Data interconnection method of intelligent computing power box supporting domestic industrial protocol

Abstract

The application discloses a data interconnection method of an intelligent computing power box supporting a domestic industrial protocol, and relates to the technical field of data processing. The technical scheme points of the method comprise the following steps: real-time monitoring of an interconnection link between the intelligent computing power box and a target detection product to obtain initial actual defect data; marking a region with an explicit defect feature in the target detection product as an explicit defect product region; extracting first historical detection data to which the explicit defect product region belongs from a historical interconnection detection database of the intelligent computing power box; processing the first historical detection data to obtain a first judgment influence coefficient; and performing error compensation on the initial actual defect data according to the first judgment influence coefficient to obtain first actual defect data. The effect is to improve the response speed of quality control and the resource utilization efficiency in industrial production.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and more specifically, to a data interconnection method for smart computing boxes that supports domestic industrial protocols. Background Technology

[0002] With the trend of industrial intelligent upgrading, the interconnection between intelligent computing boxes and target inspection products has become a key link in improving production quality. Traditional industrial interconnection relies heavily on foreign protocols, which not only pose risks such as technological dependence and data security, but also make it difficult to adapt to the communication characteristics of domestic industrial equipment. This results in poor interconnection compatibility and insufficient data transmission stability between domestically produced computing boxes and inspection products. In industrial defect detection, there are significant differences in the characteristics of explicit and implicit defects, but existing interconnection methods often use a unified detection data processing logic without distinguishing the transmission interference characteristics of the two types of defects. For example, explicit defect data is easily affected by link delay and packet loss, while implicit defect data, due to its subtle characteristics, is more easily amplified by node transmission fluctuations, directly leading to a large deviation in the initial defect data judgment, thus making it difficult to support accurate quality control. Summary of the Invention

[0003] In view of the shortcomings of existing technologies, the purpose of this invention is to provide a data interconnection method for intelligent computing boxes that supports domestic industrial protocols.

[0004] To achieve the above objectives, the present invention provides the following technical solution:

[0005] A method for data interconnection of smart computing boxes supporting domestic industrial protocols, comprising the following steps:

[0006] Initial actual defect data is obtained by real-time monitoring of the interconnection link between the intelligent computing box and the target detection product;

[0007] The area with obvious defect features in the target inspection product is marked as the obvious defect product area. The first historical inspection data of the obvious defect product area is extracted from the historical interconnection inspection database of the smart computing box. The first historical inspection data is processed to obtain the first judgment influence coefficient. The initial actual defect data is error-compensated according to the first judgment influence coefficient to obtain the first actual defect data.

[0008] The area with latent defect features in the target inspection product is marked as the latent defect product area. The second historical inspection data to which the latent defect product area belongs is extracted from the historical interconnection inspection database of the smart computing box. The second historical inspection data is processed to obtain the second judgment influence coefficient. The second judgment influence coefficient and the transmission data of the latent defect product area are processed and the initial actual defect data is error-compensated to obtain the second actual defect data.

[0009] After processing the defect warning judgment based on the first and second actual defect data, a defect warning notification is output.

[0010] Preferably, the method further includes the following steps:

[0011] The delay difference of the visible defect product area in adjacent detection cycles is obtained by acquiring the transmission delay difference of the visible defect product area.

[0012] Preferably, the first historical detection data is processed to obtain the first judgment influence coefficient, specifically including the following steps:

[0013] Determine whether the area of ​​a product with obvious defects is within the high-resolution detection area of ​​the target product.

[0014] If the area of ​​the product with obvious defects is not located in the high-resolution detection area of ​​the target product, the historical delay difference of the obvious area to be tested is extracted from the first historical detection data based on the delay difference of the obvious area to be tested, and the historical judgment error value 1 that causes defect judgment under the influence of the historical delay difference of the obvious area is obtained; the correlation change trend characteristics between the historical judgment error value 1 and the historical delay difference of the obvious area are statistically analyzed to obtain the first delay influence coefficient.

[0015] If the area of ​​the product with obvious defects is located in the high-resolution detection area of ​​the target product, the packet loss rate of the interconnection link data corresponding to the area of ​​the product with obvious defects is used to obtain the packet loss rate of the area to be tested. The packet loss rate of the area to be tested and the first historical detection data are processed to obtain the first delay and loss impact coefficient.

[0016] The first delay impact coefficient and the first loss impact coefficient are combined to form the first judgment impact coefficient.

[0017] Preferably, the packet loss rate of the visible region to be tested and the first historical detection data are processed to obtain the first delay impact coefficient, specifically including the following steps:

[0018] The packet loss rate of the visible region to be tested is the proportion of data packets lost when transmitting visible defect detection data through the interconnection link.

[0019] Based on the packet loss rate of the visible region to be tested, the historical packet loss rate 1 is extracted from the first historical detection data, and the historical packet loss rate 1 and the historical visible region delay difference are combined to form visible defect interconnection feature data.

[0020] The second historical judgment error value is extracted from the first historical detection data, which is caused by the combined influence of the delay difference in the historical explicit area and the historical packet loss rate.

[0021] The first delay influence coefficient is obtained by statistically analyzing the correlation and change trends of historical judgment error value 2 and explicit defect interconnection feature data.

[0022] Preferably, the first actual defect data is obtained by performing error compensation on the initial actual defect data according to the first judgment influence coefficient, specifically including the following steps:

[0023] The first defect compensation error value for the product area with obvious defects is obtained based on the first judgment influence coefficient.

[0024] The first actual defect data is obtained by performing error compensation processing on the initial actual defect data based on the error value that needs to be compensated for in the first defect.

