Data management method and system of refining heating furnace monitoring system

Abstract

The embodiment of the invention provides a data management method and system for a refining heating furnace monitoring system, and belongs to the technical field of data management. The method is characterized by comprising the following steps: in response to an equipment abnormity judgment signal, carrying out abnormal image anchoring on image information which is judged to be equipment abnormity in an image sequence; performing specific abnormal position positioning of the equipment in the anchored abnormal image, and performing abnormal image slicing based on an abnormal position positioning result to obtain a plurality of sub-images; background sub-image filtering is carried out on the plurality of sub-images, the remaining sub-images are used as effective images, and the effective images are stored in a preset image library; and in response to the preset image library access signal, performing effective image pushing based on the access target. According to the scheme, the problems of high system construction cost and low data access efficiency of the data management scheme of the existing refining heating furnace monitoring system are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of data management, and in particular to a data management method for a refining heating furnace monitoring system and a data management system for a refining heating furnace monitoring system. Background Art

[0002] As an important equipment in a refining enterprise, the operation stability of the refining heating furnace directly affects the operation stability of the refining enterprise. In order to ensure the stable operation of the refining heating furnace, it is necessary to monitor the operation information of the refining heating furnace in real time and issue a timely alarm when an abnormal operation is found to avoid developing into a larger technical failure and causing the entire system to crash. In the existing scheme, in order to fully perceive the operating status of the refining heating furnace and accurately discover the hidden dangers of equipment failure, it is necessary to set up a large number of sensors in the refining heating furnace to collect real-time operation information of the refining heating furnace. Among them, the combustion flame state monitoring of the refining heating furnace is a very important part. With the development of neural network technology, the efficiency and accuracy of the fault diagnosis scheme based on feature extraction have been guaranteed. The combustion flame state can be judged by the flame image features. Combined with neural network technology, the combustion flame state of the refining heating furnace can be quickly and accurately detected.

[0003] In order to ensure the accuracy of detection, a large amount of flame image information needs to be collected, and in order to read historical data later to meet the needs of fault tracing, such a large amount of image information data needs to occupy a large amount of memory space. However, in order to meet the needs, the data cannot be cleaned up at will, which makes the storage system of the existing system have to further expand its capacity, increase the construction cost, and also affect the data access efficiency. In view of the problems of high system construction cost and low data access efficiency in the data management solution of the existing refining heating furnace monitoring system, it is necessary to create a new data management solution for the refining heating furnace monitoring system. Summary of the invention

[0004] The purpose of the embodiments of the present invention is to provide a data management method and system for a refining heating furnace monitoring system, so as to at least solve the problems of high system construction cost and low data access efficiency existing in the data management scheme of the refining heating furnace monitoring system.

[0005] In order to achieve the above-mentioned purpose, the first aspect of the present invention provides a data management method for a refining heating furnace monitoring system, the method comprising: in response to an equipment abnormality determination signal, anchoring an abnormal image in an image sequence for image information determined to be equipment abnormal; locating the specific abnormal position of the equipment in the anchored abnormal image, and slicing the abnormal image based on the abnormal position location result to obtain multiple sub-images; filtering the background sub-images of the multiple sub-images, taking the remaining sub-images as valid images, and storing the valid images in a preset image library; in response to a preset image library access signal, pushing the valid image based on the access target.

[0006] Optionally, the triggering rule of the equipment abnormality judgment signal is: based on the image information of the refining and heating furnace collected in real time by the monitoring system, a preset fault diagnosis model is trained to obtain training results; based on the training results, it is determined whether there is an abnormality in the current refining and heating furnace, and if so, the equipment abnormality judgment signal is triggered.

[0007] Optionally, the abnormal image anchoring is performed on the image information determined as equipment abnormality in the image sequence, including: determining the input image indicated as abnormal by the training result as an abnormal image; and anchoring the corresponding image frame in the video image collected by the monitoring device based on the timestamp of the abnormal image as the abnormal image anchoring result.

