Industrial pipeline camera image collecting and processing method applied to complex industrial system

Abstract

The invention discloses an industrial pipeline camera image acquisition and processing method applied to a complex industrial system, relates to the technical field of image acquisition and processing, and solves the technical problem that in the prior art, error detection and analysis cannot be carried out on camera angles of multiple camera positions when camera monitoring is carried out on an industrial pipeline. The method specifically comprises the steps that after receiving, an equipment end sends an execution instruction to a multi-angle error recognition unit, and the multi-angle error recognition unit carries out cooperation angle error detection on a camera installed in an industrial pipeline; after the point locations are determined to be qualified, the camera shooting influence evaluation unit performs camera shooting influence evaluation on each point location of the industrial pipeline, acquires real-time acquisition data and non-real-time acquisition data, and performs camera shooting influence evaluation according to data comparison; after image acquisition of the industrial pipeline is completed, the pipeline image analysis unit performs image analysis on the industrial pipeline; and after the image pipeline analysis unit sets a risk level, the multi-image influence detection unit performs influence detection on the sub-frame pictures with risk positions.

Description

Technical Field

[0001] The present invention relates to the technical field of image acquisition and processing, and specifically to an industrial pipeline camera image acquisition and processing method applied to complex industrial systems. Background Art

[0002] A complex industrial system refers to an industrial system composed of a large number of interconnected and interacting components and subsystems, with characteristics such as high complexity, nonlinearity, uncertainty, and dynamics; an industrial pipeline is a tubular facility used to transport various fluid media (such as gas, liquid, slurry, etc.) in industrial production, and is an important part of connecting various industrial equipment and devices to achieve the transmission of substances and energy.

[0003] However, in the prior art, when monitoring an industrial pipeline with a camera, it is impossible to detect and analyze the error of the camera angles of multiple positions, which reduces the efficiency of position coordination, and it is also impossible to evaluate the camera impact, so that the image processing cannot use qualified pictures as the processing object, increasing the working pressure of image processing. In addition, it is impossible to analyze the impact on image acquisition, reducing the working efficiency of image processing.

[0004] In view of the above technical defects, a solution is now proposed. Summary of the Invention

[0005] The purpose of the present invention is to solve the above-mentioned problems and propose an industrial pipeline camera image acquisition and processing method applied to complex industrial systems.

[0006] The purpose of the present invention can be achieved through the following technical solutions:

[0007] An industrial pipeline camera image acquisition and processing method applied to complex industrial systems, and the process of the image acquisition and processing method is as follows:

[0008] The image acquisition platform generates device control signals and sends them to the device end;

[0009] After receiving, the device end sends an execution instruction to the multi-angle error recognition unit, and the multi-angle error recognition unit detects the cooperation angle error of the cameras installed in the industrial pipeline;

[0010] After determining that the positions are qualified, the camera impact evaluation unit evaluates the camera impact on each position of the industrial pipeline, collects real-time acquisition data and non-real-time acquisition data, and conducts camera impact evaluation based on data comparison;

[0011] After completing the image acquisition of the industrial pipeline, the pipeline image analysis unit analyzes the industrial pipeline image;

[0012] After the image pipeline analysis unit sets the risk level, the multi-image impact detection unit performs impact detection on the sub-frame pictures at the risky positions.

[0013] As a preferred embodiment of the present invention, the process of cooperating with the angle error detection is as follows:

[0014] Obtain the installed camera positions in the industrial pipeline, analyze according to the coverage areas of each camera position, mark the overlapping coverage areas of multiple camera positions as multi-angle acquisition areas, and the camera corresponding to the current multi-angle acquisition area is the area cooperation acquisition end;

[0015] Take any position on the pipe surface within the multi-angle acquisition area as the recognition subject, analyze the acquisition image of the recognition subject by the area cooperation acquisition end, obtain the display surface area of the recognition subject in the acquisition image of the area cooperation acquisition end. If the display surface area exceeds the cumulative value of the surface areas of each surface of the recognition subject, it is inferred that the cooperation coverage of the current area cooperation acquisition end is satisfied. On the contrary, if the display surface area does not exceed the cumulative value of the surface areas of each surface of the recognition subject, it is inferred that the cooperation coverage of the current area cooperation acquisition end is not satisfied; when it is not satisfied, the position adjustment of the area cooperation acquisition end is required.

