A method and apparatus for identifying a slipstream head conveyor belt zone

Abstract

The embodiment of the application provides a rear sliding machine head conveying belt area identification method and device, the method comprises the following steps: through improving frame difference method, the rear sliding machine head image obtained is identified, and a motion area is obtained; according to the motion area and the pre-set feature selection size, corner point detection and clustering are carried out on the rear sliding machine head image, and a plurality of categories of feature subgraphs are obtained; through the plurality of categories of feature subgraphs and the motion area, the rear sliding machine head image is matched and identified, and a conveying belt area is obtained; the conveying belt area is obtained by improving the frame difference method and the feature matching method for the area matching identification of the rear sliding machine head image, all related feature pixel points can be completely extracted, the internal hole of the object is filled through the multi-frame accumulation effect, the moving object is accurately detected, so that the identification accuracy of the conveying belt area is improved, and the feature matching method can avoid the motion area detection failure caused by the stop of the conveying belt area.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer, in particular to a rear sliding machine head conveyor belt area identification method and device. BACKGROUND

[0002] At present, the accurate identification of the conveyor belt area at the rear sliding machine head in the coal mine has positive significance for the position determination of large coal blocks, the early warning of dangerous areas and other applications. In related technologies, the frame difference method is usually used to identify the conveyor belt area at the rear sliding machine head. However, due to the slow position change of the moving target between adjacent frames, the overlapping part of the target inside in different frame images is difficult to detect, only the boundary can be extracted, and all related feature pixels cannot be completely extracted, which will cause the "hole" phenomenon in the moving entity. At the same time, it depends on the selected frame interval. If the selected frame interval is not appropriate, the moving object cannot be accurately detected, resulting in low accuracy of conveyor belt area identification. SUMMARY

[0003] One object of the present application is to provide a rear sliding machine head conveyor belt area identification method, which can completely extract all related feature pixels by improving the frame difference method and the feature matching method for area matching identification of the rear sliding machine head image to obtain the conveyor belt area, fill the holes in the object by the multi-frame accumulation effect, accurately detect the moving object, and thus improve the identification accuracy of the conveyor belt area. Another object of the present application is to provide a rear sliding machine head conveyor belt area identification device. Another object of the present application is to provide a computer readable medium. Another object of the present application is to provide a computer device.

[0004] In order to achieve the above objects, the present application discloses a rear sliding machine head conveyor belt area identification method, comprising:

[0005] Obtaining a rear sliding machine head video stream, the rear sliding machine head video stream comprising a plurality of rear sliding machine head images;

[0006] Identifying the area of the rear sliding machine head image by improving the frame difference method to obtain a moving area;

[0007] According to the moving area and the pre-set feature selection size, corner point detection and clustering are performed on the rear sliding machine head image to obtain a plurality of category feature subgraphs;

[0008] Matching and identifying the rear sliding machine head image by the plurality of category feature subgraphs and the moving area to obtain a conveyor belt area.

[0009] Preferably, the rear sliding machine head image is identified by improving the frame difference method to obtain the moving area, comprising:

[0010] performing pairwise difference calculation on the adjacent rear sliding machine head images of the designated order to obtain a plurality of adjacent difference results;

[0011] performing accumulation on the plurality of adjacent difference results to obtain a motion binary image;

[0012] performing frame division according to the motion binary image to obtain a motion region.

[0013] Preferably, before the step of performing corner point detection and clustering on the rear sliding machine head image according to the motion region and a pre-set feature selection size to obtain a plurality of multi-class feature subgraphs, the method further comprises:

[0014] performing region verification on the motion region through a pre-set region verification condition, the region verification condition comprising an area proportion verification condition, an image matrix verification condition, a gray mean value verification condition and a gray variance verification condition;

[0015] if the verification is passed, continuing to perform the step of performing corner point detection and clustering on the rear sliding machine head image according to the motion region and a pre-set feature selection size to obtain a plurality of multi-class feature subgraphs;

[0016] if the verification fails, re-acquiring a plurality of rear sliding machine head images and continuing to perform the step of performing region recognition on the rear sliding machine head image through the improved frame difference method to obtain a motion region.

[0017] Preferably, the step of performing corner point detection and clustering on the rear sliding machine head image according to the motion region and a pre-set feature selection size to obtain a plurality of multi-class feature subgraphs comprises:

[0018] performing region division in a designated direction side of the motion region in the rear sliding machine head image according to the feature selection size to obtain a feature selection region;

[0019] performing corner point detection in the feature selection region through a corner point detection algorithm to obtain a corner point cluster;

[0020] performing clustering on the corner point cluster through a clustering algorithm to obtain a plurality of multi-class corner point groups;

[0021] performing frame division on each multi-class corner point group to obtain a plurality of multi-class feature subgraphs.

[0022] Preferably, the rear sliding machine head image comprises a rear sliding machine head image of a current frame and a rear sliding machine head image of a previous frame of the current frame.

[0023] performing matching recognition on the rear sliding machine head image through the plurality of multi-class feature subgraphs and the motion region to obtain a conveyor belt region, comprising:

[0024] The image matching algorithm is used to match the current frame of the rear sliding head image according to the multi-category feature subgraphs, and a matching result is obtained, the matching result including a target feature subgraph matched successfully and a vertex position of each target feature subgraph;

[0025] According to the number of target feature subgraphs and the total number of multi-category feature subgraphs, a target image proportion is generated;

[0026] It is judged whether the target image proportion is greater than a preset target proportion threshold value;

[0027] If yes, the current frame of the rear sliding head image is identified according to the vertex position of the target feature subgraph and the rear sliding head image of the previous frame according to the motion region, and a conveying belt region is obtained; and the target feature subgraph matched successfully is updated, and the target feature subgraph matched unsuccessfully is deleted;

[0028] If no, after a specified number of rear sliding head images, the current frame of the rear sliding head image and the rear sliding head image of the previous frame of the current frame are re-determined, and the step of matching the current frame of the rear sliding head image according to the multi-category feature subgraph by the image matching algorithm is re-executed to obtain a matching result.

