Base station inspection video processing method and device, electronic equipment and storage medium

Abstract

The application provides a base station inspection video processing method and device, electronic equipment and storage medium, comprising: acquiring the base station inspection video frame picture of the target base station tower and inputting the picture to an image recognition model, acquiring the recognition result, the recognition result comprising the tower station type and the base station number of the target base station tower; calling the tower station registration type and the base station registration number of the target base station tower; comparing the tower station type and the tower station registration type, the base station number and the base station registration number, and generating the inspection result. The application inputs the base station inspection video frame picture to the image recognition model, acquires the tower station type and the base station number of the target base station tower, and compares the tower station registration type and the base station registration number of the target base station tower stored in the background, intelligently identifies whether the inspection result is abnormal, effectively overcomes the problems of strong personal subjectivity and heavy workload in manual identification, and effectively improves the efficiency and accuracy of the inspection result identification.

Description

Technical Field

[0001] This invention relates to the field of computer technology, and in particular to a method, apparatus, electronic device, and storage medium for processing base station inspection videos. Background Technology

[0002] With the development of communication network technology, the requirements for the quality and accuracy of communication networks are becoming increasingly higher, making the proxy maintenance work of communication networks more and more important.

[0003] Currently, the base station tower equipment that supports the propagation of communication network signals mainly relies on manual routine maintenance and inspection. This involves taking inspection videos of the base station towers and generating and uploading maintenance work orders, which are then manually inspected by quality control personnel to determine if there are any abnormalities in the base station towers. This approach suffers from problems such as excessive personal subjectivity and a large workload for video inspections. Summary of the Invention

[0004] This invention provides a method, apparatus, electronic device, and storage medium for processing base station inspection videos, in order to solve the defects of existing technologies, such as excessive personal subjectivity and large workload of video inspection.

[0005] In a first aspect, the present invention provides a base station inspection video processing method, comprising: acquiring base station inspection videos of a target base station tower from multiple angles as recorded in a maintenance work order; performing frame extraction processing on the base station inspection videos to obtain base station inspection images; inputting the base station inspection images into an image recognition model to obtain recognition results output by the image recognition model, the recognition results including the tower type of the target base station tower and the number of base stations installed on the target base station tower; retrieving resource information of the target base station tower from the identification identifier of the target base station tower recorded in the maintenance work order, the resource information including the tower registration type and the number of base stations registered; comparing the tower type with the tower registration type and the number of base stations with the number of base stations registered to generate inspection results for the target base station tower.

[0006] According to a base station inspection video processing method provided by the present invention, the base station inspection image is input into an image recognition model, comprising: performing image segmentation on the base station inspection image to obtain an initial base station tower image after removing the background image; performing image enhancement on the initial base station tower image to obtain an enhanced base station tower image; performing image thinning on the enhanced base station tower image to obtain a thinned base station tower image; performing binarization processing on the thinned base station tower image to obtain a binarized base station tower image; and inputting the binarized base station tower image into the image recognition model.

[0007] According to a base station inspection video processing method provided by the present invention, the input layer of the image recognition model extracts target features related to the shooting content and / or shooting parameters from all image features of the binarized image of the base station tower, so as to generate the recognition result; the target features related to the shooting content include image features related to the target base station tower and the number of base stations set on the target base station tower; the target features related to the shooting parameters include image features related to the lighting angle, sharpness, and shooting pitch angle related to the shooting of the base station inspection image.

[0008] According to the base station inspection video processing method provided by the present invention, the input layer of the image recognition model extracts target features related to the shooting content and / or shooting parameters from all image features of the binary image of the base station tower. This is achieved based on one of the following analysis methods: principal component analysis, linear discriminant analysis, unsupervised discriminant feature extraction based on divergence difference, principal component analysis based on image matrix, nonlinear feature extraction, and Gabor feature extraction algorithm based on feature weighting.

[0009] According to a base station inspection video processing method provided by the present invention, the step of image segmentation of the base station inspection image includes: when it is determined that the image size of the base station inspection image is smaller than a preset image size, performing image segmentation on the base station inspection image using one of a threshold-based image segmentation method and an edge-based image segmentation method; and when it is determined that the image size of the inspection image is greater than or equal to the preset image size, performing image segmentation on the base station inspection image using a region-based image segmentation method.

[0010] According to a base station inspection video processing method provided by the present invention, the image enhancement of the initial base station tower image includes: performing image enhancement on the initial base station tower image using a Gabor image filtering method or a Fourier filtering-based low-quality image enhancement method.

[0011] According to a base station inspection video processing method provided by the present invention, the step of image thinning of the enhanced base station tower image includes: thinning the enhanced base station tower image based on the OPTA thinning algorithm.

[0012] According to a base station inspection video processing method provided by the present invention, the step of obtaining the base station inspection image of the target base station tower recorded in the maintenance work order includes: extracting a digitally compressed image from a database based on the work order number of the maintenance work order; the digitally compressed image is a multi-angle image generated by extracting frames from the base station inspection video of the target base station tower, which is then compressed, uploaded, and stored in the database; and the digitally compressed image is restored to obtain the base station inspection image.

