Image identification cutting control method, system and device in power line production

Abstract

The invention relates to an image recognition cutting control method, system and device in power line production, and relates to the field of intelligent control. A preset cutting path of a cutter, a wire model, a wire specification, a cutting error threshold value, a power line zero tension length and an expected cutting length are received; retrieving a plurality of stretching ratios and cutting errors by taking the model and specification of the wire as constraints, counting a target stretching length meeting a cutting error threshold value based on the zero tension length of the power line, configuring a cutting diametral plane position in combination with an expected cutting length, receiving a real-time image of the power line, and performing recognition to obtain first and second end monitoring positions of the power line; and when the distance meets the target stretching length and the preset cutting path of the cutter belongs to the cutting diametric plane position, performing no-abnormality identification on the real-time image of the power line. The technical problems of poor cutting precision, large cutting length error, irregular cutting end face, low cutting efficiency and low automation degree in the cutting operation of the power line are solved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent control, and particularly to an image recognition and cutting control method, system and device in the production of power cords. Background Art

[0002] During the production process of power cords, the accuracy and efficiency of the cutting operation directly affect the quality of the final product and the production cost. Traditional power cord cutting methods often rely on manual operation or simple mechanical equipment, and these methods are often inefficient and of uneven quality when dealing with complex and variable wire materials and specifications. In particular, since the tension of the wire is difficult to stably control, a large number of sensors need to be deployed for real-time adjustment in traditional methods. However, this approach not only increases the complexity and maintenance cost of the production line, but also makes the control path extremely cumbersome. In addition, due to the limitations of the accuracy and response speed of the sensors themselves, as well as various external interference factors that the wire may be subjected to during the production process, such as temperature, humidity, vibration, etc., it is difficult to guarantee the cutting accuracy, and problems such as large cutting length errors and uneven cutting ends often occur.

[0003] With the rapid development of image recognition technology and intelligent control technology, new solutions have been brought to the power cord cutting operation. By introducing image recognition technology, real-time monitoring and precise positioning of the power cord position can be achieved, thus avoiding the cutting errors caused by the limitations of sensor accuracy and response speed in traditional methods. At the same time, combined with intelligent control strategies, the cutting action can be dynamically adjusted according to the real-time obtained power cord image information, further improving the cutting accuracy and automation level. However, although image recognition technology has great application potential in the field of power cord cutting, there is currently no mature method or system that can fully meet the high-precision and high-efficiency fully automated cutting requirements in the power cord production process. Therefore, it is of great significance to develop a cutting control method in power cord production based on image recognition for improving the production efficiency and product quality of power cords. Summary of the Invention

[0004] In view of the technical problems in the prior art that the cutting accuracy of the power cord cutting operation is poor, the cutting length error is large, the cutting end face is uneven, and the cutting efficiency and automation level are low, the present invention provides an image recognition and cutting control method, system and device in the production of power cords to solve these problems.

[0005] The technical solution of the present invention to solve the above technical problems is as follows:

[0006] In a first aspect, the present invention provides an image recognition and cutting control method in power cord production, including: receiving a preset cutting path of a cutting tool, wire model and wire specification, cutting error threshold, zero-tension length of the power cord, and expected cutting length from a user terminal; retrieving a number of stretching ratios and a number of cutting errors with the wire model and wire specification as constraints, and based on the zero-tension length of the power cord, counting a target stretching length that meets the cutting error threshold; configuring a cutting path plane position according to the target stretching length in combination with the expected cutting length; receiving a real-time image of the power cord before performing the power cord cutting operation, performing recognition, and obtaining a monitoring position of the first end of the power cord and a monitoring position of the second end of the power cord; when the distance between the monitoring position of the first end of the power cord and the monitoring position of the second end of the power cord meets the target stretching length, and the preset cutting path of the cutting tool belongs to the cutting path plane position, making a non-abnormal identification for the real-time image of the power cord.

[0007] In a second aspect, the present invention provides an image recognition and cutting control system in power cord production. The system includes: an information acquisition module for receiving a preset cutting path of a cutting tool, wire model and wire specification, cutting error threshold, zero-tension length of the power cord, and expected cutting length from a user terminal; a data retrieval module for retrieving a number of stretching ratios and a number of cutting errors with the wire model and wire specification as constraints, and based on the zero-tension length of the power cord, counting a target stretching length that meets the cutting error threshold; a cutting configuration module for configuring a cutting path plane position according to the target stretching length in combination with the expected cutting length; an image recognition module for receiving a real-time image of the power cord before performing the power cord cutting operation, performing recognition, and obtaining a monitoring position of the first end of the power cord and a monitoring position of the second end of the power cord; an abnormal identification module for making a non-abnormal identification for the real-time image of the power cord when the distance between the monitoring position of the first end of the power cord and the monitoring position of the second end of the power cord meets the target stretching length, and the preset cutting path of the cutting tool belongs to the cutting path plane position.

