Automatic production line management method and system based on precise control

Abstract

The invention relates to the technical field of production line management, in particular to an automatic production line management method and system based on precise control, and the method comprises the steps: carrying out the searching of product processing procedure data of collected target product text feature data and established different product processing procedure data, and generating target product processing procedure data; acquiring image data corresponding to the processing procedure data of the target product in real time through an image acquisition device, and then performing data preprocessing on the image data to generate real-time image feature data of the target product; and carrying out product image feature matching on the real-time image feature data of the target product and the established qualified image feature data of the target product, judging the operation state of the automatic production line according to a product image feature matching result, and constructing operation management data of the automatic production line and pushing the operation management data to an automatic production line management platform. And intelligent management of the automatic production line is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of production line management, and in particular to an automated production line management method and system based on precise control. Background Art

[0002] An automated production line refers to an organizational form that is formed by a series of automated equipment connected together and can realize product production. In recent years, with the continuous development and progress of automation technology, automated production lines have been applied to many modern industries.

[0003] The Chinese invention patent application with publication number CN117094603A introduces an automated management system for a medical functional food production line based on intelligent control. By monitoring, counting and processing the raw materials in the early stage of the production of the medical functional food production line, the efficiency of the subsequent medical functional food production data analysis and the accuracy of data utilization can be effectively improved; by implementing modular random inspections on the production process of the medical functional food production line and dynamically controlling the subsequent production based on the random inspection results, the abnormal production links can be actively managed and controlled in a timely and efficient manner, providing reliable data support for the subsequent production optimization and management of the medical functional food production line; however, by completing the dynamic control of automatic production through random inspections of the production process, it is impossible to accurately control the abnormal phenomena that occur in a timely manner, and it is difficult to avoid the losses caused by undetected abnormal phenomena. Summary of the invention

[0004] 1. Technical issues to be resolved

[0005] In order to solve the deficiencies in the background technology, the present invention provides an automated production line management method and system based on precise control, which realizes intelligent management of the automated production line by real-time monitoring of the operating status of the automated production line.

[0006] (II) Technical solution

[0007] An automated production line management method based on precise control comprises the following steps:

[0008] S1. Collect text feature data of target products;

[0009] S2, establishing different product processing procedure data, performing product processing procedure data search and processing according to the different product processing procedure data and the target product text feature data, and generating target product processing procedure data;

[0010] S3, collecting the target product image under the target product processing process data in the process of the automated production line in real time through an image acquisition device, and generating real-time image data of the target product;

[0011] S4, performing data preprocessing on the real-time image data of the target product to obtain real-time image feature data of the target product;

[0012] S5. Establishing qualified image feature data of target products;

[0013] S6, performing product image feature matching on the real-time image feature data of the target product and the qualified image feature data of the target product, determining the operation status of the automated production line according to the product image feature matching result, and obtaining abnormal image feature data of the target product;

[0014] S7. When the operation status of the automated production line is abnormal, the automated production line operation management data is constructed based on the abnormal image feature data of the target product and pushed to the automated production line management platform.

[0015] The present invention generates target product processing process data by performing product processing process data search and processing on the collected target product text feature data and the established processing process data of different products, collects image data corresponding to the target product processing process data in real time through an image acquisition device, and then performs data preprocessing on the image data to generate real-time image feature data of the target product, performs product image feature matching on the real-time image feature data of the target product and the established qualified image feature data of the target product, determines the operating status of the automated production line according to the product image feature matching result, and constructs automated production line operation management data and pushes it to an automated production line management platform to realize intelligent management of the automated production line.

[0016] Preferably, the specific steps of collecting target product text feature data are as follows:

[0017] S11. Collect text feature data of the products to be produced online through the production line product information management platform to generate target product text feature data A, wherein the target product text feature data includes product type, product specification, product quality standard, product quantity and product production cycle.

[0018] Preferably, the specific steps of establishing different product processing process data, performing product processing process data search and processing according to the different product processing process data and the target product text feature data, and generating the target product processing process data are as follows:

[0019] S21. Establish different product processing process data sets B = {b1, b2, ..., b i ,…,b k}, where b i represents the different product processing procedure data corresponding to the i-th type of product, k represents the total number of product types, and the different product processing procedure data represent the various processing procedures when the automated production line produces different products;

[0020] S22, using a breadth search algorithm to perform product processing procedure data search processing on the target product text feature data A and the different product processing procedure data in the different product processing procedure data sets in ascending order of serial numbers, search out the different product processing procedure data corresponding to the target product text feature data, and perform data identification to generate a target product processing procedure data set C = {c1, c2, ..., c i ,…,c l}, where c i represents the i-th processing step data of the target product, and l represents the total number of processing step data of the target product.