[0008] Optionally, the specific abnormal position of the device is located in the anchored abnormal image, including: taking the abnormal area selected by the training result in the anchor image as a candidate abnormal position, and marking the position of each candidate abnormal position to obtain the marked image information; pushing the marked image information to the user end, and collecting the user's confirmation information based on the user end; determining or deleting the candidate abnormal position based on the confirmation information, and obtaining the image information in which the abnormal position is located.

[0009] Optionally, slicing the abnormal image based on the abnormal position positioning result to obtain multiple sub-images includes: binarizing the original image of the abnormal image to obtain a binarized image; performing edge detection on the binarized image, and outlining the contours of each area based on the edge detection result to obtain a background image; projecting the image with the abnormal position positioned and the background image to obtain a contour image of the abnormal area; slicing the abnormal image based on the contour image of the abnormal area to obtain multiple sub-images.

[0010] Optionally, the abnormal image is sliced ​​based on the contour image of the abnormal area to obtain multiple sub-images, including: taking other areas except the contour image of the abnormal area as background areas; performing current image slicing based on the contour image of the abnormal area, and slicing the current image into corresponding abnormal area sub-images and background area sub-images.

[0011] Optionally, background sub-images of multiple sub-images are filtered, and the remaining sub-images are used as valid images, including: filtering all background area sub-images, and using the remaining abnormal area sub-images as initial valid images; recording the field of view information of the current image; wherein the field of view information includes viewing angle information and focal length information; establishing an association relationship between the initial valid image and the field of view information of the current image, and using the initial valid image after the association is completed as the valid image.

[0012] Optionally, the preset image library access signal includes: a user call signal and a system cleanup signal; wherein, the user call signal is a data call signal actively triggered by the user end; and the system cleanup signal is a data cleanup signal actively triggered when the preset image library meets the system cleanup rules.

[0013] Optionally, if the current access signal is a user call signal, the pushing of effective images based on the access target includes: determining effective images based on the access target, and extracting effective images from the preset image library; based on the extracted effective images, reading the associated field of view information; based on the read field of view information, extracting the current field of view image template, and combining the extracted effective image with the extracted current field of view image template to obtain a combined image; and pushing the combined image to the user end.

[0014] Optionally, if the current access signal is a system cleanup signal, the effective image push based on the access target includes: based on the system cleanup signal triggering moment, marking the data before a predetermined period of the moment as cleanup data; packaging and encrypting the cleanup data to obtain a data packet; uploading the data packet to a host computer for storage, and deleting the data packet and original data locally.

[0015] The second aspect of the present invention provides a data management system for a refining heating furnace monitoring system, the system comprising: an acquisition unit, for responding to an equipment abnormality determination signal, and anchoring an abnormal image in an image sequence for image information determined to be equipment abnormal; a processing unit, for locating the specific abnormal position of the equipment in the anchored abnormal image, and slicing the abnormal image based on the abnormal position location result to obtain multiple sub-images; a filtering unit, for filtering the background sub-images of the multiple sub-images, taking the remaining sub-images as valid images, and storing the valid images in a preset image library; an access unit, for responding to the preset image library access signal, and pushing valid images based on the access target.

[0016] On the other hand, the present invention provides a computer-readable storage medium having instructions stored thereon, which, when executed on a computer, enables the computer to execute the above-mentioned data management method for the refining heating furnace monitoring system.

[0017] Through the above technical scheme, the scheme of the present invention slices the images identified as abnormal, retains only the valid images, filters the background area, and greatly reduces the image data volume, thereby reducing the amount of data from the data source and alleviating the pressure on the storage system.