[0008] In a third aspect, the present invention provides an electronic device, including: a memory for storing a computer software program; a processor for reading and executing the computer software program to implement an image recognition and cutting control method in power cord production as described in the first aspect.

[0009] The beneficial effects of the present invention are as follows: By receiving the cropping parameters input by the user, retrieving and determining the target stretching length using the constraints of the wire model and specifications, and then configuring the cropping radial plane position. Before performing the cropping operation, the real-time monitoring positions of both ends of the power cord are obtained through image recognition technology, and the power cord is marked as normal without abnormalities under the conditions of meeting the target stretching length and the preset cropping path, thereby achieving precise cropping, significantly improving the cropping accuracy and automation level, reducing the dependence on a large number of sensors, simplifying the control path, and enhancing the production efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Figure 1 It is a schematic flow chart of an image recognition and cropping control method in the production of a power cord provided by the present invention.

[0011] Figure 2 It is a schematic structural diagram of an image recognition and cropping control system in the production of a power cord provided by the present invention.

[0012] Figure 3 It is a schematic structural diagram of an electronic device provided by the present invention.

[0013] Description of reference numerals: Information acquisition module 11, data retrieval module 12, cutting configuration module 13, image recognition module 14, anomaly marking module 15, electronic device 500, memory 510, processor 520, first computer program 511. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0014] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative efforts fall within the protection scope of the present invention.

[0015] In the description of the present invention, the terms "first" and "second" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of the described features. In the description of the present invention, "a plurality" means two or more, unless otherwise specifically defined.

[0016] In the description of the present invention, the term "for example" is used to mean "serving as an example, illustration, or explanation". Any embodiment described as "for example" in the present invention is not necessarily construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to implement and use the present invention. In the following description, details are set forth for purposes of explanation. It should be understood that those of ordinary skill in the art can recognize that the present invention can be implemented without these specific details. In other instances, well-known structures and processes are not elaborated in detail to avoid obscuring the description of the present invention with unnecessary details. Therefore, the present invention is not intended to be limited to the embodiments shown, but rather to be in line with the broadest scope consistent with the principles and features disclosed in the present invention.

[0017] Embodiment 1:

[0018] As Figure 1 shown, an image recognition and cutting control method in the production of power cords provided by an embodiment of the present invention includes:

[0019] S10: Receive, from the user side, a preset cutting path of the cutting tool, the wire model and wire specification, the cutting error threshold, the zero-tension length of the power cord, and the desired cutting length.

[0020] Exemplarily, the production process of power cords generally includes steps such as raw material preparation, wire drawing, insulation layer coating, length cutting and insulation layer stripping, quality inspection, and packaging and warehousing. Among them, wire length cutting and insulation layer stripping are important links to ensure that the power cords meet the specification requirements, and they directly affect the final use performance and appearance quality of the power cords. When performing length cutting, it is necessary to cut according to the actual use situation or specifications, and at the same time clamp both ends of the wire to reduce the length error during positioning cutting and ensure the flatness of the cutting ports. However, the conventional method cannot effectively fix both ends, resulting in a large length error during cutting, and at the same time, there will be uneven or inclined end faces during the cutting process, thus affecting the use of the wire. In addition, it is also necessary to remove a part of the insulating skin at the end face with a wire stripper, and the existing means often cannot effectively remove the end face insulating skin fully automatically, thereby realizing fully automated continuous operation, increasing the production speed, and improving the work efficiency.

[0021] In this solution, taking image recognition monitoring as the background, effective cutting control of the power cord is carried out to solve the above technical problems. Specifically, receiving a series of key parameters from the user side is the prerequisite for ensuring accurate and efficient cutting operations. These parameters include the preset cutting path of the tool, the wire model and wire specification, the cutting error threshold, the zero-tension length of the power cord, and the desired cutting length. First, receive the preset cutting path of the tool by the user, which determines the specific cutting shape and direction of the power cord, such as straight cutting or oblique cutting at a specific angle. Obtain the model and specification information of the wire, such as the AWG value or a specific diameter, which is crucial for subsequent calculation of cutting parameters. The cutting error threshold is set by the user, which represents the acceptable deviation range of the cutting length, ensuring the control of cutting accuracy. For example, the set error does not exceed ±0.5 mm. The zero-tension length of the power cord refers to the natural length of the wire when it is not under the action of external tensile force, which is the basis for calculating the target tensile length. Finally, the user also needs to input the desired cutting length, that is, the final required length of the power cord, such as 1 meter, 1.5 meters, etc., to meet the application requirements of different products. By receiving these parameters, the cutting operation can be accurately planned to ensure the accuracy and efficiency of power cord production. For example, when the user needs to produce a batch of power cords with a length of 1.5 meters, a model of AWG 18, and a required cutting error not exceeding ±0.2 mm, the system will perform subsequent cutting control based on these parameters.