[0029] Preferably, the current frame of the rear sliding head image is identified according to the vertex position of the target feature subgraph and the rear sliding head image of the previous frame according to the motion region, and a conveying belt region is obtained, including:

[0030] According to the vertex position of the target feature subgraph, a corresponding target centroid position is generated;

[0031] According to the vertex position of the rear sliding head image of the previous frame, a corresponding previous frame centroid position is generated;

[0032] According to the target centroid position and the previous frame centroid position, a moving distance and a moving direction are obtained;

[0033] According to the moving distance and the moving direction, the motion region is moved on the current frame of the rear sliding head image to obtain the conveying belt region.

[0034] Preferably, the method further comprises:

[0035] If the target image proportion is less than or equal to the target proportion threshold value, a preset failure count is incremented by 1;

[0036] It is judged whether the failure count after the increment by 1 is greater than or equal to a preset count threshold value;

[0037] If yes, a plurality of rear sliding head images are re-acquired, and the step of identifying the rear sliding head image according to the improved frame difference method to obtain the motion region is continued;

[0038] If not, continue to execute the step of determining the backwash head image of the current frame and the backwash head image of the previous frame of the current frame after waiting for the specified number of frame images.

[0039] The application further discloses a backwash head conveyor belt region identification device.

[0040] The acquisition unit is configured to acquire a backwash head video stream, the backwash head video stream comprising a plurality of backwash head images.

[0041] The motion region identification unit is configured to identify regions of the backwash head images by using an improved frame difference method to obtain motion regions.

[0042] The corner point clustering unit is configured to detect and cluster corner points of the backwash head images according to the motion regions and a pre-set feature selection size to obtain a plurality of feature subgraphs of different categories.

[0043] The matching identification unit is configured to match and identify the backwash head images by using the feature subgraphs of different categories and the motion regions to obtain the conveyor belt region.

[0044] The application further discloses a computer readable medium, which stores a computer program, and the program is executed by a processor to implement the method.

[0045] The application further discloses a computer device, which comprises a memory and a processor, the memory is configured to store information comprising program instructions, and the processor is configured to control execution of the program instructions, and the processor implements the method when executing the program.

[0046] The application acquires a backwash head video stream, the backwash head video stream comprising a plurality of backwash head images, identifies regions of the backwash head images by using an improved frame difference method to obtain motion regions, detects and clusters corner points of the backwash head images according to the motion regions and a pre-set feature selection size to obtain a plurality of feature subgraphs of different categories, and matches and identifies the backwash head images by using the feature subgraphs of different categories and the motion regions to obtain the conveyor belt region. BRIEF DESCRIPTION OF DRAWINGS

[0047] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description only constitute some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor based on these drawings.

[0048] Figure 1 A flow chart of a rear sliding machine head conveying belt area identification method provided by an embodiment of the present application is shown in the figure.

[0049] Figure 2 A flow chart of another rear sliding machine head conveying belt area identification method provided by an embodiment of the present application is shown in the figure.

[0050] Figure 3 A structural schematic diagram of a rear sliding machine head conveying belt area identification device provided by an embodiment of the present application is shown in the figure.

[0051] Figure 4 A structural schematic diagram of a computer device provided by an embodiment of the present application is shown in the figure. DETAILED DESCRIPTION

[0052] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments only constitute some embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0053] In order to facilitate understanding of the technical solutions provided by the present application, the related contents of the technical solutions of the present application will be described first. The conveying belt area at the rear sliding machine head exists left and right horizontal movement, and the method of simply manually marking the area cannot meet the demand. The relatively obvious feature of the conveying belt area compared with other objects is that the movement time is relatively long, and the area detection can adopt the frame difference method in the traditional method. The frame difference method uses pixel-based difference between adjacent two frames or three frames in an image sequence to extract the moving area in the image. The frame difference method is not sensitive to scene changes such as light, can adapt to various dynamic environments, has good stability, and the algorithm is simple to implement and has low actual program complexity.

[0054] The frame difference method includes adjacent frame difference method and three-frame difference method, which uses pixel-based difference between adjacent two frames or three frames in an image sequence to extract the moving area in the image. The basic principle is as follows:

[0055]

[0056] Wherein, D(x, y) is a binary image, I(t) and I(t-1) are adjacent two images, and T is a segmentation threshold value.

[0057] In order to solve the problem of the "hollow" phenomenon existing in the traditional frame difference method, the frame difference method is improved in the application, in order to avoid the failure of using the frame difference method when the conveyor belt area is static, only the motion area detection is carried out in a few frames at the beginning of the video stream, and the subsequent area is retrieved by using the feature matching mode.

[0058] The implementation process of the rear sliding head conveyor belt area identification method provided by the embodiment of the application will be described below by taking the rear sliding head conveyor belt area identification device as an execution subject.

[0059] Figure 1 The flowchart of the rear sliding head conveyor belt area identification method provided by the embodiment of the application is shown in the figure. Figure 1 As shown in the figure, the method comprises the following steps.

[0060] Step 101, a rear sliding head video stream is acquired, and the rear sliding head video stream comprises a plurality of rear sliding head images.

[0061] Step 102, the rear sliding head image is subjected to area identification by using the improved frame difference method, and a motion area is obtained.

[0062] Step 103, according to the motion area and a pre-set feature selection size, the rear sliding head image is subjected to corner point detection and clustering, and a plurality of categories of feature subgraphs are obtained.

[0063] Step 104, the rear sliding head image is subjected to matching identification by using the plurality of categories of feature subgraphs and the motion area, and a conveyor belt area is obtained.

