A foreign body detection method for coal mine belt conveyor

Abstract

The present invention relates to the field of image processing technology, and in particular to a method for detecting foreign matter in a coal mine belt conveyor, comprising: obtaining a coal mine RGB image and a coal mine grayscale image; obtaining the local disorder degree within the neighborhood of each pixel point according to the color deviation degree of each pixel point in each connected domain of the coal mine grayscale image; obtaining the probability of the existence of metal minerals in each connected domain according to the theoretical centroid position of the disorder degree in each connected domain of the coal mine grayscale image; obtaining the enhanced coal mine RGB image according to the enhanced color value of the first color channel of each pixel point in the coal mine RGB image; and then identifying the metal foreign matter in the coal mine conveyor belt. The present invention makes the foreign matter detection result more accurate.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and in particular to a method for detecting foreign matter in a coal mine belt conveyor. Background Art

[0002] Belt conveyors are widely used in industrial production due to their high transport capacity, high efficiency, and long operating times. With the advancement of science and technology, belt conveyors are moving towards high speeds, large inclination angles, and long distances. During the conveyor belt's transport process, foreign matter often enters the conveyor belt. When foreign matter is present on the conveyor belt, it can easily lead to production quality issues. Foreign matter on conveyor belts is typically gangue and metallic minerals. Gangue is typically sorted manually or physically, so this solution primarily targets metallic minerals for detection.

[0003] When using traditional foreign body detection algorithms to detect foreign objects in coal mine conveyor belt images, since metal minerals and coal are relatively similar in the images, detection errors are likely to occur. Therefore, image enhancement of metal minerals in coal mine conveyor belt images is necessary. Summary of the Invention

[0004] In order to solve the above problems, the present invention provides a method for detecting foreign matter in a coal mine belt conveyor, the method comprising:

[0005] Obtain coal mine RGB images and coal mine grayscale images;

[0006] Obtain several connected domains of the coal mine grayscale image; obtain the color deviation of each pixel in each connected domain of the coal mine grayscale image based on the color value difference of the three color channels of the pixel corresponding to the coal mine RGB image; obtain the local disorder degree in the neighborhood of each pixel in each connected domain based on the relative position and color deviation difference of the pixel in each connected domain;

[0007] According to the local disorder degree in the neighborhood of each pixel point in each connected domain, the theoretical centroid position of the disorder degree in each connected domain is obtained; according to the position difference between the pixel point in each connected domain and the theoretical centroid position of the disorder degree in each connected domain, the probability of the existence of metal minerals in each connected domain is obtained;

[0008] Based on the probability of the existence of metal minerals in each connected domain, the enhanced color value of the first color channel of each pixel in the coal mine RGB image is obtained; based on the enhanced color value of the first color channel of each pixel, an enhanced coal mine RGB image is obtained; based on the enhanced coal mine RGB image, metal foreign matter in the coal mine conveyor belt is identified.

[0009] Preferably, the specific method of obtaining the coal mine RGB image and the coal mine grayscale image includes:

[0010] Obtain a grayscale image of a coal mine conveyor belt; perform threshold segmentation on the grayscale image of the coal mine conveyor belt using the Otsu threshold algorithm to obtain a binary image of the coal mine conveyor belt; multiply the binary image of the coal mine conveyor belt with the RGB image of the coal mine conveyor belt and the grayscale image of the coal mine conveyor belt respectively to obtain an RGB image of the coal mine and a grayscale image of the coal mine.

[0011] Preferably, the color deviation of each pixel point in each connected domain of the coal mine grayscale image is obtained according to the color value difference of the three color channels of the pixel point of the coal mine RGB image corresponding to each pixel point in each connected domain of the coal mine grayscale image, including the specific method of:

[0012] Get the color value of each color channel of each pixel in each connected domain of the coal mine grayscale image; then The first connected domain The calculation method of the color deviation of each pixel is:

[0013]

[0014] Where, Represents the grayscale image of the coal mine The first connected domain Color deviation of each pixel; Represents the grayscale image of the coal mine The first connected domain The color value of the R color channel of each pixel; Represents the grayscale image of the coal mine The first connected domain The color value of the G color channel of each pixel; Represents the grayscale image of the coal mine The first connected domain The color value of the B color channel of each pixel; Indicates taking the absolute value.