[0022] S20: Constrained by the wire model and wire specification, retrieve a number of stretch ratios and a number of cutting errors, and based on the zero-tension length of the power cord, count the target tensile length that meets the cutting error threshold.

[0023] Optionally, during the cutting process of power cord production, to ensure the accuracy and consistency of cutting, relevant stretch ratio and cutting error data are retrieved according to the key constraint conditions of the wire model and specification. Here, the stretch ratio intuitively represents the ratio of the length of the wire after stretching to its zero-tension length (i.e., the length of the wire in its natural state), reflecting the elongation characteristics of the wire under different tensions. For example, for a certain model of thin wire, its stretch ratio may indicate that it can be extended to 1.05 times its original length under slight tension. The cutting error is an indicator that comprehensively considers the difference between the cutting record state and the desired state. Here, the Euclidean distance is used as the quantization method, which can comprehensively evaluate the quality of the cutting result. The cutting desired state not only includes the desired cutting length, that is, the specific size that the power cord should reach, such as 1 meter in length, but also covers the desired cutting neatness, that is, the flatness and perpendicularity requirements of the cutting end face, ensuring that the end face of the power cord after cutting is smooth and without inclination.

[0024] Furthermore, based on the zero-tension length of the power cord, by comparing the stretch ratio with a preset cutting error threshold, the target stretch length that can meet the cutting accuracy requirements is statistically screened out. For example, if the zero-tension length of a certain type of power cord is 1 meter, the system may find through retrieval that under the condition that the stretch ratio is 1.02 and the cutting error (measured by Euclidean distance) does not exceed 0.1 mm, the best cutting effect can be obtained. This process effectively combines the physical characteristics of the wire and the cutting accuracy requirements, provides a scientific basis for subsequent cutting operations, and ensures the accuracy and reliability of power cord cutting.

[0025] S30: Configure the cutting radial plane position according to the target stretch length and in combination with the expected cutting length.

[0026] Subsequently, according to the pre-calculated target stretch length, that is, the length that ensures the wire can meet the cutting accuracy requirements after stretching, further in combination with the user's expected cutting length (i.e., the final length that the power cord should reach), the cutting radial plane position is accurately configured. Among them, the cutting radial plane position refers to the plane position where the specific cutting path of the cutting tool is located on the power cord, which directly affects the accuracy of the cutting result and the flatness of the end face.

[0027] For example, assume that the target stretch length is 1.05 meters (obtained after slightly stretching a power cord with a zero-tension length of 1 meter), and the user's expected cutting length is 1 meter. The system will determine along which radial plane of the power cord the cutting tool should cut according to these two parameters through calculation and positioning technology to ensure that the length of the cut power cord is exactly 1 meter, and the cutting surface is flat and without inclination. This process fully considers the physical characteristics of the wire (such as stretchability) and the actual needs of the user, and realizes precise control of power cord cutting by scientifically configuring the cutting radial plane position, improving production efficiency and product quality.

[0028] S40: Before performing the power cord cutting operation, receive the real-time image of the power cord, perform recognition, and obtain the monitoring position of the first end of the power cord and the monitoring position of the second end of the power cord.

[0029] Furthermore, before performing the power cord cutting operation, receive the real-time image of the power cord. This step is to ensure the accuracy and safety of the cutting operation. The real-time image refers to the image information of the current state of the power cord captured by devices such as high-definition cameras, which contains key information such as the position and shape of the power cord.

[0030] Subsequently, image recognition technology is used to recognize the received real-time image. During this process, the two ends of the power cord, namely the first end and the second end, are concerned, and their specific positions in the image are determined through complex algorithm analysis, namely the monitoring position of the first end of the power cord and the monitoring position of the second end of the power cord. These two position information are crucial for subsequent cutting operations because they directly determine the starting point and ending point of cutting.

[0031] For example, if the power cord is in a horizontal state on the production line, the specific positions of the first end and the second end of the power cord in the horizontal direction can be accurately determined through real-time image recognition. For example, the first end is 50 cm away from the camera, and the second end is 150 cm away from the camera. With these accurate position information, the system can accurately control the moving path and cutting position of the cutting tool, thus ensuring the accuracy of the cutting operation. This process not only improves the cutting accuracy but also effectively avoids production accidents caused by misoperation, improving the overall production efficiency and safety.

[0032] S50: When the distance between the monitoring position of the first end of the power cord and the monitoring position of the second end of the power cord meets the target stretching length, and the preset cutting path of the tool belongs to the cutting path plane position, no abnormality mark is made on the real-time image of the power cord.