[0064] The technical scheme provided by the embodiment of the present application comprises the following steps: obtaining a rear sliding machine head video stream, wherein the rear sliding machine head video stream comprises a plurality of rear sliding machine head images; performing region identification on the rear sliding machine head images by improving a frame difference method to obtain a motion region; performing corner point detection and clustering on the rear sliding machine head images according to the motion region and a pre-set feature selection size to obtain a plurality of category feature subgraphs; and performing matching identification on the rear sliding machine head images through the plurality of category feature subgraphs and the motion region to obtain a conveyor belt region. The rear sliding machine head images are identified by the improved frame difference method and the feature matching method to obtain the conveyor belt region, all related feature pixels can be completely extracted, the internal holes of the object are filled through the multi-frame accumulation effect, the moving object can be accurately detected, and thus the identification accuracy of the conveyor belt region is improved, and the feature matching method can avoid the motion region detection failure caused by the stop of the conveyor belt region.

[0065] Figure 2 As shown in FIG. 1, the method comprises the following steps: Figure 2

[0066] Step 201: obtaining a rear sliding machine head video stream, wherein the rear sliding machine head video stream comprises a plurality of rear sliding machine head images.

[0067] In the embodiment of the present application, each step is executed by a rear sliding machine head conveyor belt region identification device.

[0068] In the embodiment of the present application, the rear sliding machine head video stream is photographed by an image acquisition device. In the photographing process, the movement of the conveyor belt at the rear sliding machine head of the coal mine underground is continuous and in the same direction, and the conveyor belt region at the rear sliding machine head exists left-right horizontal movement. The rear sliding machine head video stream comprises continuous and ordered multiple rear sliding machine head images.

[0069] Step 202: performing two-by-two difference calculation on the adjacent rear sliding machine head images in the specified order to obtain a plurality of adjacent difference results.

[0070] In the embodiment of the present application, the adjacent rear sliding machine head images in the specified order are the first X images of the rear sliding machine head video stream, wherein X is any positive integer. It is worth mentioning that the value of X can be set according to actual needs, and the embodiment of the present application does not limit this. As an optional solution, the adjacent rear sliding machine head images in the specified order are the first 30 images of the rear sliding machine head video stream, that is, X is 30.

[0071] For example, the adjacent rear sliding machine head images in the specified order comprise X1, X2, X3 and X4, difference calculation is performed on X1 and X2 to obtain a first adjacent difference result, difference calculation is performed on X2 and X3 to obtain a second adjacent difference result, and difference calculation is performed on X3 and X4 to obtain a third adjacent difference result. ​

[0072] Step 203, accumulate the plurality of adjacent difference results to obtain a motion binary image.

[0073] For example, initialize the "cumulative difference map" as empty, the plurality of adjacent difference results include the first adjacent difference result, the second adjacent difference result and the third adjacent difference result, add the first adjacent difference result to the "cumulative difference map" as the difference result, use the mean value of the gray scale map of the difference result as the binary threshold, threshold segment the difference result to obtain a binary image, and take the binary image as the new "cumulative difference map", add the second adjacent difference result to the new "cumulative difference map", further binary and take as the new "cumulative difference map", and repeat the above steps for the third adjacent difference result to obtain the motion binary image.

[0074] The "hole" phenomenon of the frame difference method is caused by the slow change of the position of the moving target between adjacent frames, and the difficulty in detecting the overlapping part of the target in different frame images, the application can fill the internal holes of the object by accumulating multi-frame difference of the first X frames of the video stream, and realize the detection of the motion region while avoiding excessive noise.

[0075] Specifically, the threshold segmentation compares the gray value of each pixel point in the current difference result with the binary threshold, if the gray value of the pixel point is greater than the binary threshold, the gray value of the pixel point is set to 255, if the gray value of the pixel point is less than or equal to the binary threshold, the gray value of the pixel point is set to 0. After comparing and resetting the gray value of each pixel point in the current backhaul image, a motion binary image is obtained.

[0076] Step 204, according to the motion binary image, the frame is divided to obtain the motion region.

[0077] In the embodiment of the application, according to the boundary edge in the motion binary image, a binary boundary contour is obtained, and a circumscribed rectangle or a circumscribed trapezoid of the binary boundary contour is taken to obtain the motion region.

[0078] Step 205, through the pre-set region verification condition, the motion region is verified, if the verification is passed, step 206 is executed, if the verification fails, step 201 is executed.

[0079] In the embodiment of the application, if the motion region meets the region verification condition, it indicates that the motion region is a conveyor belt region, and step 206 is continuously executed, if the motion region does not meet the region verification condition, it indicates that the motion region is not a conveyor belt region, and a plurality of backhaul images are re-acquired, that is, a plurality of backhaul images after the current backhaul image are acquired, and step 201 is continuously executed.

[0080] It is worth noting that if the conveyor belt at the rear feeder head is continuously stationary, the motion area verification will fail. While the conveyor belt at the rear feeder head is continuously stationary, step 201 is executed at specified time intervals until the conveyor belt at the rear feeder head begins to move.

[0081] In this embodiment of the invention, the region verification conditions include area proportion verification conditions, image moment verification conditions, grayscale mean verification conditions, and grayscale variance verification conditions. It is worth noting that the region verification conditions may also include other conditions, and this embodiment of the invention does not limit these.

[0082] In this embodiment of the invention, if the region verification conditions include area ratio verification conditions, step 205 specifically includes:

[0083] Step 2051: Based on the motion area and the image of the rear slide head where the motion area is located, obtain the area ratio of the motion area.

[0084] Specifically, the area occupied by the motion region is divided by the total area of ​​the rear camera head image where the motion region is located to obtain the area ratio of the motion region.

[0085] Step 2052: Determine whether the area ratio of the exercise area is within the preset area ratio range. If yes, proceed to step 2053; otherwise, proceed to step 2054.

[0086] In this embodiment of the invention, the area ratio range is set according to actual needs, and this embodiment of the invention does not limit it.

[0087] Specifically, if the area of ​​the moving area is within the area ratio range, it indicates that the area of ​​the current moving area meets the requirements of the conveyor belt area, and step 2053 is executed; if the area of ​​the moving area is outside the area ratio range, it indicates that the area of ​​the current moving area does not meet the requirements of the conveyor belt area, and step 2054 is executed.

