A multi-target real-time monitoring method for rearview mirror lens production and processing

Abstract

The present invention relates to the field of rearview mirror lens production and processing, and in particular to a multi-target real-time monitoring method for rearview mirror lens production and processing, comprising: obtaining rearview mirror lens production and processing characteristics using real-time monitoring images of rearview mirror lens production and processing; obtaining interval thresholds of rearview mirror lens production and processing characteristics according to the rearview mirror lens production and processing characteristics; using the real-time monitoring images of rearview mirror lens production and processing, rearview mirror lens production and processing characteristics and interval thresholds of rearview mirror lens production and processing characteristics to perform training and verification processing to obtain a multi-target detection feedback model; using the multi-target detection feedback model to obtain multi-target real-time monitoring results of rearview mirror lens production and processing, and on the basis of traditional visual supervision, combining the production process with multiple characteristics of the identification target itself for combined judgment, the verification results are mutually verified, the data source is not single, and the monitoring accuracy and early warning monitoring range are improved.

Description

Technical Field

[0001] The present invention relates to the field of rearview mirror lens production and processing, and in particular to a multi-target real-time monitoring method for rearview mirror lens production and processing. Background Art

[0002] During the production of automobile rearview mirrors, the mirror support plate needs to be crimped to the lens through a heating plate. At the same time, due to the differences in the left and right lenses in the processing of rearview mirrors, traditional monitoring methods that only rely on visual error prevention or manual supervision cannot ensure real-time safety during the production process. There are interferences from environmental factors in production, and the accuracy of collection methods such as robot operation or photo positioning cannot be guaranteed. The subsequent risks brought about by this cannot be controlled, and the image visual processing of fixed cameras will also have errors in certain special situations, and losses caused by failure to warn. Summary of the invention

[0003] In view of the shortcomings of the prior art, the present invention provides a multi-target real-time monitoring method for the production and processing of rearview mirror lenses. By selecting multi-level features in the production and processing, a subsequent recognition and classification model is established to quickly and accurately obtain real-time monitoring results.

[0004] To achieve the above object, the present invention provides a multi-target real-time monitoring method for rearview mirror lens production and processing, comprising:

[0005] S1. Acquire the production and processing characteristics of the rearview mirror lens using the real-time monitoring image of the production and processing of the rearview mirror lens;

[0006] S2, obtaining the interval threshold of the rearview mirror lens production and processing characteristics according to the rearview mirror lens production and processing characteristics;

[0007] S3, using the real-time monitoring image of the rearview mirror lens production and processing, the rearview mirror lens production and processing characteristics and the interval threshold of the rearview mirror lens production and processing characteristics to perform training verification processing to obtain a multi-target detection feedback model;

[0008] S4. Utilize the multi-target detection feedback model to obtain multi-target real-time monitoring results of rearview mirror lens production and processing.

[0009] Preferably, the method of obtaining the production and processing characteristics of the rearview mirror lens by using the real-time monitoring image of the production and processing of the rearview mirror lens includes:

[0010] Obtain lens model data, lens size data, production and processing type data, lens single processing time data and production and processing hardware data of rearview mirror lens production and processing as indicator data;

[0011] Using the lens model data of the index data as the first data feature of rearview mirror lens production and processing;

[0012] Using the lens size data of the index data as the second data feature of rearview mirror lens production and processing;

[0013] Using the production and processing type data and production and processing hardware data of the indicator data as priority data features for the production and processing of rearview mirror lenses;

[0014] Using the lens single processing time data of the index data as the post-processing data feature of rearview mirror lens production and processing;

[0015] The first data feature, the second data feature, the priority data feature and the post-placement data feature of the rearview mirror lens production and processing are used as the rearview mirror lens production and processing features.

[0016] Furthermore, obtaining the interval threshold of the rearview mirror lens production and processing characteristics according to the rearview mirror lens production and processing characteristics includes:

[0017] S2-1, obtaining a first abnormal threshold value according to the priority data feature of the rearview mirror lens production and processing feature;

[0018] S2-2, obtaining a second abnormal threshold value according to the post-data feature of the rearview mirror lens production and processing feature;

[0019] S2-3, obtaining the lens size data corresponding to the normal threshold value of the rearview mirror lens production and processing feature according to the second data feature of the rearview mirror lens production and processing feature;

[0020] S2-4, determining whether the first abnormal threshold and the second abnormal threshold correspond to the normal threshold of the lens size data of the rearview mirror lens production and processing characteristics, if so, executing S2-5, otherwise, using the first abnormal threshold, the second abnormal threshold and the normal threshold as the interval threshold of the rearview mirror lens production and processing characteristics;

[0021] S2-5. Adjust the boundary of the normal threshold according to the first abnormal threshold and the second abnormal threshold, and return to S2-4.