[0021] A width search algorithm is used to search for the processing procedure data of the target product, so as to achieve accurate search of the processing procedure data of the target product and ensure the accuracy of the acquired data.

[0022] Preferably, the specific steps of collecting the target product image under the target product processing process data in the process of the automated production line in real time by an image acquisition device to generate the real-time image data of the target product are as follows:

[0023] S31, installing a high-definition image acquisition device on the automated production line, and using the high-definition image acquisition device to collect in real time the target product images of the target product under each target product processing process data in the target product processing process data set during the operation of the automated production line, and generating a target product real-time image data set D = {d1, d2, ..., d i ,…,d l}, where d i Represents the real-time image data of the target product under the i-th target product processing process data.

[0024] Preferably, the specific steps of performing data preprocessing on the real-time image data of the target product to obtain the real-time image feature data of the target product are as follows:

[0025] S41, using a weighted mean filtering method to perform image data filtering and noise reduction processing on the target product real-time image data in the target product real-time image data set, to obtain a target product real-time image feature data set D'={d'1, d'2, ..., d' i ,…,d' l}, where d' i It represents the real-time image feature data of the target product obtained after the real-time image data of the target product under the i-th target product processing procedure data is subjected to image data filtering and noise reduction processing.

[0026] Preferably, the S41 comprises the following steps:

[0027] S411, randomly selecting any one of the target product real-time image data in the target product real-time image data set as sample image data;

[0028] S412, setting an m×m neighborhood window, and aligning the center position of the neighborhood window with each pixel point of the sample image data in sequence;

[0029] S413, calculating the weight coefficient of each pixel in the neighborhood window by a weighted kernel function formula; the weighted kernel function formula is as follows:

[0030]

[0031] Among them, w(i,j) represents the weight coefficient of pixel point (i,j), d(i,j) represents the Euclidean distance from pixel point (i,j) to the center point of the neighborhood window after Gaussian weighting, and a represents the smoothing parameter of the weighted kernel function;

[0032] S414, calculating the weighted average value of the pixels in the neighborhood window, and using the weighted average value as the pixel value of the center point of the neighborhood window; the weighted average value calculation formula is as follows:

[0033]

[0034] Wherein, Gray represents the weighted average value of the pixels in the neighborhood window, and f(i,j) represents the pixel value of the pixel point (i,j);

[0035] S415, repeating the operations in S411 to S414 until all the target product real-time image data in the target product real-time image data set are traversed to obtain the target product real-time image feature data set.

[0036] The weighted mean filtering method is used to filter and reduce noise on the target product image data collected in real time, generate real-time image feature data of the target product, effectively reduce the impact of image noise, ensure the quality and accuracy of the acquired image feature data, and provide a reliable data basis for subsequent operations.

[0037] Preferably, the specific steps of establishing the target product qualified image feature data are as follows:

[0038] S51, calling the image feature data of the target product under each target product processing process data in the target product processing process data set from the qualified product image feature data storage unit to generate a target product qualified image feature data set in, Represents the target product qualified image feature dataset under the i-th target product processing process data.

[0039] Preferably, the specific steps of performing product image feature matching on the real-time image feature data of the target product and the qualified image feature data of the target product, determining the operation status of the automated production line according to the product image feature matching result, and obtaining the abnormal image feature data of the target product are as follows:

[0040] S61, perform product image feature matching on each target product real-time image feature data set in the target product real-time image feature data set in order of serial number from small to large with all target product qualified image feature data sets in the target product qualified image feature data set under the corresponding target product processing procedure data. If each target product real-time image feature data set in the target product real-time image feature data set matches all target product qualified image feature data sets in the target product qualified image feature data set under the corresponding target product processing procedure data successfully, the operation status of the automated production line is determined to be normal, and no operation is performed; otherwise, the operation status of the automated production line is determined to be abnormal, and the target product real-time image feature data sets that fail to match all target product qualified image feature data sets in the target product qualified image feature data set under the corresponding target product processing procedure data are screened out, and data identification is performed to generate a target product abnormal image feature data set E = {e1, e2, …, e i ,…,e p}, where e i represents the abnormal image feature data of the i-th target product, and p represents the total number of abnormal image feature data of the target product; the product image feature matching process is as follows:

[0041] S611, constructing a pigeon population, setting the size of the pigeon population to N, the current number of iterations to t, the number of iterations in the map navigation phase to t1, the number of iterations in the landmark navigation phase to t2, and the dimension of the product image feature search space to P;

