A security control method and system based on artificial intelligence

Abstract

The present invention belongs to the field of safety management and control technology, and specifically relates to a safety management and control method and system based on artificial intelligence, the method comprising: for an image in a coal mine environment video, performing histogram equalization on all focus points to obtain an overall equalization result of each gray value; setting a mapping range according to the overall equalization result of the minimum and maximum gray values, performing histogram equalization on all pixels in the influence area to which the focus point belongs, and obtaining a local equalization result of each gray value; using the overall equalization result of the gray value of the focus point and the local equalization result of the gray value of other pixels as the enhancement result of the focus point and other pixels to form an enhancement result of the influence area, and using the enhancement results of the influence area to which all focus points belong to form an enhanced image, so as to facilitate safety management and control of the coal mine underground environment. The present invention improves the image quality of the image, thereby improving the accuracy of subsequent analysis of the coal mine environment video.

Description

Technical Field

[0001] The present invention relates to the field of security control technology. More specifically, the present invention relates to a security control method and system based on artificial intelligence. Background Art

[0002] Coal mine safety production is the primary task of coal mine enterprise safety management; with the country's increasing attention to coal mine safety production work, coal mine safety management and control systems based on sensor technology, electronic technology and computer technology have been applied in coal mine enterprises, which helps to timely alarm abnormal coal mine environment, thereby facilitating staff to take countermeasures, which can not only reduce property losses, but also help reduce the occurrence of safety accidents.

[0003] In coal mines, due to poor lighting conditions, high dust and humidity, the collected coal mine environment videos often have problems such as insufficient brightness, low contrast and blurred details, which seriously affect the image quality of the coal mine environment video and the accuracy of subsequent analysis. In order to improve this situation, it is necessary to perform histogram equalization on the images in the coal mine environment video.

[0004] In the related technology, for example, the Chinese patent document with the authorization announcement number CN107176432B discloses a machine vision-based anchor foreign body and belt tear detection system, including a belt conveyor, an industrial camera installed above the belt conveyor for collecting video images of the belt conveyor in no-load and load conditions, a video image transmission module, and a belt conveyor foreign body and belt tear monitoring platform, the platform uses bilateral filtering and histogram equalization to perform image filtering and image enhancement preprocessing respectively; uses iterative adaptive threshold segmentation method to perform image binarization processing; and completes target detection and recognition based on connected components; the patent document integrates machine vision technology, computer technology, image processing and recognition technology, and can automatically complete the detection of anchor foreign bodies and belt tears in the video monitoring image of the belt conveyor, providing necessary technical support for the safety monitoring of belt conveyors in coal mines, which is of great significance for ensuring the safe operation of belt conveyors.

[0005] Histogram equalization is based on the histogram of the entire image, which will over-enhance the contrast of some areas and ignore the details of other areas. In order to overcome the shortcomings of histogram equalization, local histogram equalization is proposed. Local histogram equalization divides the image into multiple small areas and performs histogram equalization on each area separately. It can better preserve the local details of the image and avoid the problem of over-enhancement that may occur in global equalization. However, the local area division operation of local histogram equalization will cause block effects in the enhanced image, that is, the boundaries between different areas are obvious, which in turn affects the accuracy of subsequent analysis of coal mine environment videos. Summary of the invention

[0006] In order to solve the technical problem that the quality of the enhanced image obtained by performing histogram equalization on the image of the coal mine environment video is poor, which affects the accuracy of subsequent analysis of the coal mine environment video, the present invention provides solutions in the following aspects.

[0007] In a first aspect, the present invention provides a safety management and control method based on artificial intelligence, including: collecting coal mine environment videos, performing histogram equalization on images in the coal mine environment videos, and obtaining enhanced images to facilitate safety management and control of the coal mine underground environment, including: for any pixel point, taking the pixel point as the center pixel point, obtaining multiple local areas of the center pixel point, and the sizes of the multiple local areas are different; according to the characteristic values ​​of the multiple local areas, calculating the attention degree of the center pixel point; sorting the attention degrees of all the obtained pixel points in order from large to small, and placing the sorted pixels in front and back The pixel point of Perform histogram equalization on all focus points as the mapping range to obtain the overall equalization result of each grayscale value; divide the image into regions according to the coordinates and attention of the focus points to obtain the influence area of ​​each focus point; in the influence area of ​​the focus point, As the mapping range to the gray value in The pixels within the range are histogram-equalized. As the mapping range to the gray value in The pixel points within the range are histogram equalized to obtain the local equalization results of each gray value, where: The minimum gray value in the affected area to which the focus point belongs and the maximum gray value The overall equalization result is is the gray value of the point of interest The overall equalization result of the grayscale value of the focus point is taken as the enhancement result of the focus point, and the local equalization result of the grayscale values ​​of other pixels except the focus point is taken as the enhancement result of other pixels, so as to obtain the enhancement result of the influence area to which the focus point belongs; the enhancement results of the influence areas to which all the focus points belong are combined into an enhanced image.