[0018] Other features and advantages of the embodiments of the present invention will be described in detail in the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The accompanying drawings are used to provide a further understanding of the embodiments of the present invention and constitute a part of the specification. Together with the following specific embodiments, they are used to explain the embodiments of the present invention, but do not constitute a limitation on the embodiments of the present invention. In the accompanying drawings:

[0020] Figure 1 It is a flowchart of the steps of a data management method for a refining heating furnace monitoring system provided by one embodiment of the present invention;

[0021] Figure 2 It is a system structure diagram of a data management system of a refining heating furnace monitoring system provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0022] The specific implementation of the present invention is described in detail below in conjunction with the accompanying drawings. It should be understood that the specific implementation described here is only used to illustrate and explain the present invention, and is not used to limit the present invention.

[0023] As an important equipment in a refining enterprise, the operation stability of the refining heating furnace directly affects the operation stability of the refining enterprise. In order to ensure the operation stability of the refining heating furnace, it is necessary to monitor the operation information of the refining heating furnace in real time and issue an alarm in time when an abnormal operation is found to avoid the development of a larger technical failure and cause the entire system to shut down.

[0024] In the existing scheme, in order to fully perceive the operating status of the refining and heating furnace and accurately discover the hidden dangers of equipment failure, a large number of sensors need to be installed in the refining and heating furnace to collect real-time operating information of the refining and heating furnace. Among them, the combustion flame status monitoring of the refining and heating furnace is a very important part. With the development of neural network technology, the efficiency and accuracy of the fault diagnosis scheme based on feature extraction have been guaranteed. The combustion flame status can be judged by the flame image features. Combined with neural network technology, the combustion flame status of the refining and heating furnace can be quickly and accurately detected.

[0025] In order to ensure the accuracy of detection, a large amount of flame image information needs to be collected, and in order to read historical data later to meet the needs of fault tracing, such a large amount of image information data requires a large amount of memory space. However, in order to meet the needs, data cannot be cleaned up at will, which makes the storage system of the existing system have to further expand its capacity, which increases the construction cost and affects the efficiency of data access.

[0026] In view of the problems of high system construction cost and low data access efficiency in the data management scheme of the existing refining heating furnace monitoring system, the present invention proposes a new data management method for the refining heating furnace monitoring system. The present invention performs slicing processing on images identified as abnormal, retains only valid images, filters the background area, and greatly reduces the image data volume, thereby reducing the data volume from the data source and alleviating the pressure on the storage system.

[0027] Figure 1 1 is a method flow chart of a data management method for a refining heating furnace monitoring system provided by an embodiment of the present invention. Figure 1 As shown, an embodiment of the present invention provides a data management method for a refining heating furnace monitoring system, the method comprising:

[0028] Step S10: In response to the device abnormality determination signal, abnormal image anchoring is performed in the image sequence for the image information determined as device abnormality.

[0029] Specifically, the triggering rule of the abnormality determination signal is: based on the image information of the refining and heating furnace collected in real time by the monitoring system, a preset fault diagnosis model is trained to obtain a training result; the training result determines whether there is an abnormality in the current refining and heating furnace, and if so, an abnormality determination signal is triggered.

[0030] In the embodiment of the present invention, the excellent performance of the existing neural network is utilized, and the scheme of the present invention trains the fault recognition model based on the neural network, thereby extracting the flame image features, and performing fault recognition based on the feature extraction results. When an abnormality is identified in a frame image, the image is marked as an abnormal image. In order to reduce the amount of image data, the scheme of the present invention needs to reduce the stored image information, which is achieved by reducing the stored target image on the one hand and reducing the size of the stored data on the other hand.

[0031] Embodiment 1:

[0032] The abnormal image anchoring is performed in the image sequence for the image information determined as equipment abnormality, including: determining the input image indicated as abnormal by the training result as the abnormal image; based on the timestamp of the abnormal image, anchoring the corresponding image frame in the video image collected by the monitoring device as the abnormal image anchoring result.

[0033] In the embodiment of the present invention, in order to reduce the storage target images and store only the abnormal images, it is necessary to identify and extract the corresponding abnormal image frames in the video data. Based on this, as long as the abnormality is identified, the current frame image is anchored in the video stream data, indicating that the image needs to be used as a storage target in the future. When processing and storing in the future, only the images with anchor marks need to be identified, thereby reducing the amount of images and avoiding excessive burden on the storage system.