[0088] Step 2053: Verification passed.

[0089] Step 2054: Verification failed.

[0090] In this embodiment of the invention, if the region verification condition includes the image moment verification condition, step 205 specifically includes:

[0091] Step 3051: Determine whether the shape of the motion area is vertical. If yes, proceed to step 3052; otherwise, proceed to step 3053.

[0092] Specifically, if the shape feature of the motion region is vertical, it indicates that the shape feature of the current motion region meets the requirement of the shape feature of the conveying belt, and step 3052 is executed; if the shape feature of the motion region is horizontal, it indicates that the shape feature of the current motion region does not meet the requirement of the shape feature of the conveying belt, and step 3053 is executed.

[0093] It should be noted that the image matrix checking condition can also include other image features, and the embodiments of the present application do not limit this.

[0094] It should be noted that the setting of the shape feature is related to the relative direction of the image acquisition device and the conveying belt region, and can be set as vertical or horizontal. The specific direction can be set according to actual conditions.

[0095] Step 3052, it is determined that the checking is passed.

[0096] Step 3053, it is determined that the checking fails.

[0097] In the embodiments of the present application, if the region checking condition includes the gray mean value checking condition, step 205 specifically includes:

[0098] Step 4051, it is judged whether the gray mean value of the motion region of the current rear sliding head image is located in a preset gray mean value range, if yes, step 4052 is executed; if no, step 4053 is executed.

[0099] Specifically, if the gray mean value of the motion region of the current rear sliding head image is located in the preset gray mean value range, it indicates that the gray mean value of the motion region of the current rear sliding head image meets the requirement of the gray mean value of the conveying belt, and step 4052 is executed; if the gray mean value of the motion region of the current rear sliding head image is located outside the preset gray mean value range, it indicates that the gray mean value of the motion region of the current rear sliding head image does not meet the requirement of the gray mean value of the conveying belt, and step 4053 is executed.

[0100] It should be noted that the gray mean value range can be set according to actual requirements, and the embodiments of the present application do not limit this.

[0101] Step 4052, it is determined that the checking is passed.

[0102] Step 4053, it is determined that the checking fails.

[0103] In the embodiments of the present application, if the region checking condition includes the gray mean value checking condition, step 205 specifically includes:

[0104] Step 5051, it is judged whether the gray mean value of the motion region of the current rear sliding head image is located in a preset gray mean value range, if yes, step 5052 is executed; if no, step 5053 is executed.

[0105] Specifically, if the gray variance of the motion region of the current rear roller head image is within the preset gray variance range, it indicates that the gray variance of the motion region of the current rear roller head image meets the requirement of the gray variance of the conveying belt, and step 5052 is executed; if the gray variance of the motion region of the current rear roller head image is outside the preset gray variance range, it indicates that the gray variance of the motion region of the current rear roller head image does not meet the requirement of the gray variance of the conveying belt, and step 5053 is executed.

[0106] It should be noted that the gray variance range can be set according to actual needs, and the embodiment of the present application does not limit this.

[0107] Step 5052, it is determined that the verification is passed.

[0108] Step 5053, it is determined that the verification fails.

[0109] Step 206, according to the motion region and the pre-set feature selection size, corner point detection and clustering are performed on the rear roller head image to obtain a plurality of categories of feature subgraphs.

[0110] In the embodiment of the present application, step 206 specifically includes:

[0111] Step 2061, according to the feature selection size, region division is performed on the designated direction side of the motion region in the rear roller head image to obtain a feature selection region.

[0112] In the embodiment of the present application, the feature selection size is set according to actual working conditions, and a region where the rear roller head is rigidly connected to the conveying belt is specifically selected, for example, if the rear roller head is rigidly connected to the right side region of the conveying belt, the right side of the rear roller head conveying belt region to the edge of the image is selected as the feature selection region, and the feature selection region meets the requirement of the feature selection size for subsequent corner point detection. The rear roller head image in this step is an image after the rear roller head image with a specified order in the rear roller head video stream, for example, if the rear roller head image with a specified order is the first 30 frames of the rear roller head video stream, the rear roller head image in this step is the 31st frame.

[0113] Step 2062, corner point detection is performed in the feature selection region by using a corner point detection algorithm to obtain a corner point cluster.

[0114] Specifically, the feature selection region in the rear roller head image is input into the corner point detection algorithm, corner point detection is performed in the feature selection region, and a corner point cluster is output.

[0115] It should be noted that the corner point detection algorithm is an existing mature algorithm, and the specific corner point detection process will not be described here.

[0116] In step 2063, the corner point clusters are clustered by a clustering algorithm to obtain a plurality of corner point groups of different categories.

[0117] In the embodiment of the present application, the clustering algorithm includes but is not limited to K-means (K-Means) algorithm and density-based clustering method with noise (Density-Based Spatial Clustering of Applications with Noise, referred to as: DBSCAN).

[0118] Specifically, the corner point clusters are input into the clustering algorithm for clustering, and a plurality of corner point groups of different categories are output. Each category includes at least one corner point.

[0119] It is worth noting that different categories can be represented by different colors, and corner points of the same color belong to the same category.

[0120] In step 2064, a bounding box is divided for each category of corner point group to obtain a plurality of feature subgraphs of different categories.

[0121] Specifically, a bounding box is taken for each category of corner point group to obtain a plurality of feature subgraphs of different categories. The plurality of feature subgraphs of different categories include a plurality of categories and a feature subgraph corresponding to each category.

[0122] In step 207, the image matching algorithm is used to match the back sliding machine head image of the current frame according to the plurality of feature subgraphs of different categories to obtain a matching result.

[0123] In the embodiment of the present application, the image matching algorithm is an image matching algorithm provided by a software library (OpenCV), and the back sliding machine head image includes the back sliding machine head image of the current frame and the back sliding machine head image of the previous frame of the current frame.