[0008] In the present invention, the grayscale value of the focus point is the representative grayscale value of the influence area to which the focus point belongs. Therefore, by performing histogram equalization on all focus points with a mapping range of [0, 255], an overall equalization result of each grayscale value is obtained, which is equivalent to making adjustments based on the representative grayscale value in the entire image, and can enhance the overall contrast of the image, improve the overall visual effect of the image, and make the image look clearer and more natural. The present invention determines the reference grayscale value of the enhancement result of the influence area to which each focus point belongs through the overall equalization result of the grayscale value of the focus point, and subsequently, when performing local histogram equalization on the influence area to which each focus point belongs, the reference grayscale value of the enhancement result of the focus point is determined according to the grayscale value of the focus point. The mapping range is set by the overall equalization result of the grayscale value, the overall equalization result of the minimum grayscale value and the maximum grayscale value, and histogram equalization is performed on all pixels in the affected area to obtain the local equalization result of each grayscale value. Finally, the overall equalization result of the grayscale value of the focus point is used as the enhancement result of the focus point, and the local equalization result of the grayscale values ​​of other pixels except the focus point is used as the enhancement result of other pixels. While better retaining the local details of the image, it will not cause obvious boundaries between different areas, avoid block effects in the enhanced image, improve the quality of the enhanced image, and thereby improve the accuracy of subsequent analysis of the coal mine environment video.

[0009] Preferably, the obtaining of multiple local areas of the central pixel point comprises: the local area of ​​the central pixel point is equal to The square area of is the side length of the local region, and The value of Singular in range, is the preset value, and the total number of central pixels is obtained local areas of different sizes.

[0010] Preferably, the method for obtaining the characteristic value of the local area is: the size of the central pixel is The eigenvalue of the local area The calculation formula is: ; In the formula, The size of the center pixel is In the local area The gray value of a pixel, is the gray value of the center pixel, The size of the center pixel is In the local area The distance between the gray value of each pixel and the center pixel.

[0011] Preferably, the calculation of the attention degree of the central pixel point includes: for multiple local areas of the central pixel point and the characteristic values ​​of each local area, the size and characteristic value of the local area are used as the horizontal axis and the vertical axis respectively to construct a rectangular coordinate system; according to the characteristic values ​​of local areas of different sizes, the data points corresponding to the local areas of different sizes in the rectangular coordinate system are obtained; the slope of each data point is calculated; the local area corresponding to the data point with the largest slope is obtained, and used as the target local area of ​​the central pixel point; according to the size and characteristic value of the target local area of ​​the central pixel point, the attention degree of the central pixel point is calculated.

[0012] Preferably, the calculation formula for the attention degree of the central pixel is: ; In the formula, is the attention of the central pixel, is the size of the target local area at the center pixel, is the preset value, is the characteristic value of the target local area of ​​the central pixel, is a natural exponential function.

[0013] Preferably, the obtaining of the influence area to which each focus point belongs comprises: dividing the image into a plurality of polygons by the Thiessen polygon method according to the coordinates and the attention degree of the focus point, and each obtained polygon contains only one focus point, and taking the polygon containing each focus point as the influence area to which each focus point belongs; the polygon is composed of a vertical dividing line of the connecting line segment of each two adjacent focus points; the vertical dividing line is perpendicular to the connecting line segment of the two adjacent focus points and passes through the dividing point on the connecting line segment; the dividing point is obtained by: when the two adjacent focus points are Focus and When there are points of interest, the split point to The distance of the focus point With the split point to The distance of the focus point Ratio ; , Respectively Focus and The attention of a focus point.