[0015] Preferably, the method of obtaining the color value of each color channel of each pixel in each connected domain of the coal mine grayscale image includes the following specific methods:

[0016] For the coal mine grayscale image The first connected domain pixels, get the The pixel point corresponds to the pixel point in the coal mine RGB image and is recorded as The target pixel of the pixel is The color values ​​of the R, G, and B channels of the target pixel are used as the The color value of each color channel of each pixel.

[0017] Preferably, the method of obtaining the local disorder degree within the neighborhood of each pixel point in each connected domain according to the relative positions and color deviation differences of the pixels in each connected domain includes the following specific methods:

[0018] Get the neighborhood range of each pixel point in each connected domain of the coal mine grayscale image; then The first connected domain The calculation expression of the local disorder degree within the neighborhood of a pixel is:

[0019]

[0020] Where, Represents the grayscale image of the coal mine The first connected domain The degree of local chaos within the neighborhood of a pixel; Represents the grayscale image of the coal mine The first connected domain The maximum color deviation of all pixels in the neighborhood of a pixel; Represents the grayscale image of the coal mine The first connected domain The minimum color deviation of all pixels in the neighborhood of a pixel; Represents the grayscale image of the coal mine The first connected domain In the neighborhood of pixels Pixels and The Euclidean distance between pixels; Represents the grayscale image of the coal mine The first connected domain In the neighborhood of pixels Color deviation of each pixel; Represents the grayscale image of the coal mine The first connected domain Color deviation of each pixel; Indicates taking the absolute value; represents an exponential function with a natural constant as its base; are preset parameters.

[0021] Preferably, the specific method of obtaining the neighborhood range of each pixel point in each connected domain of the coal mine grayscale image includes:

[0022] For the coal mine grayscale image The first connected domain pixels, with the Pixels are used as the window center, and the window size is obtained as and record the window as The neighborhood range of pixels.

[0023] Preferably, the specific formula for obtaining the theoretical centroid position of the degree of chaos in each connected domain according to the degree of local chaos in the neighborhood of each pixel point in each connected domain is:

[0024]

[0025]

[0026] Where, Represents the grayscale image of the coal mine The horizontal coordinate of the theoretical centroid position of the degree of chaos in a connected domain; Represents the grayscale image of the coal mine The vertical coordinate of the theoretical centroid position of the degree of chaos in a connected domain; Represents the grayscale image of the coal mine The total number of all pixels in a connected domain; Represents the grayscale image of the coal mine The first connected domain The horizontal coordinate of the pixel point; Represents the grayscale image of the coal mine The first connected domain The vertical coordinate of the pixel point; Represents the grayscale image of the coal mine The first connected domain The degree of local chaos within the neighborhood of a pixel; Indicates rounding up.

[0027] Preferably, the specific formula for obtaining the probability of the existence of metal minerals in each connected domain based on the position difference between the pixel points in each connected domain and the theoretical centroid position of the disorder degree in each connected domain is:

[0028]

[0029] Where, Represents the grayscale image of the coal mine The probability of the existence of metal minerals in a connected domain; Represents the grayscale image of the coal mine The total number of all pixels in a connected domain; Represents the grayscale image of the coal mine The first connected domain The degree of local chaos within the neighborhood of a pixel; Represents the grayscale image of the coal mine The first connected domain pixels and The Euclidean distance between the theoretical centroid positions of the degree of disorder in a connected domain; Represents an exponential function with a natural constant as its base.

[0030] Preferably, the method of obtaining the enhanced color value of the first color channel of each pixel in the coal mine RGB image according to the probability of the existence of metal minerals in each connected domain includes the following specific methods:

[0031] Get the The probability that there are metal minerals in the connected domain of the pixel point; The color channel with the largest color value of the pixel is recorded as The first color channel of the pixel point, then the first The calculation method of the color value of the first color channel of each pixel after enhancement is:

[0032]

[0033] Where, Indicates the first The enhanced color value of the first color channel of each pixel; Indicates the first The color value of the first color channel of each pixel; Indicates the first The probability that a metal mineral exists in the connected domain to which the pixel belongs.