[0013] Secondly, the present invention also provides a base station inspection video processing device, comprising: an image acquisition unit, used to acquire multi-angle base station inspection videos of a target base station tower recorded in a maintenance work order, and to perform frame extraction processing on the base station inspection videos to acquire base station inspection images; an image recognition unit, used to input the base station inspection images into an image recognition model, and to acquire recognition results output by the image recognition model, the recognition results including the tower type of the target base station tower and the number of base stations installed on the target base station tower; an information acquisition unit, used to retrieve resource information of the target base station tower from the identity identifier of the target base station tower recorded in the maintenance work order, the resource information including the tower registration type and the number of base stations registered; and an information comparison unit, used to compare the tower type with the tower registration type and the number of base stations with the number of base stations registered, to generate inspection results for the target base station tower.

[0014] Thirdly, the present invention provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of any of the base station inspection video processing methods described above.

[0015] Fourthly, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the base station inspection video processing method described above.

[0016] Fifthly, the present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of any of the base station inspection video processing methods described above.

[0017] The base station inspection video processing method, device, electronic device, and storage medium provided by this invention extract frames of base station inspection video into images and input them into an image recognition model to obtain the tower type and number of base stations of the target base station tower. The images are then compared with the tower registration type and number of base stations of the target base station tower stored in the background. This intelligently identifies whether there are any abnormalities in the inspection results, effectively overcoming the problems of excessive personal subjectivity and large workload in manual identification, and effectively improving the efficiency and accuracy of inspection result identification. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating the base station inspection video processing method provided by the present invention;

[0020] Figure 2 This is a schematic diagram of the base station inspection video processing device provided by the present invention;

[0021] Figure 3 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0023] It should be noted that in the description of the embodiments of the present invention, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element. The terms "upper," "lower," etc., indicating orientation or positional relationships based on the orientation or positional relationships shown in the accompanying drawings, are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the present invention. Unless otherwise expressly specified and limited, the terms "installed," "connected," and "linked" should be interpreted broadly, for example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two elements. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0024] The term “and / or” in this application means at least one of the connected objects, and the character “ / ” generally indicates that the preceding and following objects are in an “or” relationship.

[0025] The following is combined Figures 1 to 3 This invention describes the base station inspection video processing method, apparatus, electronic device, and storage medium provided in embodiments of the present invention.

[0026] Figure 1 This is a flowchart illustrating the base station inspection video processing method provided by the present invention, as shown below. Figure 1 As shown, including but not limited to the following steps:

[0027] Step 101: Obtain multi-angle base station inspection videos of the target base station tower recorded in the maintenance work order, perform frame extraction processing on the base station inspection videos, and obtain base station inspection images.

[0028] Specifically, the base station inspection video of the target base station tower can be manually captured by the maintenance personnel during the maintenance inspection process, or it can be captured by the maintenance personnel by operating a drone during the maintenance inspection process.

[0029] After capturing and acquiring the base station inspection video of the target base station tower, the video can be processed by frame extraction to obtain multi-angle base station inspection images of the target base station tower. The maintenance personnel can then generate a maintenance work order that includes the corresponding base station inspection images of the target base station tower and store the work order in the work order management system for unified management.

[0030] When an inspection is conducted to determine if there are any abnormalities in the target base station tower, the corresponding maintenance work order can be retrieved from the work order management system, and the base station inspection images of the target base station tower can be retrieved from it.

[0031] It should be noted that any maintenance work order related to the target base station tower may contain multiple base station inspection images. These base station inspection images may be generated by taking pictures of the target base station tower from different lighting angles, with different resolutions, and different elevation angles.

[0032] The base station inspection video processing method provided by this invention can sequentially retrieve any one of all base station inspection images for subsequent judgment of the inspection results; or it can adopt a traversal approach to sequentially retrieve all base station inspection images and perform a judgment of the inspection results for each image.

[0033] Optionally, if the inspection results of all the base station inspection images retrieved for the target base station tower are normal, it can be determined that there are no abnormalities in the target base station tower; if the inspection result of any base station inspection image retrieved for the target base station tower is abnormal, the actual condition of the target base station tower can be further confirmed, for example, through manual secondary quality inspection.

[0034] Step 102: Input the base station inspection image into the image recognition model and obtain the recognition result output by the image recognition model.

[0035] The identification results include the tower type of the target base station tower and the number of base stations installed on the target base station tower.

[0036] Specifically, after obtaining the base station inspection images of the target base station tower, the images can be input into an image recognition model, and the recognition results output by the model after merging, stitching, and comparison can be obtained. The recognition results include the tower type of the target base station tower and the number of base stations installed on the target base station tower.

[0037] Optionally, the tower type of the target base station tower can be divided into types such as angle steel tower, steel pipe tower, single pipe tower, and guyed tower according to the material used, or into types such as triangular tower, square tower, pentagonal tower, hexagonal tower, and octagonal tower according to the number of sides of the cross section.

