A big data-based multimedia intelligent meter data processing method and system

Abstract

This application relates to the field of data processing technology, and provides a data processing method and system for multimedia smart meters based on big data. The method includes: acquiring production visual data and production environment data of the coating on a coating production line collected by the smart meter; the production environment data includes production sensor data, production material data, or production operation data; performing image analysis on the production visual data based on the production environment data to obtain the production status information of the coating; and storing the production status information of the coating for product traceability. This can improve the efficiency of generating defect perception.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to a data processing method and system for multimedia smart meters based on big data. Background Technology

[0002] In modern industrial production, especially in manufacturing fields with extremely high requirements for precision and sensitivity, such as the preparation of precision functional coatings, multimedia intelligent instruments play a crucial role. These instruments are responsible for real-time acquisition and processing of multi-source heterogeneous data from the production line, including equipment sensor data, high-definition video surveillance streams, environmental parameters, raw material batch information, and operation logs.

[0003] However, existing technologies generally face the challenge of effectively identifying and diagnosing latent product defects caused by specific combinations of small, non-obvious process parameter fluctuations when processing such complex data. This is mainly because current systems lack the ability to understand and model the complex, nonlinear, and potentially time-lag-dependent deep physical logic and causal relationships between heterogeneous, multi-source data. This results in the inability to automatically extract the "contextual" information that leads to defects from massive amounts of data, thus hindering the shift from passive threshold alarms to proactive, context-aware process drift prediction and early defect diagnosis. Summary of the Invention

[0004] This application provides a data processing method and system for multimedia smart meters based on big data, which can improve the efficiency of generating defect perception.

[0005] To achieve the above objectives, this application adopts the following technical solution:

[0006] In the first aspect, this application discloses a data processing method for multimedia smart meters based on big data, including: acquiring production visual data and production environment data of the coating on the coating production line collected by the smart meter; the production environment data includes production sensor data, production material data or production operation data; performing image analysis on the production visual data based on the production environment data to obtain the production status information of the coating; and storing the production status information of the coating to enable product traceability of the coating.

[0007] Through this technical solution, this application can perform correlation analysis between production visual data and production environment data, thereby obtaining more comprehensive and accurate production status information of the coating. This effectively solves the problem that traditional methods are difficult to identify latent defects caused by the complex effects of multiple factors, and realizes refined traceability of coating product quality.

[0008] Furthermore, this application proposes that the production visual data includes a video stream, and that image analysis of the production visual data based on production environment data is performed to obtain the production status information of the coating, including: extracting multiple target video frames from the video stream at preset frame intervals; for each target video frame, determining whether there is defect information in the target video frame based on the target video frame and the two video frames before and after the target video frame; when there is no defect information in the target video, determining that the production status information of the coating indicates that there is no defect information in the target area of ​​the coating; the shooting time of the target video frame is the coating time of the target area in the coating; when there is defect information in the target video, determining the confidence level of the defect information based on the production environment data; when the confidence level of the defect information is greater than or equal to a preset confidence threshold, determining that the production status information of the coating indicates that there is defect information in the target area of ​​the coating; when the confidence level of the defect information is less than the preset confidence threshold, determining that the production status information of the coating indicates that there is no defect information in the target area of ​​the coating.

[0009] This technical solution effectively avoids misjudgment or missed judgment caused by inaccurate judgment of single visual information in traditional methods by extracting frames and judging defects from video streams and combining production environment data to evaluate the confidence of defect information, thereby improving the accuracy and reliability of defect identification.

[0010] In some preferred embodiments, when the defect information is a microbubble defect, determining whether a defect exists in the target video based on the target video frame and the two videos preceding and following the target video frame includes: determining whether there is transient optical flicker in the target video frame based on the target video frame and the two videos preceding and following the target video frame; if there is no transient optical flicker in the target video frame, determining that the production status information of the coating indicates that there is no microbubble defect in the target area of ​​the coating; if there is transient optical flicker in the target video frame, determining that the production status information of the coating indicates that there is a microbubble defect in the target area of ​​the coating.

[0011] This application provides a more targeted defect detection method by identifying microbubble defects by judging instantaneous optical scintillation, based on the characteristics of microbubble defects, thus improving the identification accuracy of specific types of defects.

[0012] Based on this, this application further proposes to determine the confidence level of defect information based on production environment data, including: taking the first moment with a preset duration before the shooting time of the target video frame as the start time of the target time period, and taking the second moment with a preset duration after the shooting time as the end time of the target time period; determining the change information of each data in the production environment data within the target time period; and determining the confidence level of defect information based on the change information of each data in the production environment data.

[0013] This technical solution allows the application to determine the confidence level of defect information by analyzing changes in production environment data within a target time period. This approach can more comprehensively consider the impact of environmental factors on defects, thereby making the confidence level assessment more accurate.

[0014] Furthermore, this application proposes to determine the confidence level of defect information based on the change information of each data in the production environment data, including: for each data in the production environment data, obtaining the preset change information of the data when the coating has microbubble defects; determining the similarity between the change information of the data in the target time period and the corresponding preset change information; and determining the confidence level of microbubble defects based on the similarity of multiple data in the production environment data.

[0015] By using this technical solution, this application can quantify the correlation between environmental factors and defects by comparing the similarity between actual environmental data changes and preset defect changes, thereby more accurately assessing the confidence level of microbubble defects.

[0016] As a technical improvement, this application also proposes to determine the confidence level of microbubble defects based on the similarity of multiple data, including: obtaining the defect influence weight of each data in multiple data; the sum of the defect influence weights of multiple data is 1; determining the weighted sum of the similarity of multiple data based on the defect influence weights of multiple data; and determining the weighted sum as the confidence level of microbubble defects.

[0017] This technical solution introduces a defect impact weight to weight the similarity of different environmental data, making environmental factors with a greater impact on defects occupy a more important position in the confidence assessment, thereby improving the accuracy and rationality of the confidence assessment.

[0018] Based on the above, this application further proposes a method for determining whether there is transient optical flicker in a target video frame based on the target video frame and the two video frames before and after it. This method includes: for each pixel in the target video frame, determining the average value of the brightness change rate of that pixel in the target video frame and the two video frames before and after it; identifying pixels whose average brightness change rate is greater than a preset brightness change rate threshold as target pixels; determining the dispersion index of the target pixels; and determining that there is transient optical flicker in the target video frame when the dispersion index is greater than a preset dispersion index threshold, otherwise determining that there is no transient optical flicker in the target video frame.

[0019] This application provides an objective and quantitative method for identifying microbubble defects by analyzing the brightness change rate and dispersion index of pixels to determine instantaneous optical flicker. This avoids errors in subjective judgment and improves the automation and accuracy of detection.

[0020] As a further improvement, this application also proposes storing coating production status information, including: obtaining the product importance index of the coating; determining a target storage device among multiple storage devices based on the target product importance index; each storage device having a different storage lifespan; and storing the coating production status information to the target storage device.