[0022] Furthermore, the multi-target detection feedback model is obtained by training and verifying the real-time monitoring image of the rearview mirror lens production and processing, the rearview mirror lens production and processing features, and the interval threshold of the rearview mirror lens production and processing features, including:

[0023] S3-1, establishing a basic data association classification model using the real-time monitoring images of the rearview mirror lens production and processing;

[0024] S3-2, using the rearview mirror lens production and processing characteristics to establish a rearview mirror lens production and processing state analysis model;

[0025] S3-3, using the interval threshold of the rearview mirror lens production and processing characteristics to sequentially verify the basic data association classification model and the rearview mirror lens production and processing status analysis model to obtain a multi-target detection feedback model.

[0026] Furthermore, establishing a basic data association classification model using the real-time monitoring image of the rearview mirror lens production and processing includes:

[0027] The lens model data, lens size data and production and processing hardware data of the real-time monitoring image of the rearview mirror lens production and processing are used to divide the first training set, the second training set and the third training set respectively;

[0028] Using the first training set as input and the lens single processing time data of the real-time monitoring image of the rearview mirror lens production and processing as output, a primary basic data model is established based on a neural network;

[0029] Using the first training set and the second training set as input, and the lens single processing time data of the real-time monitoring image of the rearview mirror lens production and processing as output, the first-level basic data model is trained based on a neural network to establish a second-level basic data model;

[0030] The first training set, the second training set and the third training set are used as input, and the single processing time data of the lens of the real-time monitoring image of the production and processing of the rearview mirror lens is output. The secondary basic data model is trained based on the neural network to obtain the tertiary patient basic data model as the basic data association classification model.

[0031] Furthermore, establishing a rearview mirror lens production and processing state analysis model using the rearview mirror lens production and processing characteristics includes:

[0032] S3-2-1. Establish standard time t;

[0033] S3-2-2, according to the standard time t, respectively obtain the priority data features corresponding to the production and processing features of the rearview mirror lens at time t+1 and the priority data features corresponding to the production and processing features of the rearview mirror lens at time t+2;

[0034] S3-2-3, using the priority data features corresponding to the production and processing features of the rearview mirror lens at the standard time t, the priority data features corresponding to the production and processing features of the rearview mirror lens at time t+1, and the priority data features corresponding to the production and processing features of the rearview mirror lens at time t+2 to establish a priority data feature time series set;

[0035] S3-2-4, according to the standard time t, respectively obtain the rear-end data features corresponding to the production and processing features of the rearview mirror lens at time t+1 and the rear-end data features corresponding to the production and processing features of the rearview mirror lens at time t+2;

[0036] S3-2-5, using the rear-view mirror lens production and processing characteristics corresponding to the post-placed data characteristics at the standard time t, the rear-view mirror lens production and processing characteristics corresponding to the post-placed data characteristics at time t+1, and the rear-view mirror lens production and processing characteristics corresponding to the post-placed data characteristics at time t+2 to establish a post-placed data characteristics time series set;

[0037] S3-2-6, obtaining the corresponding priority data feature floating trend according to the priority data feature time series set;

[0038] S3-2-7, obtaining the corresponding post-data feature floating trend according to the post-data feature time series set;

[0039] S3-2-8, using the priority data feature floating trend as input, and the data association relationship between the priority data feature floating trend and the subsequent data feature floating trend as output, establishing a hidden layer according to the number of data types of the priority data feature floating trend, and obtaining a rearview mirror lens production and processing status analysis model based on neural network training;

[0040] The time interval and the end time of the priority data feature timing set and the subsequent data feature timing set are the same.

