A method and system for intelligent monitoring of coal yards

Abstract

This invention provides an intelligent monitoring method and system suitable for coal yards, relating to the field of computer science. The method includes: enhancing a target frame image acquired via video image capture using a camera to obtain a first frame image; segmenting the first frame image based on morphological operations to obtain suspected motion regions and detecting them; if a suspected motion target exists, calling the detection information from the corresponding sensor and issuing an alarm message; if no suspected motion target exists, updating the background image and re-detecting for the presence of a motion target. By preprocessing the image acquired via the camera to improve image quality; performing motion target detection and region segmentation on the improved image to obtain suspected motion regions; detecting the presence of a motion target in the suspected motion regions and responding accordingly based on sensor detection information, intelligent monitoring is effectively achieved, thereby improving the safety early warning capabilities of the coal yard.

Description

Technical Field

[0001] This invention relates to the field of computer technology, and in particular to an intelligent monitoring method and system suitable for coal yards. Background Technology

[0002] Coal, often called "black gold," remains an indispensable energy source for human production and daily life for the foreseeable future. The supply of coal is crucial to the stability of my country's industrial development and all aspects of society, and its security is a vital element of my country's energy security.

[0003] As an important site for coal mining and storage, the safety of coal yards is of paramount importance. However, spontaneous combustion of coal, poisoning of personnel, or other losses often occur during the coal storage process. Currently, the general approach is to install sensors for monitoring. However, due to the fact that some coal yards do not meet the installation requirements, the safety assurance work is difficult to perfect and there are loopholes.

[0004] Therefore, the present invention provides an intelligent monitoring method and system for coal storage safety applicable to coal yards. Summary of the Invention

[0005] This invention provides an intelligent monitoring method and system suitable for coal yards, which improves image quality by preprocessing images acquired by cameras; detects moving targets and segments regions in the improved images to obtain suspected moving areas; detects whether moving targets exist in the suspected moving areas and makes relevant responses based on sensor detection information, effectively realizing intelligent monitoring and improving the safety early warning capability of coal yards.

[0006] This invention provides an intelligent monitoring method suitable for coal yards, comprising:

[0007] Step 1: Use a high-definition camera installed at a preset location to capture video images and obtain the target frame image;

[0008] Step 2: Enhance the target frame image to obtain the first frame image;

[0009] Step 3: Utilize video continuity to perform moving target detection on the first frame image to obtain the detected image;

[0010] Step 4: Analyze the detected image and use morphological operations to segment and obtain images of suspected motion regions;

[0011] Step 5: Detect the suspected motion area. If a suspected moving target is found, retrieve the detection information from the corresponding sensor and issue an alarm together. If no suspected moving target is found, update the background image and re-detect whether a moving target is found.

[0012] Preferably, the preset location refers to the optimal installation point for the camera at the indoor or outdoor coal storage pile.

[0013] Preferably, the target frame image is enhanced to obtain a first frame image, including:

[0014] Step 11: Convert the target frame image from RGB to YUV format;

[0015] Step 12: The Y component of the target frame image in YUV format is enhanced using weighted processing, and the U and V components are adaptively corrected.

[0016] Step 13: Convert the target frame image in YUV format after component adjustment to RGB space to obtain the first frame image.

[0017] Preferably, by utilizing video continuity, moving target detection is performed on the first frame image to obtain a detected image, including:

[0018] J i (x,y)=D i (x,y)-B i (x,y)

[0019] Among them, J i (x,y) represents the detection image; D i (x,y) represents the current i-th first frame image; B i (x,y) represents the background image;

[0020] Analyze and detect image J i The threshold Y is obtained by taking the mean and variance of the histogram corresponding to (x,y).

[0021] If the grayscale value of the detected image is greater than Y, then the detected image is determined to be a moving foreground image;

[0022] Otherwise, define it as a background image.

[0023] Preferably, the detected image is analyzed, and morphological operations are used to segment the suspected motion region image, including:

[0024] Step 21: Perform erosion and re-dilation on the detected image, and combine it with connected component analysis to obtain the first image;

[0025] Step 22: Obtain target feature information for different moving targets to be detected;

[0026] Step 23: Based on the obtained target feature information, filter and retain target pixels from the first image;

[0027] Step 24: Collect the target pixels to obtain the suspected motion region.