[0042] In the product image feature search space, N target product qualified image feature data are randomly generated, each target product qualified image feature data corresponds to a pigeon individual in the pigeon population, and the initial position set of the pigeon population is initialized as X = {X1, X2, ..., X i ,…,X N}, where X i represents the initial position of the i-th pigeon in the pigeon population;

[0043] S612, calculating the fitness function value of each individual pigeon in the pigeon population, arranging each individual pigeon in the pigeon population from large to small according to the fitness function value, and selecting the individual pigeon with the highest fitness function value as the current optimal individual; the fitness function formula is as follows:

[0044]

[0045] Among them, Fit i represents the fitness function value of the i-th pigeon in the pigeon population, x i represents the deviation between the qualified image feature data of the target product corresponding to the i-th individual pigeon in the pigeon population and the real-time image feature data of the target product for product image feature matching, and φ represents the correction value;

[0046] S613, map navigation stage, each individual pigeon in the pigeon population adjusts its own flight speed and flight direction in the product image feature search space through the map and compass operator to update its position; the position update formula is as follows:

[0047]

[0048] in, and They represent the speed and position of the i-th pigeon in the pigeon population during the t+1th iteration, and represents the speed and position of the i-th pigeon in the pigeon population during the t-th iteration, ε represents the map and compass operator, rand1 represents a random number uniformly distributed between (0,1), and X best Indicates the position of the current optimal individual;

[0049] S614, judging whether the current number of iterations t is less than the number of iterations t1 in the map navigation phase, if the current number of iterations t is less than the number of iterations t1 in the map navigation phase, then the current number of iterations t is increased by 1, and the process returns to S612; otherwise, the fitness function value of each individual pigeon in the pigeon population is recalculated, and the process proceeds to S615;

[0050] S615, landmark navigation stage, each individual pigeon in the pigeon population optimizes its own flight direction and flight speed in the product image feature search space through surrounding familiar landmarks to update its position, and at the same time, discards individual pigeons with lower fitness function values ​​in each iteration process; the position update formula is as follows:

[0051]

[0052] Among them, N t+1 and N t They represent the pigeon population size in the t+1th iteration process and the tth iteration process respectively, represents the center position of the pigeon population during the tth iteration, and rand2 represents a random number uniformly distributed between (0,1);

[0053] S616, judging whether the current number of iterations t is less than the number of iterations t2 in the landmark navigation phase, if the current number of iterations t is less than the number of iterations t2 in the landmark navigation phase, then adding 1 to the current number of iterations t, and returning to S615; otherwise, recalculating the fitness function value of each individual pigeon in the pigeon population, arranging each individual pigeon in the pigeon population from large to small according to the fitness function value, and selecting the individual pigeon with the highest fitness function value as the global optimal solution;

[0054] S617. Set a fitness function threshold. If the fitness function value of the current optimal individual is greater than the fitness function threshold, it is determined that the qualified image feature data of the target product corresponding to the global optimal solution and the real-time image feature data of the target product for image feature matching are successfully matched; otherwise, the real-time image feature data of the target product for image feature matching and all qualified image feature data of the target product in the qualified image feature data set under the corresponding target product processing procedure data are unsuccessfully matched.

[0055] The present invention realizes the product image feature matching process through the pigeon flock optimization algorithm. Compared with the traditional image similarity matching method, it effectively improves the speed and efficiency of the image feature matching process and effectively avoids the image feature matching results from falling into the local optimal solution. At the same time, the pigeon flock optimization algorithm has a certain robustness and can effectively avoid the influence of noise in the analysis data, thereby improving the reliability and stability of the image feature matching process.

[0056] Preferably, when the operation state of the automated production line is abnormal, the specific steps of constructing the automated production line operation management data based on the abnormal image feature data of the target product and pushing it to the automated production line management platform are as follows:

[0057] S71. When the operation state of the automated production line is abnormal, a bidirectional search algorithm is used to perform abnormal processing process data search processing on the target product processing process data in the target product processing process data set and the target product abnormal image feature data in the target product abnormal image feature data set, and the target product processing process data corresponding to each target product abnormal image feature data in the target product abnormal image feature data set are searched out, and data identification is performed to generate a target product abnormal processing process data set F = {f1, f2, ..., f i ,…,f p}, where f i Indicates the processing procedure data corresponding to the abnormal image feature data of the i-th target product;

[0058] S72, combining the target product abnormal image feature data set E and the target product abnormal processing process data set F to construct automated production line operation management data Q = (E, F);

[0059] S73. Push the automated production line operation management data Q=(E, F) to the automated production line management platform through the Internet of Things communication network and issue an alarm.