[0016] As a preferred embodiment of the present invention, when the cooperation coverage is satisfied, when the recognition subject operates at different vibration frequencies, further analyze the acquisition image of the area cooperation acquisition end, randomly select an acquisition picture and the display surface of the acquisition picture is lower than the set proportion threshold of the corresponding pipe surface area of the recognition subject, and mark the acquisition moment of the current acquisition picture as the error risk moment;

[0017] Obtain the acquisition pictures of the recognition subject except the current acquisition picture at the error risk moment. If the maximum deviation value of the display surface area in the currently obtained acquisition pictures exceeds the deviation threshold, or the proportion of the pictures with the display surface lower than the set value in the acquisition pictures exceeds the quantity proportion threshold, it is inferred that there is an acquisition error in the acquisition pictures at the current error risk moment;

[0018] If the maximum deviation value of the display surface area in the currently obtained acquisition pictures does not exceed the deviation threshold, and the proportion of the pictures with the display surface lower than the set value in the acquisition pictures does not exceed the quantity proportion threshold, it is inferred that there is no acquisition error in the acquisition pictures at the current error risk moment;

[0019] When the occurrence proportion of the error risk moment during the cooperation acquisition period of the area cooperation acquisition end is lower than the proportion duration threshold, it is determined that the setting of the internal camera position of the industrial pipeline is qualified.

[0020] As a preferred embodiment of the present invention, the process of camera impact evaluation is as follows:

[0021] The real-time collected data and the non-real-time collected data are respectively the continuous frequency decline span of the collected images of the recognition subject in the same multi-angle collection area during the real-time camera shooting process in the use stage of the industrial pipeline, and the proportion of the number of collection moments covered by the decline stage of the camera shooting environment parameters in the same multi-angle collection area during the non-real-time camera shooting process in the use stage of the industrial pipeline.

[0022] As a preferred embodiment of the present invention, if the real-time collected data exceeds the decline span threshold, or the non-real-time collected data exceeds the quantity proportion threshold, the current time period is marked as an inefficient camera shooting time period; if the real-time collected data does not exceed the decline span threshold and the non-real-time collected data does not exceed the quantity proportion threshold, the current time period is marked as an efficient camera shooting time period.

[0023] As a preferred embodiment of the present invention, if there is an intersection between the efficient camera shooting time period and the inefficient camera shooting time period, the collected images of the inefficient camera shooting time period are retained, and when fault detection is required, such as when the clarity is insufficient, they are deleted to reduce the image drop frame rate; if the inefficient camera shooting time period exceeds the set duration value, the current collected images are not used as the collected images for image processing.

[0024] As a preferred embodiment of the present invention, the image analysis process is as follows:

[0025] The images completed with acquisition detection by the device end are marked as processed images, the processed images are divided into several sub-frame pictures, and the sub-frame pictures are sorted according to the camera shooting acquisition time sequence, and color difference comparison is performed according to the pipeline display surface in the sub-frame pictures. The positions with color differences are marked as risk positions, and the area is determined according to the risk positions to cooperate with the acquisition end. The trajectory of the risk positions is marked as the risk form, and the predicted risk level is set according to the risk form of the sub-frame pictures;

[0026] By summarizing the sub-frame pictures of multiple angles of the risk positions according to the area to cooperate with the acquisition end, multiple-angle display surfaces of the risk trajectory are obtained according to the summarized sub-frame pictures, and the depth of the risk form is detected according to the multiple-angle display surfaces; if the numerical deviation of the corresponding risk form of the sub-frame picture display surface in the multi-angle depth detection is within the deviation range, it is inferred that the predicted risk level is fitted and marked as the set risk level, and it is sent to the image end and transferred to the image acquisition platform; the image acquisition platform makes pipeline maintenance decisions according to the set risk level;

[0027] If the numerical deviation of the corresponding risk form of the sub-frame picture display surface in the multi-angle depth detection is not within the deviation range, it is inferred that the predicted risk level is not fitted and the level is increased or decreased according to the actual deviation value. After the level adjustment is completed, it is marked as the set risk level and sent to the image end and transferred to the image acquisition platform.