[0033] Specifically, in the power cord cutting operation, the positions of both ends of the power cord, namely the first end monitoring position and the second end monitoring position, are continuously monitored. When the distance between these two positions is exactly equal to the preset target stretching length, it indicates that the power cord has reached a suitable stretching state for cutting. At the same time, check whether the preset cutting path of the tool is accurately located at the previously configured cutting path plane position, which is the key to ensuring cutting accuracy and end face quality. If the above two conditions are met simultaneously, that is, the distance between the two ends of the power cord meets the target stretching length and the tool path is correct, the system will make a no-abnormality mark on the current real-time image of the power cord. This mark means that all parameters of the power cord meet the preset requirements.

[0034] For example, assume that the target stretching length is 1.05 meters. When the system monitors that the distance between the first end and the second end of the power cord is exactly 1.05 meters, and the preset cutting path of the tool is also accurately located at the cutting path plane position, the system will mark "no abnormality" on the real-time image, indicating that cutting at this time will be able to obtain a power cord product that meets the quality requirements. This process ensures the accuracy and stability of the cutting operation through accurate position monitoring and path checking, improving production efficiency and product quality.

[0035] In a preferred embodiment, before performing the power cord cutting operation, a real-time image of the power cord is received and recognition is performed to obtain the monitoring position of the first end of the power cord and the monitoring position of the second end of the power cord, including: before performing the power cord cutting operation, fixing the first end of the power cord to the first clamping member and fixing the other end of the power cord to the second clamping member; processing the real-time image of the power cord through a clamping member identifier to obtain the real-time position of the first clamping member, which is set as the monitoring position of the first end of the power cord; processing the real-time image of the power cord through the clamping member identifier to obtain the real-time position of the second clamping member, which is set as the monitoring position of the second end of the power cord.

[0036] Preferably, before performing the power cord cutting operation, to ensure the accuracy of cutting, the two ends of the power cord are first fixed to two clamping members respectively, that is, the first end is fixed to the first clamping member and the other end is fixed to the second clamping member. These two clamping members not only stabilize the power cord but also serve as key reference points for subsequent position recognition. Subsequently, the real-time image of the power cord is processed by a clamping member identifier. In this process, the identifier accurately identifies the real-time position of the first clamping member in the image and sets it as the monitoring position of the first end of the power cord; similarly, the real-time position of the second clamping member in the image is identified and set as the monitoring position of the second end of the power cord.

[0037] For example, assume that on a production line, the power cord is firmly clamped between the first clamping member and the second clamping member, and the distance between the two can be adjusted according to production requirements. When the system activates the image recognition function, it quickly captures the exact positions of the two clamping members in the real-time image. No matter what small displacements the power cord undergoes for any reason, the system can accurately determine the positions of the two ends of the power cord by recognizing the positions of the clamping members. In this way, even if the power cord shakes during transmission, the system can ensure that the cutting operation is based on accurate position information, thus greatly improving the accuracy and stability of cutting.

[0038] In a preferred embodiment, processing the real-time image of the power cord through a clamping member identifier to obtain the real-time position of the first clamping member, which is set as the monitoring position of the first end of the power cord, includes: collecting a dataset of operation image records and a dataset of clamping member spatial position identifiers of a preset model clamping member at the cutting station where the power cord is produced; training the clamping member identifier with the dataset of clamping member spatial position identifiers as the supervision and the dataset of operation image records as the input.

[0039] Specifically, in order to accurately obtain the monitoring position of the first end of the power cord, a clamping part identifier is used to process real-time images. The realization of this process depends on two key data sets: one is the operation image recording data set of the preset model clamping parts at the cutting station, which contains a large amount of image information on the relative positions of the clamping parts and the power cord under different working conditions; the other is the clamping part spatial position identification data set, which provides the accurate spatial position information of the clamping parts corresponding to the operation image recording.

[0040] Using the clamping part spatial position identification data set as the supervision signal and the operation image recording data set as the input, the clamping part identifier is trained. During the training process, the identifier learns how to identify the features of the clamping parts from the images and associate them with the spatial position information. After sufficient training, the clamping part identifier can accurately process the real-time images of the power cord, identify the first clamping part in the images, and determine its real-time position accordingly, which is set as the monitoring position of the first end of the power cord.

[0041] For example, assume that on the production line, a certain model of clamping part is used to fix the power cord. The system first collects the image records of this model of clamping part in different operating states and simultaneously records its corresponding spatial position information. Then, the clamping part identifier is trained using these data. During the actual cutting operation, when the power cord is fixed by the clamping part, the system collects real-time images through the camera, and the identifier quickly processes the images to accurately identify the position of the first clamping part, thereby determining the monitoring position of the first end of the power cord. This process ensures the accuracy and stability of the cutting operation and improves production efficiency.