[0060] The present invention also discloses an automated production line management system based on precise control, comprising a target product text feature data acquisition module, a target product processing procedure data search module, a target product real-time image data acquisition module, a target product image data processing module, a target product qualified image feature data collection module, an automated production line operation status determination module and an automated production line management module;

[0061] The target product text feature data collection module collects text feature data of the product to be produced online through the production line product information management platform to generate target product text feature data;

[0062] The target product processing procedure data search module uses a breadth search algorithm to search and process the established different product processing procedure data and the target product text feature data in ascending order of serial numbers to generate the target product processing procedure data;

[0063] The target product real-time image data acquisition module acquires the target product image of the target product under the target product processing process data in real time during the operation of the automated production line through a high-definition image acquisition device, and generates the target product real-time image data;

[0064] The target product image data processing module performs image data filtering and noise reduction processing on the real-time image data of the target product by weighted mean filtering method to obtain the real-time image feature data of the target product;

[0065] The target product qualified image feature data collection module generates the target product qualified image feature data by calling the image feature data of the target product under the target product processing process data from the qualified product image feature data storage unit;

[0066] The automatic production line operation status determination module performs product image feature matching on the real-time image feature data of the target product and the qualified image feature data of the target product, determines the operation status of the automatic production line according to the product image feature matching result, and obtains the abnormal image feature data of the target product;

[0067] The automated production line management module searches for the target product processing procedure data corresponding to the target product abnormal image feature data through a bidirectional search algorithm, and performs data identification to generate the target product abnormal processing procedure data, combines the target product abnormal image feature data with the target product abnormal image feature data, constructs the automated production line operation management data, and pushes it to the automated production line management platform through the Internet of Things communication network.

[0068] (III) Beneficial effects

[0069] 1. The present invention generates target product processing process data by performing product processing process data search and processing on the collected target product text feature data and the established processing process data of different products, collects image data corresponding to the target product processing process data in real time through an image acquisition device, and then performs data preprocessing on the image data to generate real-time image feature data of the target product, performs product image feature matching on the real-time image feature data of the target product and the established qualified image feature data of the target product, determines the operating status of the automated production line according to the product image feature matching result, and constructs automated production line operation management data and pushes it to the automated production line management platform to realize intelligent management of the automated production line.

[0070] 2. Use the weighted mean filtering method to filter and reduce noise on the real-time collected target product image data to generate real-time image feature data of the target product, effectively reduce the impact of image noise, ensure the quality and accuracy of the acquired image feature data, and provide a reliable data basis for subsequent operations;

[0071] 3. The present invention realizes the product image feature matching process through the pigeon flock optimization algorithm. Compared with the traditional image similarity matching method, it effectively improves the speed and efficiency of the image feature matching process and effectively avoids the image feature matching results from falling into the local optimal solution. At the same time, the pigeon flock optimization algorithm has a certain robustness and can effectively avoid the influence of noise in the analysis data, thereby improving the reliability and stability of the image feature matching process. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] In order to more clearly illustrate the technical solutions of the embodiments of the invention, the drawings required for describing the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the invention. For ordinary technicians in this field, they can also obtain drawings based on these drawings without paying creative work.

[0073] Figure 1 A flowchart of an automated production line management method based on precise control provided by the present invention;

[0074] Figure 2A module schematic diagram of an automated production line management system based on precise control provided by the present invention. DETAILED DESCRIPTION

[0075] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only 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.

[0076] In the description of the present invention, it is necessary to understand that the terms "opening", "upper", "lower", "top", "middle", "inside" and the like indicating orientation or positional relationship are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the components or elements referred to must have a specific orientation, be constructed and operate in a specific orientation, and therefore cannot be understood as limiting the invention.

[0077] Embodiment 1 is as follows:

[0078] See also Figure 1 , an automated production line management method based on precise control, comprising the following steps:

[0079] S1. Collect text feature data of target products;

[0080] S11. Collect text feature data of the products to be produced online through the production line product information management platform to generate target product text feature data A, wherein the target product text feature data includes product type, product specification, product quality standard, product quantity and product production cycle.