[0028] As a preferred embodiment of the present invention, the multi-image impact detection unit operates as follows:

[0029] Screen the sorted sub-frame pictures in the processed image, use the sub-frame picture at the risk position when it first appears as the selection time point, and use the number threshold of sub-frame pictures before the selection time point as the risk picture set; and use the number threshold of sub-frame pictures after the selection time point as the impact picture set. Collect incentive acquisition data and impact acquisition data. If the incentive acquisition data exceeds the growth peak threshold, or the impact acquisition data exceeds the overlap duration threshold, mark it as risk generation information;

[0030] If the incentive acquisition data does not exceed the growth peak threshold and the impact acquisition data does not exceed the overlap duration threshold, mark it as risk-irrelevant information.

[0031] As a preferred embodiment of the present invention, the incentive acquisition data and the impact acquisition data are respectively the growth peak of the operating data showing a continuous growth trend of the industrial pipeline in the risk picture set, and the cumulative overlap duration of the risk form increase period corresponding to the risk position and the operating data floating period in the impact picture set.

[0032] Compared with the prior art, the beneficial effects of the present invention are:

[0033] 1. In the present invention, error detection and analysis are carried out through the camera angles of multiple positions, to infer whether the installation positions of the cameras in the current industrial pipeline are reasonably coordinated, avoiding errors in image acquisition by the cameras caused by improper setting of the position of the camera positions. When the image is not clear, cameras at different angles cannot cooperate to take pictures, resulting in a decrease in the image acquisition efficiency. Through error identification, the camera positions can be adjusted to improve the accuracy of image acquisition of the pipeline at the same position;

[0034] Through camera impact evaluation, infer whether there is an impact on the image acquisition of the current industrial pipeline. Through impact evaluation, the image acquisition efficiency at each moment can be inferred, so as to process the pictures at the moments without camera impact during image acquisition and processing, avoiding processing the images with low-efficiency shooting, increasing the workload of image processing, and at the same time affecting the accuracy of current image processing and recognition.

[0035] 2. In the present invention, it is inferred whether there are pipeline cracks in the industrial pipeline through industrial pipeline image analysis, improving the efficiency of image acquisition and detection in the industrial pipeline. Through multi-angle camera synchronous image processing for crack impact analysis, early warning and maintenance of the industrial pipeline are carried out in a timely manner, reducing the risk of pressure imbalance during the use of the industrial pipeline;

[0036] Perform multi-image impact analysis on the display surface of industrial pipelines through multiple sub-frame images, identify the impact on the risk patterns of industrial pipelines based on real-time acquired images, facilitate timely risk prevention and control during the use stage of industrial pipelines, and at the same time enable traceability based on impact detection and timely control to reduce the expansion of risk patterns; further improve the usage efficiency of industrial pipelines. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] For the convenience of those skilled in the art to understand, the present invention will be further described below with reference to the accompanying drawings.

[0038] Figure 1 It is the principle block diagram of the whole of the present invention;

[0039] Figure 2 It is the principle block diagram of the first embodiment of the present invention;

[0040] Figure 3 It is the principle block diagram of the second embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0041] In order to enable those skilled in the art to better understand the solution of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without making creative efforts belong to the scope of protection of the present invention.

[0042] Referring to "embodiment" herein means that the specific features, structures or characteristics described in connection with the embodiment may be included in at least one embodiment of the present invention. The phrase appears in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment mutually exclusive with other embodiments. Those skilled in the art will explicitly and implicitly understand that the embodiments described herein may be combined with other embodiments.