[0041] Furthermore, the basic data association classification model and the rearview mirror lens production and processing state analysis model are sequentially verified using the interval threshold of the rearview mirror lens production and processing characteristics to obtain a multi-target detection feedback model including:

[0042] S3-3-1, obtaining the normal threshold corresponding to the boundary value of the interval threshold of the rearview mirror lens production and processing characteristics as the first verification limit value and the second verification limit value respectively;

[0043] S3-3-2, determine whether the first verification limit value corresponds to the model output of the basic data association classification model, if so, return to S2-4, otherwise, execute S3-3-3;

[0044] S3-3-3, determine whether the second verification limit value corresponds to the model output of the basic data association classification model, if so, adjust the normal threshold value according to the interval threshold value of the rearview mirror lens production and processing characteristics corresponding to the model output of the basic data association classification model, and return to S2-4, otherwise, execute S3-3-4;

[0045] S3-3-4. Determine whether the model output of the rearview mirror lens production and processing status analysis model corresponds to the first abnormal threshold and the second abnormal threshold of the interval threshold of the rearview mirror lens production and processing characteristics. If so, use the basic data association classification model and the rearview mirror lens production and processing status analysis model as a multi-target detection feedback model. Otherwise, update the standard time t and return to S3-2-1.

[0046] Furthermore, the multi-target real-time monitoring results of rearview mirror lens production and processing obtained by using the multi-target detection feedback model include:

[0047] S4-1, using the multi-target detection feedback model to obtain a multi-target detection feedback result;

[0048] S4-2. Obtain multi-target real-time monitoring results of rearview mirror lens production and processing based on seasonal decomposition according to the multi-target detection feedback results.

[0049] Furthermore, obtaining a multi-target detection feedback result using the multi-target detection feedback model includes:

[0050] S4-1-1. Acquire real-time monitoring images of rearview mirror lens production and processing;

[0051] S4-1-2, using the real-time monitoring image of the rearview mirror lens production and processing to input the multi-target detection feedback model to obtain the basic data correlation analysis results and the rearview mirror lens production and processing status analysis results respectively;

[0052] S4-1-3, obtaining the interval threshold value corresponding to the real-time rearview mirror lens production and processing characteristics according to the real-time monitoring image of the rearview mirror lens production and processing;

[0053] S4-1-4, determine whether the first abnormal threshold and the second abnormal threshold of the interval threshold of the basic data correlation analysis result and the real-time rearview mirror lens production and processing characteristics are both corresponding, if so, use the basic data correlation analysis result and the rearview mirror lens production and processing status analysis result as the multi-target detection feedback result, otherwise, use the normal threshold of the interval threshold of the real-time rearview mirror lens production and processing characteristics as the real-time calibration value, and execute S4-1-5;

[0054] S4-1-5. Utilize the real-time calibration value, basic data correlation analysis result and rearview mirror lens production and processing status analysis result as multi-target detection feedback result.

[0055] Furthermore, obtaining multi-target real-time monitoring results of rearview mirror lens production and processing based on seasonal decomposition according to the multi-target detection feedback results includes:

[0056] S4-2-1, determine whether there is a real-time calibration value in the multi-target detection feedback result, if so, use the corresponding time of the real-time monitoring image of the rearview mirror lens production and processing as the starting time k, and execute S4-2-2, otherwise, directly output the real-time monitoring image of the rearview mirror lens production and processing and the rearview mirror lens production and processing status analysis result as the multi-target real-time monitoring result of the rearview mirror lens production and processing;

[0057] S4-2-2, respectively obtaining real-time monitoring images of rearview mirror lens production and processing corresponding to time k+1 and time k+2;

[0058] S4-2-3, using the real-time monitoring images corresponding to the production and processing of rearview mirror lenses at the starting time k, time k+1 and time k+2 to obtain the average value of the index data;

[0059] S4-2-4, using the multi-target detection feedback results to correspond to the rearview mirror lens production and processing status analysis results based on seasonal decomposition to obtain the real-time monitoring image trend and seasonal characteristics of the index data of the rearview mirror lens production and processing;

[0060] S4-2-5, judging whether the average value of the index data exceeds the real-time calibration value, if so, retaining the real-time monitoring image of the current rearview mirror lens production and processing, otherwise, retaining the average value of the index data exceeding the real-time calibration value as the updated index data, and executing S4-2-6;

[0061] S4-2-6, determining whether the indicator data trend corresponds to the real-time calibration value, if so, retaining the current indicator data trend, otherwise, retaining the non-corresponding indicator data trend as the updated indicator data trend, and executing S4-2-7;

[0062] S4-2-7. Determine whether the period corresponding to the seasonal characteristics of the index data is greater than the trend interval corresponding to the analysis result of the production and processing status of the rearview mirror lens. If so, use the period corresponding to the seasonal characteristics of the index data to update the adjacent time interval corresponding to the standard time, and return to S3-2-1. Otherwise, use the real-time monitoring image of the production and processing of the rearview mirror lens, the trend of the index data and the analysis result of the production and processing status of the rearview mirror lens as the multi-target real-time monitoring result of the production and processing of the rearview mirror lens.