[0038] In addition, the number of base stations installed on the target base station tower can be set to different numbers such as six, eight, or ten, depending on the communication capacity requirements of the target base station tower.

[0039] Furthermore, the image recognition model can be obtained by training with base station inspection image sample data as samples and pre-determined recognition result sample data corresponding to the base station inspection image sample data as sample labels.

[0040] The identification results sample data used as sample labels can be manually labeled.

[0041] Furthermore, a sample dataset can be constructed by acquiring multiple sets of base station inspection image sample data with recognition result sample data as labels. The image recognition model can be pre-trained using the sample dataset. The model parameters in the image recognition model can be adjusted according to each output result of the image recognition model, and finally the pre-training process of the image recognition model is completed.

[0042] This can be interpreted as either completing the pre-training process of the image recognition model after reaching a predetermined number of pre-training iterations, or as completing the pre-training process of the image recognition model when the training output of the image recognition model converges.

[0043] Optionally, the image recognition model can be a convolutional neural network (CNN) model, which, after pre-training, can effectively identify the recognition results corresponding to the base station inspection images of the target base station tower.

[0044] Step 103: Retrieve the resource information of the target base station tower from the identification of the target base station tower recorded in the maintenance work order.

[0045] The resource information includes the tower registration type and the number of registered base stations.

[0046] Specifically, the maintenance work order can also record the identification of the target base station tower, through which the resource information of the target base station tower can be obtained. This resource information includes the tower registration type and the number of registered base stations.

[0047] Optionally, the identification identifier can be a QR code matching the target base station tower, or it can be an equipment information table of the target base station tower, from which the tower registration type and the number of registered base stations can be obtained. The tower registration type and the number of registered base stations can be set by relevant technical personnel after the target base station tower is put into use or its use is changed.

[0048] Step 104: Compare the tower type with the tower registration type, and the number of base stations with the number of base station registrations to generate inspection results for the target base station tower.

[0049] Specifically, once the tower type and the number of base stations installed on the target base station tower are obtained through the recognition results output by the image recognition model, and the tower registration type and the number of base stations registered on the target base station tower are retrieved through the identity identifier of the target base station tower, the inspection results for the target base station tower can be determined by comparing the tower type with the tower registration type and the number of base stations with the number of base stations registered.

[0050] If the tower type and tower registration type, and the number of base stations and the number of base stations are all consistent with each other and there are no abnormal changes, then the inspection result of the target base station tower is determined to be normal.

[0051] If at least one set of comparison results is inconsistent between the tower type and the tower registration type, and between the number of base stations and the number of base stations registered, then the inspection result of the target base station tower is determined to be abnormal.

[0052] For example, if the number of base stations on the target base station tower is six, but the number of base stations registered on the target base station tower is eight, then the discrepancy between the number of base stations on the target base station tower and the number of base stations registered may indicate an abnormal situation such as a base station being damaged or falling. In this case, the inspection result of the target base station tower is determined to be abnormal, and maintenance personnel can be notified of the abnormal situation of the target base station tower to facilitate subsequent maintenance and protection of the target base station tower.

[0053] Therefore, compared to the existing method of quality inspectors manually checking base station towers for anomalies using inspection images, the base station inspection video processing method provided by this invention can not only intelligently identify whether there are anomalies in the inspection results of the target base station tower, but also achieve an accuracy rate of 77.49%, effectively meeting the daily inspection needs of base station towers. Furthermore, each maintenance work order can save an average of 2 minutes, effectively reducing workload.

[0054] The base station inspection video processing method provided by this invention extracts frames of base station inspection video into images and inputs them into an image recognition model to obtain the tower type and number of base stations of the target base station tower. It then compares these images with the tower registration type and number of base stations of the target base station tower stored in the background to intelligently identify whether there are any abnormalities in the inspection results. This method can effectively overcome the problems of excessive personal subjectivity and large workload in manual identification, and effectively improve the efficiency and accuracy of inspection result identification.

[0055] Based on the above embodiments, as an optional embodiment, the base station inspection images are input into the image recognition model, including:

[0056] Image segmentation is performed on the base station inspection images to obtain the initial base station tower images after removing the background images.

[0057] Image enhancement is performed on the initial base station tower image to obtain an enhanced base station tower image.

[0058] Image refinement is performed on the enhanced base station tower image to obtain a refined base station tower image.

[0059] The detailed base station tower image is binarized to obtain a binarized base station tower image.

[0060] The binarized image of the base station tower is input into the image recognition model.

[0061] Specifically, in order to reduce the algorithmic complexity of the image recognition model and improve its efficiency and accuracy, the acquired base station inspection images can be preprocessed.

[0062] First, image segmentation is performed on the base station inspection images to obtain the initial base station tower images after removing the background images.

[0063] Image segmentation involves dividing an image into several specific regions with unique properties (e.g., background regions, non-recoverable regions, clear regions, and recoverable regions). Therefore, image segmentation can remove background images such as the sky and obtain the initial base station tower image.