[0042] In a preferred embodiment, with the clamping member spatial position identification data set as the supervision and the operation image recording data set as the input, training the clamping member identifier includes: Any set of operation image recording data in the operation image recording data set includes operation image recording data above the work station and clamping member identification recording data in the image above the work station, operation image recording data on the first side of the work station and clamping member identification recording data in the image on the first side of the work station, operation image recording data on the second side of the work station and clamping member identification recording data in the image on the second side of the work station; With the clamping member identification recording data in the image above the work station as the supervision and the operation image recording data above the work station as the input, training a convolutional neural network to obtain the first clamping member identification branch; With the clamping member identification recording data in the image on the first side of the work station as the supervision and the operation image recording data on the first side of the work station as the input, training a convolutional neural network to obtain the second clamping member identification branch; With the clamping member identification recording data in the image on the second side of the work station as the supervision and the operation image recording data on the second side of the work station as the input, training a convolutional neural network to obtain the third clamping member identification branch; Combining the output layers of the first clamping member identification branch, the second clamping member identification branch, and the third clamping member identification branch with three input nodes of the fully connected layer respectively to obtain the clamping member identifier architecture; With the clamping member spatial position identification data set as the supervision and the operation image recording data set as the input, training the clamping member identifier architecture to obtain the clamping member identifier.

[0043] Further, in the process of training the clamping member identifier, an operation image recording data set is constructed, which covers the operation image recording data above the work station, on the first side and on the second side, as well as the corresponding clamping member identification recording data. Specifically, each set of operation image recording data includes the operation image above the work station and its clamping member identification, the operation image on the first side of the work station and its clamping member identification, and the operation image on the second side of the work station and its clamping member identification.

[0044] To make full use of this data, a step-by-step strategy is adopted in the training process. First, using the data recorded by the clamping part identifier in the image above the work station as the supervision signal and the data recorded by the operation image above the work station as the input, a convolutional neural network is trained to obtain the first clamping part identifier branch, which focuses on identifying the clamping part from the image above the work station. Second, similarly, using the data recorded by the clamping part identifier in the images on the first side and the second side of the work station as the supervision, the second and third clamping part identifier branches are trained respectively, and these two branches are responsible for identifying the clamping part from the images on the first side and the second side of the work station respectively. Next, the output layers of these three clamping part identifier branches are merged with the three input nodes of the fully connected layer to form a unified clamping part identifier architecture. This architecture can integrate information from different perspectives (above, first side, second side) to improve the recognition accuracy. Finally, using the clamping part spatial position identifier data set as the supervision signal and the entire operation image record data set as the input, the clamping part identifier architecture is trained as a whole, and finally a clamping part identifier that can accurately identify the position of the clamping part and determine the power line monitoring position accordingly is obtained.

[0045] For example, assume that on a specific production line work station, a camera captures operation images from three angles: above, the first side, and the second side, while recording the position information of the clamping part. Through the above training process, the clamping part identifier can accurately identify the clamping part in the image and determine the fixed position of the power line accordingly, providing accurate position information for subsequent cutting operations.

[0046] In a preferred embodiment, with the wire model and wire specification as constraints, several stretching ratios and several cutting errors are retrieved. Based on the zero-tension length of the power line, the target stretching length that meets the cutting error threshold is statistically analyzed, including: based on the several cutting errors, extracting a set of target stretching ratios less than or equal to the cutting error threshold from the several stretching ratios; based on the zero-tension length of the power line, traversing the set of target stretching ratios to find the ratio and obtaining an initial set of target stretching lengths; performing a central tendency analysis on the initial set of target stretching lengths to obtain a stretching length distribution interval, which is set as the target stretching length.

[0047] Specifically, to ensure the accuracy and consistency of cutting, it is necessary to retrieve and analyze the draw ratio and cutting error data with the wire model and wire specification as constraints. Based on a number of preset cutting error thresholds, those target draw ratios that are less than or equal to the threshold are selected from the retrieved draw ratios to form a set of target draw ratios. This step aims to exclude those draw ratios that may cause excessive cutting errors and ensure that the subsequent calculated draw lengths can meet the cutting accuracy requirements. Subsequently, based on the zero-tension length of the power cord (i.e., the length of the wire in the natural state), the set of target draw ratios is traversed, and the initial set of target draw lengths is obtained through ratio calculation. This set contains the target lengths that the power cord should reach under different draw ratios and is the basis for subsequent analysis. Finally, a central tendency analysis is performed on the initial set of target draw lengths, such as calculating statistics such as the mean and median, to obtain the distribution interval of the draw lengths. This distribution interval represents the reasonable range of the power cord draw lengths under different draw ratios, that is, the target draw lengths. By setting such target draw lengths, it can be ensured that during the cutting process, the length of the power cord can not only meet the design requirements but also be controlled within the allowable error range.