[0124] Specifically, a confidence threshold is preset in the image matching algorithm. The image matching algorithm matches the back sliding machine head image of the current frame according to the plurality of feature subgraphs of different categories to obtain a matching confidence. If the matching confidence is greater than or equal to the confidence threshold, it indicates that the matching is successful, and the corresponding target feature subgraph and the vertex position of the target feature subgraph are output. If the matching confidence is less than the confidence threshold, it indicates that the matching fails, and no output is made.

[0125] In the embodiment of the present application, the matching result includes the target feature subgraph of successful matching and the vertex position of each target feature subgraph.

[0126] In step 208, a target image proportion is generated according to the number of target feature subgraphs and the total number of the plurality of feature subgraphs of different categories.

[0127] Specifically, the number of target feature subgraphs is divided by the total number of the plurality of feature subgraphs of different categories to obtain the target image proportion.

[0128] Step 209, judge whether the target image proportion is greater than the preset target proportion threshold value, if yes, execute step 210; if no, execute step 211.

[0129] In the embodiment of the application, the target proportion threshold value can be set according to actual needs, and the embodiment of the application does not limit this. As an optional solution, the target proportion threshold value is 2 / 3.

[0130] Specifically, if the target image proportion is greater than the target proportion threshold value, it indicates that the current frame tailback head image matching is successful, and step 210 is continued to execute; if the target image proportion is less than or equal to the target proportion threshold value, it indicates that the current frame tailback head image matching fails, and step 211 is continued to execute.

[0131] Further, the target feature subgraph that matches successfully is updated. Specifically, the target feature subgraph is replaced with the corresponding tailback head image, and the target feature subgraph is stored for subsequent matching, thereby improving the accuracy of the conveyor belt area identification.

[0132] Further, the target feature subgraph that matches fails is deleted, and subsequent matching of the target feature subgraph is no longer performed.

[0133] Further, if the target image proportion is less than or equal to the target proportion threshold value, the preset failure count is incremented by 1; it is judged whether the failure count after the increment processing is greater than or equal to the preset count threshold value, if yes, it indicates that the number of failures is too large, and the video screen has changed greatly, a plurality of tailback head images are acquired behind the tailback head image of the current frame of the tailback head video stream, and step 210 is continued to execute; if no, it indicates that the failure reason of the feature matching may be that the feature subgraph area is blocked by fog or personnel, and step 211 is continued to execute.

[0134] It is worth noting that the initial value of the failure count and the count threshold value can be set according to actual needs, and the embodiment of the application does not limit this. As an optional solution, the initial value of the failure count is set to 0, and the count threshold value is set to 1.

[0135] Step 210, according to the vertex position of the target feature subgraph and the tailback head image of the previous frame, the tailback head image of the current frame is identified according to the motion area, and the conveyor belt area is obtained, and the process ends.

[0136] In the embodiment of the application, step 210 specifically includes:

[0137] Step 2101, generating a corresponding target centroid position according to the vertex position of the target feature subgraph.

[0138] Specifically, the vertex positions of the target feature subgraph are calculated to obtain corresponding target centroid positions. As an optional solution, the positions are represented by coordinates.

[0139] It should be noted that the method for calculating the centroid is prior art, and the embodiments of the present application will not be described again.

[0140] Step 2102, generating corresponding previous frame centroid positions according to the vertex positions of the previous frame back-haul image.

[0141] Specifically, the vertex positions of the previous frame back-haul image are calculated to obtain corresponding previous frame centroid positions. As an optional solution, the positions are represented by coordinates.

[0142] It should be noted that the method for calculating the centroid is prior art, and the embodiments of the present application will not be described again.

[0143] Step 2103, obtaining a moving distance and a moving direction according to the target centroid positions and the previous frame centroid positions.

[0144] As an optional solution, the moving distance is calculated by the Euclidean distance formula according to the target centroid positions and the previous frame centroid positions; the azimuth angle is calculated according to the target centroid positions and the previous frame centroid positions, and the azimuth angle is determined as the moving direction.

[0145] Step 2104, moving the motion region on the current frame back-haul image according to the moving distance and the moving direction to obtain a conveyor belt region.

[0146] Specifically, the vertex positions of the motion region on the current frame back-haul image are moved according to the moving distance and the moving direction to obtain the conveyor belt region, which is a region on the current frame back-haul image.

[0147] Step 211, waiting for a specified number of back-haul images.

[0148] In the embodiments of the present application, if the failure reason of feature matching may be that the feature subgraph region is blocked by fog or personnel, etc., after waiting for a specified number of back-haul images in the back-haul video stream, the feature matching is reattempted.

[0149] It should be noted that the specified number of frames can be set according to actual needs, and the embodiments of the present application do not limit this.

[0150] Step 212, re-determining the current frame back-haul image and the previous frame back-haul image of the current frame, and re-executing step 207.

[0151] In the embodiment of the present application, after waiting for the specified number of frames of the backsliding machine head image in the backsliding machine head video stream, the backsliding machine head image of the current frame and the backsliding machine head image of the previous frame of the current frame are determined, and the matching of the features is re-attempted, and step 207 is executed.

[0152] The present application can eliminate the hole phenomenon through improvement of the frame difference method, so that a relatively complete conveyor belt area is obtained; in addition, the front X frames are used for motion area detection, and the subsequent video stream is used for conveyor belt area identification through feature matching, so that the situation of conveyor belt area identification error caused by stop of the conveyor belt or interference of other moving objects can be avoided, the detection accuracy is further improved, and reliable data basis is provided for subsequent foreign matter intrusion, area demarcation, safety detection and other applications.

[0153] It should be noted that the acquisition, storage, use, processing and the like of data in the technical solutions in the present application comply with the relevant provisions of laws and regulations. The user information in the embodiments of the present application is obtained through a legal and compliant way, and the acquisition, storage, use, processing and the like of the user information are authorized and agreed by the client.