[0021] Through this technical solution, this application selects different storage devices according to the product importance index of the coating, realizes hierarchical storage of data with different levels of importance, optimizes storage resource allocation, and improves data storage efficiency and reliability.

[0022] Preferably, this application further proposes to determine the target storage device among multiple storage devices based on the target product importance index, including: obtaining a preset correspondence relationship; the preset correspondence relationship includes a one-to-one correspondence relationship between multiple product importance index ranges and multiple storage devices, wherein the maximum value of the product importance index range is positively correlated with the storage life of the corresponding storage device; and taking the storage device corresponding to the product importance index of the coating in the preset correspondence relationship as the target storage device.

[0023] Through this technical solution, this application establishes a correspondence between product importance indices and storage devices, thereby automating and intelligentizing storage strategies, ensuring that important data can be stored in devices with sufficient storage life, and further improving the efficiency and security of data management.

[0024] Secondly, this application also discloses a data processing system for a multimedia smart instrument based on big data, comprising: an acquisition device and a processing device; the acquisition device is used to acquire production visual data and production environment data of the coating on the coating production line collected by the smart instrument; the production environment data includes production sensor data, production material data or production operation data; the processing device is used to analyze the production visual data based on the production environment data to obtain the production status information of the coating; the processing device is used to store the production status information of the coating for product traceability of the coating.

[0025] Beneficial effects

[0026] This application discloses a data processing method based on big data multimedia smart instruments. It acquires production visual data and production environment data of the coating production line collected by smart instruments, and performs image analysis on the production visual data based on the production environment data to obtain the production status information of the coating. Finally, it stores the production status information of the coating for product traceability. This method effectively solves the problem in existing technologies of the difficulty in effectively identifying and diagnosing latent product defects caused by a specific combination of multiple small, non-obvious process parameter fluctuations. By deeply integrating and correlating production visual data with production environment data, this application can automatically extract the "contextual" information leading to defects from massive heterogeneous data, thus overcoming the limitations of traditional methods that rely on only a single data source or simple threshold alarms. This paradigm shift in multi-source data fusion analysis enables the system to move from passive threshold alarms to proactive, context-aware process drift prediction and early defect diagnosis, significantly improving the accuracy and timeliness of defect identification, reducing rework, scrap, and after-sales service costs, thereby enhancing production efficiency and product competitiveness. Attached Figure Description

[0027] Figure 1 A flowchart illustrating a data processing method for a multimedia smart meter based on big data, provided for this application;

[0028] Figure 2 A flowchart illustrating another data processing method for multimedia smart meters based on big data provided in this application;

[0029] Figure 3 A flowchart illustrating another data processing method for multimedia smart meters based on big data provided in this application;

[0030] Figure 4 This application provides a schematic diagram of the structure of a data processing system for a multimedia intelligent instrument based on big data. Detailed Implementation

[0031] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments. The components of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0032] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0033] In modern industrial production, especially in manufacturing fields with extremely high requirements for precision and sensitivity, such as the preparation of precision functional coatings, multimedia intelligent instruments play a crucial role. These instruments are responsible for real-time acquisition and processing of multi-source heterogeneous data from the production line, including equipment sensor data, high-definition video surveillance streams, environmental parameters, raw material batch information, and operation logs.

[0034] However, existing technologies generally face the challenge of effectively identifying and diagnosing latent product defects caused by specific combinations of small, non-obvious process parameter fluctuations when processing such complex data. This is mainly because current systems lack the ability to understand and model the complex, nonlinear, and potentially time-lag-dependent deep physical logic and causal relationships between heterogeneous, multi-source data. This results in the inability to automatically extract the "contextual" information that leads to defects from massive amounts of data, thus hindering the shift from passive threshold alarms to proactive, context-aware process drift prediction and early defect diagnosis.

[0035] In this regard, such as Figure 1 As shown, this application proposes a data processing method for multimedia smart meters based on big data, including:

[0036] S101. Acquire the production visual data and production environment data of the coating on the coating production line collected by the intelligent instrument.

[0037] Production environment data includes production sensor data, production material data, or production operation data.

[0038] S102. Based on production environment data, perform image analysis on production visual data to obtain coating production status information.

[0039] S103. Store the production status information of the coating to enable product traceability of the coating.

[0040] This application aims to provide a method that can effectively process multi-source heterogeneous data and identify and diagnose latent product defects, thereby enabling refined management of the coating production process and traceability of product quality.

[0041] To better understand the method proposed in this application, some key terms involved will be explained first.

[0042] "Smart instrument" refers to a device that integrates sensor, data acquisition, data processing and communication functions, and can monitor and record various parameters on the production line in real time.

[0043] A "coating production line" refers to an automated or semi-automated production process used to manufacture coated products, which involves multiple process steps such as coating, drying, and curing.

[0044] "Production visual data" refers to image or video information about coating appearance, surface features, etc., collected through visual sensors (such as high-definition cameras).

[0045] "Production environment data" refers to relevant data other than visual data during the coating production process, including but not limited to:

[0046] "Production sensor data": Data collected by various physical or chemical sensors, such as temperature sensors, pressure sensors, flow sensors, humidity sensors, etc., are used to monitor the operating status of production equipment and environmental parameters.

[0047] "Production Material Data": Data related to the raw materials used in coating production, such as material batch number, supplier information, material composition, viscosity, density and other physicochemical properties.

[0048] "Production Operation Data": Data that records various operational behaviors performed by operators or automated systems on the production line, such as equipment start-up and shutdown times, parameter adjustment records, and fault alarm information.

[0049] Image analysis refers to the use of computer vision and image processing technologies to process and analyze production visual data in order to extract useful information.

[0050] "Production status information" refers to the information obtained through comprehensive analysis of production visual data and production environment data regarding the quality status and defects of coated products during the production process.

[0051] "Product traceability" refers to the ability to track and query products throughout their entire lifecycle, from raw materials to finished products, by recording and storing production status information, so that the cause can be quickly located when quality problems occur.

[0052] The core of the method proposed in this application lies in the deep fusion and intelligent analysis of multi-source heterogeneous data to achieve accurate perception of coating production status and effective traceability of product quality.

[0053] Specifically, the step of "acquiring visual production data and environmental data of the coating production line from intelligent instruments" can be achieved in several ways. For example, visual production data can be acquired in real time by deploying multiple high-definition industrial cameras on the coating production line. These cameras can be fixed above key coating and drying areas to obtain continuous video streams or high-resolution images of the coating surface. This visual data can be transmitted to a local server or cloud storage system. Environmental production data can be acquired through various sensors integrated into the production equipment. For example, temperature sensors can monitor oven temperature in real time, pressure sensors can monitor coating pressure, and flow sensors can monitor the supply of coatings. Production material data can be obtained by scanning RFID tags or barcodes on raw material packaging and linked to a pre-stored material database. Production operation data can be automatically recorded by the production line control system or entered by operators through a human-machine interface. All of this data should be accompanied by accurate timestamps for subsequent synchronization and correlation analysis.