[0014] Preferably, the obtaining of the overall equalization result of each gray value includes: gray value The calculation formula for the overall equalization result is: ; In the formula, Gray value The overall equalization result is Gray value The corrected frequency, To round up.

[0015] The grayscale value of the focus point is the representative grayscale value of the influence area to which the focus point belongs. The present invention performs histogram equalization on all focus points by taking [0, 255] as the mapping range to obtain the overall equalization result of each grayscale value. By making adjustments based on the representative grayscale value in the entire image, the overall contrast of the image can be enhanced, the overall visual effect of the image can be improved, and the image can look clearer and more natural.

[0016] Preferably, the gray value Corrected frequency The calculation method is: equal the gray value to The sum of the attention of all attention points is taken as the gray value The frequency correction coefficient of The ratio of the number of all focus points to the total number of all focus points in the image is taken as the grayscale value The overall frequency of Frequency correction coefficient to gray value The overall frequency is corrected to obtain the gray value After the correction frequency, the gray value Corrected frequency Satisfies the expression: ; In the formula, Gray value The overall frequency, , Gray value and grayscale value Frequency correction factor, gray value and grayscale value The value range of .

[0017] Preferably, the obtaining of the local equalization result of each gray value comprises: Grayscale value within the range , The local equalization result The calculation formula is: , is the gray value of the point of interest The overall equalization result is is the minimum gray value The overall equalization result is , Gray value and grayscale value The local frequency, gray value The value range is , gray value The value range is , is rounded up; for Grayscale value within the range , The local equalization result The calculation formula is: , is the maximum gray value The overall equalization result is , Gray value and grayscale value The local frequency, gray value The value range is , gray value The value range is .

[0018] The present invention determines the baseline grayscale value of the enhancement result of the influence area to which each focus point belongs through the overall equalization result of the grayscale value of the focus point. While better retaining the local details of the image, it does not cause obvious boundaries between different areas, avoids the block effect in the enhanced image, improves the quality of the enhanced image, and further improves the accuracy of subsequent analysis of the coal mine environment video.

[0019] In a second aspect, the present invention provides an artificial intelligence-based security management and control system, comprising a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned artificial intelligence-based security management and control method is implemented.

[0020] By adopting the above technical solution, the above-mentioned artificial intelligence-based security management method is generated into a computer program and stored in a memory to be loaded and executed by a processor, so that a terminal device is made according to the memory and the processor for easy use.

[0021] The beneficial effects of the present invention are:

[0022] In the present invention, the grayscale value of the focus point is the representative grayscale value of the influence area to which the focus point belongs. Therefore, by performing histogram equalization on all focus points with a mapping range of [0, 255], an overall equalization result of each grayscale value is obtained, which is equivalent to making adjustments based on the representative grayscale value in the entire image, and can enhance the overall contrast of the image, improve the overall visual effect of the image, and make the image look clearer and more natural. The present invention determines the reference grayscale value of the enhancement result of the influence area to which each focus point belongs through the overall equalization result of the grayscale value of the focus point, and subsequently, when performing local histogram equalization on the influence area to which each focus point belongs, the reference grayscale value of the enhancement result of the focus point is determined according to the grayscale value of the focus point. The mapping range is set by the overall equalization result of the grayscale value, the overall equalization result of the minimum grayscale value and the maximum grayscale value, and histogram equalization is performed on all pixels in the affected area to obtain the local equalization result of each grayscale value. Finally, the overall equalization result of the grayscale value of the focus point is used as the enhancement result of the focus point, and the local equalization result of the grayscale values ​​of other pixels except the focus point is used as the enhancement result of other pixels. While better retaining the local details of the image, it will not cause obvious boundaries between different areas, avoid block effects in the enhanced image, improve the quality of the enhanced image, and thereby improve the accuracy of subsequent analysis of the coal mine environment video. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] 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:

[0024] Figure 1 is a flow chart schematically illustrating a security control method based on artificial intelligence in the present invention;

[0025] Figure 2 is a flowchart schematically showing step S2. DETAILED DESCRIPTION

[0026] 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.

[0027] The specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0028] The embodiment of the present invention discloses a security management method based on artificial intelligence, referring to Figure 1, comprising steps S1 to S3:

[0029] S1. Collect videos of coal mine environment.