[0063] Compared with the closest prior art, the present invention has the following beneficial effects:

[0064] On the basis of traditional visual supervision, the production process is combined with the multiple characteristics of the identification target itself for combined judgment, the verification results confirm each other, the data source is not single, and the monitoring accuracy and early warning monitoring range are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] Figure 1The present invention provides a flowchart of a multi-target real-time monitoring method for rearview mirror lens production and processing. DETAILED DESCRIPTION

[0066] The specific implementation modes of the present invention will be further described in detail below in conjunction with the accompanying drawings.

[0067] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0068] Embodiment 1: The present invention provides a multi-target real-time monitoring method for rearview mirror lens production and processing, such as Figure 1 As shown, including:

[0069] S1. Acquire the production and processing characteristics of the rearview mirror lens using the real-time monitoring image of the production and processing of the rearview mirror lens;

[0070] S2, obtaining the interval threshold of the rearview mirror lens production and processing characteristics according to the rearview mirror lens production and processing characteristics;

[0071] S3, using the real-time monitoring image of the rearview mirror lens production and processing, the rearview mirror lens production and processing characteristics and the interval threshold of the rearview mirror lens production and processing characteristics to perform training verification processing to obtain a multi-target detection feedback model;

[0072] S4. Utilize the multi-target detection feedback model to obtain multi-target real-time monitoring results of rearview mirror lens production and processing.

[0073] S1 specifically includes:

[0074] S1-1, obtaining lens model data, lens size data, production and processing type data, lens single processing time data and production and processing hardware data of rearview mirror lenses as indicator data;

[0075] S1-2, using the lens model data of the index data as the first data feature of rearview mirror lens production and processing;

[0076] S1-3, using the lens size data of the index data as the second data feature for rearview mirror lens production and processing;

[0077] S1-4, using the production and processing type data and production and processing hardware data of the indicator data as priority data features for rearview mirror lens production and processing;

[0078] S1-5, using the lens single processing time data of the index data as the post-processing data feature of rearview mirror lens production and processing;

[0079] S1-6, using the first data feature, the second data feature, the priority data feature and the post-data feature of the rearview mirror lens production and processing as the rearview mirror lens production and processing features.

[0080] S2 specifically includes:

[0081] S2-1, obtaining a first abnormal threshold value according to the priority data feature of the rearview mirror lens production and processing feature;

[0082] S2-2, obtaining a second abnormal threshold value according to the post-data feature of the rearview mirror lens production and processing feature;

[0083] S2-3, obtaining the lens size data corresponding to the normal threshold value of the rearview mirror lens production and processing feature according to the second data feature of the rearview mirror lens production and processing feature;

[0084] S2-4, determining whether the first abnormal threshold and the second abnormal threshold correspond to the normal threshold of the lens size data of the rearview mirror lens production and processing characteristics, if so, executing S2-5, otherwise, using the first abnormal threshold, the second abnormal threshold and the normal threshold as the interval threshold of the rearview mirror lens production and processing characteristics;

[0085] S2-5. Adjust the boundary of the normal threshold according to the first abnormal threshold and the second abnormal threshold, and return to S2-4.

[0086] S3 specifically includes:

[0087] S3-1, establishing a basic data association classification model using the real-time monitoring images of the rearview mirror lens production and processing;

[0088] S3-2, using the rearview mirror lens production and processing characteristics to establish a rearview mirror lens production and processing state analysis model;

[0089] S3-3, using the interval threshold of the rearview mirror lens production and processing characteristics to sequentially verify the basic data association classification model and the rearview mirror lens production and processing status analysis model to obtain a multi-target detection feedback model.