[0154] In the technical solution of the backsliding machine head conveyor belt area identification method provided by the embodiment of the present application, the backsliding machine head video stream is acquired, the backsliding machine head video stream includes multiple backsliding machine head images; the region of the backsliding machine head image is identified through improvement of the frame difference method, and the motion region is obtained; the corner point detection and clustering of the backsliding machine head image are performed according to the motion region and the pre-set feature selection size, and multiple categories of feature subgraphs are obtained; the backsliding machine head image is matched and identified through the multiple categories of feature subgraphs and the motion region, and the conveyor belt area is obtained; the conveyor belt area is obtained through the region matching identification of the backsliding machine head image by the improved frame difference method and the feature matching mode, all related feature pixels can be completely extracted, the internal holes of the object are filled through the multi-frame accumulation effect, the moving object is accurately detected, and thus the identification accuracy of the conveyor belt area is improved.

[0155] Figure 3 A structure diagram of a backsliding machine head conveyor belt area identification device provided by the embodiment of the present application is shown in the figure, the device is used for executing the backsliding machine head conveyor belt area identification method described above, as shown in the figure, the device includes: Figure 3

[0156] The acquisition unit 11 is used for acquiring the backsliding machine head video stream, and the backsliding machine head video stream includes multiple backsliding machine head images.

[0157] The motion region identification unit 12 is used for identifying the region of the backsliding machine head image through improvement of the frame difference method, and obtaining the motion region.

[0158] ​The corner point clustering unit 13 is configured to perform corner point detection and clustering on the back-drafting machine head image according to the motion region and a pre-set feature selection size, to obtain a plurality of feature subgraphs of different categories.

[0159] The matching recognition unit 14 is configured to perform matching recognition on the back-drafting machine head image through the plurality of feature subgraphs of different categories and the motion region, to obtain a conveyor belt region.

[0160] In the embodiment of the application, the motion region recognition unit 12 is specifically configured to perform pairwise difference calculation on the adjacent back-drafting machine head images in the specified order, to obtain a plurality of adjacent difference results; perform accumulation on the plurality of adjacent difference results, to obtain a motion binary image; perform frame division according to the motion binary image, to obtain the motion region.

[0161] In the embodiment of the application, the device further comprises a verification unit 15.

[0162] The verification unit 15 is configured to verify the motion region through a pre-set region verification condition, the region verification condition comprising an area proportion verification condition, an image matrix verification condition, a gray mean value verification condition and a gray variance verification condition, if the verification is passed, triggering the corner point clustering unit 13 to continue to perform the step of performing corner point detection and clustering on the back-drafting machine head image according to the motion region and the pre-set feature selection size, to obtain the plurality of feature subgraphs of different categories; if the verification fails, re-acquiring a plurality of back-drafting machine head images, and triggering the motion region recognition unit 12 to continue to perform the step of performing region recognition on the back-drafting machine head image through the improved frame difference method, to obtain the motion region.

[0163] In the embodiment of the application, the corner point clustering unit 13 is specifically configured to perform region division in a specified direction side of the motion region in the back-drafting machine head image according to the feature selection size, to obtain a feature selection region; perform corner point detection in the feature selection region through a corner point detection algorithm, to obtain a corner point cluster; perform clustering on the corner point cluster through a clustering algorithm, to obtain a plurality of category corner point groups; perform frame division on each category corner point group, to obtain the plurality of feature subgraphs of different categories.

[0164] In the embodiment of the present application, the rear sliding machine head image includes a rear sliding machine head image of a current frame and a rear sliding machine head image of a previous frame of the current frame; the matching recognition unit 14 is specifically configured to: according to the multi-category feature subgraph, match the rear sliding machine head image of the current frame by using an image matching algorithm, and obtain a matching result, wherein the matching result includes a target feature subgraph matched successfully and a vertex position of each target feature subgraph; generate a target image proportion according to a number of the target feature subgraphs and a total number of the multi-category feature subgraphs; determine whether the target image proportion is greater than a preset target proportion threshold; if yes, identify the rear sliding machine head image of the current frame according to the vertex position of the target feature subgraph and the rear sliding machine head image of the previous frame, and obtain a conveying belt region according to a motion region; and update the target feature subgraph matched successfully and delete the target feature subgraph matched unsuccessfully; and if no, after waiting for a specified number of rear sliding machine head images, the rear sliding machine head image of the current frame and the rear sliding machine head image of the previous frame of the current frame are determined again, and the matching recognition unit 14 is triggered to re-execute the step of matching the rear sliding machine head image of the current frame according to the multi-category feature subgraph by using the image matching algorithm to obtain the matching result.

[0165] In the embodiment of the present application, the matching recognition unit 14 is specifically configured to: generate a corresponding target centroid position according to the vertex position of the target feature subgraph; generate a corresponding previous frame centroid position according to the vertex position of the rear sliding machine head image of the previous frame; obtain a moving distance and a moving direction according to the target centroid position and the previous frame centroid position; and move the motion region on the rear sliding machine head image of the current frame according to the moving distance and the moving direction to obtain the conveying belt region.

[0166] In the embodiment of the present application, the device further includes an accumulation unit 16 and a determination unit 17.

[0167] The accumulation unit 16 is configured to: if the target image proportion is less than or equal to the target proportion threshold, perform a plus 1 processing on a preset failure count.

[0168] The determination unit 17 is configured to: determine whether the failure count after the plus 1 processing is greater than or equal to a preset count threshold; if yes, continue to execute the step of identifying the rear sliding machine head image by using the improved frame difference method to obtain the motion region by triggering the motion region identification unit 12 to re-acquire a plurality of rear sliding machine head images; and if no, continue to execute the step of determining the rear sliding machine head image of the current frame and the rear sliding machine head image of the previous frame of the current frame again after waiting for a specified number of rear sliding machine head images by triggering the matching recognition unit 14.