[0081] S2, establishing different product processing procedure data, performing product processing procedure data search and processing according to the different product processing procedure data and the target product text feature data, and generating target product processing procedure data;

[0082] S21. Establish different product processing process data sets B = {b1, b2, ..., b i ,…,b k}, where b i represents the different product processing procedure data corresponding to the i-th type of product, k represents the total number of product types, and the different product processing procedure data represent the various processing procedures when the automated production line produces different products;

[0083] S22, using a breadth search algorithm to perform product processing procedure data search processing on the target product text feature data A and the different product processing procedure data in the different product processing procedure data sets in ascending order of serial numbers, search out the different product processing procedure data corresponding to the target product text feature data, and perform data identification to generate a target product processing procedure data set C = {c1, c2, ..., c i ,…,c l}, where c i represents the i-th processing step data of the target product, and l represents the total number of processing step data of the target product.

[0084] S3, collecting the target product image under the target product processing process data in the process of the automated production line in real time through an image acquisition device, and generating real-time image data of the target product;

[0085] S31, installing a high-definition image acquisition device on the automated production line, and using the high-definition image acquisition device to collect in real time the target product images of the target product under each target product processing process data in the target product processing process data set during the operation of the automated production line, and generating a target product real-time image data set D = {d1, d2, ..., d i ,…,d l}, where d i Represents the real-time image data of the target product under the i-th target product processing process data.

[0086] S4, performing data preprocessing on the real-time image data of the target product to obtain real-time image feature data of the target product;

[0087] S41, using a weighted mean filtering method to perform image data filtering and noise reduction processing on the target product real-time image data in the target product real-time image data set, to obtain a target product real-time image feature data set D'={d'1, d'2, ..., d' i ,…,d' l}, where d' i represents the target product real-time image feature data obtained after the real-time image data of the target product under the i-th target product processing process data is subjected to image data filtering and noise reduction processing;

[0088] S411, randomly selecting any one of the target product real-time image data in the target product real-time image data set as sample image data;

[0089] S412, setting an m×m neighborhood window, and aligning the center position of the neighborhood window with each pixel point of the sample image data in sequence;

[0090] S413, calculating the weight coefficient of each pixel in the neighborhood window by a weighted kernel function formula; the weighted kernel function formula is as follows:

[0091]

[0092] Among them, w(i,j) represents the weight coefficient of pixel point (i,j), d(i,j) represents the Euclidean distance from pixel point (i,j) to the center point of the neighborhood window after Gaussian weighting, and a represents the smoothing parameter of the weighted kernel function;

[0093] S414, calculating the weighted average value of the pixels in the neighborhood window, and using the weighted average value as the pixel value of the center point of the neighborhood window; the weighted average value calculation formula is as follows:

[0094]

[0095] Wherein, Gray represents the weighted average value of the pixels in the neighborhood window, and f(i,j) represents the pixel value of the pixel point (i,j);

[0096] S415, repeating the operations in S411 to S414 until all the target product real-time image data in the target product real-time image data set are traversed to obtain the target product real-time image feature data set.

[0097] S5. Establishing qualified image feature data of target products;

[0098] S51, calling the image feature data of the target product under each target product processing process data in the target product processing process data set from the qualified product image feature data storage unit to generate a target product qualified image feature data set in, Represents the target product qualified image feature dataset under the i-th target product processing process data.

[0099] S6, performing product image feature matching on the real-time image feature data of the target product and the qualified image feature data of the target product, determining the operation status of the automated production line according to the product image feature matching result, and obtaining abnormal image feature data of the target product;

[0100] S61, perform product image feature matching on each target product real-time image feature data set in the target product real-time image feature data set in order of serial number from small to large with all target product qualified image feature data sets in the target product qualified image feature data set under the corresponding target product processing procedure data; if each target product real-time image feature data set in the target product real-time image feature data set matches all target product qualified image feature data sets in the target product qualified image feature data set under the corresponding target product processing procedure data successfully, the operation status of the automated production line is determined to be normal, and no operation is performed; otherwise, the operation status of the automated production line is determined to be abnormal, and the target product real-time image feature data sets that fail to match all target product qualified image feature data sets in the target product qualified image feature data set under the corresponding target product processing procedure data are screened out, and data identification is performed to generate a target product abnormal image feature data set E = {e1, e2, …, e i ,…,e p}, where e i represents the abnormal image feature data of the i-th target product, and p represents the total number of abnormal image feature data of the target product; the product image feature matching process is as follows:

[0101] S611, constructing a pigeon population, setting the size of the pigeon population to N, the current number of iterations to t, the number of iterations in the map navigation phase to t1, the number of iterations in the landmark navigation phase to t2, and the dimension of the product image feature search space to P;