[0090] S3-1 specifically includes:

[0091] S3-1-1, using the lens model data, lens size data and production and processing hardware data of the real-time monitoring image of the rearview mirror lens production and processing to divide the first training set, the second training set and the third training set respectively;

[0092] S3-1-2, using the first training set as input and the lens single processing time data of the real-time monitoring image of the rearview mirror lens production and processing as output, establishing a primary basic data model based on a neural network;

[0093] S3-1-3, using the first training set and the second training set as input, and the lens single processing time data of the real-time monitoring image of the rearview mirror lens production and processing as output, training the first-level basic data model based on a neural network to establish a second-level basic data model;

[0094] S3-1-4. Using the first training set, the second training set and the third training set as input, and the single processing time data of the lens of the real-time monitoring image of the production and processing of the rearview mirror lens as output, the secondary basic data model is trained based on the neural network to obtain the tertiary patient basic data model as the basic data association classification model.

[0095] S3-2 specifically includes:

[0096] S3-2-1. Establish standard time t;

[0097] S3-2-2, according to the standard time t, respectively obtain the priority data features corresponding to the production and processing features of the rearview mirror lens at time t+1 and the priority data features corresponding to the production and processing features of the rearview mirror lens at time t+2;

[0098] S3-2-3, using the priority data features corresponding to the production and processing features of the rearview mirror lens at the standard time t, the priority data features corresponding to the production and processing features of the rearview mirror lens at time t+1, and the priority data features corresponding to the production and processing features of the rearview mirror lens at time t+2 to establish a priority data feature time series set;

[0099] S3-2-4, according to the standard time t, respectively obtain the rear-end data features corresponding to the production and processing features of the rearview mirror lens at time t+1 and the rear-end data features corresponding to the production and processing features of the rearview mirror lens at time t+2;

[0100] S3-2-5, using the rear-view mirror lens production and processing characteristics corresponding to the post-placed data characteristics at the standard time t, the rear-view mirror lens production and processing characteristics corresponding to the post-placed data characteristics at time t+1, and the rear-view mirror lens production and processing characteristics corresponding to the post-placed data characteristics at time t+2 to establish a post-placed data characteristics time series set;

[0101] S3-2-6, obtaining the corresponding priority data feature floating trend according to the priority data feature time series set;

[0102] S3-2-7, obtaining the corresponding post-data feature floating trend according to the post-data feature time series set;

[0103] S3-2-8, using the priority data feature floating trend as input, and the data association relationship between the priority data feature floating trend and the subsequent data feature floating trend as output, establishing a hidden layer according to the number of data types of the priority data feature floating trend, and obtaining a rearview mirror lens production and processing status analysis model based on neural network training;

[0104] The time interval and the end time of the priority data feature timing set and the subsequent data feature timing set are the same.

[0105] S3-3 specifically includes:

[0106] S3-3-1, obtaining the normal threshold corresponding to the boundary value of the interval threshold of the rearview mirror lens production and processing characteristics as the first verification limit value and the second verification limit value respectively;

[0107] S3-3-2, determine whether the first verification limit value corresponds to the model output of the basic data association classification model, if so, return to S2-4, otherwise, execute S3-3-3;

[0108] S3-3-3, determine whether the second verification limit value corresponds to the model output of the basic data association classification model, if so, adjust the normal threshold value according to the interval threshold value of the rearview mirror lens production and processing characteristics corresponding to the model output of the basic data association classification model, and return to S2-4, otherwise, execute S3-3-4;

[0109] S3-3-4. Determine whether the model output of the rearview mirror lens production and processing status analysis model corresponds to the first abnormal threshold and the second abnormal threshold of the interval threshold of the rearview mirror lens production and processing characteristics. If so, use the basic data association classification model and the rearview mirror lens production and processing status analysis model as a multi-target detection feedback model. Otherwise, update the standard time t and return to S3-2-1.

[0110] S4 specifically includes:

[0111] S4-1, using the multi-target detection feedback model to obtain a multi-target detection feedback result;

[0112] S4-2. Obtain multi-target real-time monitoring results of rearview mirror lens production and processing based on seasonal decomposition according to the multi-target detection feedback results.