[0048] For example, assume that the zero-tension length of a certain model of power cord is 1 meter. After retrieval and analysis, the set of target draw ratios obtained is {1.02, 1.03, 1.04}, and the cutting error threshold is set to 0.01 meter. By traversing the set of target draw ratios and calculating the ratios, the initial set of target draw lengths obtained is {1.02 meters, 1.03 meters, 1.04 meters}. After further central tendency analysis, the distribution interval of the draw lengths is determined to be from 1.02 meters to 1.04 meters, that is, the target draw lengths. In this way, during the cutting process, the length of the power cord can be controlled within this range to ensure the accuracy and consistency of cutting.

[0049] In a preferred embodiment, with the wire model and wire specification as constraints, a number of draw ratios and a number of cutting errors are retrieved, including: with the wire model and the wire specification as constraints, a number of wire cutting logs are retrieved, wherein any one of the number of wire cutting logs includes a recorded value of the zero-tension length of the power cord, a recorded value of the drawn length of the power cord, a recorded value of the included angle between the cutting cross-section and the preset radial plane, and a recorded value of the cutting length deviation; traversing the number of wire cutting logs, calculating the ratio of the recorded value of the zero-tension length of the power cord to the recorded value of the drawn length of the power cord to obtain the number of draw ratios; configuring a first weight for the normalized parameter of the recorded value of the included angle between the cutting cross-section and the preset radial plane, and configuring a second weight for the normalized parameter of the recorded value of the cutting length deviation; according to the first weight and the second weight, traversing the number of wire cutting logs, and performing a weighted mean calculation on the normalized values of the recorded value of the included angle between the cutting cross-section and the preset radial plane and the recorded value of the cutting length deviation to obtain the number of cutting errors.

[0050] Exemplarily, the specific retrieval execution includes: retrieving a number of wire cutting logs according to the wire model and specifications. These logs detail the key parameters of each cutting operation, including the recorded value of the zero-tension length of the power cord (i.e., the length of the wire in its natural state), the recorded value of the stretched length of the power cord (the length after stretching), the recorded value of the angle between the cutting cross-section and the preset diameter plane (reflecting the neatness of the cutting), and the recorded value of the cutting length deviation (the difference between the actual cutting length and the expected length). Furthermore, traverse these cutting logs. By calculating the ratio of the recorded value of the zero-tension length of the power cord to the recorded value of the stretched length of the power cord, a number of stretch ratios are obtained. These stretch ratios reflect the stretching degree of the wire in different cutting operations. At the same time, in order to quantify the cutting error, weights need to be configured for the normalization parameters of the recorded value of the angle between the cutting cross-section and the preset diameter plane and the recorded value of the cutting length deviation respectively. The first weight is used to adjust the relative importance of the angle between the cutting cross-section and the preset diameter plane, while the second weight is used to adjust the relative importance of the cutting length deviation. According to these weights, traverse the cutting logs again, and calculate the weighted mean of the normalized values of the recorded value of the angle between the cutting cross-section and the preset diameter plane and the recorded value of the cutting length deviation, so as to obtain a number of cutting error values. These cutting error values comprehensively reflect the accuracy and consistency of the cutting operation.

[0051] For example, assume that the specification of a certain model of power cord is a diameter of 1 millimeter, and the retrieved cutting logs record the data of multiple cutting operations. By calculation, a set of stretch ratios such as {1.02, 1.03, 1.04} is obtained, indicating the stretching degree of the wire in different operations. At the same time, according to the configured weights, a set of cutting errors such as {0.005, 0.007, 0.006} is calculated. These error values reflect the accuracy level of the cutting operation. By analyzing these stretch ratios and cutting errors, the cutting process can be optimized to improve production efficiency and product quality.

[0052] In a preferred embodiment, when the distance between the first monitoring position and the second monitoring position of the power cord meets the target stretching length, and the preset cutting path of the tool belongs to the cutting plane position, an anomaly-free label is applied to the real-time image of the power cord, including: when the distance between the first monitoring position and the second monitoring position of the power cord meets the target stretching length, and the preset cutting path of the tool belongs to the cutting plane position, a target peeling lateral distance is received from the user terminal, and based on the peeling depth calibration database, the wire type and the wire specification are processed to obtain a target peeling radial distance; the cut power cord is conveyed to the insulation layer peeling station, and a radial tool movement image and a lateral tool movement image are extracted from the real-time image of the power cord; the radial tool movement image is processed to obtain a radial cutting-in distance, and the lateral tool movement image is processed to obtain a lateral peeling distance; when the target peeling lateral distance is consistent with the lateral peeling distance, and the radial cutting-in distance is consistent with the target peeling radial distance, an anomaly-free label is applied to the real-time image of the power cord.