[0030] It should be noted that the camera is a key device in the industrial television system to carry out real-time image monitoring of the main production links, important equipment and key positions, and outputs the collected video images in the form of wired network signals, electrical signals, optical signals or wireless signals.

[0031] Specifically, the coal mine environment video underground in the coal mine is collected by the camera of the coal mine monitoring system for safe management and control of the underground environment of the coal mine. The duration of the coal mine environment video is equal to 1 minute. The camera outputs the collected video signal in the form of an electrical signal or an optical signal, and can observe the conditions of various monitoring positions underground in the coal mine in real time, which is convenient for the staff to safely manage and control the coal mine.

[0032] Among them, the coal mine monitoring system is divided into a ground part and an underground part; the ground part is composed of monitoring equipment, storage equipment, and information transmission interface, and the underground part is composed of substations, cameras, and power supplies. The ground part and the underground part are connected and communicated through cables or optical cables; for the ground part, the input end of the monitoring equipment is connected to the underground part through cables and information transmission interfaces, and the output end of the monitoring equipment is connected to a storage device; for the underground part, the substations are connected in series through cables or optical cables, and the substations are connected to multiple cameras through cables or optical cables, and each substation and camera are connected to a power supply;

[0033] The coal mine monitoring system has the functions of collecting, transmitting and displaying environmental videos in coal mines. It mainly collects coal mine environmental videos through cameras in the underground part, transmits the coal mine environmental videos to the ground part through substations, and displays the coal mine environmental videos through monitoring equipment in the ground part and stores them in storage devices.

[0034] S2. Perform histogram equalization on the image in the coal mine environment video to obtain an enhanced image.

[0035] It should be noted that, in coal mines, due to poor lighting conditions, high dust, high humidity and other factors, the collected coal mine environment videos often have problems such as insufficient brightness, low contrast, and blurred details, which seriously affect the image quality of the coal mine environment video and the accuracy of subsequent analysis; in order to improve this situation, it is necessary to perform histogram equalization on the images in the coal mine environment video; histogram equalization is a method of enhancing the image contrast by adjusting the grayscale distribution of the image so that the histogram of the image is evenly distributed. It can effectively improve the brightness and contrast of the image, make the details clearer, and facilitate safety risk assessment for staff or automation systems; through this processing, not only the visual effect of the image is improved, but also more reliable data support is provided for coal mine safety monitoring, which helps to timely discover potential safety hazards.

[0036] It should be further explained that histogram equalization is an image enhancement method. Since it is adjusted based on the histogram of the entire image, it can enhance the overall contrast of the image, improve the overall visual effect of the image, and make the image look clearer and more natural. However, since it is adjusted based on the histogram of the entire image, it will excessively enhance the contrast of certain areas and ignore the details of other areas. In order to overcome the shortcomings of histogram equalization, local histogram equalization is proposed. Local histogram equalization divides the image into multiple small areas and performs histogram equalization on each area separately. It can better retain the local details of the image and avoid the problem of over-enhancement that may occur in global equalization. However, the local area division operation of local histogram equalization will cause block effects in the enhanced image, that is, the boundaries between different areas are obvious, which in turn affects the accuracy of subsequent analysis of coal mine environment videos.

[0037] In summary, the present invention performs histogram equalization on the image in the coal mine environment video to obtain an enhanced image. In this process, [0, 255] is used as the mapping range to perform histogram equalization on all focus points to obtain the overall equalization result of each grayscale value. In the influence area to which each focus point belongs, the mapping range is set according to the overall equalization result of the minimum grayscale value and the maximum grayscale value, and the histogram equalization is performed on all pixels in the influence area to obtain the local equalization result of each grayscale value; the overall equalization result of the grayscale value of the focus point and the local equalization result of the grayscale value of other pixels are respectively used as the enhancement results of the focus point and other pixels, so as to obtain the enhancement result of the influence area to which the focus point belongs, and the enhancement results of the influence areas to which all focus points belong are used to form an enhanced image.

[0038] See the flowchart of step S2. Figure 2 , including steps S201 to S205, specifically:

[0039] S201. Calculate the attention degree of the pixel point according to the feature value of the local area of ​​the pixel point.

[0040] Specifically, for any pixel point in the image, taking the pixel point as the central pixel point, multiple local areas of the central pixel point are obtained, and the sizes of the multiple local areas are different; the local area is a square area, is the side length of the local region, then the size of the local region is ,and The value of Singular in range, is the preset value, so the total number of central pixels is local areas of different sizes.