[0113] S4-1 specifically includes:

[0114] S4-1-1. Acquire real-time monitoring images of rearview mirror lens production and processing;

[0115] S4-1-2, using the real-time monitoring image of the rearview mirror lens production and processing to input the multi-target detection feedback model to obtain the basic data correlation analysis results and the rearview mirror lens production and processing status analysis results respectively;

[0116] S4-1-3, obtaining the interval threshold value corresponding to the real-time rearview mirror lens production and processing characteristics according to the real-time monitoring image of the rearview mirror lens production and processing;

[0117] S4-1-4, determine whether the first abnormal threshold and the second abnormal threshold of the interval threshold of the basic data correlation analysis result and the real-time rearview mirror lens production and processing characteristics are both corresponding, if so, use the basic data correlation analysis result and the rearview mirror lens production and processing status analysis result as the multi-target detection feedback result, otherwise, use the normal threshold of the interval threshold of the real-time rearview mirror lens production and processing characteristics as the real-time calibration value, and execute S4-1-5;

[0118] S4-1-5. Utilize the real-time calibration value, basic data correlation analysis result and rearview mirror lens production and processing status analysis result as multi-target detection feedback result.

[0119] S4-2 specifically includes:

[0120] S4-2-1, determine whether there is a real-time calibration value in the multi-target detection feedback result, if so, use the corresponding time of the real-time monitoring image of the rearview mirror lens production and processing as the starting time k, and execute S4-2-2, otherwise, directly output the real-time monitoring image of the rearview mirror lens production and processing and the rearview mirror lens production and processing status analysis result as the multi-target real-time monitoring result of the rearview mirror lens production and processing;

[0121] S4-2-2, respectively obtaining real-time monitoring images of rearview mirror lens production and processing corresponding to time k+1 and time k+2;

[0122] S4-2-3, using the real-time monitoring images corresponding to the production and processing of rearview mirror lenses at the starting time k, time k+1 and time k+2 to obtain the average value of the index data;

[0123] S4-2-4, using the multi-target detection feedback results to correspond to the rearview mirror lens production and processing status analysis results based on seasonal decomposition to obtain the real-time monitoring image trend and seasonal characteristics of the index data of the rearview mirror lens production and processing;

[0124] S4-2-5, judging whether the average value of the index data exceeds the real-time calibration value, if so, retaining the real-time monitoring image of the current rearview mirror lens production and processing, otherwise, retaining the average value of the index data exceeding the real-time calibration value as the updated index data, and executing S4-2-6;

[0125] S4-2-6, determining whether the indicator data trend corresponds to the real-time calibration value, if so, retaining the current indicator data trend, otherwise, retaining the non-corresponding indicator data trend as the updated indicator data trend, and executing S4-2-7;

[0126] S4-2-7. Determine whether the period corresponding to the seasonal characteristics of the index data is greater than the trend interval corresponding to the analysis result of the production and processing status of the rearview mirror lens. If so, use the period corresponding to the seasonal characteristics of the index data to update the adjacent time interval corresponding to the standard time, and return to S3-2-1. Otherwise, use the real-time monitoring image of the production and processing of the rearview mirror lens, the trend of the index data and the analysis result of the production and processing status of the rearview mirror lens as the multi-target real-time monitoring result of the production and processing of the rearview mirror lens.