[0054] In the step of "analyzing production visual data based on production environment data to obtain coating production status information," the following methods can be used. For example, deep learning models can be used to analyze production visual data in real time to identify various defects on the coating surface, such as bubbles, scratches, and foreign objects. These models can be trained on a large amount of labeled image data. Simultaneously, production environment data can serve as auxiliary information for image analysis. For instance, when image analysis identifies a suspected defect, the system can combine environmental parameters such as the oven temperature, coating speed, and material viscosity to more accurately determine the type and severity of the defect. For example, if visual data shows the presence of tiny bubbles on the coating surface, while production environment data shows a low oven temperature or excessively high coating viscosity, it can be inferred that these bubbles may be due to insufficient drying or poor coating flowability. This multi-source data analysis approach can improve the accuracy and robustness of defect identification.

[0055] In the step of "storing coating production status information for product traceability," the following methods can be used. For example, the analyzed coating production status information, including defect type, location, severity, occurrence time, and related production environment parameters, can be stored in a structured database. This database can be linked to product batch numbers, serial numbers, etc. When quality traceability is needed for a specific batch or product, its complete production status information can be obtained by querying this database. For instance, if a customer reports a quality problem with a product, all visual data, environmental data, and operational data from its production process can be quickly retrieved using the product serial number, thereby pinpointing the cause and stage of the problem. This traceability capability is crucial for quality control, defect analysis, and liability determination.

[0056] The data processing method based on big data multimedia smart instruments proposed in this application acquires production visual data and production environment data of the coating on the coating production line collected by smart instruments. Based on the production environment data, image analysis is performed on the production visual data to obtain the production status information of the coating. Finally, the production status information of the coating is stored for product traceability. This method can effectively solve the problem in existing technologies of difficulty in identifying and diagnosing latent product defects caused by a specific combination of multiple small, non-obvious process parameter fluctuations.

[0057] Compared to existing technologies, the innovation of this application lies in its deep fusion and intelligent analysis capabilities for multi-source heterogeneous data. Traditional methods often treat video information, sensor information, and log information as independent data streams weakly correlated only by timestamps. This makes it difficult for engineers to manually sift through massive amounts of data, compare, and identify complex correlation patterns that are cross-modal, multi-factorial, and may have time lag effects. This application achieves a deeper understanding of production visual data by using production environment data as auxiliary information for image analysis, thereby enabling more accurate identification and diagnosis of latent defects. For example, in traditional systems, a tiny bubble defect may be overlooked by the visual inspection system because it is not obvious, or even if it is detected, its cause cannot be accurately determined. However, the method of this application, by combining environmental parameters such as temperature, humidity, and material viscosity, can more accurately determine the type, severity, and potential process causes of the bubble defect. This context-aware analysis capability enables the system to shift from passive threshold alarms to proactive, context-aware process drift prediction and early defect diagnosis, thereby significantly improving the efficiency and accuracy of product quality control and reducing rework, scrap, and after-sales service costs.

[0058] This application further proposes a data processing method for multimedia intelligent instruments based on big data. Specifically, the method involves image analysis of production visual data based on production environment data to obtain coating production status information, and includes the following steps:

[0059] Production visual data includes video streams. Image analysis is performed on the production visual data based on production environment data to obtain coating production status information, including:

[0060] Multiple target video frames are extracted from the video stream at preset frame intervals. For each target video frame, the presence of defect information is determined based on the target video frame and the two preceding and following video frames. If no defect information is found in the target video, the production status information of the coating indicates that the target area in the coating is free of defects. The capture time of the target video frame is the coating time of the target area in the coating. If defect information is found in the target video, the confidence level of the defect information is determined based on production environment data. If the confidence level of the defect information is greater than or equal to a preset confidence threshold, the production status information of the coating indicates that the target area in the coating is free of defects. If the confidence level of the defect information is less than the preset confidence threshold, the production status information of the coating indicates that the target area in the coating is free of defects.

[0061] Specifically, a video stream can be understood as a data sequence consisting of a series of consecutive image frames, used to record dynamic visual information on the coating production line, such as real-time image changes on the coating surface. The preset frame interval can be set according to the actual production line speed, the expected size of coating defects, and the real-time requirements of image analysis. For example, one frame can be extracted every N frames, or the extraction frequency can be increased when a potential anomaly is detected. This aims to reduce the amount of data that needs to be processed while ensuring that key changes in production status can be captured.

[0062] Furthermore, by combining the analysis of the target video frame with the frames before and after it, information in the temporal dimension can be used to distinguish between transient noise and real defects, improving the accuracy and robustness of defect detection. For example, a transient flicker might be misjudged as a defect in a single frame, but if it does not persist in the frames before and after it, it can be excluded.

[0063] When visual analysis does not detect any defects, the area is directly deemed to be defect-free, and the capture time of that frame is taken as the coating time of that area, facilitating subsequent traceability.

[0064] Production environment data includes production sensor data, production material data, and production operation data. This data can provide additional verification information for potential defects detected visually. For example, if visual inspection detects a bubble defect, and production environment data shows abnormal paint viscosity or spraying pressure fluctuations, the confidence level for that bubble defect can be increased.

[0065] The preset confidence threshold is a configurable parameter used to balance the sensitivity and false alarm rate of defect detection. When the confidence level reaches or exceeds this threshold, it indicates that the defect has a high degree of realism and should be recognized as an actual defect. This means that even if a potential defect is visually detected, if the confidence level fails to meet the preset standard when combined with production environment data, the potential defect is considered a false alarm or a non-critical anomaly and is not recognized as an actual defect.

[0066] This application's solution concretizes production visual data into video streams and introduces a defect judgment mechanism based on temporal context (previous and subsequent video frames), effectively avoiding misjudgments caused by instantaneous noise in single-frame image analysis. Furthermore, when a potential defect is detected visually, this application does not directly confirm it as a defect, but innovatively introduces a step to determine the confidence level of defect information based on production environment data. This mechanism allows visual inspection results to be cross-validated with changes in physical and chemical parameters during the actual production process, thereby significantly improving the accuracy and reliability of defect judgment. For example, if a suspected defect is detected visually, but production environment data (such as temperature, humidity, material batch, equipment operating parameters, etc.) are all within the normal range, the confidence level of the suspected defect may be low, thus being judged as non-defect, effectively avoiding false alarms. Conversely, if the environmental data also shows abnormalities, the defect confidence level will increase, ensuring that real defects are not missed.