[0028] Preferably, based on the obtained target feature information, target pixels are selected and retained from the first image, including:

[0029] Based on different target feature information, the importance value of each first pixel in the first image is obtained, and the pixels are grouped according to different moving targets to obtain several first sets D = {d cj ,j=1,2,3,...,n}, where d cj Let represent the importance value of the j-th first pixel for the c-th type of moving target, and n represent the total number of first pixels corresponding to the c-th type of moving target;

[0030]

[0031] Where, N 1j Represented as the maximum value among the R, G, and B channels for the j-th first pixel; N 2j Represented as the minimum value of the j-th first pixel based on the R, G, and B channels; N 3j Let P be the average value of the j-th first pixel based on the R, G, and B channels; Δ1 represents the influence coefficient of the difference between the maximum and minimum values ​​of the corresponding first pixel based on the R, G, and B channels on the importance value of that first pixel; Δ2 represents the influence coefficient of the average value of the corresponding first pixel based on the R, G, and B channels on the importance value of that first pixel; c Let Δ3 represent the target feature value of the c-th type of moving target; Δ3 represents the comparison coefficient of the importance value of the corresponding target feature value to the first matched pixel, and Δ3>Δ2>Δ1; This is represented as the loss factor for calculating the importance value of the first pixel.

[0032] Determine the percentage of significant values ​​greater than a preset threshold in all first set D;

[0033] Based on the obtained percentage results, the first set whose percentage results are greater than the preset percentage threshold is selected as the target set;

[0034] Pixels corresponding to all significant values ​​greater than a preset threshold within the target set are extracted from the first image and retained as suspected motion regions.

[0035] Preferably, the suspected motion area is detected. If a suspected moving target is found, an effective response is made based on the alarm status of the corresponding sensor and the detection information. If no suspected moving target is found, the background image is updated, and the detection for a moving target is performed again, including:

[0036] Calculate the gradient of each pixel within the suspected motion region;

[0037] Divide the suspected motion area into image blocks of a preset size;

[0038] Define the image patch as a window, and compare the grayscale values ​​of the center pixel with those of the surrounding pixels.

[0039] If the surrounding pixel values ​​are all greater than the center pixel value, then the position of that pixel is marked as 1; otherwise, it is marked as 0. The generated binary numbers are arranged in sequence to obtain the binary numbers in the window.

[0040] The obtained binary number is converted into a decimal number and used as the first feature value of the center pixel of the window;

[0041] The frequency of occurrence of each first feature value is determined and normalized to obtain a statistical histogram;

[0042] The statistical histograms of each window are connected to obtain the first feature vector;

[0043] Calculate the gradient histogram for each image patch;

[0044] A preset number of image blocks are grouped into connected intervals, and the corresponding gradient histograms are normalized to obtain the gradient histogram within the interval.

[0045] By combining all the obtained gradient histograms within the intervals, we obtain the second feature vector;

[0046] The key feature information reflected by the first and second feature vectors is compared with the target feature information of the suspected moving target to obtain the similarity result;

[0047] If the similarity result is greater than the preset comparison threshold, it is determined that there is a suspected moving target in the suspected motion area. At this time, the relevant detection information of the corresponding sensor is called to transmit the data back and the alarm action is activated.

[0048] Otherwise, update the background image, and then extract the moving foreground based on the corresponding detection image to re-detect whether there is a moving target.

[0049] Preferably, an effective response is made based on the alarm status and detection information of the corresponding sensor, including:

[0050] Step 31: After determining that there is a suspected moving target in the suspected motion area, combine the preset location of the camera set and retrieve the key location information of the nearby position sensor to confirm the existence of the suspected moving target;

[0051] Step 32: Based on the morphological category of the suspected moving target, select the associated first sensor to obtain key data information and alarm information;

[0052] Step 33: Analyze the alarm information of the first sensor. If there are no historical alarm records before the preset time period, the key location information and key data information are sent back to the server and corresponding effective measures are taken in a timely manner.

[0053] This invention provides an intelligent monitoring system suitable for coal yards, comprising:

[0054] Image acquisition module: Uses a high-definition camera installed at a preset location to acquire video images and obtain the target frame image;

[0055] Image processing module: Enhances the target frame image to obtain the first frame image;

[0056] Target detection module: Utilizes video continuity to perform moving target detection on the first frame image to obtain a detected image;

[0057] Motion region segmentation module: Analyzes the detected image and uses morphological operations to segment and obtain images of suspected motion regions;

[0058] Alarm module: Detects the suspected motion area. If a suspected moving target is found, it calls the detection information of the corresponding sensor and issues an alarm together with the alarm information. If no suspected moving target is found, it updates the background image and re-detects whether a moving target is found.