[0034] Step S20: locating the specific abnormal position of the device in the anchored abnormal image, and slicing the abnormal image based on the abnormal position locating result to obtain multiple sub-images.

[0035] Specifically, the specific abnormal position of the device is located in the anchored abnormal image, including: taking the abnormal area selected by the training result in the anchored image as the candidate abnormal position, and marking the position of each candidate abnormal position to obtain the marked image information; pushing the marked image information to the user end, and collecting the user's confirmation information based on the user end; determining or deleting the candidate abnormal position based on the confirmation information, and obtaining the image information in which the abnormal position is located.

[0036] In an embodiment of the present invention, the scheme of the present invention performs abnormality identification through a fault diagnosis model. Because it is an abnormality identification performed through special diagnosis extraction, it will select the abnormal position based on the abnormal special diagnosis. The scheme of the present invention directly uses the training results and uses the abnormal area selected by the training results as the candidate abnormal position. However, this data is only the data automatically identified by the system, which may have certain errors. In order to avoid errors, the data needs to be pushed to the user end for the user to confirm. After the confirmation is passed, the accurate abnormal position positioning result is determined.

[0037] Embodiment 2:

[0038] In order to avoid the user having to confirm too much information, which increases the user's workload, it is preferred that if an abnormality occurs in the device, its abnormal state is the same within a certain period of time, that is, other image frames attached to an abnormal image frame may have little difference. When confirming the image push, only one confirmation message needs to be pushed within a sampling period, thereby reducing the image push amount and the user's confirmation workload.

[0039] Preferably, the abnormal image slicing based on the abnormal position positioning result to obtain multiple sub-images includes: binarizing the original image of the abnormal image to obtain a binarized image; performing edge detection in the binarized image, and outlining the contours of each area based on the edge detection result to obtain a background image; projecting the image after the abnormal position positioning and the background image to obtain a contour image of the abnormal area; slicing the abnormal image based on the contour image of the abnormal image to obtain multiple sub-images.

[0040] In an embodiment of the present invention, in a refining heating furnace, there are multiple nozzles in the furnace, and the nozzles will produce flames. The furnace is a closed space, so its background is pure black. Under such a simple background condition, only the flame characteristics of the nozzle have changed. That is, the background is relatively simple, and the foreground is only a flame image. The scheme of the present invention makes full use of this characteristic, distinguishes the foreground and the background, and slices the image based on the distinction result. Existing slicing schemes are often sliced ​​based on regular size information. In order to ensure that the saved data is valid data as much as possible, the scheme of the present invention directly slices the image based on the foreground (that is, the abnormal area), removes the background, and only retains the foreground information. When the user needs to extract valid information later, the single background information is directly used as the background of the extracted image, and the image is fused and pushed to the user. That is, only foreground information is stored in the system, and the background information will be stored as template information to ensure that a large amount of stored data does not have background information, thereby reducing the amount of data for each stored image.

[0041] Based on this, the solution of the present invention will identify the contour of the abnormal area so as to slice the image based on the contour and delete all invalid background information.

[0042] Specifically, the abnormal image is sliced ​​based on the contour image of the abnormal image to obtain multiple sub-images, including: all areas other than the contour image of all abnormal areas are used as background areas; based on the contour image of the abnormal area, the current image is sliced ​​into corresponding abnormal area sub-images and background area sub-images.

[0043] Step S30: filtering the background sub-images of the plurality of sub-images, taking the remaining sub-images as valid images, and storing the valid images into a preset image library.

[0044] Embodiment 3:

[0045] All background area sub-images are filtered, and the remaining abnormal area sub-images are used as initial valid images; the field of view information of the current image is recorded; wherein the field of view information includes viewing angle information and focal length information; an association relationship is established between the initial valid image and the field of view information of the current image, and the initial valid image after the association is completed is used as the valid image.