[0041] The preset value The specific value of can be set according to the actual application scenario and requirements. Set to 11.

[0042] In one embodiment, the variance of the grayscale values ​​of all pixels in the local areas of different sizes of the central pixel is calculated as the feature value of the local areas of different sizes of the central pixel.

[0043] In another embodiment, the calculation formula of the characteristic values ​​of local areas of different sizes of the central pixel is:

[0044] ;

[0045] In the formula, The size of the center pixel is The eigenvalues ​​of the local area, The size of the center pixel is In the local area The gray value of a pixel, is the gray value of the center pixel, The size of the center pixel is In the local area The distance between the gray value of each pixel and the center pixel.

[0046] Furthermore, for multiple local areas of the central pixel and the eigenvalues ​​of each local area, a rectangular coordinate system is constructed with the size and eigenvalue of the local area as the horizontal axis and the vertical axis respectively; according to the eigenvalues ​​of local areas of different sizes, the corresponding data points of local areas of different sizes in the rectangular coordinate system are obtained; the slope of each data point is calculated; the local area corresponding to the data point with the largest slope is obtained, and used as the target local area of ​​the central pixel.

[0047] Furthermore, according to the size and characteristic value of the target local area of ​​the central pixel, the attention degree of the central pixel is calculated. The specific calculation formula is:

[0048] ;

[0049] In the formula, is the attention of the central pixel, is the size of the target local area at the center pixel, is the preset value, is the characteristic value of the target local area of ​​the central pixel, is a natural exponential function.

[0050] It should be noted that due to the size of the target local area of ​​the central pixel ,therefore ; Due to the characteristic value of the target local area of ​​the central pixel ,therefore In summary .

[0051] It should be further explained that the eigenvalue of the local area reflects the difference of the pixels in the local area. The larger the eigenvalue of the local area, the larger the difference of the pixels in the local area, and the smaller the eigenvalue of the local area, the smaller the difference of the pixels in the local area. Therefore, in the rectangular coordinate system constructed by the size and eigenvalue of the local area, the larger the slope of the data point, the greater the eigenvalue of the next local area of ​​the local area corresponding to the data point suddenly becomes larger than the eigenvalue of the local area corresponding to the data point, that is, the difference of the pixels in the local area suddenly becomes larger. Therefore, for the data point with the largest slope, it shows that the difference of the pixels in the local area corresponding to the data point is relatively stable, but the difference of the pixels in the next local area changes dramatically. Therefore, the local area corresponding to the data point with the largest slope is the local similar area with the largest area of ​​the central pixel, and the larger the size of the local similar area with the largest area, that is, the larger the size of the target local area of ​​the central pixel, the greater the attention of the central pixel. When the eigenvalue of the target local area of ​​the central pixel is smaller, it means that the difference of the pixels in the target local area of ​​the central pixel is smaller, and the central pixel is more likely to be located at the center of a local area. Therefore, the attention of the central pixel is greater.

[0052] S202, sorting all pixels in descending order of attention degree, taking the top 5% of pixels as attention points, dividing the image into regions according to the coordinates and attention degrees of the attention points, and obtaining the influence region to which each attention point belongs.

[0053] Specifically, all pixels are sorted in descending order of attention, and the top 5% of pixels after sorting are taken as attention points.

[0054] Furthermore, according to the coordinates and attention degree of the attention points, the image is divided into multiple polygons by using the Thiessen polygon method, and each obtained polygon contains only one attention point, and the polygon containing each attention point is used as the influence area to which each attention point belongs.

[0055] The polygon is composed of vertical dividing lines of the connecting line segments of every two adjacent attention points; the vertical dividing line is perpendicular to the connecting line segments of the two adjacent attention points and passes through the dividing point on the connecting line segments; the dividing point is obtained by: when the two adjacent attention points are Focus and When there are points of interest, the split point to The distance of the focus point With the split point to The distance of the focus point Ratio ; , Respectively Focus and The attention of a focus point.

[0056] S203, Histogram equalization is performed on all focus points as the mapping range to obtain the overall equalization result of each grayscale value.

[0057] Specifically, all pixels are sorted in descending order of attention, and the top 5% of pixels after sorting are taken as attention points.