[0059] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description, claims, and drawings.

[0060] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0061] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0062] Figure 1 This is a flowchart illustrating an intelligent monitoring method applicable to coal yards in an embodiment of the present invention;

[0063] Figure 2 This is a structural diagram of an intelligent monitoring system suitable for coal yards, as described in an embodiment of the present invention. Detailed Implementation

[0064] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0065] This invention provides an intelligent monitoring method suitable for coal yards, such as... Figure 1 As shown, it includes:

[0066] Step 1: Use a high-definition camera installed at a preset location to capture video images and obtain the target frame image;

[0067] Step 2: Enhance the target frame image to obtain the first frame image;

[0068] Step 3: Utilize video continuity to perform moving target detection on the first frame image to obtain the detected image;

[0069] Step 4: Analyze the detected image and use morphological operations to segment and obtain images of suspected motion regions;

[0070] Step 5: Detect the suspected motion area. If a suspected moving target is found, retrieve the detection information from the corresponding sensor and issue an alarm together. If no suspected moving target is found, update the background image and re-detect whether a moving target is found.

[0071] In this embodiment, the preset location refers to the optimal installation point for the camera at the indoor or outdoor coal storage pile; the target frame image is obtained by the camera.

[0072] In this embodiment, the enhancement process is used to improve the image acquisition quality and enhance the contrast and visibility of moving targets in the image. The moving targets include smoke, fire, and vehicles. The first frame image is obtained by enhancing the target frame image.

[0073] In this embodiment, moving target detection refers to extracting objects that have undergone displacement, deformation, or other movements from the background when the camera is stationary. The detection image is the image of the first frame image after moving target detection, which is mainly used to segment suspected moving regions using morphological operations. The morphological operations mainly include two operation methods: dilation and erosion.

[0074] In this embodiment, the suspected motion area refers to the area that is highly similar to the feature information of a suspected moving target. The suspected moving target is the moving target to be detected. The sensors include position sensors, temperature sensors, and gas sensors, etc. The relevant sensors can be selected according to the category of the suspected moving target, and the alarm information of the sensor and the uploaded detection information can be analyzed.

[0075] In this embodiment, the alarm information includes alarm time record and alarm reason; different sensors have different detection information. For example, the detection information of a position sensor includes position coordinate information; the detection information of a gas sensor includes gas name and gas concentration; and the detection information of a temperature sensor includes current temperature.

[0076] The beneficial effects of the above technical solution are as follows: image quality is improved by preprocessing the images acquired by the camera; suspected moving areas are obtained by moving target detection and region segmentation of the improved images; the presence of moving targets is detected in the suspected moving areas, and relevant responses are made in combination with the detection information of the sensors, which effectively realizes intelligent monitoring and improves the safety early warning capability of the coal yard.

[0077] This invention provides an intelligent monitoring method suitable for coal yards, which enhances the target frame image to obtain a first frame image, including:

[0078] Step 11: Convert the target frame image from RGB to YUV format;

[0079] Step 12: The Y component of the target frame image in YUV format is enhanced using weighted processing, and the U and V components are adaptively corrected.

[0080] Step 13: Convert the target frame image in YUV format after component adjustment to RGB space to obtain the first frame image.

[0081] In this embodiment, the Y component refers to the luminance component in the YUV color space. Weighting is used to emphasize the relative importance of the Y component, thereby enhancing it. The U and V components represent the chrominance components of the image. Adaptive correction is used to ensure that the image after Y component enhancement is not distorted.

[0082] The beneficial effects of the above technical solution are: by enhancing the Y component of the target frame image in the YUV space, adaptively correcting the U and V components and converting them to the RGB space to obtain the first frame image, the color of the target frame image is preserved, and the contrast and visibility of the moving target in the image are improved, which is more conducive to subsequent moving target detection.

[0083] This invention provides an intelligent monitoring method suitable for coal yards, which utilizes video continuity to perform moving target detection on the first frame image to obtain a detected image, including:

[0084] J i (x,y)=D i (x,y)-B i (x,y)

[0085] Among them, J i(x,y) represents the detection image; D i (x,y) represents the current i-th first frame image; B i (x,y) represents the background image;

[0086] Analyze and detect image J i The threshold Y is obtained by taking the mean and variance of the histogram corresponding to (x,y).