[0127] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0128] The present invention is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of the processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0129] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0130] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0131] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A multi-target real-time monitoring method for rearview mirror lens production and processing, characterized in that: include: S1. Acquire the production and processing characteristics of the rearview mirror lens using the real-time monitoring image of the production and processing of the rearview mirror lens; S1-1, obtaining lens model data, lens size data, production and processing type data, lens single processing time data and production and processing hardware data of rearview mirror lenses as indicator data; S1-2, using the lens model data of the index data as the first data feature of rearview mirror lens production and processing; S1-3, using the lens size data of the index data as the second data feature for rearview mirror lens production and processing; S1-4, using the production and processing type data and production and processing hardware data of the indicator data as priority data features for rearview mirror lens production and processing; S1-5, using the lens single processing time data of the index data as the post-processing data feature of rearview mirror lens production and processing; S1-6, using the first data feature, the second data feature, the priority data feature and the rear data feature of the rearview mirror lens production and processing as the rearview mirror lens production and processing features; S2, obtaining the interval threshold of the rearview mirror lens production and processing characteristics according to the rearview mirror lens production and processing characteristics; S2-1, obtaining a first abnormal threshold value according to the priority data feature of the rearview mirror lens production and processing feature; S2-2, obtaining a second abnormal threshold value according to the post-data feature of the rearview mirror lens production and processing feature; S2-3, obtaining the lens size data corresponding to the normal threshold value of the rearview mirror lens production and processing feature according to the second data feature of the rearview mirror lens production and processing feature; S2-4, determining whether the first abnormal threshold and the second abnormal threshold correspond to the normal threshold of the lens size data of the rearview mirror lens production and processing characteristics, if so, executing S2-5, otherwise, using the first abnormal threshold, the second abnormal threshold and the normal threshold as the interval threshold of the rearview mirror lens production and processing characteristics; S2-5, adjusting the boundary of the normal threshold according to the first abnormal threshold and the second abnormal threshold, and returning to S2-4; S3, using the real-time monitoring image of the rearview mirror lens production and processing, the rearview mirror lens production and processing characteristics and the interval threshold of the rearview mirror lens production and processing characteristics to perform training verification processing to obtain a multi-target detection feedback model; S3-1, establishing a basic data association classification model using the real-time monitoring images of the rearview mirror lens production and processing; S3-1-1, using the lens model data, lens size data and production and processing hardware data of the real-time monitoring image of the rearview mirror lens production and processing to divide the first training set, the second training set and the third training set respectively; S3-1-2, using the first training set as input and the lens single processing time data of the real-time monitoring image of the rearview mirror lens production and processing as output, establishing a primary basic data model based on a neural network; S3-1-3, using the first training set and the second training set as input, and the lens single processing time data of the real-time monitoring image of the rearview mirror lens production and processing as output, training the first-level basic data model based on a neural network to establish a second-level basic data model; S3-1-4, using the first training set, the second training set and the third training set as input, and the lens single processing time data of the real-time monitoring image of the rearview mirror lens production and processing as output, training the secondary basic data model based on the neural network to obtain the tertiary patient basic data model as the basic data association classification model; S3-2, using the rearview mirror lens production and processing characteristics to establish a rearview mirror lens production and processing state analysis model; S3-2-1. Establish standard time t; S3-2-2, according to the standard time t, respectively obtain the priority data features corresponding to the production and processing features of the rearview mirror lens at time t+1 and the priority data features corresponding to the production and processing features of the rearview mirror lens at time t+2; S3-2-3, using the priority data features corresponding to the production and processing features of the rearview mirror lens at the standard time t, the priority data features corresponding to the production and processing features of the rearview mirror lens at time t+1, and the priority data features corresponding to the production and processing features of the rearview mirror lens at time t+2 to establish a priority data feature time series set; S3-2-4, according to the standard time t, respectively obtain the rear-end data features corresponding to the production and processing features of the rearview mirror lens at time t+1 and the rear-end data features corresponding to the production and processing features of the rearview mirror lens at time t+2; S3-2-5, using the rear-view mirror lens production and processing characteristics corresponding to the post-placed data characteristics at the standard time t, the rear-view mirror lens production and processing characteristics corresponding to the post-placed data characteristics at time t+1, and the rear-view mirror lens production and processing characteristics corresponding to the post-placed data characteristics at time t+2 to establish a post-placed data characteristics time series set; S3-2-6, obtaining the corresponding priority data feature floating trend according to the priority data feature time series set; S3-2-7, obtaining the corresponding post-data feature floating trend according to the post-data feature time series set; S3-2-8, using the priority data feature floating trend as input, and the data association relationship between the priority data feature floating trend and the subsequent data feature floating trend as output, establishing a hidden layer according to the number of data types of the priority data feature floating trend, and obtaining a rearview mirror lens production and processing status analysis model based on neural network training; The time interval and the end time of the priority data feature time series set and the subsequent data feature time series set are the same; S3-3, using the interval threshold of the rearview mirror lens production and processing characteristics to sequentially verify the basic data association classification model and the rearview mirror lens production and processing state analysis model to obtain a multi-target detection feedback model; S3-3-1, obtaining the normal threshold corresponding to the boundary value of the interval threshold of the rearview mirror lens production and processing characteristics as the first verification limit value and the second verification limit value respectively; S3-3-2, determine whether the first verification limit value corresponds to the model output of the basic data association classification model, if so, return to S2-4, otherwise, execute S3-3-3; S3-3-3, determine whether the second verification limit value corresponds to the model output of the basic data association classification model, if so, adjust the normal threshold value according to the interval threshold value of the rearview mirror lens production and processing characteristics corresponding to the model output of the basic data association classification model, and return to S2-4, otherwise, execute S3-3-4; S3-3-4, judging whether the model output of the rearview mirror lens production and processing state analysis model corresponds to the first abnormal threshold and the second abnormal threshold of the interval threshold of the rearview mirror lens production and processing characteristics, if so, using the basic data association classification model and the rearview mirror lens production and processing state analysis model as a multi-target detection feedback model, otherwise, updating the standard time t, and returning to S3-2-1; S4. Utilize the multi-target detection feedback model to obtain multi-target real-time monitoring results of rearview mirror lens production and processing.