[0067] Through the above technical solution, this application overcomes the limitations of traditional image analysis in processing video streams, which is susceptible to transient interference. Specifically, by extracting the target video frame and combining it with preceding and following video frames for judgment, the robustness of defect detection is effectively improved. More importantly, by introducing a mechanism to determine defect confidence based on production environment data, defect judgment no longer relies solely on visual information but integrates multi-source data, thereby significantly improving the accuracy of defect detection and reducing false alarm and false negative rates. Consequently, the obtained coating production status information is more accurate and reliable, providing a solid data foundation for subsequent product traceability and ultimately improving the overall production quality management level.

[0068] like Figure 2 As shown, this application further proposes a method for determining whether a defect exists in a target video frame when the defect information is a microbubble defect, based on the target video frame and the two video frames before and after the target video frame. Specifically, the method includes:

[0069] S201. Based on the target video frame and the two video frames before and after the target video frame, determine whether there is instantaneous optical flicker in the target video frame.

[0070] S202. When there is no instantaneous optical flicker in the target video frame, determine that the production status information of the coating indicates that there are no microbubble defects in the target area of ​​the coating.

[0071] S203. When there is transient optical flicker in the target video frame, determine the production status information of the coating to indicate that there is a microbubble defect in the target area of ​​the coating.

[0072] Specifically, microbubble defects refer to tiny bubbles that form on or inside the coating surface during the production process due to improper gas mixing or release. These microbubbles typically manifest visually as optical flickering, appearing and disappearing within a short timeframe. This is because the bursting or movement of bubbles on the coating surface causes drastic changes in local light reflection and refraction. Therefore, detecting transient optical flicker can effectively identify microbubble defects. Transient optical flicker can be understood as a drastic and brief change in brightness or color of a certain area in a video frame within an extremely short period; this change is usually associated with the formation, bursting, or movement of bubbles. Determining the presence of transient optical flicker aims to capture the unique dynamic visual characteristics of microbubble defects, thereby distinguishing them from background noise or other types of defects. When transient optical flicker is detected, the target area is considered to have a microbubble defect; conversely, the absence of transient optical flicker indicates the absence of a microbubble defect. This judgment mechanism enables the system to provide more accurate identification results for microbubble defects.

[0073] This application's solution effectively addresses the limitations of traditional general defect identification methods in recognizing microbubble defects by combining the assessment of microbubble defects with the detection of transient optical flicker. One of the essential characteristics of microbubble defects is that their formation, rupture, or movement on or within a coating surface causes rapid changes in local light reflection and refraction, resulting in brief but intense optical flicker in video frames. By analyzing the target video frame and the two frames preceding and following it, the system can capture this transient visual change.

[0074] Specifically, when microbubbles appear or disappear in a coating, their light scattering and reflection characteristics change instantaneously, causing significant, non-linear fluctuations in the brightness or color of local pixels between consecutive frames. This fluctuation is the visual manifestation of transient optical flicker. Therefore, by specifically detecting this transient optical flicker, this application can accurately locate and identify microbubble defects, avoiding misjudging other types of defects or minor fluctuations in normal production processes as microbubble defects, thereby improving the accuracy and reliability of defect identification.

[0075] Through the above technical solution, this application provides a highly specific and accurate detection method for microbubble defects on coating production lines. Compared to relying solely on general image analysis methods, this solution significantly improves the identification accuracy of microbubble defects and reduces false alarm and false negative rates by focusing on the unique instantaneous optical scintillation phenomenon of microbubble defects. This allows the coating production status information to more accurately reflect the actual defect situation, providing more reliable data support for subsequent product traceability and quality control. Furthermore, this optimized judgment mechanism for specific defect types helps improve the robustness and practicality of the entire intelligent instrument data processing method, enabling it to function better in complex industrial production environments.

[0076] In some preferred embodiments, a specific example is given below. Suppose that on a coating production line, a video stream captured by a smart instrument shows a tiny bubble on the coating surface. When this bubble bursts or moves on the coating surface, the light reflection characteristics of the surrounding area change instantaneously, causing the pixel brightness in that area to suddenly increase or decrease in consecutive video frames, creating a brief "flash" effect. This method analyzes the target video frame and the two preceding and following video frames. For example, if the brightness of a pixel in the target video frame suddenly increases in the current frame, while the brightness remains relatively stable or rapidly decreases in the previous and following frames, it can be determined that there is a momentary optical flicker in that pixel area. When such a momentary optical flicker is detected, the system determines that the coating production status information indicates the presence of a microbubble defect in the target area. Conversely, if no such momentary, drastic optical change is detected in the video frame, it is assumed that there is no microbubble defect. In this way, even tiny and transient microbubble defects can be effectively identified, thereby ensuring comprehensive monitoring of coating quality.

[0077] This application further proposes a more refined and robust method for determining the confidence level of defect information. By analyzing changes in production environment data over a period of time before and after the shooting time of the target video frame, the possibility of defect existence can be assessed more comprehensively.

[0078] According to the above method, such as Figure 3 As shown, when the defect information is a microbubble defect, the confidence level of the defect information is determined based on production environment data, including:

[0079] S301. Take the first moment of a preset duration before the shooting time of the target video frame as the start moment of the target time period, and take the second moment of a preset duration after the shooting time as the end moment of the target time period.

[0080] S302. For each data point in the production environment data, determine the changes in the data within the target time period.

[0081] S303. Determine the confidence level of defect information based on the change information of each data in the production environment data.

[0082] Specifically, the capture time of the aforementioned target video frame refers to the moment when the target area in the coating is applied, serving as the central point for time analysis. The setting of the first and second moments aims to construct a target time period around this capture time, the length of which is determined by a preset duration. For example, the preset duration can be empirically set based on the characteristics of the coating production process and the timescale of defect formation to ensure that key environmental changes related to defect formation can be captured. This target time period setting allows the analysis of production environment data to extend beyond a single instantaneous point to a time span, thus providing a more comprehensive reflection of the dynamic changes in the production process.

[0083] Each piece of data in the production environment data can be understood as any or a combination of production sensor data, production material data, or production operation data. Determining the changes in this data over a target time period involves analyzing its trends, fluctuations, extreme values, or rates of change within that period. For example, this might involve calculating the average rate of change, standard deviation, the difference between the maximum and minimum values, or identifying any abrupt changes or persistent deviations. This information can reveal potential abnormal fluctuations or specific trends in production environment parameters before and after a defect occurs.

[0084] In practical applications, determining the confidence level of defect information based on changes in each data point in the production environment refers to quantifying the likelihood of a defect's existence by establishing a correlation model between the changing information and the probability of defect occurrence. For example, a mapping relationship between different environmental data change patterns and the confidence level of a specific defect (such as a microbubble defect) can be established in advance through historical data analysis or expert experience. When a specific change in environmental data is detected, the corresponding defect confidence level can be derived based on this mapping relationship.