[0169] In the scheme of the embodiment of the present application, the rear sliding machine head video stream is acquired, and the rear sliding machine head video stream includes multiple rear sliding machine head images; the rear sliding machine head image is subjected to region identification through the improved frame difference method, and a motion region is obtained; the rear sliding machine head image is subjected to corner point detection and clustering according to the motion region and a pre-set feature selection size, and multiple categories of feature subgraphs are obtained; the rear sliding machine head image is subjected to matching identification through the multiple categories of feature subgraphs and the motion region, and a conveying belt region is obtained; the conveying belt region is obtained through the region matching identification of the rear sliding machine head image by the improved frame difference method and the feature matching method, all related feature pixel points can be completely extracted, the internal holes of the object are filled through the multi-frame accumulation effect, the moving object is accurately detected, and thus the identification accuracy of the conveying belt region is improved, and the feature matching method can avoid the motion region detection failure caused by the stop of the conveying belt region.

[0170] The system, device, module or unit illustrated in the above embodiments can be specifically implemented by a computer chip or entity, or by a product with certain functions. A typical implementation device is a computer device, and specifically, the computer device can be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

[0171] The embodiment of the present application provides a computer device, including a memory and a processor, the memory is used for storing information including program instructions, the processor is used for controlling the execution of the program instructions, when the program instructions are loaded and executed by the processor, each step of the embodiment of the rear sliding machine head conveying belt region identification method is implemented, and specific description can be referred to the above-mentioned embodiment of the rear sliding machine head conveying belt region identification method.

[0172] Reference is made below to Figure 4 which shows a structural schematic diagram of a computer device 600 suitable for being used to implement the embodiments of the present application.

[0173] As shown in Figure 4 , the computer device 600 includes a central processing unit (CPU) 601, which can perform various appropriate work and processing according to a program stored in a read-only memory (ROM) 602 or a program loaded from a storage portion 608 to a random access memory (RAM) 603. In the RAM 603, various programs and data required for the operation of the computer device 600 are also stored. The CPU 601, the ROM 602, and the RAM 603 are connected to each other through a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0174] The following components are connected to the I / O interface 605: an input part 606 including a keyboard, a mouse, etc.; an output part 607 including a display such as a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage part 608 including a hard disk, etc.; and a communication part 609 including a network interface card such as a LAN card, a modem, etc. The communication part 609 performs communication processing via a network such as the Internet. A drive 610 is also connected to the I / O interface 605 as necessary. A removable medium 611 such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc. is attached to the drive 610 as necessary, so that a computer program read out therefrom is installed in the storage part 608 as necessary.

[0175] In particular, the processes described above with reference to the flowcharts can be implemented as a computer software program according to embodiments of the present application. For example, embodiments of the present application include a computer program product comprising a computer program tangibly embodied on a machine-readable medium, the computer program comprising program code for executing the methods illustrated by the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via the communication part 609, and / or installed from the removable medium 611.

[0176] Computer readable media includes permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology for storing information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read only memory (ROM), electrically erasable programmable read only memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassette, magnetic tape disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information accessible by a computing device. According to the definition herein, computer readable media does not include transitory media such as modulated data signals and carriers.

[0177] For the convenience of description, the above apparatus is described in various units by function. Of course, the functions of each unit can be implemented in the same or more software and / or hardware when implementing the present application.

[0178] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 one or more flow or blocks Figure 1 means for functionally implementing the steps listed in the flowchart block or blocks.

[0179] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instructions which implement the function specified in the flowchart block or blocks. Figure 1 one or more flow or blocks Figure 1 means for functionally implementing the steps listed in the flowchart block or blocks.

[0180] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 one or more flow or blocks Figure 1 means for functionally implementing the steps listed in the flowchart block or blocks.

[0181] It should also be noted that the terms "comprising", "comprises", or any other variation thereof are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements in the list, but can also include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element preceded by "comprises a" does not, without more constraints, foreclose the existence of additional identical elements in the process, method, article, or apparatus that includes the recited element.

[0182] The acquisition, storage, use, processing, etc. of data in the technical solutions of the present application comply with relevant provisions of national laws and regulations.

[0183] Those skilled in the art will appreciate that embodiments of the present application can be readily used as a method, a system or a computer program product. Accordingly, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application can take the form of a computer program product on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage devices, etc.) embodying computer readable program code.

[0184] The present application can be described in the general context of computer- executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. The present application can also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules can be located in both local and remote computer storage media including memory storage devices.

[0185] Each of the embodiments described in this specification has been described taking a progressive approach, and parts that are the same or similar between the embodiments can be mutually referred to. Each of the embodiments focuses on the differences from other embodiments. In particular, the system embodiments are described relatively simply because they are basically similar to the method embodiments, and the relevant parts can be referred to the parts of the method embodiments.

[0186] The embodiments described above are only for the purpose of illustrating the present application, and are not intended to limit the present application. The present application can have various modifications and changes for those skilled in the art. Any modification, equivalent replacement, improvement, etc. within the spirit and principle of the present application shall be included in the scope of the claims of the present application.

Claims

1. A method for identifying the area of ​​a rear conveyor belt, characterized in that, The method comprises: acquiring a rear sliding machine head video stream, the rear sliding machine head video stream comprising a plurality of rear sliding machine head images; performing region identification on the rear sliding machine head images by means of an improved frame difference method to obtain a motion region; performing corner point detection and clustering on the rear sliding machine head images according to the motion region and a pre-set feature selection size to obtain a plurality of category feature subgraphs; performing matching identification on the rear sliding machine head images by means of the plurality of category feature subgraphs and the motion region to obtain a conveyor belt region; the rear sliding machine head images comprise a current frame rear sliding machine head image and a rear sliding machine head image of a previous frame of the current frame; the matching identification on the rear sliding machine head images by means of the plurality of category feature subgraphs and the motion region to obtain a conveyor belt region comprises: performing matching on the current frame rear sliding machine head image according to the plurality of category feature subgraphs by means of an image matching algorithm to obtain a matching result, the matching result comprising a target feature subgraph matched successfully and a vertex position of each target feature subgraph; generating a target image proportion according to a number of the target feature subgraphs and a total number of the plurality of category feature subgraphs; determining whether the target image proportion is greater than a pre-set target proportion threshold value; if yes, performing identification on the current frame rear sliding machine head image according to the vertex position of the target feature subgraph and the rear sliding machine head image of the previous frame in the motion region to obtain the conveyor belt region; and updating the target feature subgraph matched successfully and deleting the target feature subgraph matched unsuccessfully; if no, waiting for a specified number of rear sliding machine head images, re-determining the current frame rear sliding machine head image and the rear sliding machine head image of the previous frame of the current frame, and re-performing the matching on the current frame rear sliding machine head image according to the plurality of category feature subgraphs by means of the image matching algorithm to obtain the matching result.