[0046] In an embodiment of the present invention, the corresponding background area is different under different field of view conditions. Therefore, corresponding template information is preset for various field of view information, so that when the image is subsequently merged, the template suitable for the current field of view can be extracted to ensure that the image information pushed to the user, even if it is a merged image, is consistent with the expected image information.

[0047] Step S40: In response to the preset image library access signal, effective image push is performed based on the access target.

[0048] Specifically, the preset image library access signal includes: a user call signal and a system cleanup signal; wherein the user call signal is a data call signal actively triggered by the user end; and the system cleanup signal is a data cleanup signal actively triggered when the preset image library meets the system cleanup rule.

[0049] Preferably, if the current access signal is a user call signal, the effective image push based on the access target includes: determining the effective image based on the access target, and extracting the effective image from the preset image library; based on the extracted effective image, reading its associated field of view information; based on the read field of view information, extracting the current field of view image template, and combining the extracted effective image with the extracted current field of view image template to obtain a combined image; and pushing the combined image to the user end.

[0050] Preferably, if the current access signal is a system cleanup signal, the effective image push based on the access target includes: based on the system cleanup signal triggering moment, marking the data before a predetermined period before the moment as cleanup data; packaging and encrypting the cleanup data to obtain a data packet; uploading the data packet to the host computer for storage, and deleting the data packet and the original data locally.

[0051] In an embodiment of the present invention, even if the amount of data is reduced, the storage system may still be full. In order to relieve the pressure on the storage system, data cleaning needs to be performed automatically to avoid full storage. Therefore, when a certain amount of data is stored and the cleaning rule is triggered, the data with the longest storage time will be automatically used as cleaning data. Because this part of data has been stored for too long, the possibility of being called is even lower. After compressing and encrypting this part of data, it is stored in the cloud, and the local data is deleted to relieve the pressure on the local storage system. Even if the user has a calling demand later, the data can be downloaded directly from the cloud, and after decompression, it is also combined with the template and pushed to the user. Enable users to query data throughout the entire data life cycle.

[0052] Figure 2 1 is a system structure diagram of a data management system for a refining heating furnace monitoring system provided by an embodiment of the present invention. Figure 2 As shown, an embodiment of the present invention provides a data management system for a refining heating furnace monitoring system, the system comprising:

[0053] The acquisition unit is used to respond to the device abnormality determination signal and anchor the abnormal image in the image sequence for the image information determined as the device abnormality.

[0054] Embodiment 4:

[0055] The triggering rule of the abnormality determination signal is: based on the image information of the refining and heating furnace collected in real time by the monitoring system, a preset fault diagnosis model is trained to obtain a training result; the training result determines whether there is an abnormality in the current refining and heating furnace, and if so, an abnormality determination signal is triggered.

[0056] In the embodiment of the present invention, the excellent performance of the existing neural network is utilized, and the scheme of the present invention trains the fault recognition model based on the neural network, thereby extracting the flame image features, and performing fault recognition based on the feature extraction results. When an abnormality is identified in a frame image, the image is marked as an abnormal image. In order to reduce the amount of image data, the scheme of the present invention needs to reduce the stored image information, which is achieved by reducing the stored target image on the one hand and reducing the size of the stored data on the other hand.

[0057] Specifically, the abnormal image anchoring is performed on the image information determined as equipment abnormality in the image sequence, including: determining the input image indicated as abnormal by the training result as the abnormal image; based on the timestamp of the abnormal image, anchoring the corresponding image frame in the video image collected by the monitoring device as the abnormal image anchoring result.

[0058] In the embodiment of the present invention, in order to reduce the storage target images and store only the abnormal images, it is necessary to identify and extract the corresponding abnormal image frames in the video data. Based on this, as long as the abnormality is identified, the current frame image is anchored in the video stream data, indicating that the image needs to be used as a storage target in the future. When processing and storing in the future, only the images with anchor marks need to be identified, thereby reducing the amount of images and avoiding excessive burden on the storage system.