[0085] This application's solution addresses the limitations of traditional methods that rely solely on instantaneous environmental data or single visual features to determine defect confidence by setting a target time period before and after the capture of the target video frame and analyzing changes in production environment data within that period. Many coating defects, especially microbubble defects, are not instantaneous but are induced by cumulative or dynamic changes in environmental parameters during production, making the analysis of environmental data changes crucial. By capturing these dynamic changes, such as abnormal fluctuations or specific trends in temperature, humidity, material viscosity, or coating speed within the target time period, stronger evidence or counter-evidence can be provided for visually detected instantaneous optical flicker. This method combines visual information with the dynamic evolution of process parameters, leading to a deeper understanding of defect causes and improving the accuracy of defect assessment.

[0086] Through the above technical solution, this application can significantly improve the accuracy and robustness of determining the confidence level of defect information. By analyzing the changes in production environment data within the target time period, environmental factors closely related to defect formation can be identified more effectively, thereby avoiding misjudgments caused by data fluctuations or visual noise at a single moment. For example, when a momentary optical flicker is detected, if abnormal changes in environmental parameters strongly correlated with microbubble defects are simultaneously found within the target time period, the confidence level for the existence of the defect can be significantly increased; conversely, if the environmental parameters remain stable, the confidence level can be reduced, effectively decreasing false positives. This comprehensive judgment mechanism makes the assessment of coating production status information more reliable, thereby improving the effectiveness of product traceability and quality control.

[0087] This application further proposes a more refined method for determining the confidence level of defect information based on the change information of each data point in the production environment data.

[0088] Specifically, the confidence level of defect information is determined based on the changes in each data point in the production environment data, including:

[0089] For each data point in the production environment data, preset change information of the data when microbubble defects exist in the coating is obtained; the similarity between the change information of the data in the target time period and the corresponding preset change information is determined; the confidence level of microbubble defects is determined based on the similarity of multiple data points in the production environment data.

[0090] "Preset variation information" refers to the typical change patterns or ranges exhibited by specific production environment data (such as temperature, humidity, and pressure) when microbubble defects occur during the coating production process. This preset variation information can be obtained based on historical production data analysis, expert experience, or experimental verification. Its purpose is to provide a known environmental characteristic benchmark strongly correlated with the defect for defect identification. For example, if historical data shows that microbubble defects are often accompanied by a sudden drop in temperature in the coating area, then the magnitude and duration of this temperature drop can be defined as preset variation information for the temperature data.

[0091] Furthermore, "similarity" refers to comparing the observed changes in production environment data within a target time period with pre-defined changes related to microbubble defects to quantify the degree of matching between the two. This similarity can be calculated using various mathematical or statistical methods, such as correlation coefficients, Euclidean distance, and dynamic time warping (DTW). The most suitable similarity measurement method can be selected based on the data type and change characteristics. Higher similarity indicates a closer match between the current environmental data's change pattern and the typical inducing conditions for microbubble defects.

[0092] Therefore, the solution proposed in this application solves the problem of accurately determining defect confidence based solely on data change information by introducing "preset change information" as a reference standard and calculating the "similarity" between the actually observed changes in production environment data and this preset change information. Specifically, when the system detects a potential microbubble defect in a target video frame, it no longer relies solely on the changes in production environment data within the target time period, but compares these changes with known typical environmental data change patterns that lead to microbubble defects. By quantifying this similarity, the correlation between current environmental conditions and the occurrence of microbubble defects can be assessed more objectively and accurately. For example, if the trend of a certain environmental parameter is highly similar to the preset trend that leads to microbubble defects, then that parameter will contribute more to the confidence of microbubble defects. This similarity-based assessment mechanism makes the determination of defect confidence more scientific and data-driven, avoiding the limitations of subjective judgment or simple threshold determination.

[0093] The above technical solution enables a more accurate and reliable determination of the confidence level of microbubble defects in coatings. By comparing changes in actual production environment data with preset defect-related change patterns, environmental fluctuations unrelated to defects can be effectively eliminated, thereby improving the accuracy of defect identification. This method allows for a more refined assessment of coating production status information, helping to identify potential quality problems earlier and more accurately. It provides a more solid data foundation for subsequent product traceability and process optimization, thereby improving the efficiency and effectiveness of product quality control.

[0094] This application further proposes a method for determining the confidence level of microbubble defects based on the similarity of multiple data. By introducing defect influence weights, the calculation of confidence level becomes more refined and accurate.

[0095] Specifically, the confidence level of microbubble defects is determined based on the similarity of the above-mentioned data, including:

[0096] Obtain the defect impact weight for each data point from multiple datasets; sum the defect impact weights of multiple datasets to 1; determine the weighted sum of the similarities of multiple datasets based on their defect impact weights; and determine the confidence level of the microbubble defect based on the weighted sum.

[0097] The "defect impact weight" refers to the degree of influence of different production environment data on defect formation or detection results when judging microbubble defects in coatings. For example, certain production environment parameters (such as coating temperature and humidity) may have a more direct and significant impact on microbubble formation, while the impact of other parameters (such as ambient light intensity) may be relatively small. These weights can be determined based on historical data analysis, expert experience, or machine learning models to reflect the relative importance of each data point in microbubble defect judgment. The sum of the defect impact weights of multiple data points is set to 1 to ensure that the weighted sum calculation result is within a reasonable range and to facilitate weight allocation and normalization.

[0098] Furthermore, after obtaining the defect impact weight for each production environment data point, a weighted sum of the similarities of multiple data points is determined by multiplying the similarity of each data point by its corresponding defect impact weight and summing all the multiplications. This weighted sum comprehensively considers the degree of influence of each production environment data point on microbubble defects, giving greater weight to data points that are more critical to defect judgment in the confidence calculation. Ultimately, this weighted sum is directly determined as the confidence score for microbubble defects, used to quantify the probability of the presence of microbubble defects in the target area of ​​the coating.

[0099] The proposed solution assigns defect impact weights to different production environment data and calculates a weighted sum of similarities based on these weights, thereby making the confidence assessment of microbubble defects more scientific and accurate. This avoids the drawback of treating all environmental data the same in traditional methods, and can more accurately reflect the influence of various factors on defect formation, thus improving the reliability of defect judgment.

[0100] The above technical solution enables a more accurate assessment of the confidence level of microbubble defects, especially when faced with diverse and complex production environment data. It effectively distinguishes the importance of different data points in defect identification, thereby improving the accuracy and reliability of defect detection. This helps the production line to promptly identify and correct the root causes of microbubble defects, improving product quality.