[0087] If the grayscale value of the detected image is greater than Y, then the detected image is determined to be a moving foreground image;

[0088] Otherwise, define it as a background image.

[0089] In this embodiment, the histogram reflects the gray-level distribution pattern in the image, and the image quality can be inferred from the shape of the histogram.

[0090] In this embodiment, the threshold Y is no greater than 35.

[0091] In this embodiment, for example, if the grayscale value of the detected image A is 55, which is greater than the threshold Y, then the detected image A is determined to be a moving foreground image.

[0092] The beneficial effects of the above technical solution are: by performing moving target detection on the first frame image, a detection image is obtained, which lays the foundation for subsequent segmentation of suspected moving regions, thereby achieving effective detection of moving targets.

[0093] This invention provides an intelligent monitoring method suitable for coal yards, which analyzes the detected image and uses morphological operations to segment and obtain images of suspected motion regions, including:

[0094] Step 21: Perform erosion and re-dilation on the detected image, and combine it with connected component analysis to obtain the first image;

[0095] Step 22: Obtain target feature information for different moving targets to be detected;

[0096] Step 23: Based on the obtained target feature information, filter and retain target pixels from the first image;

[0097] Step 24: Collect the target pixels to obtain the suspected motion region.

[0098] In this embodiment, erosion is used to shrink image boundaries and eliminate small, meaningless targets; dilation is used to fill certain gaps in the target area and eliminate small particle noise contained in the target area; connected component analysis refers to finding and marking adjacent pixels with the same pixel value in the image; target feature information includes shape, appearance, and texture information.

[0099] In this embodiment, the suspected motion region is composed of target pixels, which are selected and retained from the first image based on target feature information.

[0100] The beneficial effects of the above technical solution are: by performing morphological operations and connected component analysis on the detected image to obtain the first image; by filtering and retaining target pixels in the first image based on target feature information, the suspected moving region can be accurately obtained, laying the foundation for subsequent moving target detection.

[0101] This invention provides an intelligent monitoring method suitable for coal yards, which involves filtering and retaining target pixels from a first image based on obtained target feature information, including:

[0102] Based on different target feature information, the importance value of each first pixel in the first image is obtained, and the pixels are grouped according to different moving targets to obtain several first sets D = {d cj ,j=1,2,3,...,n}, where d cj Let represent the importance value of the j-th first pixel for the c-th type of moving target, and n represent the total number of first pixels corresponding to the c-th type of moving target;

[0103]

[0104] Where, N 1j Represented as the maximum value among the R, G, and B channels for the j-th first pixel; N 2j Represented as the minimum value of the j-th first pixel based on the R, G, and B channels; N 3j Let P be the average value of the j-th first pixel based on the R, G, and B channels; Δ1 represents the influence coefficient of the difference between the maximum and minimum values ​​of the corresponding first pixel based on the R, G, and B channels on the importance value of that first pixel; Δ2 represents the influence coefficient of the average value of the corresponding first pixel based on the R, G, and B channels on the importance value of that first pixel; c Let Δ3 represent the target feature value of the c-th type of moving target; Δ3 represents the comparison coefficient of the importance value of the corresponding target feature value to the first matched pixel, and Δ3>Δ2>Δ1; This is represented as the loss factor for calculating the importance value of the first pixel.

[0105] Determine the percentage of significant values ​​greater than a preset threshold in all first set D;

[0106] Based on the obtained percentage results, the first set whose percentage results are greater than the preset percentage threshold is selected as the target set;

[0107] Extract and retain all pixels in the target set that have important values ​​greater than a preset threshold from the first image.

[0108] In this embodiment, the value range of the importance value of the first pixel is (0, 1), and the preset threshold is set in advance, usually 0.6.

[0109] In this embodiment, for example, there are first sets D1, D2, and D3, whose corresponding important value percentages greater than a preset threshold are 55%, 60%, and 89%, respectively. Based on the preset percentage threshold of 70%, the first set D3 is used as the target set, and the pixels corresponding to the important values ​​greater than the preset threshold in the first set are extracted as target pixels and retained.

[0110] The beneficial effects of the above technical solution are as follows: by obtaining the importance value of each first pixel in different first images based on different target feature information, and a first set is obtained by grouping different moving targets; the proportion of importance values ​​in the first set is analyzed, and a target set is obtained from it; and all pixels in the target set whose importance values ​​are greater than a preset threshold are retained, so as to effectively obtain the suspected moving region.