[0064] By separating base station tower images from background images through image segmentation, feature extraction can be avoided in areas lacking useful information, thereby improving the efficiency of subsequent image processing and the accuracy of image feature extraction and image classification matching.

[0065] Optionally, image segmentation can employ a threshold-based image segmentation method, an edge-based image segmentation method, a region-based image segmentation method, or a variable model-based image segmentation method.

[0066] Then, after obtaining the initial base station tower image, image enhancement can be performed on the initial base station tower image to obtain an enhanced base station tower image.

[0067] Image enhancement refers to filtering the initial base station tower image by selecting an appropriate filtering algorithm. The main purpose is to remove the stuck, broken, and blurred parts of the image while keeping the edges of the texture in the initial base station tower image as intact as possible, thereby improving the image quality.

[0068] Alternatively, image enhancement can employ texture filtering, Fourier analysis, or wavelet analysis.

[0069] Furthermore, after obtaining the enhanced base station tower image, the enhanced base station tower image can be refined to obtain a refined base station tower image.

[0070] The main purpose of image thinning is to remove unnecessary line width information in the enhanced base station tower image, thereby reducing the amount of data in the enhanced base station tower image and making the connection structure simpler and clearer. This facilitates the extraction of detailed features from the base station tower image in the future, and ultimately improves the processing efficiency of the base station tower image.

[0071] Alternatively, image thinning can be performed using a serial thinning method, a parallel thinning method, or a hybrid thinning method.

[0072] Furthermore, after obtaining the refined base station tower image, the refined base station tower image can be binarized to obtain a binarized base station tower image.

[0073] Among them, image binarization refers to transforming a refined base station tower image into a binary image that uses only two values ​​to represent target information and background information by setting a threshold. The specific processing procedure can be shown in the following formula (1):

[0074]

[0075] Where f(i,j) represents the image value in the i-th row and j-th column of the refined base station tower image; T is the threshold.

[0076] Optionally, image binarization can employ either global thresholding or local thresholding. Global thresholding involves applying a single threshold to the entire refined base station tower image for binarization; for example, a thresholding method based on amplitude histograms can be used. Local thresholding involves dividing the refined base station tower image into several smaller blocks and setting a suitable threshold for each block, which can effectively improve the quality of the binarized base station tower image; for example, a nonlinear dynamic window thresholding method can be used.

[0077] After obtaining the binarized image of the base station tower, the binarized image of the base station tower is input into the image recognition model for subsequent image feature extraction and image classification matching.

[0078] The base station inspection video processing method provided by the present invention preprocesses the base station inspection images by performing image segmentation, image enhancement, image thinning and binarization, and then inputs the obtained binarized base station tower image into the image recognition model. This can reduce the algorithm complexity of the image recognition model and effectively improve the recognition efficiency and accuracy of the image recognition model.

[0079] Based on the above embodiments, as an optional embodiment, the input layer of the image recognition model extracts target features related to the shooting content and / or shooting parameters from all image features of the binary image of the base station tower, so as to generate recognition results.

[0080] Target features related to the captured content include image features related to the target base station tower and the number of base stations installed on the target base station tower.

[0081] Target features related to shooting parameters include image features related to lighting angle, sharpness, and shooting pitch angle related to the shooting of base station inspection images.

[0082] Specifically, when the binarized image of the base station tower is input into the image recognition model, the input layer of the image recognition model will extract target features related to the shooting content and target features related to the shooting parameters from all image features of the binarized image of the base station tower.

[0083] Among them, the target features related to the shooting content mainly include image features related to the target base station tower and the number of base stations set on the target base station tower; the target features related to the shooting parameters mainly include image features related to the light angle, clarity, and shooting pitch angle related to the shooting base station inspection images.

[0084] By acquiring target features related to the shooting content and shooting parameters, image classification and matching can be used to determine the tower type of the target base station tower and the number of base stations installed on the target base station tower.

[0085] This can be achieved by comparing the extracted image features of the current input base station inspection image with a pre-saved set of template image features. The similarity between the two images can be used to determine whether they are consistent. If they are inconsistent, they can be compared with the next set of template image features.

[0086] By iteratively executing the above steps, until a template image feature matching the image feature of the currently input base station inspection image is obtained, the tower type of the target base station tower and the number of base stations set on the target base station tower are determined.

[0087] In image matching, detail matching can be used. The set of detail feature vectors for the template image and the set of detail feature vectors for the input base station inspection image can be represented as follows:

[0088] P = {F i p |i=1,...,M};Q={F j p |j=1,...,N}

[0089] The template image feature vector set P contains M detail features; the input base station inspection image feature vector set Q contains N detail features.

[0090] Furthermore, based on the above representation, by searching for the best correspondence between all detailed features in set P and set Q, the matching score MS obtained under the best correspondence is compared with a pre-set threshold R. If the matching score MS≥R, the template image features are considered to match the image features of the base station inspection image; otherwise, they are considered not to match.

[0091] Optionally, the above detail matching method can adopt the SIFT (Scale-Invariant Feature Transform) feature matching method.