[0101] As a specific implementation method, it is assumed that the main production environment data affecting microbubble defects during the coating production process include coating temperature data, coating humidity data, and coating speed data. Through historical data analysis and expert experience, the influence weights of these three data on microbubble defects can be determined as follows: coating temperature data 0.5, coating humidity data 0.3, and coating speed data 0.2. Obviously, the sum of these weights is 1. When analyzing a target video frame and determining the confidence level of defect information based on production environment data, the changes in coating temperature data, coating humidity data, and coating speed data within the target time period are first obtained, and their similarity to preset changes is calculated. For example, assuming the calculated coating temperature similarity is 0.8, coating humidity similarity is 0.7, and coating speed similarity is 0.9, the confidence level of microbubble defects will be calculated as: 0.8*0.5+0.7*0.3+0.9*0.2=0.4+0.21+0.18=0.79. This value of 0.79 was determined as the confidence level for the microbubble defect. Through this weighting method, the coating temperature data, due to its higher weight, has a more significant impact on the final confidence level, thus making the confidence level assessment more consistent with the influence of various factors on defect formation in actual production.

[0102] In some of the embodiments described above in this application, when the defect information is a microbubble defect, it is proposed to determine whether there is transient optical flicker in the target video frame based on the target video frame and the two video frames before and after the target video frame. However, in practical applications, more detailed and precise steps are needed to specifically determine the existence of transient optical flicker.

[0103] In this regard, this application further proposes a step for determining whether there is transient optical flicker in the target video frame based on the target video frame and the two video frames before and after the target video frame, including:

[0104] For each pixel in a target video frame, the average value of the brightness change rate of the pixel in the target video frame and the two preceding and following video frames is determined; pixels whose average brightness change rate is greater than a preset brightness change rate threshold are identified as target pixels; the dispersion index of the target pixels is determined; if the dispersion index is greater than a preset dispersion index threshold, it is determined that there is transient optical flicker in the target video frame; otherwise, it is determined that there is no transient optical flicker in the target video frame.

[0105] Specifically, when determining whether transient optical flicker exists in a target video frame, a pixel-level analysis of the target video frame and the two preceding and following frames is first required. For each pixel in the target video frame, the average rate of brightness change of that pixel across the target video frame and the two preceding and following frames needs to be determined. The average rate of brightness change can be understood as the degree of brightness fluctuation of that pixel across three consecutive video frames, aiming to capture rapid brightness changes that microbubbles may cause in the video. For example, the average rate of brightness change can be obtained by calculating the absolute value of the brightness difference of the same pixel between adjacent frames and averaging these differences.

[0106] Furthermore, pixels whose average brightness change rate exceeds a preset brightness change rate threshold are identified as target pixels. The preset brightness change rate threshold is an empirical value or a value obtained through model training, used to distinguish between normal brightness fluctuations and drastic brightness changes that may be caused by microbubbles. When the average brightness change rate of a pixel exceeds this threshold, it indicates that the pixel may be located in an area where microbubbles appear or disappear, and is therefore marked as a target pixel.

[0107] Based on this, it is necessary to determine the dispersion index of the target pixels. The dispersion index can be understood as the degree of concentration or dispersion of the target pixels in spatial distribution, and its purpose is to identify the typical spatial characteristics of transient optical flicker. For example, the dispersion index can be obtained by calculating the variance, standard deviation, or density index based on clustering algorithms of the spatial distribution of target pixels in the target video frame. When microbubbles appear, they usually manifest as rapid flickering in local areas in the video, and these flickering points will show a certain degree of concentration or specific distribution patterns in space.

[0108] Ultimately, if the dispersion index is greater than a preset dispersion index threshold, transient optical flicker is determined to exist in the target video frame; otherwise, transient optical flicker is determined not to exist in the target video frame. The preset dispersion index threshold is a key parameter used to distinguish between normal image noise or brightness changes not caused by microbubbles and actual transient optical flicker. When the dispersion index of a target pixel exceeds this threshold, it indicates that these pixels with drastic brightness changes exhibit a specific spatial distribution pattern related to microbubbles, thus accurately determining the presence of transient optical flicker.

[0109] This application's solution effectively captures the transient optical flicker characteristics exhibited by microbubble defects in video by averaging the brightness change rate of pixels in a video frame and combining this with the dispersion index of the target pixels. Specifically, when microbubbles form on the coating surface, their refraction and reflection properties cause rapid brightness changes in localized areas of the video, i.e., transient optical flicker. This brightness fluctuation can be quantified by calculating the average brightness change rate of the pixels. Furthermore, microbubbles typically appear in a certain spatial distribution rather than as random isolated points; therefore, by analyzing the dispersion index of the target pixels, it can be further confirmed whether these brightness changes conform to the spatial characteristics of microbubbles. This analysis method, combining temporal and spatial characteristics, makes the detection of transient optical flicker more accurate and reliable.

[0110] The above technical solution enables more accurate determination of whether transient optical flicker exists in the target video frame, thereby improving the accuracy of microbubble defect detection. This solution avoids false alarms or missed alarms that may result from judging based on a single frame or a single feature. By comprehensively considering the brightness change rate and spatial dispersion index of pixels, the defect identification process is more robust, contributing to improved product quality control in coating production lines.

[0111] In some embodiments described above, a scheme for storing coating production status information to enable product traceability of the coating was proposed. However, in practical applications, the production status information of different coating products may have different levels of importance and traceability requirements. If all production status information adopts a uniform storage strategy, it may lead to wasted storage resources or insufficient data storage lifespan for important data. To address this, this application further proposes an optimized storage scheme. By obtaining a product importance index for the coating and determining a target storage device among multiple storage devices based on this index, differentiated storage of coating production status information can be achieved.

[0112] Stores production status information for the coating, including:

[0113] Obtain the product importance index of the coating; determine the target storage device among multiple storage devices based on the target product importance index; each storage device has a different storage life; store the coating's production status information to the target storage device.

[0114] Specifically, the product importance index can be understood as an indicator measuring the criticality or value level of a coating product in its production, sales, use, and after-sales processes. This index is derived through a comprehensive evaluation based on various factors such as product type, application, market value, customer requirements, and regulatory requirements. For example, high-value, high-risk products, or products requiring long-term traceability, can be assigned a higher product importance index. Multiple storage devices refer to storage media or systems with different storage characteristics and costs, such as high-speed solid-state drives, large-capacity hard disk drives, tape libraries, and cloud storage services. These storage devices have different lifespans; for example, some storage media may be suitable for short-term high-speed access, while others are suitable for long-term archiving. After obtaining the coating's product importance index, the system selects the most suitable storage device as the target storage device based on this index. For example, for products with a high importance index, their production status information may need to be stored in a storage device with a longer lifespan and higher reliability; while for products with a low importance index, a lower-cost storage device with a moderate lifespan can be selected. Thus, the coating's production status information is stored in the determined target storage device.

[0115] This application's solution introduces a product importance index, transforming the storage of coating production status information from a singular, homogenous approach. Specifically, after obtaining the coating's product importance index, the system can quantitatively assess the importance of the production status information based on this index. Subsequently, based on this quantitative assessment, it intelligently selects a target storage device matching the importance index from multiple storage devices with different storage lifetimes. For example, for coating products with high importance and requiring long-term traceability, their production status information is directed to storage media with longer lifespans and higher reliability; while for products with relatively low importance and shorter traceability cycles, they are stored in more cost-effective storage media. This differentiated storage strategy ensures that critical data is properly and long-term preserved, while avoiding over-storage of non-critical data, thereby optimizing the utilization of storage resources.