[0111] This invention provides an intelligent monitoring method suitable for coal yards, which detects suspected movement areas. If a suspected moving target is found, an effective response is made based on the alarm status of the corresponding sensor and the detection information. If no suspected moving target is found, the background image is updated, and the detection of the presence of a moving target is repeated. The method includes:

[0112] Calculate the gradient of each pixel within the suspected motion region;

[0113] Divide the suspected motion area into image blocks of a preset size;

[0114] Define the image patch as a window, and compare the grayscale values ​​of the center pixel with those of the surrounding pixels.

[0115] If the surrounding pixel values ​​are all greater than the center pixel value, then the position of that pixel is marked as 1; otherwise, it is marked as 0. The generated binary numbers are arranged in sequence to obtain the binary numbers in the window.

[0116] The obtained binary number is converted into a decimal number and used as the first feature value of the center pixel of the window;

[0117] The frequency of occurrence of each first feature value is determined and normalized to obtain a statistical histogram;

[0118] The statistical histograms of each window are connected to obtain the first feature vector;

[0119] Calculate the gradient histogram for each image patch;

[0120] A preset number of image blocks are grouped into connected intervals, and then the gradient histograms corresponding to the image blocks are normalized to obtain the gradient histograms within the intervals.

[0121] By combining all the obtained gradient histograms within the intervals, we obtain the second feature vector;

[0122] The key feature information reflected by the first and second feature vectors is compared with the target feature information of the suspected moving target to obtain the similarity result;

[0123] If the similarity result is greater than the preset comparison threshold, it is determined that there is a suspected moving target in the suspected motion area. At this time, the relevant detection information of the corresponding sensor is called to transmit the data back and the alarm action is activated.

[0124] Otherwise, update the background image, and then extract the moving foreground based on the corresponding detection image to re-detect whether there is a moving target.

[0125] In this embodiment, gradient refers to a vector, including magnitude and direction. Calculating the gradient of each pixel is mainly to capture contour information and reduce the interference of lighting.

[0126] In this embodiment, for example, the grayscale value of the center pixel of window B1 is 80, and the grayscale values ​​of the surrounding adjacent pixels b1, b2, and b3 are 87, 92, and 50, respectively. At this time, the position labels of pixels b1, b2, and b3 are 1, 1, and 0, respectively.

[0127] In this embodiment, the first feature value can be used to reflect the texture information of the corresponding window; the statistical histogram can represent the local texture features of each window; the first feature vector is the texture feature vector of the entire suspected motion area, which is obtained by connecting all statistical histograms.

[0128] In this embodiment, the gradient histogram can represent the local contour features of each image block; the connected interval is composed of a preset number of image blocks; the gradient histogram within the interval is obtained by normalizing the gradient histogram corresponding to the image block, wherein the normalization process is to reduce the range of gradient intensity variation and further compress illumination, shadows and edges.

[0129] In this embodiment, the second feature vector refers to the shape and appearance features of the entire suspected motion area; the key feature information includes the texture feature information, appearance information and shape information of the suspected motion area.

[0130] In this embodiment, the preset comparison threshold is generally set to 0.85.

[0131] In this embodiment, the sensors include position sensors, temperature sensors, and gas sensors, etc.

[0132] The beneficial effects of the above technical solution are: by extracting features from the suspected moving area, key feature information is obtained; by comparing the similarity between the key feature information and the target feature information, it is determined whether there is a suspected moving target in the suspected moving area, eliminating interference from other similar targets and effectively improving the accuracy of moving target detection, thereby ensuring intelligent and accurate monitoring.

[0133] This invention provides an intelligent monitoring method suitable for coal yards, which effectively responds based on alarm status and detection information from corresponding sensors, including:

[0134] Step 31: After determining that there is a suspected moving target in the suspected movement area, combine the camera preset location and retrieve the key location information of the nearby position sensor to confirm the existence of the suspected moving target;

[0135] Step 32: Based on the morphological category of the suspected moving target, select the associated first sensor to obtain key data information and alarm information;

[0136] Step 33: Analyze the alarm information of the first sensor. If there are no historical alarm records before the preset time period, the key location information and key data information are sent back to the server and corresponding effective measures are taken in a timely manner.