[0025] Preferably, the method further includes the following steps:

[0026] The transmission delay data of each detection node in the interconnection link corresponding to the product area with hidden defects is detected separately to obtain node transmission delay data and link terminal transmission delay data.

[0027] The average node transmission delay is obtained by averaging the two adjacent sets of node delay data.

[0028] The set of hidden region delay differences is obtained by subtracting the average transmission delay of a node from the transmission delay data of the corresponding link terminal of the adjacent node.

[0029] The average delay difference of the hidden region is obtained by averaging all the hidden region delay difference values ​​in the set of hidden region delay difference values.

[0030] Preferably, the second historical detection data is processed to obtain the second judgment influence coefficient, specifically including the following steps:

[0031] Determine whether the area of ​​the product with hidden defects is within the high-resolution detection area of ​​the target product.

[0032] If the area of ​​the latent defect product is not located in the high-resolution detection area of ​​the target product, the historical latent area delay difference is extracted from the second historical detection data based on the average delay difference of the latent area to be tested, and the historical judgment error value three is obtained under the influence of the historical latent area delay difference; the correlation and change trend of the historical judgment error value three and the historical latent area delay difference are statistically analyzed to obtain the second delay influence coefficient.

[0033] If the latent defect product area is located in the high-resolution detection area of ​​the target product, the packet loss rate of the interconnection link data corresponding to the latent defect product area is used to obtain the packet loss rate of the latent area to be tested. The second historical packet loss rate is extracted from the second historical detection data. The second historical packet loss rate and the historical latent area delay difference are combined to form the latent defect interconnection feature data.

[0034] The historical judgment error value four, which is jointly affected by the historical latent area delay difference and the historical packet loss rate, is extracted from the second historical detection data; the second delay and packet loss influence coefficient is obtained by statistically analyzing the correlation and change trend characteristics of the historical judgment error value four and the interconnection feature data of latent defects.

[0035] The second delay impact coefficient and the second loss impact coefficient are combined to form the second judgment impact coefficient.

[0036] Preferably, the second actual defect data is obtained by processing the transmission data of the second judgment influence coefficient and the latent defect product area and then performing error compensation on the initial actual defect data. Specifically, this includes the following steps:

[0037] The first defect compensation error value for the hidden defect product area is obtained based on the second judgment influence coefficient and the average transmission delay difference in the hidden defect product area.

[0038] The second defect compensation error value is obtained based on the second judgment influence coefficient and the transmission delay difference between each adjacent detection node in the product area with hidden defects;

[0039] The second actual defect data is obtained by performing error compensation on the initial actual defect data based on the first defect compensation error value and the second defect compensation error value.

[0040] Preferably, the second defect compensation error value is obtained based on the second judgment influence coefficient and the transmission delay difference between adjacent detection nodes belonging to the latent defect product area, specifically including the following steps:

[0041] The packet loss rate of the hidden region to be tested is divided into several sub-proportions to obtain the packet loss rate dataset of the sub-region to be tested;

[0042] Based on the test sub-packet loss rate dataset, the hidden area delay difference set, and the second judgment influence coefficient, the defect compensation sub-error dataset generated by the defects in the hidden defect product area under the influence of transmission delay and data packet loss rate is obtained.

[0043] The second defect compensation error value is obtained by summing up the sub-error data in the defect compensation sub-error dataset.

[0044] The third defect compensation error value is obtained by averaging the first defect compensation error value and the second defect compensation error value.

[0045] The second actual defect data is obtained by performing error compensation processing on the initial actual defect data based on the third defect error value.

[0046] Preferably, after processing the defect warning judgment based on the first actual defect data and the second actual defect data, the defect warning notification information is output, specifically including the following steps:

[0047] The first actual defect data and the second actual defect data are compared with the preset warning trigger threshold to determine whether the warning trigger requirements are met.

[0048] If the warning triggering conditions are met, the warning level is determined according to the corresponding warning triggering threshold, and then the defect warning notification information is output.

[0049] If the warning triggering conditions are not met, no defect warning notification information will be output.

[0050] Compared with the prior art, the present invention has the following beneficial effects:

[0051] This invention distinguishes between areas of products with visible defects and areas of products with hidden defects. It calculates a first and a second judgment influence coefficient based on historical inspection data, and then performs targeted error compensation. This effectively corrects deviations caused by transmission interference. For example, after compensation, visible defect data more closely reflects the actual defect severity. Hidden defects are further refined through sub-proportional segmentation and node delay difference analysis, reducing the risk of misjudging minor defects. This results in more accurate first and second actual defect data, providing a reliable basis for subsequent quality control. A tiered early warning mechanism compares the corrected defect data with preset thresholds, quickly distinguishing the severity of defects and triggering corresponding warnings. This avoids production interruptions caused by excessive warnings and prevents the underestimation of serious defects, allowing staff to allocate resources more effectively and improving the response speed and resource utilization efficiency of quality control in industrial production. Attached Figure Description

[0052] Figure 1 A schematic diagram illustrating the steps of the intelligent computing box data interconnection method supporting domestic industrial protocols proposed in this invention;

[0053] Figure 2 This is a schematic diagram illustrating the steps in obtaining the first delay impact coefficient in the intelligent computing box data interconnection method supporting domestic industrial protocols proposed in this invention. Detailed Implementation

[0054] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0055] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0056] Secondly, the term "an embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places throughout this specification does not necessarily refer to the same embodiment, nor is it a single embodiment or an embodiment selectively excluded from other embodiments.

[0057] Reference Figures 1-2 As shown.

[0058] The embodiments further illustrate the data interconnection method for smart computing boxes that supports domestic industrial protocols proposed in this invention.