[0043] Please refer to Figure 1 As shown, an industrial pipeline camera image acquisition and processing method applied to a complex industrial system includes an image acquisition platform. It can be understood that the image acquisition platform acquires and analyzes images of industrial pipelines in the industrial system, and the image acquisition platform is composed of a device end and an image end; it should be explained that the device end represents the device integration for acquiring images of industrial pipelines in the industrial system, such as devices like cameras, and the image end represents the device integration for analyzing the images acquired by industrial pipelines in the industrial system, such as devices like image processors;

[0044] First Embodiment

[0045] This embodiment conducts platform operation from the perspective of the device end. Please refer to Figure 2 As shown, the image acquisition platform is communicatively connected to a multi-angle error recognition unit and a camera imaging evaluation unit corresponding to the device end; the specific method is as follows:

[0046] The image acquisition platform generates a device control signal and sends it to the device end;

[0047] After receiving it, the device end sends an execution instruction to the multi-angle error recognition unit;

[0048] The multi-angle error recognition unit detects the angular error of the installed camera in the industrial pipeline. Through the camera angles of multiple positions, error detection and analysis are carried out to infer whether the installation position of the camera in the current industrial pipeline is reasonably coordinated, avoiding errors in image acquisition of the camera caused by improper setting of the position of the camera. When the image is not clear, cameras at different angles cannot cooperate in shooting, resulting in a decrease in image acquisition efficiency. Through error recognition, the position of the camera can be adjusted to improve the accuracy of image acquisition of the pipeline at the same position;

[0049] Obtain the installed camera positions in the industrial pipeline, and analyze according to the coverage areas of each camera position. Mark the overlapping coverage areas of multiple camera positions as multi-angle acquisition areas, and the camera corresponding to the current multi-angle acquisition area is the area cooperative acquisition end;

[0050] Take any position on the pipe surface within the multi-angle acquisition area as the recognition subject, and analyze the acquired image of the recognition subject by the area cooperative acquisition end to obtain the display surface area of the recognition subject in the acquired image by the area cooperative acquisition end. If the display surface area exceeds the cumulative value of the surface areas of each surface of the recognition subject, it is inferred that the cooperative coverage of the current area cooperative acquisition end is satisfied. On the contrary, if the display surface area does not exceed the cumulative value of the surface areas of each surface of the recognition subject, it is inferred that the cooperative coverage of the current area cooperative acquisition end is not satisfied; when it is not satisfied, the position adjustment of the area cooperative acquisition end is required;

[0051] When the fitting coverage is satisfied and the recognition subject operates at different vibration frequencies, the acquisition images of the regional fitting acquisition end are further analyzed. Arbitrarily select an acquisition picture, and the display surface of the acquisition picture is lower than the set proportion threshold of the corresponding pipe surface area of the recognition subject. Mark the acquisition moment of the current acquisition picture as the error risk moment. Obtain the acquisition pictures of the recognition subject except the current acquisition picture at the error risk moment. If the maximum deviation value of the display surface area in the currently obtained acquisition pictures exceeds the deviation threshold, or the proportion of the number of pictures with the display surface lower than the set value in the acquisition pictures exceeds the proportion threshold, it is inferred that there is an acquisition error in the acquisition picture at the current error risk moment. It should be noted that the display surface lower than the set value means that the current picture cannot display the entire structure of the pipe surface or the pipeline. Therefore, the current fitting acquisition end should perform recognition subject image acquisition according to the coverage area, rather than having a large deviation in the corresponding display surface during fitting acquisition, so that most of the acquisition images at the current moment are pictures with a low display surface during image detection, greatly delaying the image detection efficiency. For example, when there is a crack on the pipe surface, early warning cannot be given in time;

[0052] If the maximum deviation value of the display surface area in the currently obtained acquisition pictures does not exceed the deviation threshold, and the proportion of the number of pictures with the display surface lower than the set value in the acquisition pictures does not exceed the proportion threshold, it is inferred that there is no acquisition error in the acquisition picture at the current error risk moment;

[0053] When the occurrence proportion of the error risk moment during the fitting acquisition period of the regional fitting acquisition end is lower than the proportion duration threshold, it is determined that the setting of the internal camera position of the industrial pipeline is qualified;