[0137] In this embodiment, the position sensor is used to determine the displacement and position of the moving target; the key position information is mainly composed of latitude and longitude coordinates.

[0138] In this embodiment, the morphological category of the suspected moving target mainly refers to solids and solid-liquid mixtures; key data information is related to sensors. For example, the key data information corresponding to the gas sensor includes the name and concentration of the gas that the suspected moving target may generate, and the key data information corresponding to the temperature sensor includes the current temperature and temperature change trend that the suspected moving target may generate; alarm information includes alarm time records and alarm reasons; the preset time period is set in advance, generally 30 seconds.

[0139] The beneficial effects of the above technical solution are: by selecting the associated first sensor based on the suspected moving target; by analyzing whether the first sensor has historical alarm records, it is determined whether to transmit data back, thereby ensuring that effective measures are taken in a timely manner to solve safety hazards, effectively realizing intelligent monitoring, and improving the safety early warning capability of the coal yard.

[0140] This invention provides an intelligent monitoring system suitable for coal yards, such as... Figure 2 As shown, it includes:

[0141] Image acquisition module: Uses a high-definition camera installed at a preset location to acquire video images and obtain the target frame image;

[0142] Image processing module: Enhances the target frame image to obtain the first frame image;

[0143] Target detection module: Utilizes video continuity to perform moving target detection on the first frame image to obtain a detected image;

[0144] Motion region segmentation module: Analyzes the detected image and uses morphological operations to segment and obtain images of suspected motion regions;

[0145] Alarm module: Detects the suspected motion area. If a suspected moving target is found, it calls the detection information of the corresponding sensor and issues an alarm together with the alarm information. If no suspected moving target is found, it updates the background image and re-detects whether a moving target is found.

[0146] The beneficial effects of the above technical solution are as follows: image quality is improved by preprocessing the images acquired by the camera; suspected moving areas are obtained by moving target detection and region segmentation of the improved images; the presence of moving targets is detected in the suspected moving areas, and relevant responses are made in combination with the detection information of the sensors, which effectively realizes intelligent monitoring and improves the safety early warning capability of the coal yard.

[0147] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. An intelligent monitoring method suitable for a coal yard, characterized in that, The method comprises the following steps: Step 1: video image acquisition is performed by using a high-definition camera installed at a preset position to obtain a target frame image; Step 2: enhancement processing is performed on the target frame image to obtain a first frame image; Step 3: motion target detection is performed on the first frame image by using video continuity to obtain a detection image; Step 4: the detection image is analyzed, and a suspected motion region image is obtained by using morphological operation segmentation; Step 5: the suspected motion region is detected, if there is a suspected motion target, detection information of a corresponding sensor is called, and alarm information is combined and sent, if there is no suspected motion target, a background image is updated, and whether there is a motion target is detected again; The detection image is analyzed, and a suspected motion region image is obtained by using morphological operation segmentation, which comprises the following steps: Step 21: erosion and then expansion are performed on the detection image, and a first image is obtained by combining connected domain analysis processing; Step 22: target feature information of different motion targets required to be detected is obtained; Step 23: target pixel points are screened from the first image according to the obtained target feature information and are reserved; Step 24: the suspected motion region is obtained by collecting the target pixel points; The target pixel points are screened from the first image according to the obtained target feature information and are reserved, which comprises the following steps: Based on different target feature information, an importance value of each first pixel point in the first image is acquired, and is collected according to different moving targets to obtain a plurality of first collections D={d cj ,j=1,2,3,...,n}, wherein d cj represents the importance value of the jth first pixel point for the cth moving target, and n represents the total number of the first pixel points corresponding to the cth moving target. wherein, N 1j represents the maximum value in the R, G, and B channels based on the jth first pixel point; N 2j represents the minimum value in the R, G, and B channels based on the jth first pixel point; N 3j represents the average value in the R, G, and B channels based on the jth first pixel point; Δ1 represents the influence coefficient of the difference between the maximum value and the minimum value in the R, G, and B channels based on the corresponding first pixel point on the importance value of the first pixel point; Δ2 represents the influence coefficient of the average value in the R, G, and B channels based on the corresponding first pixel point on the importance value of the first pixel point; P c represents the target feature value of the cth moving target; Δ3 represents the comparison coefficient of the target feature value on the importance value of the matched first pixel point, and Δ3>Δ2>Δ1; represents the loss factor for calculating the importance value of the first pixel point; A proportion result of important values greater than a preset threshold in all first sets D is determined; According to the obtained proportion result, a first set with a proportion result greater than a preset proportion threshold is selected as a target set; All pixel points corresponding to important values greater than the preset threshold in the target set are extracted from the first image and are reserved as the suspected motion region.