[0059] A method for data interconnection of smart computing boxes supporting domestic industrial protocols, comprising the following steps:

[0060] Initial actual defect data is obtained by real-time monitoring of the interconnection link between the intelligent computing box and the target detection product;

[0061] The intelligent computing box continuously interacts with the target inspection product, while simultaneously activating a link monitoring mechanism to collect various data in real time during the interconnection process. This data covers link transmission latency, data packet loss, and the integrity of defect feature transmission. During this process, the target inspection product transmits its detected defect-related data to the intelligent computing box via a domestic industrial protocol. The intelligent computing box synchronously records the status of this data during transmission, ultimately integrating it to form the initial actual defect data.

[0062] Assuming the target product to be inspected is a vision inspection device on an industrial production line, it detects a scratch on the surface of a product. The device then sends data such as the location and size of the scratch to a smart computing box via an interconnected link. Upon receiving this data, the smart computing box simultaneously records the time difference between the data transmission from the device and reception, as well as any data packet loss information. This data, combined with the scratch detection data itself, forms the initial actual defect data.

[0063] The area with obvious defect features in the target inspection product is marked as the obvious defect product area. The first historical inspection data of the obvious defect product area is extracted from the historical interconnection inspection database of the smart computing box. The first historical inspection data is processed to obtain the first judgment influence coefficient. The initial actual defect data is error-compensated according to the first judgment influence coefficient to obtain the first actual defect data.

[0064] The area with latent defect features in the target inspection product is marked as the latent defect product area. The second historical inspection data to which the latent defect product area belongs is extracted from the historical interconnection inspection database of the smart computing box. The second historical inspection data is processed to obtain the second judgment influence coefficient. The second judgment influence coefficient and the transmission data of the latent defect product area are processed and the initial actual defect data is error-compensated to obtain the second actual defect data.

[0065] Visible defect areas are defective areas that can be directly and intuitively identified by the target product being inspected. These defects usually have obvious appearance features, such as cracks, dents, and stains on the surface of industrial products. Their features differ significantly from normal areas and can be captured by inspection equipment without additional analysis or magnification. On the other hand, defects corresponding to latent defect areas do not have intuitive appearance features, such as micro-cracks inside the product or latent stress damage in the material. The features of these defects differ very slightly from normal areas and may even require specific detection algorithms or high-sensitivity sensors to identify.

[0066] For example, obvious scratches on the surface of a product are considered visible defects, which can be clearly captured by a regular lens of the inspection equipment; however, micro-areas with uneven thickness inside the product are considered hidden defects, which require ultrasonic testing combined with algorithms for identification. Furthermore, the latency fluctuations of the link nodes during the transmission of the test data can more easily cause deviations in the defect parameters.

[0067] After processing the defect warning judgment based on the first and second actual defect data, a defect warning notification is output.

[0068] It also includes the following steps:

[0069] The delay difference of the visible defect product area in adjacent detection cycles is obtained by acquiring the transmission delay difference of the visible defect product area.

[0070] The detection cycle is a fixed time interval for the smart computing box to collect a complete set of data from the interconnection link, such as 10 seconds per detection cycle. The transmission delay is the time required for the detection data of the product area with obvious defects to be sent from the target product to the smart computing box. In real-world scenarios, network load and temporary equipment failures can cause differences in transmission delays between different cycles, and these differences can interfere with the accuracy of the defect data. Therefore, it is necessary to calculate the delay difference between adjacent cycles.

[0071] In the current testing cycle, the intelligent computing box records the transmission delay T1 of the data corresponding to the product area with obvious defects; in the next consecutive testing cycle, the transmission delay of the data in the same area is also recorded as T2, and the delay difference of the obvious area to be tested is T2-T1.

[0072] Assume that in the first detection cycle, it takes 200 milliseconds (T1 = 200ms) for data from a visible defect area (such as a noticeable crack on the product surface) to be transmitted from the product to the computing box; in the second detection cycle, the transmission time for this area becomes 250 milliseconds (T2 = 250ms). Then the latency difference for the visible defect area is 250ms - 200ms = 50ms. This 50ms difference reflects the magnitude of the change in data transmission latency for this area between adjacent cycles.

[0073] The first historical detection data is processed to obtain the first judgment influence coefficient, which specifically includes the following steps:

[0074] Determine whether the area of ​​a product with obvious defects is within the high-resolution detection area of ​​the target product.

[0075] If the area of ​​the product with obvious defects is not located in the high-resolution detection area of ​​the target product, the historical delay difference of the obvious area to be tested is extracted from the first historical detection data based on the delay difference of the obvious area to be tested, and the historical judgment error value 1 that causes defect judgment under the influence of the historical delay difference of the obvious area is obtained; the correlation change trend characteristics between the historical judgment error value 1 and the historical delay difference of the obvious area are statistically analyzed to obtain the first delay influence coefficient.

[0076] If the area of ​​the product with obvious defects is located in the high-resolution detection area of ​​the target product, the packet loss rate of the interconnection link data corresponding to the area of ​​the product with obvious defects is used to obtain the packet loss rate of the area to be tested. The packet loss rate of the area to be tested and the first historical detection data are processed to obtain the first delay and loss impact coefficient.

[0077] The first delay impact coefficient and the first loss impact coefficient are combined to form the first judgment impact coefficient.

[0078] First, the target inspection product pre-divides the inspection area and records the resolution parameters of each area. For example, industrial vision inspection equipment divides the area covered by the lens into different sub-areas and stores the pixel density corresponding to each sub-area. High-resolution inspection areas usually correspond to higher pixel densities, such as 1,000 pixels per square centimeter. These parameters are pre-entered into the configuration database of the target inspection product.