[0034] Preferably, the obtaining The probability of the existence of metal minerals in the connected domain of each pixel point is calculated, including the following specific methods:

[0035] Get the The pixel point corresponds to the pixel point in the coal mine grayscale image and is recorded as The first pixel of the pixel The connected domain where the first pixel of the pixels is located is recorded as the first connected domain, and the probability of the existence of metal minerals in the first connected domain is taken as the first connected domain. The probability that a metal mineral exists in the connected domain to which the pixel belongs.

[0036] The beneficial effect of the technical solution of the present invention is: the present invention enhances the color value of the first color channel of each pixel in the coal mine RGB image according to the probability of the existence of metal minerals in each connected domain, and obtains the enhanced coal mine RGB image, so that the color difference between the metal minerals and the coal mine in the enhanced coal mine RGB image is large, and it is easier to distinguish between the metal minerals and the coal mine in the enhanced coal mine RGB image, thereby making the foreign matter detection result more accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. 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.

[0038] Figure 1 The present invention is a flow chart of the steps of a method for detecting foreign matter on a coal mine belt conveyor. DETAILED DESCRIPTION

[0039] 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 a method for detecting foreign matter in a coal mine belt conveyor according to the present invention. In the following description, references to "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.

[0040] 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.

[0041] The specific scheme of the method for detecting foreign matter in a coal mine belt conveyor provided by the present invention is described in detail below with reference to the accompanying drawings.

[0042] See also Figure 1 , which shows a flow chart of a method for detecting foreign matter in a coal mine belt conveyor provided by one embodiment of the present invention, the method comprising the following steps:

[0043] Step S001: Acquire a coal mine RGB image and a coal mine grayscale image.

[0044] It should be noted that in the process of coal collection, belt conveyors are used to transport the collected coal, and some metal minerals are often mixed in the transportation process. Coal is generally transported in mines. The environment of the mines is relatively dim, and lighting is generally provided by lights. Under the influence of lights, the surface of the coal is relatively smooth due to its surface structure, and it appears brighter under the light. The metal minerals also have a strong ability to reflect light.

[0045] Specifically, in order to implement the coal mine belt conveyor foreign body detection method proposed in this embodiment, it is first necessary to collect the coal mine RGB image and the coal mine grayscale image. The specific process is as follows:

[0046] Lights and fixed cameras are arranged above the coal mine conveyor belt to obtain the RGB image of the coal mine conveyor belt, and the RGB image of the coal mine conveyor belt is grayscaled to obtain the coal mine conveyor belt grayscale image; the coal mine conveyor belt grayscale image is threshold segmented using the Otsu threshold algorithm to obtain the coal mine conveyor belt binary image, and the coal mine conveyor belt binary image is multiplied with the coal mine conveyor belt RGB image and the coal mine conveyor belt grayscale image respectively to obtain the coal mine RGB image and the coal mine grayscale image.

[0047] The coal mine RGB image and the coal mine grayscale image only contain coal mines and metal minerals; the Otsu threshold algorithm and grayscale operation are existing technologies and will not be described in detail in this embodiment.

[0048] So far, the coal mine RGB image and coal mine grayscale image are obtained through the above method.

[0049] Step S002: According to the color deviation of each pixel in each connected domain of the coal mine grayscale image, the local disorder degree in the neighborhood of each pixel is obtained.

[0050] It should be noted that although the coal mine grayscale image only contains coal mines and metal minerals, because both coal mines and metal minerals have reflective phenomena, the connected domain of the coal mine grayscale image may be coal mines or metal minerals, so it is necessary to judge the category of the connected domain.

[0051] Specifically, a gradient image of the coal mine grayscale image is obtained, and a watershed segmentation algorithm is performed on the gradient image of the coal mine grayscale image to obtain multiple regions of the coal mine grayscale image. Each region is used as a connected domain, and then several connected domains of the coal mine grayscale image are obtained.

[0052] 1. Obtain the color deviation of each pixel in each connected domain of the coal mine grayscale image.