2. The intelligent monitoring method for a coal yard according to claim 1, wherein, The preset position refers to the best installation point of the camera at the indoor or outdoor coal storage site.

3. The intelligent monitoring method for coal yard of claim 1, wherein, The target frame image is enhanced to obtain the first frame image, which comprises the following steps: Step 11: the target frame image is converted from RGB to YUV format; Step 12: the Y component of the target frame image in YUV format is enhanced by using weighting, and the U and V components are adaptively corrected; Step 13: the target frame image in YUV format after adjustment of the components is converted to RGB space to obtain the first frame image.

4. The intelligent monitoring method for coal yard of claim 1, wherein, Motion target detection is performed on the first frame image by using video continuity to obtain a detection image, Which comprises the following steps: J i (x,y) = D i (x,y) - B i (x,y) wherein J i (x,y) represents a detection image; D i (x,y) represents a current i-th first frame image; B i (x,y) represents a background image; Analyzing the detection image J i (x, y) corresponding to the mean and variance of the histogram, the threshold Y is obtained; If the gray value of the detection image is greater than Y, it is determined that the detection image is a motion foreground image; Otherwise, it is defined as a background image.

5. The intelligent monitoring method for a coal yard of claim 1, wherein, The suspected motion region is detected, if there is a suspected motion target, effective response is performed according to the alarm condition and the detection information of the corresponding sensor, if there is no suspected motion target, a background image is updated, and whether there is a motion target is detected again, which comprises the following steps: The gradient of each pixel in the suspected motion region is calculated; The suspected motion region is divided into image blocks of a preset size; The image block is defined as a window, and the size relationship of the gray values of the center pixel point and the surrounding pixel points of the window is compared, If any surrounding pixel value is greater than the center pixel value, the position of the surrounding pixel value greater than the center pixel value is marked as 1, otherwise, it is marked as 0, and a binary number is sequentially arranged to obtain a binary number in the window. Convert the obtained binary number into a decimal number as a first feature value of the center pixel point of the window; Determine the frequency of occurrence of each first feature value and obtain a statistical histogram after normalization processing; Connect the statistical histogram of each window obtained to obtain a first feature vector; Statistical histogram of each image block gradient; Group a preset number of image blocks into a connected interval, and then normalize the corresponding gradient histogram to obtain an interval gradient histogram; Collect all the obtained interval gradient histograms to obtain a second feature vector; Compare the key feature information reflected by the first feature vector and the second feature vector with the target feature information of the suspected moving target to obtain a similarity result; If the similarity result is greater than a preset comparison threshold, it is determined that there is a suspected moving target in the suspected moving region, at which time the relevant detection information of the corresponding sensor is called to perform data back transmission, and an alarm action is started; Otherwise, update the background image, and then extract the moving foreground from the corresponding detection image to re-detect whether there is a moving target.

6. The intelligent monitoring method for a coal yard of claim 5, wherein, According to the alarm situation and detection information of the corresponding sensor, effective responses are performed, including: Step 31: After determining that the suspected moving region has a suspected moving target, the key position information of the suspected moving target is confirmed by calling the adjacent position sensor in combination with the preset camera; Step 32: According to the shape category of the suspected moving target, select the associated first sensor to obtain key data information and alarm information; Step 33: Analyze the alarm information of the first sensor, if there is no historical alarm record before a preset time period, then the key position information and the key data information are back transmitted to the server, and corresponding effective measures are taken for processing in time.

7. An intelligent monitoring system suitable for a coal yard, characterized in that, The system uses the method of any one of claims 1 to 6, and the system comprises: An image acquisition module: a high-definition camera installed at the preset position is used to acquire video images to obtain target frame images; An image processing module: the target frame images are enhanced to obtain first frame images; A target detection module: using video continuity, the first frame images are detected for moving targets to obtain detection images; A moving region segmentation module: analyzing the detection images, using morphological operations to segment to obtain suspected moving region images; An alarm module: detecting the suspected moving region, if there is a suspected moving target, calling the detection information of the corresponding sensor, and then combining the alarm information to issue, if there is no suspected moving target, updating the background image, and re-detecting whether there is a moving target.

Images