[0116] Through the above technical solution, this application enables refined management and optimized storage of coating production status information. On the one hand, by storing important data in storage devices with longer storage lifespans, the reliability and traceability of critical production status information are significantly improved, reducing the risk of data loss or damage, thereby better supporting product quality traceability and liability determination. On the other hand, by storing non-critical data in more cost-effective storage devices, storage resources and operating costs are effectively saved, improving the overall efficiency of the storage system. This differentiated storage strategy based on product importance indices allows the storage system to respond more flexibly and efficiently to the traceability needs of different products, providing a more economical and reliable solution for big data multimedia intelligent instrument data processing.

[0117] In some preferred embodiments, assume a coating production line produces two products: a high-performance coating A for aerospace applications and a coating B for ordinary civilian furniture. Due to the criticality of its application and stringent quality requirements, high-performance coating A is assigned a high product importance index, for example, set to 90 points (out of 100). Ordinary civilian furniture coating B, on the other hand, has a relatively lower importance index and is assigned a lower product importance index, for example, set to 30 points. The system pre-configures two storage devices: the first is an enterprise-grade solid-state drive array with a storage lifespan of up to 10 years, but at a higher cost; the second is a conventional hard disk drive array with a storage lifespan of 3 years, but at a lower cost. When production status information for coating A is obtained, since its product importance index is 90 points, the system will store it in the first storage device according to preset rules (e.g., selecting the first storage device if the importance index is higher than 80 points). When production status information for coating B is obtained, since its product importance index is 30 points, the system will store it in the second storage device according to preset rules (e.g., selecting the second storage device if the importance index is lower than 80 points). In this way, the production status information of high-value, high-risk coating A is preserved reliably for a long period, meeting stringent traceability requirements; while the production status information of coating B is stored in a more economical manner, avoiding unnecessary storage costs. This achieves the rational allocation and efficient utilization of storage resources.

[0118] In some of the above embodiments, when storing the production status information of the coating, it is necessary to determine the target storage device among multiple storage devices based on the product importance index of the coating. Specifically, determining the target storage device among multiple storage devices may include the following steps:

[0119] Obtain the preset correspondence; the preset correspondence includes a one-to-one correspondence between multiple product importance index ranges and multiple storage devices, and the maximum value of the product importance index range is positively correlated with the storage life of the corresponding storage device; the storage device corresponding to the product importance index of the coating in the preset correspondence is taken as the target storage device.

[0120] The pre-defined mapping relationship refers to a pre-established set of rules or tables used to match the product importance index of a coating with suitable storage devices. This mapping relationship can be configured based on factors such as actual production needs, data importance levels, storage costs, and the performance of storage devices. The product importance index range can divide the coating's product importance index into different level intervals; for example, it can be divided into low, medium, and high ranges, or more finely detailed ranges. Multiple storage devices can include storage media of different types, performance levels, or storage lifetimes, such as high-speed solid-state drives, large-capacity hard disk drives, tape libraries, or cloud storage services. Storage lifetime refers to the length of time a storage device can reliably retain data. The one-to-one mapping relationship ensures that each product importance index range uniquely corresponds to a specific storage device. The maximum value of the product importance index range is positively correlated with the storage lifetime of the corresponding storage device, meaning that coatings with higher product importance indices will have their production status information stored in storage devices with longer storage lifetimes.

[0121] This application's solution establishes a correlation between product importance indices and storage device lifespan, enabling intelligent selection of storage devices. When the product importance index of a coating is obtained, the system searches for the range to which the product importance index belongs based on a preset correlation, and thus determines the target storage device that matches it. This mechanism ensures that coating production status information of different importance levels can be stored on storage media that match their importance, thereby optimizing the utilization of storage resources.

[0122] The above technical solution enables refined management of coating production status information storage. This solution automatically selects storage devices with appropriate lifespans based on the coating's product importance index, avoiding the risk of losing critical data due to insufficient storage media lifespan, while also preventing the waste of high-cost, long-lifespan storage resources for non-critical data. This improves the reliability and economy of data storage and provides a more robust data foundation for subsequent product traceability.

[0123] This application proposes a data processing system for a multimedia smart instrument based on big data, comprising: an acquisition device and a processing device; the acquisition device is used to acquire production visual data and production environment data of the coating on the coating production line collected by the smart instrument; the production environment data includes production sensor data, production material data or production operation data; the processing device is used to analyze the production visual data based on the production environment data to obtain the production status information of the coating; the processing device is used to store the production status information of the coating for product traceability of the coating.

[0124] The core of the system proposed in this application lies in the deep fusion and intelligent analysis of multi-source heterogeneous data through the collaborative work of the acquisition device and the processing device, so as to achieve accurate perception of the coating production status and effective traceability of product quality.

[0125] Specifically, the data processing method for multimedia smart meters based on big data has already been described in the above embodiments, and will not be repeated here. It should be emphasized that the system proposed in this application implements the above method steps through a specific device.

[0126] The acquisition device can be configured in various forms to adapt to different production environments and data acquisition needs. For example, the acquisition device may include a series of industrial-grade high-definition cameras, infrared sensors, temperature sensors, pressure sensors, flow sensors, humidity sensors, and other physical sensors. These sensors are deployed at key locations on the coating production line to collect real-time visual data of the coating surface and environmental parameters during the production process. Furthermore, the acquisition device can integrate RFID readers or barcode scanners for automatic identification and acquisition of production material data, such as raw material batch information and supplier information. The acquisition device can also connect to the production line control system via a data interface to acquire production operation data, such as equipment start / stop records and parameter adjustment logs. These different data acquisition modules can be integrated into a unified hardware platform or deployed in a distributed manner, communicating with the processing device via wired or wireless networks. As a preferred implementation, the acquisition device can be designed with a modular structure to facilitate flexible configuration and expansion according to the specific needs of the production line.

[0127] The processing device can be implemented as one or more computer servers, embedded systems, cloud computing platforms, or edge computing devices. The core function of the processing device is to efficiently analyze and process the massive amounts of multi-source heterogeneous data collected by the acquisition device. Specifically, the processing device can include a data preprocessing module, an image analysis module, a data fusion module, and a storage management module. The data preprocessing module is responsible for cleaning, format conversion, and time synchronization of the raw data to ensure data consistency and availability. The image analysis module can use deep learning algorithms to perform real-time analysis of production visual data to identify potential defects on the coating surface. The data fusion module is responsible for associating and comprehensively analyzing the image analysis results with production environment data to generate more accurate coating production status information. The storage management module is responsible for structurally storing this production status information in a database and establishing a link with product batch numbers or serial numbers to facilitate subsequent product traceability. The processing device can employ high-performance processors, large-capacity memory, and high-speed storage media to meet the computing and storage requirements of big data processing. In some implementations, the processing device can also be configured with distributed computing capabilities to address the challenges of ultra-large-scale data processing.