[0054] After determining that the position is qualified, the camera impact assessment unit conducts a camera impact assessment on each position of the industrial pipeline. Through the camera impact assessment, it can be inferred whether there is an impact on the image acquisition of the current industrial pipeline. Through the impact assessment, the image acquisition efficiency at each moment can be inferred, so as to process the pictures at the moment without camera impact during image acquisition and processing, avoid processing the images with low-efficiency camera shooting, increase the workload of image processing, and at the same time affect the accuracy of current image processing and recognition;

[0055] Collect the continuous frequency decline span of the collected images of the recognition subject within the same multi-angle collection area during the real-time camera shooting process in the industrial pipeline usage stage, and obtain the proportion of the number of acquisition moments covered by the decline stage of the camera shooting environment parameters within the same multi-angle collection area during the non-real-time camera shooting process in the industrial pipeline usage stage. Mark the continuous frequency decline span of the collected images of the recognition subject within the same multi-angle collection area during the real-time camera shooting process in the industrial pipeline usage stage and the proportion of the number of acquisition moments covered by the decline stage of the camera shooting environment parameters within the same multi-angle collection area during the non-real-time camera shooting process in the industrial pipeline usage stage as real-time acquisition data and non-real-time acquisition data respectively. It can be understood that real-time acquisition means continuous acquisition is required at each operation moment of the industrial pipeline operation, and non-real-time acquisition means selective acquisition at the operation moment of the industrial pipeline operation;

[0056] If the continuous frequency decline span of the collected images of the recognition subject within the same multi-angle collection area during the real-time camera shooting process in the industrial pipeline usage stage exceeds the decline span threshold, or the proportion of the number of acquisition moments covered by the decline stage of the camera shooting environment parameters within the same multi-angle collection area during the non-real-time camera shooting process in the industrial pipeline usage stage exceeds the proportion threshold, then it is inferred that the camera shooting effect is abnormal during the industrial pipeline image acquisition stage, and mark the current time period as an inefficient camera shooting time period;

[0057] If the continuous frequency decline span of the collected images of the recognition subject within the same multi-angle collection area during the real-time camera shooting process in the industrial pipeline usage stage does not exceed the decline span threshold, and the proportion of the number of acquisition moments covered by the decline stage of the camera shooting environment parameters within the same multi-angle collection area during the non-real-time camera shooting process in the industrial pipeline usage stage does not exceed the proportion threshold, then it is inferred that the camera shooting effect is normal during the industrial pipeline image acquisition stage, and mark the current time period as an efficient camera shooting time period;

[0058] The device terminal sends the collected images corresponding to the efficient camera shooting time period and the inefficient camera shooting time period to the image acquisition platform. If there is an overlap between the efficient camera shooting time period and the inefficient camera shooting time period, retain the collected images of the inefficient camera shooting time period, and perform deletion processing when fault detection is required, such as when the clarity is insufficient, to reduce the image drop frame rate; if the inefficient camera shooting time period exceeds the set duration value, the currently collected images are not used as the collected images for image processing;

[0059] Embodiment 2

[0060] The previous embodiment analyzed and controlled the industrial pipeline image acquisition device. This embodiment performs image processing on the basis of the previous embodiment. Please refer to Figure 3 As shown, the image terminal of the image acquisition platform is communicatively connected to a pipeline image analysis unit and a multi-image impact detection unit; the specific operation method of the image terminal is as follows:

[0061] After the image acquisition of the industrial pipeline is completed, the pipeline image analysis unit analyzes the image of the industrial pipeline, infers whether there are pipeline cracks in the industrial pipeline through the analysis of the industrial pipeline image, improves the efficiency of image acquisition and detection in the industrial pipeline, conducts crack impact analysis based on the synchronous image processing of multi-angle cameras, and gives early warnings and maintenance to the industrial pipeline in a timely manner, reducing the risk of pressure imbalance during the use of the industrial pipeline;

[0062] The image marked as the processed image according to the image collected and detected by the device end is divided into several sub-frame pictures, and the sub-frame pictures are sorted according to the time sequence of camera acquisition. Color difference comparison is carried out according to the pipeline display surface in the sub-frame pictures, and the positions with color difference are marked as risk positions. And according to the risk positions, the area is coordinated with the acquisition end. The trajectory of the risk position is marked as the risk form, and the predicted risk level is set according to the risk form of the sub-frame pictures. It should be explained that the risk form can reflect the area and length of the trajectory, and the risk level can be divided according to the set values;