[0079] The target detection product then calls the region segmentation data in the configuration database, extracts the resolution parameters corresponding to the region, and determines whether it meets the high resolution threshold standard, such as pixel density ≥ 800 pixels / square centimeter.

[0080] If the coordinates of a visible defect area fall within the preset area A, and the pixel density of area A is 1200 pixels / square centimeter, which meets the high resolution threshold, then the visible defect area can be determined to be in the high resolution detection area; if the defect area falls within area B, and the pixel density of area B is only 300 pixels / square centimeter, then it is determined not to be in the high resolution detection area.

[0081] Assuming the visible defect area is a clear crack on the product surface, first determine if this area is within the high-resolution detection area. If the crack area is not within the high-resolution detection area, then the focus is on the impact of transmission delay. Extract the historical visible region delay difference corresponding to the current visible region delay difference from the first historical detection data. For example, historical data may contain records with delay differences of 40ms, 50ms, and 60ms. Obtain the historical judgment error value corresponding to these historical delay differences. For example, the error value is 0.02 for a delay difference of 40ms, 0.03 for a delay difference of 50ms, and 0.04 for a delay difference of 60ms. Analyze the correlation trend between the two. For example, for every 10ms increase in delay difference, the error value increases by 0.01. Calculate the first delay influence coefficient based on this trend. Assuming that through fitting, the first delay influence coefficient = 0.001 × historical visible region delay difference, the first delay influence coefficient for a 50ms delay difference is 0.05.

[0082] If the cracked area is within a high-resolution detection zone, the impact of packet loss must also be considered. The packet loss rate of the interconnecting link corresponding to this area is used to obtain the packet loss rate of the visible area under test, for example, a packet loss rate of 5%. Historical packet loss rate one matching this rate is extracted from the first historical detection data. This is combined with the corresponding historical visible area delay difference to form visible defect interconnection feature data. For example, there may be a historical record with a packet loss rate of 5% and a delay difference of 50ms. The historical judgment error value two corresponding to this feature data is obtained, for example, a historical judgment error value two of 0.04. The correlation trend between historical judgment error value two and this feature data is statistically analyzed. For example, for every 1% increase in the packet loss rate and a 10ms increase in the delay difference, the error value increases by 0.01. Based on this, the first delay impact coefficient is obtained. For example, corresponding to a 5% packet loss rate and a 50ms delay difference, the first delay impact coefficient is 0.04.

[0083] The first judgment influence coefficient is composed of the first delay influence coefficient and the first delay loss influence coefficient. If it is in a non-high resolution area, the first judgment influence coefficient is the first delay influence coefficient; if it is in a high resolution area, the first judgment influence coefficient is the first delay loss influence coefficient. Subsequently, the first judgment influence coefficient will be used to correct the error of the initial defect data.

[0084] The first delay / loss impact coefficient is obtained by processing the packet loss rate of the visible region to be tested and the first historical detection data, specifically including the following steps:

[0085] The packet loss rate of the visible area under test is the proportion of data packets lost when transmitting visible defect detection data through the interconnection link;

[0086] Based on the packet loss rate of the visible region to be tested, the historical packet loss rate 1 is extracted from the first historical detection data, and the historical packet loss rate 1 and the historical visible region delay difference are combined to form visible defect interconnection feature data.

[0087] The second historical judgment error value is extracted from the first historical detection data, which is caused by the combined influence of the delay difference in the historical explicit area and the historical packet loss rate.

[0088] The first delay influence coefficient is obtained by statistically analyzing the correlation and change trends of historical judgment error value 2 and explicit defect interconnection feature data.

[0089] The packet loss rate of the visible area under test is the ratio of the number of data packets that fail to be delivered to the total number of data packets sent when the interconnection link transmits the detection data of the visible defect. It directly reflects the stability of the link transmission. The higher the packet loss rate, the more information of the defect data is lost during transmission.

[0090] Assuming the visible defect area is a crack on the product surface and is within a high-resolution detection area, and the current detected packet loss rate for the visible area under test is 5%, extract a historical packet loss rate matching this packet loss rate from the first historical detection data. For example, if there is a record with a packet loss rate of 5% in the historical data, extract the corresponding historical visible area delay difference, such as 50ms in the previous example. Combine these two parameters to form the visible defect interconnection feature data, namely, the packet loss rate of 5% and the delay difference of 50ms. This feature data integrates the two core interference factors of packet loss and delay during transmission.

[0091] Extract the second historical judgment error value corresponding to the interconnection feature data of the explicit defect from the first historical detection data. That is, under the combined influence of a packet loss rate of 5% and a latency difference of 50ms, the deviation between the defect judgment result and the actual defect situation. For example, the second historical judgment error value corresponding to this feature is 0.04.

[0092] It is necessary to statistically analyze the correlation and changing trend of historical judgment error value 2 with the interconnection feature data of explicit defects. Specifically, this involves determining the error value change pattern corresponding to different combinations of packet loss rate and latency difference. When the packet loss rate increases by 1% and the latency difference increases by 10ms, the historical judgment error value 2 increases by 0.01. Based on this trend and combined with the current interconnection feature data of explicit defects, the first latency impact coefficient is calculated. Taking the current packet loss rate of 5% and latency difference of 50ms as an example, the first latency impact coefficient = 0.001 × packet loss rate + 0.0008 × latency difference. Substituting the parameters, we get the first latency impact coefficient = 0.001 × 5 + 0.0008 × 50 = 0.005 + 0.04 = 0.045. The first latency impact coefficient quantifies the degree of interference of the current transmission state on the judgment of explicit defects, and will be used to correct the error of the initial actual defect data.