[0059] The processing unit is used to locate the specific abnormal position of the device in the anchored abnormal image, and slice the abnormal image based on the abnormal position positioning result to obtain multiple sub-images.

[0060] Specifically, the specific abnormal position of the device is located in the anchored abnormal image, including: taking the abnormal area selected by the training result in the anchored image as the candidate abnormal position, and marking the position of each candidate abnormal position to obtain the marked image information; pushing the marked image information to the user end, and collecting the user's confirmation information based on the user end; determining or deleting the candidate abnormal position based on the confirmation information, and obtaining the image information in which the abnormal position is located.

[0061] In an embodiment of the present invention, the scheme of the present invention performs abnormality identification through a fault diagnosis model. Because it is an abnormality identification performed through special diagnosis extraction, it will select the abnormal position based on the abnormal special diagnosis. The scheme of the present invention directly uses the training results and uses the abnormal area selected by the training results as the candidate abnormal position. However, this data is only the data automatically identified by the system, which may have certain errors. In order to avoid errors, the data needs to be pushed to the user end for the user to confirm. After the confirmation is passed, the accurate abnormal position positioning result is determined.

[0062] In another possible implementation, in order to avoid the user having to confirm too much information, which increases the user's workload, preferably, if an abnormality occurs in the device, its abnormal state is the same within a certain period of time, that is, other image frames attached to an abnormal image frame may have little difference. When confirming the image push, only one confirmation message needs to be pushed within a sampling period, thereby reducing the image push amount and the user's confirmation workload.

[0063] Embodiment 5:

[0064] The method of slicing the abnormal image based on the abnormal position positioning result to obtain multiple sub-images includes: binarizing the original image of the abnormal image to obtain a binarized image; performing edge detection in the binarized image, and outlining the contours of each area based on the edge detection result to obtain a substrate image; projecting the image after the abnormal position positioning and the substrate image to obtain a contour image of the abnormal area; slicing the abnormal image based on the contour image of the abnormal image to obtain multiple sub-images.

[0065] In an embodiment of the present invention, in a refining heating furnace, there are multiple nozzles in the furnace, and the nozzles will produce flames. The furnace is a closed space, so its background is pure black. Under such a simple background condition, only the flame characteristics of the nozzle have changed. That is, the background is relatively simple, and the foreground is only a flame image. The scheme of the present invention makes full use of this characteristic, distinguishes the foreground and the background, and slices the image based on the distinction result. Existing slicing schemes are often sliced ​​based on regular size information. In order to ensure that the saved data is valid data as much as possible, the scheme of the present invention directly slices the image based on the foreground (that is, the abnormal area), removes the background, and only retains the foreground information. When the user needs to extract valid information later, the single background information is directly used as the background of the extracted image, and the image is fused and pushed to the user. That is, only foreground information is stored in the system, and the background information will be stored as template information to ensure that a large amount of stored data does not have background information, thereby reducing the amount of data for each stored image.

[0066] Based on this, the solution of the present invention will identify the contour of the abnormal area so as to slice the image based on the contour and delete all invalid background information.

[0067] Specifically, the abnormal image is sliced ​​based on the contour image of the abnormal image to obtain multiple sub-images, including: all areas other than the contour image of all abnormal areas are used as background areas; based on the contour image of the abnormal area, the current image is sliced ​​into corresponding abnormal area sub-images and background area sub-images.

[0068] The filtering unit is used to filter the background sub-images of the multiple sub-images, take the remaining sub-images as valid images, and store the valid images into a preset image library.

[0069] Specifically, all background area sub-images are filtered, and the remaining abnormal area sub-images are used as initial valid images; the field of view information of the current image is recorded; wherein the field of view information includes viewing angle information and focal length information; an association relationship between the initial valid image and the field of view information of the current image is established, and the initial valid image after the association is completed is used as the valid image.