[0128] The big data multimedia intelligent instrument data processing system proposed in this application collects production visual data and production environment data of the coating on the coating production line through an acquisition device. The processing device then analyzes the production visual data based on the production environment data to obtain the production status information of the coating. Finally, the processing device stores the production status information of the coating for product traceability. This system can effectively solve the problem in existing technologies of difficulty in identifying and diagnosing latent product defects caused by specific combinations of small, non-obvious process parameter fluctuations.

[0129] Compared with existing technologies, the innovation of this application lies in its ability to deeply integrate and intelligently analyze multi-source heterogeneous data. Traditional methods often treat video information, sensor information, and log information as independent data streams with weak correlations only through timestamps. This makes it difficult for engineers to manually sift through, compare, and identify complex correlation patterns across modalities, multiple factors, and potentially time-lag effects from massive amounts of data. This application, through the collaborative work of the acquisition and processing devices, uses production environment data as auxiliary information for image analysis, achieving a deeper understanding of production visual data, thereby enabling more accurate identification and diagnosis of latent defects. For example, in traditional systems, a tiny bubble defect may be overlooked by the visual inspection system because it is not obvious, or even if it is detected, its cause cannot be accurately determined. However, the system of this application, by combining environmental parameters such as temperature, humidity, and material viscosity, can more accurately determine the type, severity, and potential process causes of the bubble defect. This context-aware analytical capability enables the system to shift from passive threshold alarms to proactive, context-aware process drift prediction and early defect diagnosis, significantly improving the quality control level and efficiency of coating production lines and reducing rework, scrap, and after-sales service costs.

[0130] The above are merely embodiments of this application and are not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A data processing method for multimedia smart meters based on big data, characterized in that, include: Acquire production visual data and production environment data of the coating on the coating production line collected by smart instruments; The production environment data includes production sensor data, production material data, or production operation data. Based on the production environment data, image analysis is performed on the production visual data to obtain the production status information of the coating; The production status information of the coating is stored to enable product traceability of the coating.

2. The data processing method for a multimedia intelligent instrument based on big data according to claim 1, characterized in that, The production visual data includes a video stream. Based on the production environment data, image analysis is performed on the production visual data to obtain the production status information of the coating, including: Multiple target video frames are extracted from the video stream at preset frame intervals; For each of the multiple target video frames, determine whether there is defect information in the target video frame based on the target video frame and the two video frames before and after the target video frame; When no defect information is found in the target video frame, the production status information of the coating is determined to indicate that no defect information is found in the target area of ​​the coating; the shooting time of the target video frame is the coating time of the target area in the coating. When defect information exists in the target video, the confidence level of the defect information is determined based on the production environment data; When the confidence level of the defect information is greater than or equal to a preset confidence threshold, the production status information of the coating indicates that there is defect information in the target area of ​​the coating. When the confidence level of the defect information is less than a preset confidence threshold, the production status information of the coating indicates that there is no defect information in the target area of ​​the coating.

3. The data processing method for a multimedia intelligent instrument based on big data according to claim 2, characterized in that, When the defect information is a microbubble defect, determining whether a defect exists in the target video based on the target video frame and the two videos preceding and following the target video frame includes: Based on the target video frame and the two video frames before and after the target video frame, determine whether there is transient optical flicker in the target video frame; When there is no transient optical flicker in the target video frame, the production status information of the coating is determined to indicate that there are no microbubble defects in the target area of ​​the coating; When transient optical flicker is present in the target video frame, the production status information of the coating is determined to indicate that microbubble defects exist in the target area of ​​the coating.

4. The data processing method for a multimedia intelligent instrument based on big data according to claim 3, characterized in that, Determining the confidence level of the defect information based on the production environment data includes: The first moment, a preset duration before the shooting time of the target video frame, is taken as the start time of the target time period, and the second moment, a preset duration after the shooting time, is taken as the end time of the target time period. For each data point in the production environment data, determine the changes in that data within the target time period; The confidence level of the defect information is determined based on the change information of each data point in the production environment data.

5. The data processing method for a multimedia intelligent instrument based on big data according to claim 4, characterized in that, The confidence level of the defect information is determined based on the change information of each data point in the production environment data, including: For each data point in the production environment data, obtain preset change information of the data when the coating has microbubble defects; Determine the similarity between the changes in the data within the target time period and the corresponding preset changes; The confidence level of microbubble defects is determined based on the similarity of multiple data points in the production environment data.

6. The data processing method for a multimedia intelligent instrument based on big data according to claim 5, characterized in that, Determining the confidence level of microbubble defects based on the similarity of the multiple data sets includes: Obtain the defect impact weight of each data point from the plurality of data; the sum of the defect impact weights of the plurality of data points is 1; The weighted sum of the similarity of multiple data points is determined based on the impact weight of defects in multiple data points; The weighted sum is determined as the confidence level of the microbubble defect.

7. The data processing method for a multimedia intelligent instrument based on big data according to claim 3, characterized in that, Based on the target video frame and the two video frames preceding and following the target video frame, determine whether there is transient optical flicker in the target video frame, including: For each pixel in the target video frame, determine the average value of the brightness change rate of the pixel in the target video frame and the two preceding and following video frames; Pixels whose average brightness change rate is greater than a preset brightness change rate threshold are identified as target pixels. Determine the dispersion index of the target pixel; When the dispersion index is greater than a preset dispersion index threshold, it is determined that there is transient optical flicker in the target video frame; otherwise, it is determined that there is no transient optical flicker in the target video frame.

8. The data processing method for a multimedia intelligent instrument based on big data according to claim 1, characterized in that, The production status information of the coating is stored, including: Obtain the product importance index of the coating; The target storage device is determined from multiple storage devices based on the importance index of the target product; each storage device has a different storage lifespan. The production status information of the coating is stored in the target storage device.

9. The data processing method for a multimedia intelligent instrument based on big data according to claim 8, characterized in that, The target storage device among multiple storage devices is determined based on the target product importance index, including: Obtain a preset correspondence; the preset correspondence includes a one-to-one correspondence between multiple product importance index ranges and multiple storage devices, and the maximum value of the product importance index range is positively correlated with the storage life of the corresponding storage device; The storage device corresponding to the product importance index of the coating in the preset correspondence is taken as the target storage device.

10. A data processing system for multimedia intelligent meters based on big data, characterized in that, include: Acquisition device and processing device; The acquisition device is used to acquire production visual data and production environment data of the coating on the coating production line collected by the smart instrument. The production environment data includes production sensor data, production material data, or production operation data. The processing device is used to analyze the production visual data based on the production environment data to obtain the production status information of the coating. The processing device is used to store the production status information of the coating in order to trace the coating as a product.

Images