[0102] In the product image feature search space, N target product qualified image feature data are randomly generated, each target product qualified image feature data corresponds to a pigeon individual in the pigeon population, and the initial position set of the pigeon population is initialized as X = {X1, X2, ..., X i ,…,X N}, where X i represents the initial position of the i-th pigeon in the pigeon population;

[0103] S612, calculating the fitness function value of each individual pigeon in the pigeon population, arranging each individual pigeon in the pigeon population from large to small according to the fitness function value, and selecting the individual pigeon with the highest fitness function value as the current optimal individual; the fitness function formula is as follows:

[0104]

[0105] Among them, Fit i represents the fitness function value of the i-th pigeon in the pigeon population, x irepresents the deviation between the qualified image feature data of the target product corresponding to the i-th individual pigeon in the pigeon population and the real-time image feature data of the target product for product image feature matching, and φ represents the correction value;

[0106] S613, map navigation stage, each individual pigeon in the pigeon population adjusts its own flight speed and flight direction in the product image feature search space through the map and compass operator to update its position; the position update formula is as follows:

[0107]

[0108] in, and They represent the speed and position of the i-th pigeon in the pigeon population during the t+1th iteration, and represents the speed and position of the i-th pigeon in the pigeon population during the t-th iteration, ε represents the map and compass operator, rand1 represents a random number uniformly distributed between (0,1), and X best Indicates the position of the current optimal individual;

[0109] S614, judging whether the current number of iterations t is less than the number of iterations t1 in the map navigation phase, if the current number of iterations t is less than the number of iterations t1 in the map navigation phase, then the current number of iterations t is increased by 1, and the process returns to S612; otherwise, the fitness function value of each individual pigeon in the pigeon population is recalculated, and the process proceeds to S615;

[0110] S615, landmark navigation stage, each individual pigeon in the pigeon population optimizes its own flight direction and flight speed in the product image feature search space through surrounding familiar landmarks to update its position, and at the same time, discards individual pigeons with lower fitness function values ​​in each iteration process; the position update formula is as follows:

[0111]

[0112] Among them, N t+1 and N t They represent the pigeon population size in the t+1th iteration process and the tth iteration process respectively, represents the center position of the pigeon population during the tth iteration, and rand2 represents a random number uniformly distributed between (0,1);

[0113] S616, judging whether the current number of iterations t is less than the number of iterations t2 in the landmark navigation phase, if the current number of iterations t is less than the number of iterations t2 in the landmark navigation phase, then adding 1 to the current number of iterations t, and returning to S615; otherwise, recalculating the fitness function value of each individual pigeon in the pigeon population, arranging each individual pigeon in the pigeon population from large to small according to the fitness function value, and selecting the individual pigeon with the highest fitness function value as the global optimal solution;

[0114] S617. Set a fitness function threshold. If the fitness function value of the current optimal individual is greater than the fitness function threshold, it is determined that the qualified image feature data of the target product corresponding to the global optimal solution and the real-time image feature data of the target product for image feature matching are successfully matched; otherwise, the real-time image feature data of the target product for image feature matching and all qualified image feature data of the target product in the qualified image feature data set under the corresponding target product processing procedure data are unsuccessfully matched.

[0115] S7. When the operation state of the automated production line is abnormal, the automated production line operation management data is constructed based on the abnormal image feature data of the target product, and pushed to the automated production line management platform;

[0116] S71. When the operation state of the automated production line is abnormal, a bidirectional search algorithm is used to perform abnormal processing process data search processing on the target product processing process data in the target product processing process data set and the target product abnormal image feature data in the target product abnormal image feature data set, and the target product processing process data corresponding to each target product abnormal image feature data in the target product abnormal image feature data set are searched out, and data identification is performed to generate a target product abnormal processing process data set F = {f1, f2, ..., f i ,…,f p}, where f i Indicates the processing procedure data corresponding to the abnormal image feature data of the i-th target product;

[0117] S72, combining the target product abnormal image feature data set E and the target product abnormal processing process data set F to construct automated production line operation management data Q = (E, F);

[0118] S73. Push the automated production line operation management data Q=(E, F) to the automated production line management platform through the Internet of Things communication network and issue an alarm.