[0092] The base station inspection video processing method provided by this invention extracts target features related to the shooting content and target features related to the shooting parameters in the input layer of the image recognition model, so as to facilitate subsequent image feature classification and matching, improve image recognition accuracy, and achieve better identification of the tower type and number of base stations of the target base station tower.

[0093] Based on the above embodiments, as an optional embodiment, the input layer of the image recognition model extracts target features related to the captured content and / or shooting parameters from all image features of the binary image of the base station tower. This is achieved based on one of the following analysis methods:

[0094] Principal component analysis, linear discriminant analysis, unsupervised discriminant feature extraction based on divergence difference, principal component analysis based on image matrix, nonlinear feature extraction, and Gabor feature extraction algorithm based on feature weighting.

[0095] Specifically, the input layer of the image recognition model extracts target features related to the shooting content and / or shooting parameters from all image features of the binary image of the base station tower. This can be achieved through any of the following methods: principal component analysis, linear discriminant analysis, unsupervised discriminant feature extraction based on divergence difference, principal component analysis based on image matrix, nonlinear feature extraction, or Gabor feature extraction algorithm based on feature weighting.

[0096] Among them, Principal Component Analysis (PCA) can determine the target features by calculating the covariance matrix of all image features in the binarized image of the base station tower, which can effectively improve the speed of image feature extraction.

[0097] The basic idea of ​​Linear Discriminant Analysis (LDA) is to take the vector where the Fisher criterion function reaches its extreme value as the optimal projection direction, and project all image features of the binarized base station tower image onto this direction, so as to maximize the inter-class dispersion and minimize the intra-class dispersion. This can be used to reduce the dimensionality of the features, thereby improving the accuracy of base station tower image classification and matching.

[0098] The unsupervised discriminative feature extraction method based on divergence difference is based on the large divergence difference criterion, which can effectively reduce the algorithm complexity and has strong robustness and stability. It is not easily affected by noise and outliers, so it can help improve the image feature extraction efficiency of binary images of base station towers.

[0099] Principal component analysis based on image matrices mainly reduces the dimensionality of the image matrix in the binarized image of the base station tower, thereby reducing the data dimensionality while retaining the main information of the binarized image of the base station tower.

[0100] Nonlinear feature extraction can effectively overcome the problem of nonlinear features being difficult to process. It can map the original data in the binary image of the base station tower to a new high-dimensional space through a mapping function, making the original data easier to process in the new space.

[0101] The Gabor feature extraction algorithm based on feature weighting weights the Gabor feature vectors according to the degree of dispersion of their neighboring components. It has strong robustness and class representation ability, which can effectively reduce the error rate of image recognition and thus make the recognition results of the image recognition model more accurate.

[0102] Alternatively, image feature extraction can also be performed using algorithms based on local detail features. For example, feature extraction can be accomplished using the Local Binary Patterns (LBP) algorithm, the steps of which are as follows:

[0103] 1. Select a pixel, as well as the size and shape of its neighborhood.

[0104] 2. Compare the value of the neighboring pixels with the value of the center pixel. Mark the value that is larger than the center pixel as 1, and otherwise mark it as 0.

[0105] 3. Based on the binary encoding result, convert it into a decimal number as the LBP feature value of the pixel.

[0106] 4. Repeat the above steps to process each pixel in the image to obtain the LBP feature representation of the entire image.

[0107] By traversing all image feature points (i.e. pixels) of the binarized image of the base station tower, the type and location of each feature point are determined, which facilitates the subsequent extraction of target features related to the shooting content and shooting parameters.

[0108] The base station inspection video processing method provided by this invention extracts image features by employing any one of the following methods: principal component analysis, linear discriminant analysis, unsupervised discriminant feature extraction based on divergence difference, principal component analysis based on image matrix, nonlinear feature extraction, or Gabor feature extraction algorithm based on feature weighting. This method can effectively extract target features related to the shooting content and shooting parameters from all image features of the binary image of the base station tower, facilitating subsequent image feature classification and matching. It can also improve the speed of image feature extraction and reduce the recognition time required by the image recognition model.

[0109] Based on the above embodiments, as an optional embodiment, image segmentation is performed on base station inspection images, including:

[0110] When the image size of the base station inspection image is determined to be smaller than the preset image size, one of the threshold-based image segmentation method and the edge-based image segmentation method is used to segment the base station inspection image.

[0111] When the image size of the inspection image is determined to be greater than or equal to the preset image size, a region-based image segmentation method is used to segment the base station inspection image.

[0112] Specifically, when performing image segmentation on base station inspection images, different image segmentation methods can be determined based on the comparison results by comparing the image size of the base station inspection image with a preset image size. The preset image size can be pre-set according to the specific usage requirements of the scenario.