[0063] By summarizing the sub-frame pictures of multiple angles of the risk positions according to the area coordinated with the acquisition end, and obtaining the multi-angle display surfaces of the risk trajectory according to the summarized sub-frame pictures, and detecting the depth of the risk form according to the multi-angle display surfaces. It can be understood that the depth of the pipe surface side cannot be directly detected, but the depth display surface of the current risk position can be obtained by adjusting the side angle and collecting images; it is sufficient to infer whether the current level is underestimated; if the numerical deviation of the risk form corresponding to the multi-angle depth detection sub-frame picture display surface is within the deviation range, it is inferred that the predicted risk level is consistent and marked as the set risk level, and sent to the image end and then transferred to the image acquisition platform; the image acquisition platform makes pipeline maintenance decisions according to the set risk level;

[0064] If the numerical deviation of the risk form corresponding to the multi-angle depth detection sub-frame picture display surface is not within the deviation range, it is inferred that the predicted risk level is inconsistent and the level is increased or decreased according to the actual deviation value. After the level adjustment is completed, it is marked as the set risk level and sent to the image end and then transferred to the image acquisition platform;

[0065] After the image pipeline analysis unit sets the risk level, the multi-image impact detection unit conducts impact detection on the sub-frame pictures with risk positions, conducts multi-image impact analysis on the industrial pipeline display surface through multiple sub-frame pictures, and conducts impact recognition on the industrial pipeline risk form according to the real-time collected images, facilitating timely risk prevention and control during the use stage of the industrial pipeline. At the same time, according to the impact detection, tracing can be carried out to control and reduce the expansion of the risk form in a timely manner; further improving the use efficiency of the industrial pipeline;

[0066] Screen the sub-frame pictures sorted within the processed image, use the sub-frame picture at the risk position that appears for the first time as the selection time point, and use the number threshold of sub-frame pictures before the selection time point as the risk picture set; and use the number threshold of sub-frame pictures after the selection time point as the influence picture set, where the number threshold of sub-frame pictures is a threshold number set artificially according to the size of the risk position.

[0067] Obtain the growth peak of the operation data showing a continuous growth trend of the industrial pipeline in the risk picture set. At the same time, obtain the cumulative overlapping duration of the risk form increase period corresponding to the risk position in the influence picture set and the operation data floating period, and mark the growth peak of the operation data showing a continuous growth trend of the industrial pipeline in the risk picture set and the cumulative overlapping duration of the risk form increase period corresponding to the risk position in the influence picture set and the operation data floating period as the incentive acquisition data and the influence acquisition data respectively, and compare them with the growth peak threshold and the overlapping duration threshold respectively: where the operation data is represented by the operation parameters of the industrial pipeline, such as vibration frequency, operation duration and other parameters.

[0068] If the growth peak of the operation data showing a continuous growth trend of the industrial pipeline in the risk picture set exceeds the growth peak threshold, or the cumulative overlapping duration of the risk form increase period corresponding to the risk position in the influence picture set and the operation data floating period exceeds the overlapping duration threshold, it is inferred that there is an impact on the operation data corresponding to the industrial pipeline, and it is marked as risk generation information and sent to the image end. The image end transfers it to the image acquisition platform, and compares the risk generation information according to the operation of the industrial pipeline. If the probability of the risk position appearing in the operation period with floating risk production line information exceeds the set probability threshold, the risk production line information is used as the operation red line value and matched with the current type of industrial pipeline.

[0069] If the growth peak of the operation data showing a continuous growth trend of the industrial pipeline in the risk picture set does not exceed the growth peak threshold, and the cumulative overlapping duration of the risk form increase period corresponding to the risk position in the influence picture set and the operation data floating period does not exceed the overlapping duration threshold, it is inferred that there is no impact on the operation data corresponding to the industrial pipeline, and it is marked as risk irrelevant information and sent to the image end. The image end transfers it to the image acquisition platform.