[0053] It should be noted that coal mines are black in color, but due to lighting conditions, they appear as black, gray, and white in images. Their common characteristic is that their RGB colors are relatively similar, but metallic minerals are mixed with color, and the RGB color values ​​of some pixels in the metallic mineral area vary significantly. Therefore, a preliminary judgment can be made based on the RGB color value deviation. The less similar the RGB channel values ​​of a single pixel are, the further away it is from the color values ​​of black, gray, and white, and the greater the color deviation.

[0054] Specifically, for the coal mine grayscale image The first connected domain pixels, get the The pixel point corresponds to the pixel point in the coal mine RGB image and is recorded as The target pixel of the pixel is The grayscale values ​​of the R, G, and B channels of the target pixel are used as the first The color value of each color channel of the pixel; then the first The first connected domain The calculation method of the color deviation of each pixel is:

[0055]

[0056] Where, Represents the grayscale image of the coal mine The first connected domain Color deviation of each pixel; Represents the grayscale image of the coal mine The first connected domain The color value of the R color channel of each pixel; Represents the grayscale image of the coal mine The first connected domain The color value of the G color channel of each pixel; Represents the grayscale image of the coal mine The first connected domain The color value of the B color channel of each pixel; Indicates taking the absolute value.

[0057] For any pixel, the grayscale values ​​of the R, G, and B channels of the pixel are used as the color value of each color channel of the pixel.

[0058] At this point, the color deviation of each pixel in each connected domain of the coal mine grayscale image is obtained.

[0059] 2. Obtain the local disorder degree in the neighborhood of each pixel point in each connected domain of the coal mine grayscale image.

[0060] It's important to note that judging the presence of metallic minerals based on the color deviation of a single pixel is unreliable. Therefore, further determination of whether a pixel is a metallic mineral is made based on its local color deviation. Metallic minerals generally have complex colors, with common surface colors such as yellow, red, and brown. Pixels of different colors have different color deviations, resulting in significant variations in the color deviations corresponding to pixels on the surface of metallic minerals. However, the RGB color values ​​of coal mine surfaces are nearly uniform, resulting in relatively small and evenly distributed color deviations across the pixels.

[0061] Preset an order , wherein this embodiment is based on This example is described as an example, and this embodiment is not specifically limited. It depends on the specific implementation situation.

[0062] Specifically, for the coal mine grayscale image The first connected domain pixels, with the Pixels are used as the window center, and the window size is obtained as and record the window as The neighborhood range of pixels; then the first pixel of the coal mine grayscale image The first connected domain The calculation expression of the local disorder degree within the neighborhood of a pixel is:

[0063]

[0064] Where, Represents the grayscale image of the coal mine The first connected domain The degree of local chaos within the neighborhood of a pixel; Represents the grayscale image of the coal mine The first connected domain The maximum color deviation of all pixels in the neighborhood of a pixel; Represents the grayscale image of the coal mine The first connected domain The minimum color deviation of all pixels in the neighborhood of a pixel; Represents the grayscale image of the coal mine The first connected domain In the neighborhood of pixels Pixels and The Euclidean distance between pixels; Represents the grayscale image of the coal mine The first connected domain In the neighborhood of pixels Color deviation of each pixel; Represents the grayscale image of the coal mine The first connected domain Color deviation of each pixel; Indicates taking the absolute value; represents an exponential function with a natural constant as its base; are preset parameters.

[0065] It should be noted that within a pixel's neighborhood, when the difference between the maximum and minimum color deviation values ​​within the neighborhood is large, the confidence level in the confusion between the pixel's color deviation and the color deviations of the pixels in its neighborhood is higher. The smaller the Euclidean distance between any pixel in the pixel's neighborhood and that pixel, the greater the difference in their color deviations, indicating greater color confusion at that location.

[0066] So far, the local disorder degree in the neighborhood of each pixel point in each connected domain of the coal mine grayscale image is obtained through the above method.

[0067] Step S003: according to the theoretical centroid position of the disorder degree in each connected domain of the coal mine grayscale image, the probability of the existence of metal minerals in each connected domain is obtained.