2. A multi-target real-time monitoring method for rearview mirror lens production and processing as claimed in claim 1, characterized in that: The multi-target real-time monitoring results of rearview mirror lens production and processing obtained by using the multi-target detection feedback model include: S4-1, using the multi-target detection feedback model to obtain a multi-target detection feedback result; S4-2. Obtain multi-target real-time monitoring results of rearview mirror lens production and processing based on seasonal decomposition according to the multi-target detection feedback results.

3. A multi-target real-time monitoring method for rearview mirror lens production and processing as claimed in claim 2, characterized in that: The multi-target detection feedback result obtained by using the multi-target detection feedback model includes: S4-1-1. Acquire real-time monitoring images of rearview mirror lens production and processing; S4-1-2, using the real-time monitoring image of the rearview mirror lens production and processing to input the multi-target detection feedback model to obtain the basic data correlation analysis results and the rearview mirror lens production and processing status analysis results respectively; S4-1-3, obtaining the interval threshold value corresponding to the real-time rearview mirror lens production and processing characteristics according to the real-time monitoring image of the rearview mirror lens production and processing; S4-1-4, determine whether the first abnormal threshold and the second abnormal threshold of the interval threshold of the basic data correlation analysis result and the real-time rearview mirror lens production and processing characteristics are both corresponding, if so, use the basic data correlation analysis result and the rearview mirror lens production and processing status analysis result as the multi-target detection feedback result, otherwise, use the normal threshold of the interval threshold of the real-time rearview mirror lens production and processing characteristics as the real-time calibration value, and execute S4-1-5; S4-1-5. Utilize the real-time calibration value, basic data correlation analysis result and rearview mirror lens production and processing status analysis result as multi-target detection feedback result.

4. A multi-target real-time monitoring method for rearview mirror lens production and processing as claimed in claim 3, characterized in that: According to the multi-target detection feedback results, the multi-target real-time monitoring results of rearview mirror lens production and processing are obtained based on seasonal decomposition, including: S4-2-1, determine whether there is a real-time calibration value in the multi-target detection feedback result, if so, use the corresponding time of the real-time monitoring image of the rearview mirror lens production and processing as the starting time k, and execute S4-2-2, otherwise, directly output the real-time monitoring image of the rearview mirror lens production and processing and the rearview mirror lens production and processing status analysis result as the multi-target real-time monitoring result of the rearview mirror lens production and processing; S4-2-2, respectively obtaining real-time monitoring images of rearview mirror lens production and processing corresponding to time k+1 and time k+2; S4-2-3, using the real-time monitoring images corresponding to the production and processing of rearview mirror lenses at the starting time k, time k+1 and time k+2 to obtain the average value of the index data; S4-2-4, using the multi-target detection feedback results to correspond to the rearview mirror lens production and processing status analysis results based on seasonal decomposition to obtain the real-time monitoring image trend and seasonal characteristics of the index data of the rearview mirror lens production and processing; S4-2-5, judging whether the average value of the index data exceeds the real-time calibration value, if so, retaining the real-time monitoring image of the current rearview mirror lens production and processing, otherwise, retaining the average value of the index data exceeding the real-time calibration value as the updated index data, and executing S4-2-6; S4-2-6, determining whether the indicator data trend corresponds to the real-time calibration value, if so, retaining the current indicator data trend, otherwise, retaining the non-corresponding indicator data trend as the updated indicator data trend, and executing S4-2-7; S4-2-7. Determine whether the period corresponding to the seasonal characteristics of the index data is greater than the trend interval corresponding to the analysis result of the production and processing status of the rearview mirror lens. If so, use the period corresponding to the seasonal characteristics of the index data to update the adjacent time interval corresponding to the standard time, and return to S3-2-1. Otherwise, use the real-time monitoring image of the production and processing of the rearview mirror lens, the trend of the index data and the analysis result of the production and processing status of the rearview mirror lens as the multi-target real-time monitoring result of the production and processing of the rearview mirror lens.

Images