[0119] Embodiment 2 is as follows:

[0120] See also Figure 2, an automated production line management system based on precise control, including a target product text feature data acquisition module, a target product processing procedure data search module, a target product real-time image data acquisition module, a target product image data processing module, a target product qualified image feature data collection module, an automated production line operation status determination module and an automated production line management module;

[0121] The target product text feature data collection module collects text feature data of the product to be produced online through the production line product information management platform to generate target product text feature data;

[0122] The target product processing procedure data search module uses a breadth search algorithm to search and process the established different product processing procedure data and the target product text feature data in ascending order of serial numbers to generate the target product processing procedure data;

[0123] The target product real-time image data acquisition module acquires the target product image of the target product under the target product processing process data in real time during the operation of the automated production line through a high-definition image acquisition device, and generates the target product real-time image data;

[0124] The target product image data processing module performs image data filtering and noise reduction processing on the real-time image data of the target product by weighted mean filtering method to obtain the real-time image feature data of the target product;

[0125] The target product qualified image feature data collection module generates the target product qualified image feature data by calling the image feature data of the target product under the target product processing process data from the qualified product image feature data storage unit;

[0126] The automatic production line operation status determination module performs product image feature matching on the real-time image feature data of the target product and the qualified image feature data of the target product, determines the operation status of the automatic production line according to the product image feature matching result, and obtains the abnormal image feature data of the target product;

[0127] The automated production line management module searches for the target product processing procedure data corresponding to the target product abnormal image feature data through a bidirectional search algorithm, and performs data identification to generate the target product abnormal processing procedure data, combines the target product abnormal image feature data with the target product abnormal image feature data, constructs the automated production line operation management data, and pushes it to the automated production line management platform through the Internet of Things communication network.

[0128] In the description of this specification, the description with reference to the terms "one embodiment", "example", "specific example", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0129] The preferred embodiments of the invention disclosed above are only used to help explain the invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to the specific implementation methods described. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the invention, so that those skilled in the art can understand and use the invention well.

Claims

1. An automated production line management method based on precise control, characterized in that: The steps include: S1. Collect text feature data of target products; S2, establishing different product processing procedure data, performing product processing procedure data search and processing according to the different product processing procedure data and the target product text feature data, and generating target product processing procedure data; S3, collecting the target product image under the target product processing process data in the process of the automated production line in real time through an image acquisition device, and generating real-time image data of the target product; S4, performing data preprocessing on the real-time image data of the target product to obtain real-time image feature data of the target product; S5. Establishing qualified image feature data of target products; S6, performing product image feature matching on the real-time image feature data of the target product and the qualified image feature data of the target product, determining the operation status of the automated production line according to the product image feature matching result, and obtaining abnormal image feature data of the target product; S7. When the operation status of the automated production line is abnormal, the automated production line operation management data is constructed based on the abnormal image feature data of the target product and pushed to the automated production line management platform.

2. The method for managing an automated production line based on precise control according to claim 1, characterized in that: The S1 comprises the following steps: S11. Collect text feature data of the products to be produced online through the production line product information management platform to generate target product text feature data A, wherein the target product text feature data includes product type, product specification, product quality standard, product quantity and product production cycle.

3. The method for managing an automated production line based on precise control according to claim 2, characterized in that: The S2 comprises the following steps: S21. Establish different product processing process data sets B = {b1, b2, ..., b i ,…,b k }, where b i represents the different product processing procedure data corresponding to the i-th type of product, k represents the total number of product types, and the different product processing procedure data represent the various processing procedures when the automated production line produces different products; S22, using a breadth search algorithm to perform product processing procedure data search processing on the target product text feature data A and the different product processing procedure data in the different product processing procedure data sets in ascending order of serial numbers, search out the different product processing procedure data corresponding to the target product text feature data, and perform data identification to generate a target product processing procedure data set C = {c1, c2, ..., c i ,…,c l }, where c i represents the i-th processing step data of the target product, and l represents the total number of processing step data of the target product.

4. The method for managing an automated production line based on precise control according to claim 3, characterized in that: The S3 comprises the following steps: S31. Install a high-definition image acquisition device on the automated production line, and use the high-definition image acquisition device to collect in real time the target product images of the target product under each target product processing process data in the target product processing process data set during the operation of the automated production line, to generate a target product real-time image data set.

5. The method for managing an automated production line based on precise control according to claim 4, characterized in that: The 4 steps include the following steps: S41, using a weighted mean filtering method to perform image data filtering and noise reduction processing on the target product real-time image data in the target product real-time image data set to obtain a target product real-time image feature data set.

6. The method for managing an automated production line based on precise control according to claim 5, characterized in that: The S41 comprises the following steps: S411, randomly selecting any one of the target product real-time image data in the target product real-time image data set as sample image data; S412, setting an m×m neighborhood window, and aligning the center position of the neighborhood window with each pixel point of the sample image data in sequence; S413, calculating the weight coefficient of each pixel in the neighborhood window by using a weighted kernel function; S414, calculating the weighted average value of the pixels in the neighborhood window, and using the weighted average value as the pixel value of the center point of the neighborhood window; S415, repeating the operations in S411 to S414 until all the target product real-time image data in the target product real-time image data set are traversed to obtain the target product real-time image feature data set.