[0058] Further, the gray value is equal to The sum of the attention of all attention points is taken as the gray value The frequency correction coefficient of The ratio of the number of all focus points to the total number of all focus points in the image is taken as the grayscale value The overall frequency of Frequency correction coefficient to gray value The overall frequency is corrected to obtain the gray value After the correction frequency, the gray value Corrected frequency Satisfies the expression:

[0059] ;

[0060] In the formula, Gray value The corrected frequency, Gray value The overall frequency, , Gray value and grayscale value Frequency correction factor, gray value and grayscale value The value range of .

[0061] Further, according to all gray values ​​and the frequencies corrected by all gray values, As the mapping range, all the focus points are histogram-equalized to obtain the overall equalization result of each gray value. The calculation formula for the overall equalization result is:

[0062] ;

[0063] In the formula, Gray value The overall equalization result is Gray value The corrected frequency, is the overall minimum gray value, To round up.

[0064] It should be noted that the grayscale value of the focus point is the representative grayscale value of the influence area to which the focus point belongs. The present invention performs histogram equalization on all focus points by taking [0, 255] as the mapping range to obtain the overall equalization result of each grayscale value. By making adjustments based on the representative grayscale values ​​in the entire image, the overall contrast of the image can be enhanced, the overall visual effect of the image can be improved, and the image can look clearer and more natural.

[0065] S204: In the influence area to which each focus point belongs, a mapping range is set according to the overall equalization result of the minimum grayscale value and the maximum grayscale value, and histogram equalization is performed on all pixels in the influence area to obtain a local equalization result of each grayscale value.

[0066] Among them, the gray value of the focus point is recorded as , the gray value of the focus point The overall equalization result is recorded as ; Get the minimum grayscale value and the maximum grayscale value in the affected area of ​​the focus point, and record them as and , the minimum gray value and the maximum gray value The overall equalization results are recorded as , .

[0067] Specifically, in the affected area to which the focus point belongs, the gray value is equal to The ratio of the number of all pixels in the affected area to the total number of all pixels in the affected area is taken as the grayscale value. The local frequency of The value range is .

[0068] Furthermore, in the influence area to which the focus point belongs, As the mapping range to the gray value in The pixels within the range are histogram-equalized. As the mapping range to the gray value in The pixel points within the range are histogram equalized to obtain the local equalization results of each gray value.

[0069] Among them, for Grayscale value within the range , gray value The local equalization result The calculation formula is:

[0070] ;

[0071] In the formula, Gray value The local equalization result is is the gray value of the point of interest The overall equalization result is is the minimum gray value The overall equalization result is , Gray value and grayscale value The local frequency, gray value The value range is , gray value The value range is , To round up.

[0072] Among them, for Grayscale value within the range , gray value The local equalization result The calculation formula is:

[0073] ;

[0074] In the formula, Gray value The local equalization result is is the maximum gray value The overall equalization result is is the gray value of the point of interest The overall equalization result is , Gray value and grayscale value The local frequency, gray value The value range is , gray value The value range is , To round up.

[0075] It should be noted that the present invention determines the reference grayscale value of the enhancement result of the influence area to which each focus point belongs through the overall equalization result of the grayscale value of the focus point. When the influence area to which each focus point belongs is subsequently subjected to local histogram equalization, the mapping range is set according to the overall equalization result of the grayscale value of the focus point, the minimum grayscale value and the maximum grayscale value. The histogram equalization is performed on all pixels in the influence area to obtain the local equalization result of each grayscale value. Finally, the overall equalization result of the grayscale value of the focus point is used as the enhancement result of the focus point, and the local equalization result of the grayscale value of other pixel points except the focus point is used as the enhancement result of other pixel points. While better retaining the local details of the image, it will not cause obvious boundaries between different areas, avoid the block effect in the enhanced image, improve the quality of the enhanced image, and thereby improve the accuracy of subsequent analysis of the coal mine environment video.

[0076] S205. The overall equalization result of the grayscale value of the focus point and the local equalization result of the grayscale values ​​of other pixels are respectively used as the enhancement results of the focus point and other pixels to form the enhancement result of the influence area to which the focus point belongs. The enhancement results of the influence areas to which all the focus points belong are combined into an enhanced image.