[0093] The first actual defect data is obtained by performing error compensation on the initial actual defect data based on the first judgment influence coefficient, specifically including the following steps:

[0094] The first defect compensation error value for the product area with obvious defects is obtained based on the first judgment influence coefficient.

[0095] The first actual defect data is obtained by performing error compensation processing on the initial actual defect data based on the error value that needs to be compensated for in the first defect.

[0096] The first judgment influence coefficient is a quantitative result of the degree of interference of obvious defect data with factors such as delay and packet loss during transmission. The first defect compensation error value is the specific value of the error that needs to be corrected calculated based on this coefficient.

[0097] Assuming the first judgment influence coefficient is 0.045, and the initial actual defect data is the defect severity value directly obtained from the target inspection product, for example, the initial actual defect data is 0.8. The first defect compensation error value = initial actual defect data × first judgment influence coefficient. Substituting the parameters, we get: 0.8 × 0.045 = 0.036, which is the amount of defect data deviation caused by the current transmission interference.

[0098] Next, error compensation is performed. Since the first judgment influence coefficient corresponds to the positive error caused by transmission interference, meaning the initial data is too high or too low due to interference, the initial data needs to be adjusted according to the direction of the error during actual compensation. Assuming the first judgment influence coefficient corresponds to the error of the initial data being too high, the first actual defect data = initial actual defect data - first defect compensation error value. Substituting the parameters, we get the first actual defect data as 0.8 - 0.036 = 0.764.

[0099] This process uses a quantified interference coefficient to correct the deviation of the initial data, making the first actual defect data closer to the true state of the defect. For example, the original initial data of 0.8 was too high due to transmission packet loss and delay. After compensation, the resulting 0.764 more accurately reflects the actual severity of the product crack area, providing a reliable data foundation for subsequent defect early warning.

[0100] It also includes the following steps:

[0101] The transmission delay data of each detection node in the interconnection link corresponding to the product area with hidden defects is detected separately to obtain node transmission delay data and link terminal transmission delay data.

[0102] The average node transmission delay is obtained by averaging the two adjacent sets of node delay data.

[0103] The set of hidden region delay differences is obtained by subtracting the average transmission delay of a node from the transmission delay data of the corresponding link terminal of the adjacent node.

[0104] The average delay difference of the hidden region is obtained by averaging all the hidden region delay difference values ​​in the set of hidden region delay difference values.

[0105] First, it is necessary to clarify the interconnection structure corresponding to the latent defect area. This type of link typically includes multiple detection nodes and link terminals. The transmission delay data of each detection node and the transmission delay data of each link terminal are then detected: the node transmission delay data is the transmission time of the latent defect detection data as it passes through each node, and the link terminal transmission delay data is the total time it takes for the data to travel from the target detection product to the intelligent computing box.

[0106] Assuming the latent defect area is a micro-gap area inside the product, its interconnection link contains 3 detection nodes. The detected node transmission delay data is 10ms for node 1, 12ms for node 2, and 11ms for node 3; the transmission delay data at the link end is 35ms.

[0107] The average transmission delay of a node is obtained by averaging the delay data of two adjacent sets of nodes. The average transmission delay of a node is calculated as (adjacent node delay data 1 + adjacent node delay data 2) / 2. Taking nodes 1 and 2 in the example as an example, the average transmission delay of nodes 1 and 2 is (10ms + 12ms) / 2 = 11ms; the average transmission delay of nodes 2 and 3 is (12ms + 11ms) / 2 = 11.5ms.

[0108] The implicit region delay difference set is obtained by subtracting the average transmission delay of a node from the corresponding link termination transmission delay data of its neighboring nodes. "Neighboring node" refers to the termination delay of a link segment ending at that node. The implicit region delay difference = link termination transmission delay data - average node transmission delay. For example, with the average transmission delay of nodes 1 and 2 (11ms), the corresponding implicit region delay difference = 35ms - 11ms = 24ms; with the average transmission delay of nodes 2 and 3 (11.5ms), the difference = 35ms - 11.5ms = 23.5ms. Therefore, the implicit region delay difference sets are 24ms and 23.5ms.

[0109] The average delay difference of the hidden region is obtained by averaging all the differences in the set of hidden region delay differences. The average delay difference of the hidden region is equal to the sum of all values ​​in the set of hidden region delay differences divided by the number of differences. Substituting the example data, the average delay difference of the hidden region is (24ms + 23.5ms) / 2 = 23.75ms.

[0110] The second historical detection data is processed to obtain the second judgment influence coefficient, which specifically includes the following steps:

[0111] Determine whether the area of ​​the product with hidden defects is within the high-resolution detection area of ​​the target product.

[0112] If the area of ​​the latent defect product is not located in the high-resolution detection area of ​​the target product, the historical latent area delay difference is extracted from the second historical detection data based on the average delay difference of the latent area to be tested, and the historical judgment error value three is obtained under the influence of the historical latent area delay difference; the correlation and change trend of the historical judgment error value three and the historical latent area delay difference are statistically analyzed to obtain the second delay influence coefficient.

[0113] If the latent defect product area is located in the high-resolution detection area of ​​the target product, the packet loss rate of the interconnection link data corresponding to the latent defect product area is used to obtain the packet loss rate of the latent area to be tested. The second historical packet loss rate is extracted from the second historical detection data. The second historical packet loss rate and the historical latent area delay difference are combined to form the latent defect interconnection feature data.

[0114] The historical judgment error value four, which is jointly affected by the historical latent area delay difference and the historical packet loss rate, is extracted from the second historical detection data; the second delay and packet loss influence coefficient is obtained by statistically analyzing the correlation and change trend characteristics of the historical judgment error value four and the interconnection feature data of latent defects.