[0070] When the present invention is in use, an image acquisition platform generates a device control signal and sends it to the device end; after receiving it, the device end sends an execution instruction to the multi-angle error recognition unit, and the multi-angle error recognition unit detects the angle error in cooperation with the camera installed in the industrial pipeline; after determining that the position is qualified, the camera image evaluation unit evaluates the camera image of each position of the industrial pipeline, collects real-time acquisition data and non-real-time acquisition data, and conducts camera image evaluation based on data comparison; after completing the image acquisition of the industrial pipeline, the pipeline image analysis unit analyzes the image of the industrial pipeline; after the image pipeline analysis unit sets the risk level, the multi-image impact detection unit detects the impact on the sub-frame pictures of the positions with risks.

[0071] The preferred embodiments of the present invention disclosed above are only used to help explain the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the present invention to only the specific implementation manners. Obviously, many modifications and variations can be made according to the content of this specification. These embodiments are selected and specifically described in this specification to better explain the principle and practical application of the present invention, so that those skilled in the art can understand and utilize the present invention well. The present invention is only limited by the claims and their full scope and equivalents.

Claims

1. An industrial pipeline camera image acquisition and processing method applied to complex industrial systems, characterized in that: The image acquisition and processing method process is as follows: The image acquisition platform generates a device control signal and sends it to the device end; After receiving the command, the device sends the execution command to the multi-angle error recognition unit, and the multi-angle error recognition unit performs angle error detection on the camera installed in the industrial pipeline; After determining that the points are qualified, the camera impact assessment unit conducts a camera impact assessment on each point of the industrial pipeline, collects real-time data and non-real-time data, and conducts a camera impact assessment based on data comparison; After completing the image acquisition of the industrial pipeline, the pipeline image analysis unit performs image analysis on the industrial pipeline; After the image pipeline analysis unit sets the risk level, the multi-image impact detection unit performs impact detection on the sub-frame images where the risk position exists.

2. The industrial pipeline camera image acquisition and processing method applied to complex industrial systems according to claim 1 is characterized in that: The process of matching angle error detection is as follows: The camera points set in the industrial pipeline are obtained, and the coverage areas of each camera point are analyzed. The overlapping coverage areas of multiple camera points are marked as multi-angle acquisition areas, and the camera corresponding to the current multi-angle acquisition area is the regional cooperation acquisition end; The pipe surface at any position in the multi-angle acquisition area is taken as the identification subject, and the acquisition image of the identification subject is analyzed according to the regional cooperation acquisition end to obtain the display surface area of ​​the identification subject in the acquisition image of the regional cooperation acquisition end. If the display surface area exceeds the cumulative value of the surface areas of the identification subject, it is inferred that the coordination coverage of the current regional cooperation acquisition end is satisfied. Conversely, if the display surface area does not exceed the cumulative value of the surface areas of the identification subject, it is inferred that the coordination coverage of the current regional cooperation acquisition end is not satisfied. If it is not satisfied, the position of the regional cooperation acquisition end needs to be adjusted.

3. The industrial pipeline camera image acquisition and processing method applied to complex industrial systems according to claim 2 is characterized in that: When the matching coverage is satisfied, the recognition subject operates at different vibration frequencies, and further analyzes the collected images of the regional matching collection end, selects any collected image whose display surface is lower than the set ratio threshold of the corresponding tube surface area of ​​the recognition subject, and marks the collection time of the current collected image as the error risk time; Obtain the subject's collected pictures other than the current collected picture at the error risk moment. If the maximum deviation value of the display surface area in the currently obtained collected pictures exceeds the deviation threshold, or the proportion of pictures with display surfaces below the set value in the collected pictures exceeds the number proportion threshold, it is inferred that there is a collection error in the collected pictures at the current error risk moment. If the maximum deviation value of the display surface area in the currently acquired collected image does not exceed the deviation threshold, and the proportion of the number of images in the collected image whose display surface is lower than the set value does not exceed the number proportion threshold, it is inferred that there is no collection error in the collected image at the current error risk moment; When the proportion of error risk moments during the cooperative collection period of the regional cooperative collection end is lower than the proportion duration threshold, it is determined that the internal camera point settings of the industrial pipeline are qualified.