7. The method for managing an automated production line based on precise control according to claim 6, characterized in that: The S5 comprises the following steps: S51. Call the image feature data of the target product under each target product processing procedure data in the target product processing procedure data set from the qualified product image feature data storage unit to generate a target product qualified image feature data set.

8. The method for managing an automated production line based on precise control according to claim 7, characterized in that: The S6 comprises the following steps: S61, perform product image feature matching on each target product real-time image feature data set in the target product real-time image feature data set in order of serial number from small to large with all target product qualified image feature data sets in the target product qualified image feature data set under the corresponding target product processing procedure data; if each target product real-time image feature data set in the target product real-time image feature data set matches all target product qualified image feature data sets in the target product qualified image feature data set under the corresponding target product processing procedure data successfully, the operation status of the automated production line is determined to be normal, and no operation is performed; otherwise, the operation status of the automated production line is determined to be abnormal, and the target product real-time image feature data sets that fail to match all target product qualified image feature data sets in the target product qualified image feature data set under the corresponding target product processing procedure data are screened out, and data identification is performed to generate a target product abnormal image feature data set E; the product image feature matching process is as follows: S611, constructing a pigeon population, setting the size of the pigeon population to N, the current number of iterations to t, the number of iterations in the map navigation phase to t1, the number of iterations in the landmark navigation phase to t2, and the dimension of the product image feature search space to P; Randomly generate N target product qualified image feature data in the product image feature search space, each target product qualified image feature data corresponds to a pigeon individual in the pigeon population, and initialize the initial position set of the pigeon population; S612, calculating the fitness function value of each individual pigeon in the pigeon population, arranging the individual pigeons in the pigeon population from large to small according to the fitness function value, and selecting the individual pigeon with the highest fitness function value as the current optimal individual; S613, in the map navigation stage, each individual pigeon in the pigeon population adjusts its own flight speed and flight direction in the product image feature search space through a map and compass operator to update its position; S614, judging whether the current number of iterations t is less than the number of iterations t1 in the map navigation phase, if the current number of iterations t is less than the number of iterations t1 in the map navigation phase, then the current number of iterations t is increased by 1, and the process returns to S612; otherwise, the fitness function value of each individual pigeon in the pigeon population is recalculated, and the process proceeds to S615; S615, landmark navigation stage, each individual pigeon in the pigeon population optimizes its own flight direction and flight speed in the product image feature search space through surrounding familiar landmarks to update its position, and at the same time, discards individual pigeons with lower fitness function values ​​in each iteration process; S616, judging whether the current number of iterations t is less than the number of iterations t2 in the landmark navigation phase, if the current number of iterations t is less than the number of iterations t2 in the landmark navigation phase, then adding 1 to the current number of iterations t, and returning to S615; otherwise, recalculating the fitness function value of each individual pigeon in the pigeon population, arranging each individual pigeon in the pigeon population from large to small according to the fitness function value, and selecting the individual pigeon with the highest fitness function value as the global optimal solution; S617. Set a fitness function threshold. If the fitness function value of the current optimal individual is greater than the fitness function threshold, it is determined that the qualified image feature data of the target product corresponding to the global optimal solution and the real-time image feature data of the target product for image feature matching are successfully matched; otherwise, the real-time image feature data of the target product for image feature matching and all qualified image feature data of the target product in the qualified image feature data set under the corresponding target product processing procedure data are unsuccessfully matched.

9. The method for managing an automated production line based on precise control according to claim 8, characterized in that: The S7 comprises the following steps: S71. When the operation state of the automated production line is abnormal, a bidirectional search algorithm is used to perform abnormal processing process data search processing on the target product processing process data in the target product processing process data set and the target product abnormal image feature data in the target product abnormal image feature data set, and the target product processing process data corresponding to each target product abnormal image feature data in the target product abnormal image feature data set are searched out, and data identification is performed to generate a target product abnormal processing process data set F; S72, combining the target product abnormal image feature data set E and the target product abnormal processing process data set F to construct automated production line operation management data Q = (E, F); S73, push the automated production line operation management data Q = (E, F) through the Internet of Things communication network The data is sent to the automated production line management platform and an alarm is issued.

10. A system for implementing the automated production line management method based on precise control as described in any one of claims 1 to 9.

Images