[0053] Specifically, in the power cord cutting operation, when the distance between the first monitoring position and the second monitoring position of the power cord meets the target stretching length, and the preset cutting path of the tool is accurately located at the cutting plane position, the system will further perform an anomaly-free label verification on the real-time image of the power cord, especially for the subsequent insulation layer peeling process.

[0054] First, after receiving the target peeling lateral distance at the user terminal, based on the peeling depth calibration database, combined with the wire type and wire specification being processed currently, the target peeling radial distance is calculated. Among them, the peeling depth calibration database mainly contains depth calibration data related to the power cord peeling operation, such as peeling depth data, peeling process parameters, etc. This step ensures the accuracy of the peeling operation because different wire types and specifications may require different peeling parameters. Then, the cut power cord is conveyed to the insulation layer peeling station. At this station, the system extracts a radial tool movement image and a lateral tool movement image from the real-time image of the power cord. By processing these images, the system can obtain a radial cutting-in distance and a lateral peeling distance respectively, and these two parameters are the key indicators for evaluating whether the peeling operation is accurate. Finally, a verification is carried out: when the target peeling lateral distance is consistent with the lateral peeling distance, and the radial cutting-in distance is consistent with the target peeling radial distance, it indicates that the peeling operation is carried out completely according to the preset parameters without deviation. At this time, the system will apply an anomaly-free label to the real-time image of the power cord, indicating that the insulation layer peeling process of this power cord is qualified.

[0055] For example, assume that the target lateral peeling distance of a certain type of power cord is 5 mm, and the target radial peeling distance is 2 mm. At the peeling station, the actual lateral peeling distance obtained by the system through image processing is 5 mm, and the actual radial cutting-in distance is 2 mm. Since these two actual values are exactly the same as the preset values, the system will mark the real-time image of the power cord as normal without any abnormalities, confirming that the insulation layer peeling process is correct. This process ensures the accuracy and consistency of the power cord during the cutting and peeling processes, improving the product quality.

[0056] An image recognition and cutting control method in the production of power cords provided by an embodiment of the present invention has at least the following technical effects:

[0057] 1. By combining multi-dimensional parameters such as wire type, wire specification, cutting error threshold, and zero-tension length of the power cord, the target stretching length is accurately calculated, and the cutting radial plane position is configured accordingly. Before performing the cutting operation, the monitoring positions at both ends of the power cord are accurately obtained using image recognition technology, ensuring high-precision control of the cutting length and position. This way of comprehensively considering various factors and performing accurate calculations significantly improves the accuracy and consistency of cutting, reducing the scrap rate caused by improper cutting.

[0058] 2. By training the clamping part identifier, the positions of the clamping parts at both ends of the power cord can be accurately identified and located, thereby accurately obtaining the monitoring positions of the first end and the second end of the power cord. This not only improves the recognition efficiency but also enhances the recognition accuracy, and can work stably even in a complex working environment. In addition, through image acquisition and training from multiple perspectives (above, first side, second side), the robustness and reliability of the clamping part recognition are further improved.

[0059] 3. After the cutting operation is completed, a verification function for insulation layer peeling is also integrated. By receiving the target lateral peeling distance and calculating the target radial peeling distance in combination with the wire type and specification, the actual peeling distance is verified at the insulation layer peeling station. Only when the target lateral peeling distance is consistent with the actual lateral peeling distance, and the radial cutting-in distance is consistent with the target radial peeling distance, the real-time image of the power cord is marked as normal without any abnormalities. This integrated verification mechanism ensures the continuity and accuracy of the cutting and peeling processes, improving the production efficiency and product quality.

[0060] Embodiment 2:

[0061] As Figure 2 shown, based on the same inventive concept as the image recognition and cutting control method in the production of power cords provided in Embodiment 1, an embodiment of the present invention further provides an image recognition and cutting control system in the production of power cords, and the system includes:

[0062] An information acquisition module 11, configured to receive a preset cutting path of a tool, a wire model, a wire specification, a cutting error threshold, a zero-tension length of a power cord, and a desired cutting length from a user terminal.

[0063] A data retrieval module 12, configured to retrieve a number of stretching ratios and a number of cutting errors with the wire model and the wire specification as constraints, and based on the zero-tension length of the power cord, to statistically calculate a target stretching length that meets the cutting error threshold.

[0064] A cutting configuration module 13, configured to configure a cutting path plane position according to the target stretching length and in combination with the desired cutting length.

[0065] An image recognition module 14, configured to receive a real-time image of a power cord before performing a power cord cutting operation, perform recognition, and obtain a monitoring position of a first end of the power cord and a monitoring position of a second end of the power cord.