[0113] When the image size of the base station inspection image is determined to be smaller than a preset image size, one of the following image segmentation methods can be used to segment the base station inspection image: a threshold-based image segmentation method or an edge-based image segmentation method. The threshold-based image segmentation method can be any one of the following: histogram method, histogram transform method, maximum class space variance method, minimum error method, and mean error method. The edge-based image segmentation method can be any one of the following: Canny edge detection method, multi-scale method, multi-resolution method, and boundary curve fitting method.

[0114] When the image size of the base station inspection image is determined to be greater than or equal to a preset image size, a region-based image segmentation method can be used to segment the base station inspection image. This region-based image segmentation method can be any of the following: region growing method, region splitting method, or split-merge method.

[0115] The base station inspection video processing method provided by the present invention compares the image size of the base station inspection image with the preset image size, and then selects a more suitable image segmentation method based on the comparison result. This not only saves subsequent image processing time, but also effectively improves the reliability of image segmentation.

[0116] Based on the above embodiments, as an optional embodiment, image enhancement is performed on the initial base station tower image, including:

[0117] Image enhancement is performed on the initial base station tower image using either Gabor image filtering or Fourier filtering-based low-quality image enhancement methods.

[0118] Specifically, when performing image enhancement on the initial base station tower image, either Gabor-based image filtering or low-quality image enhancement based on Fourier filtering can be used.

[0119] Among them, the Gabor-based image filtering method mainly uses Gabor filters, and the definition of a Gabor filter is shown in formula (2):

[0120]

[0121] Where G(x,y) represents the spatial scale factor; δx represents the spatial scale factor representing the horizontal orientation; δy represents the spatial scale factor representing the vertical orientation; i represents the center frequency; w x Indicates location.

[0122] If Gabor filters are to be used for image enhancement, the Gabor function needs to be changed to a digital filter, as shown in formula (3):

[0123]

[0124] Where G'(u,v) represents a specific spatial frequency; f x This represents the frequency parameter.

[0125] Although the time required for frequency calculation and filtering calculation in the Gabor image filtering method accounts for a large proportion of the total time required for the preprocessing of base station inspection images, the filtering effect is good and can effectively improve image quality, thus facilitating subsequent image processing.

[0126] In addition, the low-quality image enhancement method based on Fourier filtering mainly involves converting the initial base station tower image into a frequency domain image and using a high-pass filter to remove low-frequency information from the frequency domain image, since low-frequency information is often the blur and noise in the image.

[0127] The filtered frequency domain image is subjected to inverse Fourier transform to convert it into a filtered base station tower image. Furthermore, a histogram equalization algorithm can be used to enhance the contrast and brightness of the base station tower image, thereby obtaining an enhanced base station tower image.

[0128] The base station inspection video processing method provided by the present invention enhances the initial base station tower image by using a Gabor image filtering method or a low-quality image enhancement method based on Fourier filtering. This effectively removes noise and blur from the initial base station tower image, thereby improving the image quality of the initial base station tower image.

[0129] Based on the above embodiments, as an optional embodiment, image refinement is performed on the enhanced base station tower image, including:

[0130] Based on the OPTA thinning algorithm, image thinning is performed on the enhanced base station tower image.

[0131] Specifically, when performing image thinning on enhanced base station tower images, the OPTA thinning algorithm can be used to complete the image thinning.

[0132] Among them, the OPTA thinning algorithm is a serial thinning algorithm. This algorithm can ensure that the ridge width in the enhanced base station tower image is a single pixel. In addition, the degree and effect of image thinning can be adjusted by adjusting the parameter settings in the OPTA thinning algorithm, so as to effectively remove unnecessary pixels in the enhanced base station tower image and obtain more accurate shape features and contour features.

[0133] The base station inspection video processing method provided by the present invention uses the OPTA thinning algorithm to thin the enhanced base station tower image, thereby effectively removing unnecessary pixels in the enhanced base station tower image to obtain more accurate shape and contour features.

[0134] Based on the above embodiments, as an optional embodiment, obtaining base station inspection images of the target base station tower recorded in the maintenance work order includes:

[0135] Based on the work order number of the maintenance work order, digital compressed images are extracted from the database. The digital compressed images are multi-angle images generated by extracting frames from the base station inspection video of the target base station tower, compressing the images, and then uploading and storing them in the database.

[0136] Image restoration is performed on the digitally compressed image to obtain base station inspection images.

[0137] Specifically, during image transmission, due to the large amount of image information, it is necessary to compress the image information during storage and transmission.

[0138] Therefore, after capturing a base station inspection video of the target base station tower and generating multi-angle images through frame extraction, the images can be compressed, labeled with the corresponding maintenance work order number, and stored in the database of the work order management system.

[0139] Image compression can employ any of the following methods: wavelet transform coding, fractal coding, and model-based coding. This can effectively save image space, reduce transmission losses, and improve database storage efficiency.

[0140] When the corresponding digital compressed image is extracted from the database based on the work order number of the maintenance work order, image compression will cause the image quality to degrade or degrade. Therefore, it is necessary to restore the digital compressed image to improve the image. This process of reconstructing or restoring the degraded digital compressed image is to improve the image fidelity and obtain the base station inspection image used as input to the image recognition model.