[0068] 1. Obtain the theoretical centroid position of the chaos degree in each connected domain of the coal mine grayscale image.

[0069] It should be noted that image enhancement for metal minerals requires determining the probability of the presence of metal minerals within a single connected domain. A single connected domain may contain both coal and metal minerals, with the metal minerals occupying only a small area. Directly using the mean value for calculations may result in an underestimation of the probability of the presence of metal minerals within a single connected domain. Therefore, the theoretical centroid position of the degree of chaos is calculated, and the distance from the theoretical centroid position is used as a weight to calculate the probability of metal minerals within a single connected domain.

[0070] Specifically, the first grayscale image of the coal mine The calculation method of the theoretical centroid position of the degree of chaos in a connected domain is:

[0071]

[0072]

[0073] Where, Represents the grayscale image of the coal mine The horizontal coordinate of the theoretical centroid position of the degree of chaos in a connected domain; Represents the grayscale image of the coal mine The vertical coordinate of the theoretical centroid position of the degree of chaos in a connected domain; Represents the grayscale image of the coal mine The total number of all pixels in a connected domain; Represents the grayscale image of the coal mine The first connected domain The horizontal coordinate of the pixel point; Represents the grayscale image of the coal mine The first connected domain The vertical coordinate of the pixel point; Represents the grayscale image of the coal mine The first connected domain The degree of local chaos within the neighborhood of a pixel; Indicates rounding up.

[0074] At this point, the theoretical centroid position of the chaos degree in each connected domain of the coal mine grayscale image is obtained.

[0075] 2. Obtain the probability of the existence of metal minerals in each connected domain of the coal mine grayscale image.

[0076] It should be noted that for any pixel point in the connected domain, the closer the pixel point is to the theoretical centroid of the disorder degree in the connected domain, the more likely it is to be a metal mineral pixel point, and the greater the weight it occupies when calculating the probability of the existence of metal minerals in the connected domain.

[0077] Specifically, the first grayscale image of the coal mine The probability of the existence of metal minerals in a connected domain is calculated as follows:

[0078]

[0079] Where, Represents the grayscale image of the coal mine The probability of the existence of metal minerals in a connected domain; Represents the grayscale image of the coal mine The total number of all pixels in a connected domain; Represents the grayscale image of the coal mine The first connected domain The degree of local chaos within the neighborhood of a pixel; Represents the grayscale image of the coal mine The first connected domain pixels and The Euclidean distance between the theoretical centroid positions of the degree of disorder in a connected domain; Represents an exponential function with a natural constant as its base.

[0080] So far, the probability of the existence of metal minerals in each connected domain of the coal mine grayscale image is obtained through the above method.

[0081] Step S004: obtaining an enhanced RGB image of the coal mine according to the enhanced color value of the first color channel of each pixel in the RGB image of the coal mine; and then identifying metal foreign matter in the conveyor belt of the coal mine.

[0082] It should be noted that in the acquired RGB images of coal mines, metal minerals and coal have similar colors, which can easily lead to misdetection when using neural networks to segment and identify metal minerals. Therefore, based on the probability of the presence of metal minerals in connected domains, the RGB color values ​​of pixels within the connected domains are adjusted to increase the color saturation of the metal minerals, thereby improving the accuracy of the neural network's recognition of metal minerals.

[0083] Specifically, for the first pixels, get the The pixel point corresponds to the pixel point in the coal mine grayscale image and is recorded as The first pixel of the pixel The connected domain where the first pixel of the pixels is located is recorded as the first connected domain, and the probability of the existence of metal minerals in the first connected domain is taken as the first connected domain. The probability that there are metal minerals in the connected domain of the pixel point; The color channel with the largest color value of the pixel is recorded as The first color channel of the pixel point, then the first The calculation method of the color value of the first color channel of each pixel after enhancement is:

[0084]

[0085] Where, Indicates the first The enhanced color value of the first color channel of each pixel; Indicates the first The color value of the first color channel of each pixel; Indicates the first The probability that a metal mineral exists in the connected domain to which the pixel belongs.