4. The industrial pipeline camera image acquisition and processing method applied to complex industrial systems according to claim 1 is characterized in that: The camera impact assessment process is as follows: The real-time acquisition data and non-real-time acquisition data are respectively the continuous frequency reduction span of the subject recognition image acquisition in the same multi-angle acquisition area during the real-time camera process in the use stage of the industrial pipeline, and the proportion of acquisition moments covered by the reduction stage of the camera environment parameters in the same multi-angle acquisition area during the non-real-time camera process in the use stage of the industrial pipeline.

5. The industrial pipeline camera image acquisition and processing method applied to complex industrial systems according to claim 4 is characterized in that: If the real-time collected data exceeds the descending span threshold, or the non-real-time collected data exceeds the quantity ratio threshold, the current period is marked as an inefficient camera period; If the real-time collected data does not exceed the descending span threshold, and the non-real-time collected data does not exceed the quantity ratio threshold, the current period is marked as an efficient camera period.

6. The industrial pipeline camera image acquisition and processing method applied to complex industrial systems according to claim 5 is characterized in that: If there is an overlap between the efficient camera period and the inefficient camera period, the captured images of the inefficient camera period will be retained, and if the clarity is not enough when fault detection is required, they will be deleted to reduce the frame drop rate of the image; if the inefficient camera period exceeds the set duration value, the current captured image will not be used as the captured image for image processing.

7. The industrial pipeline camera image acquisition and processing method applied to complex industrial systems according to claim 1 is characterized in that: The image analysis process is as follows: The image collected and detected by the device end is marked as a processed image, and the processed image is divided into several sub-frame images. The sub-frame images are sorted according to the camera acquisition time sequence, and the color difference is compared according to the pipeline display surface in the sub-frame image. The position with color difference is marked as a risk position, and the area is determined according to the risk position. Cooperate with the acquisition end, mark the risk position as a risk form according to the trajectory of the risk position, and set the predicted risk level according to the risk form of the sub-frame image; By summarizing the sub-frame images of multiple angles of the risk position according to the area and the acquisition end, and obtaining the multiple angle display surfaces of the risk trajectory according to the summarized sub-frame images, and detecting the depth of the risk form according to the multiple angle display surfaces; if the numerical deviation of the risk form corresponding to the multi-angle depth detection sub-frame image display surface is within the deviation range, it is inferred that the predicted risk level is matched and marked as the set risk level, and sent to the image end and forwarded to the image acquisition platform; The image acquisition platform makes pipeline maintenance decisions based on the set risk level; If the numerical deviation of the risk morphology corresponding to the multi-angle depth detection sub-frame image display surface is not within the deviation range, it is inferred that the predicted risk level is not consistent and the level is increased or decreased according to the actual deviation value. After the level adjustment is completed, it is marked as the set risk level, sent to the image end and forwarded to the image acquisition platform.

8. The industrial pipeline camera image acquisition and processing method applied to complex industrial systems according to claim 1 is characterized in that: The multi-image impact detection unit process is as follows: The sub-frame images sorted in the processed image are screened, and the sub-frame image where the risk position appears for the first time is used as the selection time point, and the threshold of the number of sub-frame images before the selection time point is used as the risk image set; The threshold of the number of sub-frame images after the selected time point is used as the impact image set, and the cause collection data and the impact collection data are collected. If the cause collection data exceeds the growth peak threshold, or the impact collection data exceeds the overlap duration threshold, it is marked as risk generation information; If the cause collection data does not exceed the growth peak threshold, and the impact collection data does not exceed the overlapping duration threshold, it will be marked as risk-irrelevant information.

9. The industrial pipeline camera image acquisition and processing method applied to complex industrial systems according to claim 8 is characterized in that: The inducement collection data and the impact collection data are respectively the peak growth value of the operating data of the industrial pipeline in the risk picture set showing a continuous growth trend, and the cumulative overlapping time of the risk position in the impact picture set corresponding to the risk form increase period and the operating data floating period.

Images