, the blocking threshold is 0.55, and the pixels with a blocking index greater than 0.55 are regarded as blocked pixels.

[0041] In the actual operation of the sludge dewatering machine, the edges of the potential blocking areas in the filter area are more chaotic and the texture is more complex, while the texture of the sewage drop area is smoother and the edges are clearer and more regular. Therefore, by analyzing the texture characteristics of the pixels in the area to which each pixel belongs in the filter area, the anomaly index of the pixel is optimized to obtain the potential blocking index of the corresponding pixel. Among them, the more chaotic the gradient amplitude change in the area to which the pixel belongs and the more complex the gradient change direction, the greater the possibility that the area to which the corresponding pixel belongs is a potential blocking area, and the greater the blocking index of the pixel.

[0042] In one embodiment, the expression of the congestion index is: ; In the formula, represents the potential blocking index of the m-th pixel, represents the anomaly index of the mth pixel, The information entropy of the cosine value of the angle between the gradient direction and the horizontal direction of all edge pixels in the region to which the m-th pixel belongs, It represents the sum of the absolute values ​​of the differences between two adjacent chain code values ​​in the chain code in the area to which the m-th pixel belongs, and norm represents the normalization function.

[0043] The larger it is, the more chaotic the gradient amplitude change of the edge pixels in the area to which the m-th pixel belongs is, which means that the possibility that the area to which the m-th pixel belongs belongs to a potential blockage area is greater, and the possibility that it belongs to a sewage drop area is smaller, and the corresponding m-th pixel has a larger potential blockage index.

[0044] S3: using an unsharp mask algorithm to enhance the blocked pixels to obtain an enhanced image, performing anomaly detection on the enhanced image, and issuing a warning signal in response to the presence of an abnormal area in the enhanced image.

[0045] The grayscale image is blurred using the Gaussian blur method to obtain a blurred image, and the grayscale image and the blurred image are differentiated to obtain a difference image; in the difference image, the pixels corresponding to the blocked pixels in the grayscale image are used as enhanced pixels, and then the enhanced pixels are multiplied by a preset enhancement factor; the difference image and the grayscale image are added to obtain an enhanced image.

[0046] For example, ; In the formula, represents the pixel in the difference image, Represents a pixel in a grayscale image. represents the pixel in the blurred image, represents the enhanced pixel in the difference image, represents the pixel point in the enhanced image, x, y represent the horizontal and vertical coordinates of the pixel point in the corresponding image, and k represents the enhancement factor, which is set manually according to the actual situation. For example, k=1.8.

[0047] In one embodiment, a template matching method is used to perform abnormal detection on the blocked area in the enhanced image to determine whether there is a blocked area in the filter area. When a blocked area exists in the enhanced image, a warning signal is issued. The warning signal is one or more combinations of voice prompts, information prompts, and traffic light prompts. The template matching method is a prior art, and the specific steps are not repeated here.

[0048] In one embodiment, a convolutional neural network model is used to detect the enhanced image to detect whether there is a blocked area in the filter area.

[0049] By calculating the potential blockage index of each pixel, we can understand the blockage situation in the grayscale image, provide a theoretical basis for the enhancement process of the unsharp mask algorithm, and ensure that the unsharp mask algorithm only enhances the potential blockage areas in the image, reducing the interference of sewage drops or other irrelevant factors in image enhancement, and improving the accuracy of filter detection.

[0050] An embodiment of the present invention also discloses a sludge dewatering machine risk warning system based on image recognition, including a processor and a memory, wherein the memory stores computer program instructions. When the computer program instructions are executed by the processor, a sludge dewatering machine risk warning method based on image recognition according to the present invention is implemented.

[0051] 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.

[0052] 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.

[0053] 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.

[0054] 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. A sludge dewatering machine risk warning method based on image recognition, characterized in that: Includes steps: Obtaining a grayscale image of the filter screen of the sludge dewatering machine at the current moment, filtering out blocked pixels in the grayscale image, enhancing the blocked pixels using an unsharp masking algorithm to obtain an enhanced image, performing anomaly detection on the enhanced image, and issuing an early warning signal in response to the presence of an abnormal area in the enhanced image; Among them, the method for screening out blocked pixels in the grayscale image is as follows: calculating the anomaly index of each pixel, the anomaly index represents the degree of difference between the corresponding pixel and the remaining pixels; calculating the blocking index of each pixel, the blocking index represents the possibility of blocking in the corresponding pixel area, and the blocking index is positively correlated with the anomaly index; and taking the pixels whose blocking index is greater than the preset blocking threshold as blocked pixels.

2. The sludge dewatering machine risk early warning method based on image recognition according to claim 1 is characterized in that: The expression of the anomaly index is: ； In the formula, represents the anomaly index of the mth pixel, Indicates the maximum grayscale value of a pixel in a grayscale image. Represents the gray value of the mth pixel, Represents the grayscale mean value of the mth pixel in the grayscale image at the historical moment, is a hyperparameter, and norm represents the normalization function.

3. The sludge dewatering machine risk early warning method based on image recognition according to claim 1 is characterized in that: The method also includes: A window is constructed with any pixel point as the center, and the absolute value of the difference between the anomaly index of other pixels in the window area and the central pixel point is calculated. In response to the absolute value of the difference being less than a preset threshold, the corresponding pixel point is taken as the first pixel point; a window is constructed with the first pixel point as the center, and the absolute value of the difference between the anomaly index of other pixels in the window area and the central pixel point is repeatedly calculated until the absolute value of the difference between the anomaly index of all pixels in the window area and the central pixel point is greater than or equal to the preset threshold; the area composed of all first pixels is taken as the area to which the corresponding pixel point belongs; edge detection is performed on the belonging area to obtain the edge of the belonging area, and the chain code of the edge is calculated.

4. The sludge dewatering machine risk warning method based on image recognition according to claim 3 is characterized in that: The expression of the congestion index is: ； In the formula, represents the potential blocking index of the m-th pixel, represents the anomaly index of the mth pixel, The information entropy representing the gradient amplitude of all edge pixels in the region to which the m-th pixel belongs, It represents the sum of the absolute values ​​of the differences between two adjacent chain code values ​​in the chain code in the area to which the m-th pixel belongs, and norm represents the normalization function.

5. The sludge dewatering machine risk early warning method based on image recognition according to claim 3 is characterized in that: The expression of the congestion index is: ； In the formula, represents the potential blocking index of the m-th pixel, represents the anomaly index of the mth pixel, The information entropy of the cosine value of the angle between the gradient direction and the horizontal direction of all edge pixels in the region to which the m-th pixel belongs, It represents the sum of the absolute values ​​of the differences between two adjacent chain code values ​​in the chain code in the area to which the m-th pixel belongs, and norm represents the normalization function.

6. The sludge dewatering machine risk early warning method based on image recognition according to claim 1 is characterized in that: Use the Sobel operator to perform edge detection on the area to obtain the edge of the area.

7. The sludge dewatering machine risk warning method based on image recognition according to claim 1 is characterized in that: The method of enhancing blocked pixels using the unsharp mask algorithm is: The grayscale image is blurred to obtain a blurred image, and the grayscale image and the blurred image are differentiated to obtain a difference image; in the difference image, the pixel points corresponding to the blocked pixel points in the grayscale image are used as enhanced pixel points, and then the enhanced pixel points are multiplied by a preset enhancement factor; The difference image is added to the grayscale image to obtain the enhanced image.

8. A sludge dewatering machine risk warning system based on image recognition, characterized in that: include: A processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, a sludge dewatering machine risk warning method based on image recognition according to any one of claims 1 to 7 is implemented.

Images