[0086] According to the calculation method of the enhanced color value of the first color channel of each pixel in the coal mine RGB image, the color value of the first color channel of each pixel in the coal mine RGB image is updated to obtain the enhanced coal mine RGB image.

[0087] Furthermore, the specific method for identifying metal minerals through semantic segmentation U-net neural network is as follows: obtain a large number of previously collected RGB images of conveyor belts containing metal minerals, perform the above-mentioned enhancement operation on the conveyor belt RGB images, use the enhanced RGB images as a data set, manually annotate the category information in the image, wherein the background pixel category is marked as 0 and the metal mineral is marked as 1; and use the cross-entropy loss function to supervise the training. After the neural network training is completed, the enhanced coal mine RGB image is sent to the neural network, and the corresponding semantic segmentation image is obtained by network inference. Then, the metal mineral target is obtained through the pixel category label, and after identifying the position of the metal mineral, the morphological position of the metal mineral is visualized on the corresponding display, thereby identifying the metal mineral foreign matter in the coal mine conveyor belt.

[0088] At this point, this embodiment is completed.

[0089] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for detecting foreign matter in a coal mine belt conveyor, characterized in that: The method comprises the following steps: Obtain coal mine RGB images and coal mine grayscale images; Obtain several connected domains of the coal mine grayscale image; obtain the color deviation of each pixel point in each connected domain of the coal mine grayscale image according to the color value difference of the three color channels of the pixel point of the coal mine RGB image corresponding to each pixel point in each connected domain of the coal mine grayscale image; obtain the local disorder degree within the neighborhood range of each pixel point in each connected domain according to the relative position and color deviation difference of the pixel points in each connected domain; According to the local disorder degree in the neighborhood of each pixel point in each connected domain, the theoretical centroid position of the disorder degree in each connected domain is obtained; according to the position difference between the pixel point in each connected domain and the theoretical centroid position of the disorder degree in each connected domain, the probability of the existence of metal minerals in each connected domain is obtained; According to the probability of the existence of metal minerals in each connected domain, the enhanced color value of the first color channel of each pixel in the coal mine RGB image is obtained; according to the enhanced color value of the first color channel of each pixel, the enhanced coal mine RGB image is obtained; according to the enhanced coal mine RGB image, the metal foreign matter in the coal mine conveyor belt is identified; The specific method of obtaining the local disorder degree within the neighborhood of each pixel point in each connected domain according to the difference in relative position and color deviation of the pixel points in each connected domain is as follows: Get the neighborhood range of each pixel point in each connected domain of the coal mine grayscale image; then The first The calculation expression of the local disorder degree within the neighborhood of a pixel is: In the formula, Represents the grayscale image of the coal mine The first The degree of local disorder within the neighborhood of a pixel; Represents the grayscale image of the coal mine The first The maximum color deviation of all pixels in the neighborhood of a pixel; Represents the grayscale image of the coal mine The first The minimum color deviation of all pixels in the neighborhood of a pixel; Represents the grayscale image of the coal mine The first In the neighborhood of pixels Pixels and The Euclidean distance between pixels; Represents the grayscale image of the coal mine The first In the neighborhood of pixels Color deviation of each pixel; Represents the grayscale image of the coal mine The first Color deviation of each pixel; Indicates taking the absolute value; represents an exponential function with a natural constant as base; is the preset parameter; The specific method of obtaining the neighborhood range of each pixel point in each connected domain of the coal mine grayscale image includes: For the coal mine grayscale image The first pixels, with the pixels as the window center, and the window size is and record the window as The neighborhood range of pixels.

2. A method for detecting foreign matter in a coal mine belt conveyor according to claim 1, characterized in that: The specific method of obtaining the coal mine RGB image and the coal mine grayscale image includes: Acquire the grayscale image of the coal mine conveyor belt; use the Otsu threshold algorithm to perform threshold segmentation on the coal mine conveyor belt grayscale image to obtain the coal mine conveyor belt binary image; multiply the coal mine conveyor belt binary image with the coal mine conveyor belt RGB image and the coal mine conveyor belt grayscale image to obtain the coal mine RGB image and the coal mine grayscale image.