[0070] In an embodiment of the present invention, the corresponding background area is different under different field of view conditions. Therefore, corresponding template information is preset for various field of view information, so that when the image is subsequently merged, the template suitable for the current field of view can be extracted to ensure that the image information pushed to the user, even if it is a merged image, is consistent with the expected image information.

[0071] The access unit is used to respond to the preset image library access signal and push effective images based on the access target.

[0072] Specifically, the preset image library access signal includes: a user call signal and a system cleanup signal; wherein the user call signal is a data call signal actively triggered by the user end; and the system cleanup signal is a data cleanup signal actively triggered when the preset image library meets the system cleanup rule.

[0073] Preferably, if the current access signal is a user call signal, the effective image push based on the access target includes: determining the effective image based on the access target, and extracting the effective image from the preset image library; based on the extracted effective image, reading its associated field of view information; based on the read field of view information, extracting the current field of view image template, and combining the extracted effective image with the extracted current field of view image template to obtain a combined image; and pushing the combined image to the user end.

[0074] Preferably, if the current access signal is a system cleanup signal, the effective image push based on the access target includes: based on the system cleanup signal triggering moment, marking the data before a predetermined period before the moment as cleanup data; packaging and encrypting the cleanup data to obtain a data packet; uploading the data packet to the host computer for storage, and deleting the data packet and the original data locally.

[0075] In an embodiment of the present invention, even if the amount of data is reduced, the storage system may still be full. In order to relieve the pressure on the storage system, data cleaning needs to be performed automatically to avoid full storage. Therefore, when a certain amount of data is stored and the cleaning rule is triggered, the data with the longest storage time will be automatically used as cleaning data. Because this part of data has been stored for too long, the possibility of being called is even lower. After compressing and encrypting this part of data, it is stored in the cloud, and the local data is deleted to relieve the pressure on the local storage system. Even if the user has a calling demand later, the data can be downloaded directly from the cloud, and after decompression, it is also combined with the template and pushed to the user. Enable users to query data throughout the entire data life cycle.

[0076] The embodiment of the present invention further provides a computer-readable storage medium, on which instructions are stored, which, when executed on a computer, enable the computer to execute the above-mentioned data management method for the refining heating furnace monitoring system.

[0077] Those skilled in the art will understand that all or part of the steps in the method for implementing the above-mentioned embodiments can be completed by instructing the relevant hardware through a program, and the program is stored in a storage medium, including several instructions for making a single-chip microcomputer, a chip or a processor (processor) perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0078] The optional embodiments of the present invention are described in detail above in conjunction with the accompanying drawings. However, the embodiments of the present invention are not limited to the specific details in the above embodiments. Within the technical concept of the embodiments of the present invention, the technical scheme of the embodiments of the present invention can be subjected to a variety of simple modifications, and these simple modifications all belong to the protection scope of the embodiments of the present invention. It should also be noted that the various specific technical features described in the above specific embodiments can be combined in any suitable manner without contradiction. In order to avoid unnecessary repetition, the embodiments of the present invention will not further describe various possible combinations.

[0079] In addition, various embodiments of the present invention may be arbitrarily combined, and as long as they do not violate the concept of the embodiments of the present invention, they should also be regarded as the contents disclosed in the embodiments of the present invention.

Claims

1. A data management method for a refining heating furnace monitoring system, characterized in that: The method comprises: In response to a device abnormality determination signal, anchoring abnormal images in an image sequence for image information determined to be device abnormal; Locating the specific abnormal position of the device in the anchored abnormal image, and slicing the abnormal image based on the abnormal position positioning result to obtain multiple sub-images; Perform background sub-image filtering on multiple sub-images, use the remaining sub-images as valid images, and store the valid images into a preset image library; In response to a preset image library access signal, effective image push is performed based on the access target.

2. The method according to claim 1, characterized in that: The triggering rule of the device abnormality determination signal is: Based on the image information of the refining heating furnace collected in real time by the monitoring system, a preset fault diagnosis model is trained to obtain training results; Based on the training results, determine whether there is any abnormality in the current refining and heating furnace. If so, trigger the equipment abnormality determination signal.