[0141] Image restoration can be performed using wavelet transform time-domain analysis. Applying wavelet analysis to image restoration can provide timely feedback of image feature values ​​to guide the image restoration operation, and it has good localization properties in both the time and frequency domains.

[0142] The base station inspection video processing method provided by this invention compresses images of the target base station tower to obtain digital compressed images, which are then used for image transmission and database storage. The extracted digital compressed images are then restored to base station inspection images through image reconstruction, so that they can be input into an image recognition model to obtain recognition results. This method can effectively save storage space for digital compressed images and improve the transmission efficiency of digital compressed images, thereby improving the working efficiency of the base station inspection video processing method.

[0143] Figure 2 This is a schematic diagram of the base station inspection video processing device provided by the present invention, as shown below. Figure 2 As shown, it mainly includes: an image acquisition unit 21, an image recognition unit 22, an information acquisition unit 23, and an information comparison unit 24, wherein:

[0144] Image acquisition unit 21 is used to acquire base station inspection videos of the target base station tower from multiple angles as recorded in the maintenance work order, and to perform frame extraction processing on the base station inspection videos to acquire base station inspection images.

[0145] Image recognition unit 22 is used to input the base station inspection image into the image recognition model and obtain the recognition result output by the image recognition model. The recognition result includes the tower type of the target base station tower and the number of base stations set on the target base station tower.

[0146] The information acquisition unit 23 is used to retrieve the resource information of the target base station tower from the identity identifier of the target base station tower recorded in the maintenance work order. The resource information includes the tower registration type and the number of base station registrations.

[0147] The information comparison unit 24 is used to compare the tower type with the tower registration type and the number of base stations with the number of base station registrations, so as to generate inspection results for the target base station tower.

[0148] It should be noted that the base station inspection video processing device provided in this embodiment of the invention can execute the base station inspection video processing method described in any of the above embodiments during specific operation, and this embodiment will not elaborate on this.

[0149] The base station inspection video processing device provided by the present invention extracts frames of base station inspection video into images and inputs them into an image recognition model to obtain the tower type and number of base stations of the target base station tower. It then compares these images with the tower registration type and number of base stations of the target base station tower stored in the background to intelligently identify whether there are any abnormalities in the inspection results. This effectively overcomes the problems of excessive personal subjectivity and large workload in manual identification, and effectively improves the efficiency and accuracy of inspection result identification.

[0150] Figure 3 This is a schematic diagram of the structure of the electronic device provided by the present invention, such as... Figure 3 As shown, the electronic device may include: a processor 310, a communication interface 320, a memory 330, and a communication bus 340, wherein the processor 310, the communication interface 320, and the memory 330 communicate with each other through the communication bus 340. The processor 310 can call logic instructions in the memory 330 to execute a base station inspection video processing method. This method includes: acquiring multi-angle base station inspection videos of the target base station tower recorded in the maintenance work order; performing frame extraction processing on the base station inspection videos to obtain base station inspection images; inputting the base station inspection images into an image recognition model to obtain recognition results output by the image recognition model, the recognition results including the tower type of the target base station tower and the number of base stations installed on the target base station tower; retrieving the resource information of the target base station tower from the identification identifier of the target base station tower recorded in the maintenance work order, the resource information including the tower registration type and the number of registered base stations; comparing the tower type with the tower registration type, and the number of base stations with the number of registered base stations, to generate inspection results for the target base station tower.

[0151] Furthermore, the logical instructions in the aforementioned memory 330 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0152] On the other hand, the present invention also provides a computer program product, the computer program product including a computer program stored on a non-transitory computer-readable storage medium, the computer program including program instructions, when the program instructions are executed by a computer, the computer can execute the base station inspection video processing method provided in the above embodiments, the method including: acquiring base station inspection videos of the target base station tower from multiple angles recorded in the maintenance work order; performing frame extraction processing on the base station inspection video to acquire base station inspection images; inputting the base station inspection images into an image recognition model to acquire recognition results output by the image recognition model, the recognition results including the tower type of the target base station tower and the number of base stations installed on the target base station tower; retrieving the resource information of the target base station tower from the identity identifier of the target base station tower recorded in the maintenance work order, the resource information including the tower registration type and the number of base stations registered; comparing the tower type with the tower registration type, and the number of base stations with the number of base stations registered, to generate inspection results for the target base station tower.

[0153] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon. When executed by a processor, the computer program performs the base station inspection video processing method provided in the above embodiments. The method includes: acquiring base station inspection videos of a target base station tower from multiple angles as recorded in a maintenance work order; performing frame extraction processing on the base station inspection videos to obtain base station inspection images; inputting the base station inspection images into an image recognition model to obtain a recognition result output by the image recognition model, the recognition result including the tower type of the target base station tower and the number of base stations installed on the target base station tower; retrieving the resource information of the target base station tower from the identification of the target base station tower recorded in the maintenance work order, the resource information including the tower registration type and the number of base stations registered; comparing the tower type with the tower registration type and the number of base stations with the number of base stations registered to generate an inspection result for the target base station tower.