[0077] Specifically, for the influence area to which any focus point belongs, the overall equalization result of the grayscale value of the focus point is used as the enhancement result of the focus point, and the local equalization result of the grayscale values ​​of other pixel points except the focus point is used as the enhancement result of other pixel points, so as to obtain the enhancement result of the influence area to which the focus point belongs.

[0078] Furthermore, the enhancement results of the affected areas to which all focus points belong are combined to form an enhanced image.

[0079] It should be noted that, in the present invention, the grayscale value of the focus point is the representative grayscale value of the influence area to which the focus point belongs. Therefore, by performing histogram equalization on all focus points as the mapping range [0, 255], the overall equalization result of each grayscale value is obtained, which is equivalent to adjusting based on the representative grayscale value in the entire image, which can enhance the overall contrast of the image, improve the overall visual effect of the image, and make the image look clearer and more natural. The present invention determines the reference grayscale value of the enhancement result of the influence area to which each focus point belongs through the overall equalization result of the grayscale value of the focus point, and subsequently performs local histogram equalization on the influence area to which each focus point belongs according to the focus point. The mapping range is set by the overall equalization result of the grayscale value, the overall equalization result of the minimum grayscale value and the maximum grayscale value, and histogram equalization is performed on all pixels in the affected area to obtain the local equalization result of each grayscale value. Finally, the overall equalization result of the grayscale value of the focus point is used as the enhancement result of the focus point, and the local equalization result of the grayscale values ​​of other pixels except the focus point is used as the enhancement result of other pixels. While better retaining the local details of the image, it will not cause obvious boundaries between different areas, avoid block effects in the enhanced image, improve the quality of the enhanced image, and then improve the accuracy of subsequent analysis of the coal mine environment video.

[0080] S3. Through enhanced images, safety control of the underground environment of coal mines is carried out.

[0081] Specifically, all enhanced images are combined into an enhanced coal mine environment video, and the enhanced coal mine environment video is input into the trained neural network to determine the probability of safety risks in the underground environment of the coal mine. When the probability of safety risks in the underground environment of the coal mine is greater than a preset threshold, it indicates that there are safety hazards in the underground environment of the coal mine, and the staff is notified to conduct inspections and investigations.

[0082] The specific value of the preset threshold can be set according to actual application scenarios and requirements. The present invention sets the preset threshold to 0.8.

[0083] The training process of the neural network includes: performing histogram equalization on the historical images in the historical coal mine environment video according to the method of step S2 to obtain enhanced historical images, and composing all the enhanced historical images into an enhanced historical coal mine environment video; using the historical coal mine environment video as a training sample, labeling the training sample, wherein the label is a numerical value in the range of [0,1], which is used to indicate the probability that there is a safety risk in the coal mine underground environment corresponding to the training sample; and training the neural network through all the training samples and their labels, wherein the neural network adopts an RNN network structure.

[0084] An embodiment of the present invention further discloses an artificial intelligence-based security management and control 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 artificial intelligence-based security management and control method according to the present invention is implemented.

[0085] 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.

[0086] In the description of this specification, "plurality" or "several" means at least two, such as two, three or more, etc., unless otherwise clearly and specifically defined.

[0087] 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, alternatives to the embodiments of the present invention described herein may be employed.

Claims

1. A security management and control method based on artificial intelligence, characterized in that: include: Collect coal mine environment videos, perform histogram equalization on the images in the coal mine environment videos, and obtain enhanced images to facilitate safety management and control of the coal mine underground environment, including: For any pixel point, take the pixel point as the central pixel point, obtain multiple local areas of the central pixel point, and the sizes of the multiple local areas are different; calculate the attention degree of the central pixel point according to the characteristic values ​​of the multiple local areas; sort the attention degrees of all the obtained pixel points in order from large to small, and sort the ordered attention degrees. The pixel point of Perform histogram equalization on all focus points as the mapping range to obtain the overall equalization result of each grayscale value; According to the coordinates and attention degree of the focus point, the image is divided into regions to obtain the influence area of ​​each focus point; in the influence area of ​​the focus point, As the mapping range to the gray value in The pixels within the range are histogram-equalized. As the mapping range to the gray value in The pixel points within the range are histogram equalized to obtain the local equalization results of each gray value, where: The minimum gray value in the affected area to which the focus point belongs and the maximum gray value The overall equalization result is is the gray value of the point of interest The overall equalization result; The overall equalization result of the grayscale value of the focus point is used as the enhancement result of the focus point, and the local equalization result of the grayscale values ​​of other pixels except the focus point is used as the enhancement result of other pixels, so as to obtain the enhancement result of the influence area to which the focus point belongs; The enhancement results of the affected areas of all focus points are combined to form an enhanced image.