3. A method for detecting foreign matter in a coal mine belt conveyor according to claim 1, characterized in that: The method of obtaining the color deviation of each pixel point in each connected domain of the coal mine grayscale image according to the color value difference of the three color channels of the pixel point of the coal mine RGB image corresponding to each pixel point in each connected domain of the coal mine grayscale image includes the following specific methods: Get the color value of each color channel of each pixel in each connected domain of the coal mine grayscale image; then The first The calculation method of the color deviation of a pixel is: In the formula, Represents the grayscale image of the coal mine The first Color deviation of each pixel; Represents the grayscale image of the coal mine The first The color value of the R color channel of each pixel; Represents the grayscale image of the coal mine The first The color value of the G color channel of each pixel; Represents the grayscale image of the coal mine The first The color value of the B color channel of each pixel; Indicates taking the absolute value.

4. A method for detecting foreign matter in a coal mine belt conveyor according to claim 3, characterized in that: The specific method of obtaining the color value of each color channel of each pixel in each connected domain of the coal mine grayscale image includes: For the coal mine grayscale image The first pixels, get the The pixel point corresponds to the pixel point in the coal mine RGB image and is recorded as The target pixel of pixels is The color values ​​of the R, G, and B channels of the target pixel are taken as the The color value of each color channel of each pixel.

5. A method for detecting foreign matter in a coal mine belt conveyor according to claim 1, characterized in that: The specific formula for obtaining the theoretical centroid position of the degree of chaos in each connected domain according to the degree of local chaos in the neighborhood of each pixel point in each connected domain is: In the formula, Represents the grayscale image of the coal mine The horizontal coordinate of the theoretical centroid position of the degree of chaos in a connected domain; Represents the grayscale image of the coal mine The ordinate of the theoretical centroid position of the degree of chaos in a connected domain; Represents the grayscale image of the coal mine The total number of all pixels in a connected domain; Represents the grayscale image of the coal mine The first The horizontal coordinate of the pixel point; Represents the grayscale image of the coal mine The first The vertical coordinate of the pixel; Represents the grayscale image of the coal mine The first The degree of local disorder within the neighborhood of a pixel; Indicates rounding up.

6. A method for detecting foreign matter in a coal mine belt conveyor according to claim 1, characterized in that: The specific formula for obtaining the probability of the existence of metal minerals in each connected domain based on the position difference between the pixel point in each connected domain and the theoretical centroid position of the disorder degree in each connected domain is: In the formula, Represents the grayscale image of the coal mine The probability of the existence of metal minerals in a connected domain; Represents the grayscale image of the coal mine The total number of all pixels in a connected domain; Represents the grayscale image of the coal mine The first The degree of local disorder within the neighborhood of a pixel; Represents the grayscale image of the coal mine The first pixels and The Euclidean distance between the theoretical centroid positions of the degree of disorder in a connected domain; Represents an exponential function with a natural constant as base.

7. A method for detecting foreign matter in a coal mine belt conveyor according to claim 1, characterized in that: The method of obtaining the enhanced color value of the first color channel of each pixel in the coal mine RGB image according to the probability of the existence of metal minerals in each connected domain includes the following specific methods: Get the The probability that a metal mineral exists in the connected domain to which the pixel belongs; The color channel with the largest color value of the pixel is recorded as The first color channel of the pixel point, then the first The calculation method of the color value of the first color channel of a pixel after enhancement is: In the formula, Indicates the first The color value of the first color channel of each pixel after enhancement; Indicates the first The color value of the first color channel of a pixel; Indicates the first The probability that a metal mineral exists in the connected domain to which the pixel belongs.

8. A method for detecting foreign matter in a coal mine belt conveyor according to claim 7, characterized in that: The acquisition The probability of the existence of metal minerals in the connected domain to which the pixel belongs includes the following specific methods: Get the The pixel point corresponds to the pixel point in the coal mine grayscale image and is recorded as The first pixel of the pixels The connected domain where the first pixel of the pixels is located is recorded as the first connected domain, and the probability of the existence of metal minerals in the first connected domain is taken as the The probability that a metal mineral exists in the connected domain to which the pixel belongs.

Images