[0154] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0155] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0156] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for processing base station inspection video, characterized in that, include: Obtain multi-angle base station inspection videos of the target base station tower recorded in the maintenance work order, perform frame extraction processing on the base station inspection videos, and obtain base station inspection images. The base station inspection image is input into the image recognition model to obtain the recognition result output by the image recognition model. The recognition result includes the tower type of the target base station tower and the number of base stations installed on the target base station tower. From the identity identifier of the target base station tower recorded in the maintenance work order, retrieve the resource information of the target base station tower, including the tower registration type and the number of base station registrations; The tower type is compared with the tower registration type, and the number of base stations is compared with the number of base stations registered, to generate an inspection result for the target base station tower. This includes: if at least one set of comparison results between the tower type and the tower registration type, and between the number of base stations and the number of base stations registered, is inconsistent, then the inspection result of the target base station tower is determined to be abnormal. Inputting the base station inspection images into the image recognition model includes: The base station inspection images are segmented to obtain initial base station tower images after removing background images; Image enhancement is performed on the initial base station tower image to obtain an enhanced base station tower image; The enhanced base station tower image is thinned to obtain a thinned base station tower image; The refined base station tower image is binarized to obtain a binarized base station tower image; The binarized image of the base station tower is input into the image recognition model.

2. The base station inspection video processing method according to claim 1, characterized in that, The input layer of the image recognition model extracts target features related to the shooting content and / or shooting parameters from all image features of the binary image of the base station tower, so as to generate the recognition result; Target features related to the captured content include image features related to the target base station tower and the number of base stations installed on the target base station tower; The target features related to the shooting parameters include image features related to the light angle, sharpness, and shooting pitch angle related to the shooting of the base station inspection images.

3. The base station inspection video processing method according to claim 2, characterized in that, The input layer of the image recognition model extracts target features related to the captured content and / or shooting parameters from all image features of the binary image of the base station tower. This is achieved based on one of the following analysis methods: Principal component analysis, linear discriminant analysis, unsupervised discriminant feature extraction based on divergence difference, principal component analysis based on image matrix, nonlinear feature extraction, and Gabor feature extraction algorithm based on feature weighting.

4. The base station inspection video processing method according to claim 1, characterized in that, The image segmentation of the base station inspection images includes: When it is determined that the image size of the base station inspection image is smaller than the preset image size, one of the threshold-based image segmentation method and the edge-based image segmentation method is used to segment the base station inspection image. When the image size of the inspection image is determined to be greater than or equal to the preset image size, a region-based image segmentation method is used to segment the base station inspection image.

5. The base station inspection video processing method according to claim 1, characterized in that, The image enhancement of the initial base station tower image includes: The initial base station tower image is enhanced using either Gabor-based image filtering or a low-quality image enhancement method based on Fourier filtering.

6. The base station inspection video processing method according to claim 1, characterized in that, The image refinement of the enhanced base station tower image includes: The enhanced base station tower image is thinned based on the OPTA thinning algorithm.

7. The base station inspection video processing method according to claim 1, characterized in that, The acquisition of base station inspection images of the target base station tower recorded in the maintenance work order includes: Based on the work order number of the maintenance work order, a digital compressed image is extracted from the database; the digital compressed image is generated by extracting frames from the base station inspection video of the target base station tower, compressing the images, and then uploading and storing them in the database. The digitally compressed image is restored to obtain the base station inspection image.

8. A base station inspection video processing device, characterized in that, include: The image acquisition unit is used to acquire base station inspection videos from multiple angles of the target base station tower recorded in the maintenance work order, and to perform frame extraction processing on the base station inspection videos to acquire base station inspection images. An image recognition unit is used to input the base station inspection image into an image recognition model and obtain the recognition result output by the image recognition model. The recognition result includes the tower type of the target base station tower and the number of base stations installed on the target base station tower. The process of inputting the base station inspection image into an image recognition model includes: performing image segmentation on the base station inspection image to obtain an initial base station tower image after removing the background image; performing image enhancement on the initial base station tower image to obtain an enhanced base station tower image; performing image thinning on the enhanced base station tower image to obtain a thinned base station tower image; performing binarization processing on the thinned base station tower image to obtain a binarized base station tower image; and inputting the binarized base station tower image into the image recognition model. The information acquisition unit is used to retrieve the resource information of the target base station tower from the identity identifier of the target base station tower recorded in the maintenance work order. The resource information includes the tower registration type and the number of base station registrations. An information comparison unit is used to compare the tower type with the tower registration type and the number of base stations with the number of base stations registered, in order to generate an inspection result for the target base station tower, including: if at least one set of comparison results between the tower type and the tower registration type and between the number of base stations and the number of base stations registered is inconsistent, then the inspection result of the target base station tower is determined to be abnormal.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the base station inspection video processing method as described in any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the base station inspection video processing method as described in any one of claims 1 to 7.

11. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the base station inspection video processing method as described in any one of claims 1 to 7.

Images