3. The method according to claim 2, characterized in that The anchoring of abnormal images in the image sequence for the image information determined to be abnormal for the device includes: determining an input image indicated as abnormal by the training result as an abnormal image; Based on the timestamp of the abnormal image, the corresponding image frame is anchored in the video image collected by the monitoring device as the abnormal image anchoring result.

4. The method according to claim 2, characterized in that: The specific abnormal position location of the device in the anchored abnormal image includes: The abnormal area selected by the training result in the anchor image is used as a candidate abnormal position, and each candidate abnormal position is annotated to obtain the annotated image information; Pushing the annotated image information to a user terminal, and collecting confirmation information from the user based on the user terminal; Based on the confirmation information, candidate abnormal positions are determined or deleted to obtain image information in which the abnormal position is located.

5. The method according to claim 1, characterized in that The abnormal image is sliced ​​based on the abnormal position positioning result to obtain multiple sub-images, including: Binarization is performed on the original image of the abnormal image to obtain a binarized image; Perform edge detection on the binarized image, and outline each area based on the edge detection result to obtain a background image; Performing projection mapping on the image where the abnormal position is located and the substrate image to obtain a contour image of the abnormal area; The abnormal image is sliced ​​based on the contour image of the abnormal area to obtain multiple sub-images.

6. The method according to claim 5, characterized in that The abnormal image is sliced ​​based on the contour image of the abnormal area to obtain a plurality of sub-images, including: The other areas except the contour image of the abnormal area are regarded as the background area; The current image is sliced ​​based on the contour image of the abnormal area, and the current image is sliced ​​into corresponding abnormal area sub-images and background area sub-images.

7. The method according to claim 6, characterized in that Perform background sub-image filtering on multiple sub-images and use the remaining sub-images as valid images, including: Filter all background area sub-images and use the remaining abnormal area sub-images as initial valid images; Recording the field of view information of the current image; wherein the field of view information includes viewing angle information and focal length information; An association relationship between the field of view information of the initial effective image and the current image is established, and the initial effective image after the association is completed is used as the effective image.

8. The method according to claim 7, characterized in that The preset image library access signal includes: a user call signal and a system cleanup signal; wherein, The user call signal is a data call signal actively triggered by the user end; The system cleaning signal is a data cleaning signal actively triggered by a preset image library when the system cleaning rule is met.

9. The method according to claim 8, characterized in that If the current access signal is a user call signal, the effective image push based on the access target includes: Determine a valid image based on the access target, and extract the valid image from the preset image library; Based on the extracted effective image, read its associated field of view information; Based on the read field of view information, extract the current field of view image template, and combine the extracted valid image with the extracted current field of view image template to obtain a combined image; The combined image is pushed to the user end.

10. The method according to claim 8, characterized in that If the current access signal is a system cleanup signal, the effective image push based on the access target includes: Based on the system cleaning signal triggering time, the data before the predetermined period of time is marked as cleaning data; Packing and encrypting the cleaned data to obtain a data packet; Upload the data packet to the host computer for storage, and delete the data packet and original data locally.

11. A data management system for a refining heating furnace monitoring system, characterized in that: The system comprises: A collection unit, for responding to a device abnormality determination signal, anchoring abnormal images in an image sequence for image information determined to be device abnormal; A processing unit, used to locate the specific abnormal position of the device in the anchored abnormal image, and slice the abnormal image based on the abnormal position location result to obtain multiple sub-images; A filtering unit, used for filtering the background sub-images of the multiple sub-images, taking the remaining sub-images as valid images, and storing the valid images into a preset image library; The access unit is used to respond to the preset image library access signal and push effective images based on the access target.

12. A computer-readable storage medium having instructions stored thereon, which, when executed on a computer, enables the computer to execute the data management method for a refining heating furnace monitoring system as claimed in any one of claims 1 to 10.

Images