[0115] The second delay impact coefficient and the second loss impact coefficient are combined to form the second judgment impact coefficient.

[0116] Assuming the latent defect area is a micro-gap region inside the product, the average delay difference of the latent area to be tested is 23.75ms. If this area is not within the high-resolution detection area, extract the historical latent area delay difference corresponding to 23.75ms from the second historical detection data. For example, there are historical records of delay difference values ​​of 23ms and 24ms. Obtain the historical judgment error values ​​corresponding to these historical delay differences, for example, 23ms corresponds to an error value of 0.025, and 24ms corresponds to an error value of 0.026. Analyze the correlation trend between the two. For example, for every 1ms increase in the delay difference, the error value increases by 0.001. Based on this, obtain the second delay influence coefficient, which is 0.001 × historical latent area delay difference, thus obtaining a second delay influence coefficient of 0.02375.

[0117] If the micro-gap region is within the high-resolution detection area, the packet loss rate of the corresponding interconnect link data is used to obtain the packet loss rate of the hidden region under test, for example, the packet loss rate of the hidden region under test is 4%. From the second historical detection data, extract the second historical packet loss rate that matches the packet loss rate of the hidden region under test (4%), and extract the corresponding historical hidden region latency difference, for example, a historical hidden region latency difference of 23.75ms. Combine these two to form the packet loss rate of the hidden defect interconnect feature data (4%) and the latency difference of 23.75ms. From the second historical detection data, extract the fourth historical judgment error value corresponding to this feature, for example, a historical judgment error value of 0.03. The correlation trend between the statistical error value and this feature is as follows: for example, for every 1% increase in packet loss rate and 5ms increase in latency difference, the error value increases by 0.008. Based on this, the second latency impact coefficient is obtained. The second latency impact coefficient = 0.002 × packet loss rate + 0.0003 × latency difference. Substituting the parameters, we get 0.002 × 4 + 0.0003 × 23.75 = 0.008 + 0.007125 = 0.015125.

[0118] The second judgment influence coefficient is a combination of the second delay influence coefficient and the second delay loss influence coefficient: if the region is not in a high-resolution region, the second judgment influence coefficient is the second delay influence coefficient; if it is in a high-resolution region, the second judgment influence coefficient is the second delay loss influence coefficient. This coefficient will be used to correct the error of the latent defect data.

[0119] After processing the transmission data of the second judgment influence coefficient and the latent defect product area, error compensation is performed on the initial actual defect data to obtain the second actual defect data. The specific steps include:

[0120] The first defect compensation error value for the hidden defect product area is obtained based on the second judgment influence coefficient and the average transmission delay difference in the hidden defect product area.

[0121] The second defect compensation error value is obtained based on the second judgment influence coefficient and the transmission delay difference between each adjacent detection node in the product area with hidden defects;

[0122] The second actual defect data is obtained by performing error compensation on the initial actual defect data based on the first defect compensation error value and the second defect compensation error value.

[0123] The first defect compensation error value is based on the overall average transmission delay difference, while the second defect compensation error value focuses on the delay difference between each node. The combination of the two can more comprehensively cover the transmission interference of hidden defect data.

[0124] Assuming the latent defect area is a micro-gap area inside the product, the second judgment influence coefficient is 0.015125, which corresponds to the second delay influence coefficient of the high-resolution area. The average delay difference of the latent area to be tested is 23.75ms, and the initial actual defect data is 0.5, representing the severity of the latent defect of the micro-gap.

[0125] The first defect compensation error value = the second judgment influence coefficient × the average delay difference of the hidden area to be tested. Substituting the parameters, we can get: 0.015125 × 23.75 = 0.35921875. The first defect compensation error value reflects the degree of interference of the overall delay fluctuation on the judgment of the hidden defect.

[0126] The second defect compensation error value = the second judgment influence coefficient × the sum of the delay differences of each adjacent node. Substituting the parameters, we get: 0.015125 × (24 + 23.5) = 0.015125 × 47.5 = 0.7184375. The second defect compensation error value reflects the additional interference of the delay difference between nodes on the latent defect data.

[0127] Finally, error compensation processing is performed. Since the error of the latent defect data is usually the superposition of multi-dimensional interference, the second actual defect data = the initial actual defect data - (the error value to be compensated for the first defect + the error value to be compensated for the second defect) / 2.

[0128] The second defect compensation error value is obtained based on the second judgment influence coefficient and the transmission delay difference between adjacent detection nodes in the product area with hidden defects. The specific steps include:

[0129] The packet loss rate of the hidden region to be tested is divided into several sub-proportions to obtain the packet loss rate dataset of the sub-region to be tested;

[0130] Based on the test sub-packet loss rate dataset, the hidden area delay difference set, and the second judgment influence coefficient, the defect compensation sub-error dataset generated by the defects in the hidden defect product area under the influence of transmission delay and data packet loss rate is obtained.

[0131] The second defect compensation error value is obtained by summing up the sub-error data in the defect compensation sub-error dataset.

[0132] The third defect compensation error value is obtained by averaging the first defect compensation error value and the second defect compensation error value.

[0133] The second actual defect data is obtained by performing error compensation processing on the initial actual defect data based on the third defect error value.

[0134] First, the packet loss rate of the hidden region to be tested is divided into sub-proportions. The overall packet loss rate is broken down into multiple sub-proportions to cover the differentiated impact of different packet loss intervals on the defect data. For example, if the packet loss rate of the hidden region to be tested is 4%, it is divided into two equal-proportion sub-packet loss rates, resulting in the packet loss rate datasets to be tested as 2% and 2%.