2. The sledding head conveyor belt zone identification method of claim 1, wherein, the region identification on the rear sliding machine head images by means of the improved frame difference method to obtain the motion region comprises: performing pairwise difference calculation on adjacent rear sliding machine head images in a specified order to obtain a plurality of adjacent difference results; accumulating the plurality of adjacent difference results to obtain a motion binary image; performing frame division according to the motion binary image to obtain the motion region.

3. The sledding head conveyor belt zone identification method of claim 1, wherein, before the corner point detection and clustering on the rear sliding machine head images according to the motion region and the pre-set feature selection size to obtain the plurality of category feature subgraphs, the method further comprises: verifying the motion region by means of a pre-set region verification condition, the region verification condition comprising an area proportion verification condition, an image matrix verification condition, a gray mean value verification condition and a gray variance verification condition; if the verification is passed, continuing to perform the corner point detection and clustering on the rear sliding machine head images according to the motion region and the pre-set feature selection size to obtain the plurality of category feature subgraphs; if the verification fails, re-acquiring a plurality of rear sliding machine head images and continuing to perform the region identification on the rear sliding machine head images by means of the improved frame difference method to obtain the motion region.

4. The sledding head conveyor belt zone identification method of claim 1, wherein, The corner point detection and clustering are performed on the back sliding machine head image according to the motion region and a pre-set feature selection size, and a plurality of category feature subgraphs are obtained, including: The feature selection region is divided in a specified direction side of the motion region in the back sliding machine head image according to the feature selection size, and a feature selection region is obtained; The corner point detection is performed in the feature selection region by a corner point detection algorithm, and a corner point cluster is obtained; The corner point cluster is clustered by a clustering algorithm, and a plurality of category corner point groups are obtained; The bounding box is divided for each category corner point group, and a plurality of category feature subgraphs are obtained.

5. The sledding head conveyor belt zone identification method of claim 1, wherein, The back sliding machine head image of the current frame and the back sliding machine head image of the previous frame are identified according to the target feature subgraph vertex position and the motion region, and a conveyor belt region is obtained, including: The corresponding target centroid position is generated according to the target feature subgraph vertex position; The corresponding previous frame centroid position is generated according to the vertex position of the back sliding machine head image of the previous frame; The movement distance and movement direction are obtained according to the target centroid position and the previous frame centroid position; The motion region is moved on the back sliding machine head image of the current frame according to the movement distance and movement direction, and a conveyor belt region is obtained.

6. The sledding head conveyor belt zone identification method of claim 1, wherein, The method further includes: If the target image proportion is less than or equal to the target proportion threshold, the pre-set failure count is incremented by 1; It is judged whether the failure count after the increment processing is greater than or equal to the pre-set count threshold; If yes, a plurality of back sliding machine head images are re-acquired, and the step of identifying the motion region by improving the frame difference method is continued to be performed on the back sliding machine head image; If no, the step of re-determining the back sliding machine head image of the current frame and the back sliding machine head image of the previous frame of the current frame is continued to be performed after waiting for the specified number of back sliding machine head images.

7. A sled head conveyor belt zone identification apparatus, characterized by, The device includes: An acquisition unit is configured to acquire a back sliding machine head video stream, wherein the back sliding machine head video stream includes a plurality of back sliding machine head images; A motion region identification unit is configured to identify a motion region by improving a frame difference method on the back sliding machine head image; A corner point clustering unit is configured to perform corner point detection and clustering on the back sliding machine head image according to the motion region and a pre-set feature selection size, and obtain a plurality of category feature subgraphs; A matching identification unit is configured to perform matching identification on the back sliding machine head image by the plurality of category feature subgraphs and the motion region, and obtain a conveyor belt region; The back sliding machine head image includes a back sliding machine head image of the current frame and a back sliding machine head image of the previous frame of the current frame. The matching recognition unit is specifically configured to: perform matching on the rear sliding nose image of the current frame according to the multi-category feature subgraphs by using an image matching algorithm to obtain a matching result, the matching result including target feature subgraphs that are successfully matched and vertex positions of each target feature subgraph; generate a target image proportion according to a number of the target feature subgraphs and a total number of the multi-category feature subgraphs; determine whether the target image proportion is greater than a preset target proportion threshold; if yes, perform recognition on the rear sliding nose image of the current frame according to the vertex positions of the target feature subgraphs and the rear sliding nose image of the previous frame according to a motion region to obtain a conveyor belt region; and update the target feature subgraphs that are successfully matched and delete the target feature subgraphs that are unsuccessfully matched; and if no, wait for a specified number of rear sliding nose images, then re-determine the rear sliding nose image of the current frame and the rear sliding nose image of the previous frame of the current frame, and trigger the matching recognition unit to re-perform the step of performing matching on the rear sliding nose image of the current frame according to the multi-category feature subgraphs by using the image matching algorithm to obtain the matching result.

8. A computer readable medium having stored thereon a computer program, characterized in that, The program is executed by the processor to implement the rear sliding nose conveyor belt region recognition method according to any one of claims 1 to 6.

9. A computer device comprising a memory for storing information including program instructions, and a processor for controlling execution of the program instructions, characterized in that, The program instruction is loaded and executed by the processor to implement the rear sliding nose conveyor belt region recognition method according to any one of claims 1 to 6.

Images