2. According to the security management method based on artificial intelligence in claim 1, it is characterized in that: The step of obtaining multiple local areas of a central pixel point includes: The local area of ​​the central pixel is equal to The square area of is the side length of the local region, and The value of Singular in range, is the preset value, and the total number of central pixels is obtained local areas of different sizes.

3. According to the security management method based on artificial intelligence in claim 1, it is characterized in that: The method for obtaining the characteristic value of the local area is: The size of the center pixel is The eigenvalue of the local area The calculation formula is: ； In the formula, The size of the center pixel is In the local area The gray value of a pixel, is the gray value of the center pixel, The size of the center pixel is In the local area The distance between the gray value of each pixel and the center pixel.

4. According to the security management method based on artificial intelligence in claim 1, it is characterized in that: The calculating the attention degree of the central pixel point includes: For multiple local areas of the central pixel and the characteristic values ​​of each local area, the size and characteristic value of the local area are used as the horizontal axis and the vertical axis respectively to construct a rectangular coordinate system; according to the characteristic values ​​of local areas of different sizes, the data points corresponding to the local areas of different sizes in the rectangular coordinate system are obtained; the slope of each data point is calculated; the local area corresponding to the data point with the largest slope is obtained and used as the target local area of ​​the central pixel; According to the size and feature value of the target local area of ​​the central pixel, the attention degree of the central pixel is calculated.

5. The artificial intelligence-based security management method according to claim 4 is characterized in that: The calculation formula of the attention degree of the central pixel is: ； In the formula, is the attention of the central pixel, is the size of the target local area at the center pixel, is the preset value, is the characteristic value of the target local area of ​​the central pixel, is a natural exponential function.

6. The artificial intelligence-based security management method according to claim 1, characterized in that: The obtaining of the influence area to which each focus point belongs includes: According to the coordinates and attention degree of the attention points, the image is divided into multiple polygons by using the Thiessen polygon method, and each obtained polygon contains only one attention point, and the polygon containing each attention point is regarded as the influence area of ​​each attention point; The polygon is composed of a vertical dividing line of a connecting line segment of each two adjacent attention points; the vertical dividing line is perpendicular to the connecting line segment of the two adjacent attention points and passes through a dividing point on the connecting line segment; the dividing point is obtained by: when the two adjacent attention points are Focus and When there are points of interest, the split point to The distance of the focus point With the split point to The distance of the focus point Ratio ; , Respectively Focus and The attention of a focus point.

7. The artificial intelligence-based security management method according to claim 1, characterized in that: The obtaining of the overall equalization result of each gray value includes: Gray value The calculation formula for the overall equalization result is: ； In the formula, Gray value The overall equalization result is Gray value The corrected frequency, To round up.

8. The artificial intelligence-based security management method according to claim 7, characterized in that: The gray value Corrected frequency The calculation method is: The gray value is equal to The sum of the attention of all attention points is taken as the gray value The frequency correction coefficient of The ratio of the number of all focus points to the total number of all focus points in the image is taken as the grayscale value The overall frequency of Frequency correction coefficient to gray value The overall frequency is corrected to obtain the gray value After the correction frequency, the gray value Corrected frequency Satisfies the expression: ； In the formula, Gray value The overall frequency, , Gray value and grayscale value Frequency correction factor, gray value and grayscale value The value range of .

9. The artificial intelligence-based security management method according to claim 1, characterized in that: The obtaining of the local equalization result of each gray value includes: for Grayscale value within the range , The local equalization result The calculation formula is: , is the gray value of the point of interest The overall equalization result is is the minimum gray value The overall equalization result is , Gray value and grayscale value The local frequency, gray value The value range is , gray value The value range is , To round up; for Grayscale value within the range , The local equalization result The calculation formula is: , is the maximum gray value The overall equalization result is , Gray value and grayscale value The local frequency, gray value The value range is , gray value The value range is .

10. A security management and control system based on artificial intelligence, 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, an artificial intelligence-based security management and control method according to any one of claims 1 to 9 is implemented.

Images