[0135] Calculating the defect compensation sub-error dataset requires combining the test sub-packet loss rate dataset, the hidden region delay difference set, and the second judgment influence coefficient. The defect compensation sub-error = second judgment influence coefficient × test sub-packet loss rate × the hidden region delay difference of the corresponding node. Taking the example data, the first sub-packet loss rate of 2% corresponds to the first delay difference of 24ms, so the defect compensation sub-error = 0.015125 × 2% × 24 = 0.015125 × 0.02 × 24 = 0.00726; the second sub-packet loss rate of 2% corresponds to the second delay difference of 23.5ms, so the defect compensation sub-error = 0.015125 × 0.02 × 23.5 = 0.00711375. Therefore, the defect compensation sub-error dataset is 0.00726 and 0.00711375.

[0136] The second defect compensation error value is obtained by summing the values ​​in the defect compensation sub-error dataset. The second defect compensation error value is equal to the sum of all sub-error data. Substituting the example data, we get: 0.00726 + 0.00711375 = 0.01437375.

[0137] The third defect compensation error value is obtained by averaging the first defect compensation error value and the second defect compensation error value. The third defect compensation error value = (first defect compensation error value + second defect compensation error value) / 2. Substituting the data, we get: (0.35921875 + 0.01437375) / 2 = 0.3735925 / 2 = 0.18679625.

[0138] The initial actual defect data is corrected using the third defect compensation error value. The second actual defect data = initial actual defect data - third defect compensation error value. This is more accurate than the previous calculation because the impact of packet loss rate is refined by sub-proportional segmentation, and combined with the difference in node delay, the error compensation better matches the transmission interference characteristics of the latent defect data.

[0139] After processing the defect warning judgment based on the first and second actual defect data, the defect warning notification information is output, specifically including the following steps:

[0140] The first actual defect data and the second actual defect data are compared with the preset warning trigger threshold to determine whether the warning trigger requirements are met.

[0141] If the warning triggering conditions are met, the warning level is determined according to the corresponding warning triggering threshold, and then the defect warning notification information is output.

[0142] If the warning triggering conditions are not met, no defect warning notification information will be output.

[0143] First, it is necessary to clarify the preset warning trigger threshold, which is a critical value for the severity of defects set according to production needs or product quality standards. Usually, different levels of thresholds are divided to correspond to different warning levels. For example, in an industrial scenario, the preset warning trigger threshold is divided into three levels: Level 1 threshold 0.4 (mild warning), Level 2 threshold 0.6 (moderate warning), and Level 3 threshold 0.8 (severe warning). The higher the value, the stricter the warning standard for the severity of defects.

[0144] Assume the first actual defect data is 0.764 after correction for explicit defects (product surface cracks), and the second actual defect data is 0.31320375 after correction for implicit defects (micro-cracks inside the product). Compare these two data points with preset warning trigger thresholds to determine if the triggering requirements are met: if the first actual defect data 0.764 is greater than the second-level threshold 0.6 and less than the third-level threshold 0.8, the warning triggering condition is met; if the second actual defect data 0.31320375 is less than the first-level threshold 0.4, the warning triggering condition is not met.

[0145] If the data meets the warning trigger conditions, the warning level is determined according to the corresponding warning trigger threshold. For example, the first actual defect data of 0.764 corresponds to the second-level threshold, so the warning level is determined to be a moderate warning. Subsequently, the defect warning notification information is output, which includes the defect type (explicit defect), the warning level (moderate warning), and the defect severity (0.764), so that staff can handle it in a timely and targeted manner.

[0146] If the data does not meet the warning triggering conditions, such as the second actual defect data 0.31320375 not reaching the lowest level threshold, no defect warning notification information will be output, indicating that the current defect is within an acceptable range and no additional intervention is required.

[0147] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0148] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0149] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for data interconnection of intelligent computing boxes supporting domestic industrial protocols, characterized in that, The method includes the following steps: Initial actual defect data is obtained by real-time monitoring of the interconnection link between the intelligent computing box and the target detection product; The areas in the target product with obvious defect characteristics are marked as obvious defect product areas. The first historical detection data of the obvious defect product areas is extracted from the historical interconnection detection database of the smart computing box. The first historical detection data is processed to obtain the first judgment influence coefficient, which specifically includes the following steps: Determine whether the area of ​​a product with obvious defects is within the high-resolution detection area of ​​the target product. If the area of ​​the product with obvious defects is not located in the high-resolution detection area of ​​the target product, the historical delay difference of the obvious area to be tested is extracted from the first historical detection data based on the delay difference of the obvious area to be tested, and the historical judgment error value 1 that causes defect judgment under the influence of the historical delay difference of the obvious area is obtained; the correlation change trend characteristics between the historical judgment error value 1 and the historical delay difference of the obvious area are statistically analyzed to obtain the first delay influence coefficient. If the area of ​​the product with obvious defects is located in the high-resolution detection area of ​​the target product, the packet loss rate of the interconnection link data corresponding to the area of ​​the product with obvious defects is used to obtain the packet loss rate of the area to be tested. The packet loss rate of the area to be tested and the first historical detection data are processed to obtain the first delay and loss impact coefficient. Among them, the first delay impact coefficient and the first loss impact coefficient are combined to form the first judgment impact coefficient; The first actual defect data is obtained by performing error compensation on the initial actual defect data based on the first judgment influence coefficient. The areas in the target product containing latent defect features are marked as latent defect product areas. Secondary historical detection data belonging to these latent defect product areas is extracted from the historical interconnected detection database of the smart computing box. The secondary historical detection data is then processed to obtain a second judgment influence coefficient. This process includes the following steps: The transmission delay data of each detection node in the interconnection link corresponding to the product area with hidden defects is detected separately to obtain node transmission delay data and link terminal transmission delay data. The average node transmission delay is obtained by averaging the two adjacent sets of node delay data. The set of hidden region delay differences is obtained by subtracting the average transmission delay of a node from the transmission delay data of the corresponding link terminal of the adjacent node. The average delay difference of the hidden region is obtained by averaging all the delay difference values ​​in the hidden region set. Determine whether the area of ​​the product with hidden defects is within the high-resolution detection area of ​​the target product. If the area of ​​the latent defect product is not located in the high-resolution detection area of ​​the target product, the historical latent area delay difference is extracted from the second historical detection data based on the average delay difference of the latent area to be tested, and the historical judgment error value three is obtained under the influence of the historical latent area delay difference; the correlation and change trend of the historical judgment error value three and the historical latent area delay difference are statistically analyzed to obtain the second delay influence coefficient. If the latent defect product area is located in the high-resolution detection area of ​​the target product, the packet loss rate of the interconnection link data corresponding to the latent defect product area is used to obtain the packet loss rate of the latent area to be tested. The second historical packet loss rate is extracted from the second historical detection data. The second historical packet loss rate and the historical latent area delay difference are combined to form the latent defect interconnection feature data. The historical judgment error value four, which is jointly affected by the historical latent area delay difference and the historical packet loss rate, is extracted from the second historical detection data; the second delay and packet loss influence coefficient is obtained by statistically analyzing the correlation and change trend characteristics of the historical judgment error value four and the interconnection feature data of latent defects. The second delay impact coefficient and the second loss impact coefficient are combined into the second judgment impact coefficient; After processing the second judgment influence coefficient and the transmission data of the product area with hidden defects, the initial actual defect data is compensated for errors to obtain the second actual defect data. After processing the defect warning judgment based on the first and second actual defect data, a defect warning notification is output.

2. The method for data interconnection of intelligent computing boxes supporting domestic industrial protocols according to claim 1, characterized in that, It also includes the following steps: The delay difference of the visible defect product area in adjacent detection cycles is obtained by acquiring the transmission delay difference of the visible defect product area.

3. The data interconnection method for intelligent computing boxes supporting domestic industrial protocols according to claim 1, characterized in that, The first delay / loss impact coefficient is obtained by processing the packet loss rate of the visible region to be tested and the first historical detection data, specifically including the following steps: The packet loss rate of the visible region to be tested is the proportion of data packets lost when transmitting visible defect detection data through the interconnection link. Based on the packet loss rate of the visible region to be tested, the historical packet loss rate 1 is extracted from the first historical detection data, and the historical packet loss rate 1 and the historical visible region delay difference are combined to form visible defect interconnection feature data. The second historical judgment error value is extracted from the first historical detection data, which is caused by the combined influence of the delay difference in the historical explicit area and the historical packet loss rate. The first delay influence coefficient is obtained by statistically analyzing the correlation and change trends of historical judgment error value 2 and explicit defect interconnection feature data.

4. The data interconnection method for intelligent computing boxes supporting domestic industrial protocols according to claim 1, characterized in that, The first actual defect data is obtained by performing error compensation on the initial actual defect data based on the first judgment influence coefficient, specifically including the following steps: The first defect compensation error value for the product area with obvious defects is obtained based on the first judgment influence coefficient. The first actual defect data is obtained by performing error compensation processing on the initial actual defect data based on the error value that needs to be compensated for in the first defect.

5. The method for data interconnection of intelligent computing boxes supporting domestic industrial protocols according to claim 1, characterized in that, After processing the transmission data of the second judgment influence coefficient and the latent defect product area, error compensation is performed on the initial actual defect data to obtain the second actual defect data. The specific steps include: The first defect compensation error value for the hidden defect product area is obtained based on the second judgment influence coefficient and the average transmission delay difference in the hidden defect product area. The second defect compensation error value is obtained based on the second judgment influence coefficient and the transmission delay difference between each adjacent detection node in the product area with hidden defects; The second actual defect data is obtained by performing error compensation on the initial actual defect data based on the first defect compensation error value and the second defect compensation error value.

6. The data interconnection method for intelligent computing boxes supporting domestic industrial protocols according to claim 5, characterized in that, The second defect compensation error value is obtained based on the second judgment influence coefficient and the transmission delay difference between adjacent detection nodes in the product area with hidden defects. The specific steps include: The packet loss rate of the hidden region to be tested is divided into several sub-proportions to obtain the packet loss rate dataset of the sub-region to be tested; Based on the test sub-packet loss rate dataset, the hidden area delay difference set, and the second judgment influence coefficient, the defect compensation sub-error dataset generated by the defects in the hidden defect product area under the influence of transmission delay and data packet loss rate is obtained. The second defect compensation error value is obtained by summing up the sub-error data in the defect compensation sub-error dataset. The third defect compensation error value is obtained by averaging the first defect compensation error value and the second defect compensation error value. The second actual defect data is obtained by performing error compensation processing on the initial actual defect data based on the third defect error value.

7. The method for data interconnection of intelligent computing boxes supporting domestic industrial protocols according to claim 1, characterized in that, After processing the defect warning judgment based on the first and second actual defect data, the defect warning notification information is output, specifically including the following steps: The first actual defect data and the second actual defect data are compared with the preset warning trigger threshold to determine whether the warning trigger requirements are met. If the warning triggering conditions are met, the warning level is determined according to the corresponding warning triggering threshold, and then the defect warning notification information is output. If the warning triggering conditions are not met, no defect warning notification information will be output.

Images