[0066] An abnormality identification module 15, configured to, when the distance between the monitoring position of the first end of the power cord and the monitoring position of the second end of the power cord meets the target stretching length and the preset cutting path of the tool belongs to the cutting path plane position, perform a non-abnormality identification on the real-time image of the power cord.

[0067] Furthermore, the image recognition module 14 is further configured to perform the following steps:

[0068] Before performing a power cord cutting operation, fix a first end of the power cord to a first clamping member, and fix the other end of the power cord to a second clamping member; process the real-time image of the power cord through a clamping member identifier to obtain a real-time position of the first clamping member, and set it as the monitoring position of the first end of the power cord; process the real-time image of the power cord through the clamping member identifier to obtain a real-time position of the second clamping member, and set it as the monitoring position of the second end of the power cord.

[0069] Furthermore, the image recognition module 14 is further configured to perform the following steps:

[0070] Collect a dataset of operation image records and a dataset of clamping member spatial position identifiers of a preset model clamping member at a cutting station for power cord production; train a clamping member identifier with the dataset of clamping member spatial position identifiers as supervision and the dataset of operation image records as input.

[0071] Furthermore, the image recognition module 14 is further configured to perform the following steps:

[0072] Any set of job image recording data in the job image recording data set includes job image recording data above the work station and identification record data of the image clamping member above the work station, job image recording data on the first side of the work station and identification record data of the image clamping member on the first side of the work station, and job image recording data on the second side of the work station and identification record data of the image clamping member on the second side of the work station. Using the identification record data of the image clamping member above the work station as supervision and the job image recording data above the work station as input, a convolutional neural network is trained to obtain a first clamping member identification branch. Using the identification record data of the image clamping member on the first side of the work station as supervision and the job image recording data on the first side of the work station as input, a convolutional neural network is trained to obtain a second clamping member identification branch. Using the identification record data of the image clamping member on the second side of the work station as supervision and the job image recording data on the second side of the work station as input, a convolutional neural network is trained to obtain a third clamping member identification branch. The output layers of the first clamping member identification branch, the second clamping member identification branch, and the third clamping member identification branch are respectively merged with three input nodes of the fully connected layer to obtain a clamping member recognizer architecture. Using the clamping member spatial position identification data set as supervision and the job image recording data set as input, the clamping member recognizer architecture is trained to obtain the clamping member recognizer.

[0073] Furthermore, the data retrieval module 12 is further configured to perform the following steps:

[0074] Based on the several cropping errors, a set of target stretching ratios less than or equal to the cropping error threshold is extracted from the several stretching ratios. Based on the zero-tension length of the power cord, the set of target stretching ratios is traversed to obtain an initial set of target stretching lengths. Central tendency analysis is performed on the initial set of target stretching lengths to obtain a stretching length distribution interval, which is set as the target stretching length.

[0075] Furthermore, the data retrieval module 12 is further configured to perform the following steps:

[0076] Constrained by the wire model and the wire specification, retrieve several wire cutting logs. Any one of the several wire cutting logs includes a zero-tension length record value of the power cord, a stretched length record value of the power cord, an included angle record value between the cutting section and the preset diameter plane, and a cutting length deviation record value. Traverse the several wire cutting logs, calculate the ratio of the zero-tension length record value of the power cord to the stretched length record value of the power cord to obtain the several stretch ratios. Configure a first weight for the normalized parameter of the included angle record value between the cutting section and the preset diameter plane, and configure a second weight for the normalized parameter of the cutting length deviation record value. According to the first weight and the second weight, traverse the several wire cutting logs and perform a weighted mean calculation on the normalized values of the included angle record value between the cutting section and the preset diameter plane and the cutting length deviation record value to obtain the several cutting errors.

[0077] Furthermore, the anomaly identification module 15 is further configured to perform the following steps:

[0078] When the distance between the first monitoring position of the power cord and the second monitoring position of the power cord meets the target stretched length, and the preset cutting path of the tool belongs to the cutting diameter plane position, receive a target peeling lateral distance from the user terminal, and process the wire model and the wire specification based on the peeling depth calibration database to obtain a target peeling radial distance. Transport the cut power cord to the insulation layer peeling station, and extract a radial tool movement image and a lateral tool movement image from the real-time image of the power cord. Process the radial tool movement image to obtain a radial cutting-in distance, and process the lateral tool movement image to obtain a lateral peeling distance. When the target peeling lateral distance is consistent with the lateral peeling distance, and the radial cutting-in distance is consistent with the target peeling radial distance, no anomaly is identified for the real-time image of the power cord.

[0079] Through the foregoing detailed description of an image recognition and cutting control method in the production of a power cord in this specification, those skilled in the art can clearly know an image recognition and cutting control system in the production of a power cord in this embodiment. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the description of the method part.

[0080] Embodiment III:

[0081] Please refer to Figure 3 , Figure 3 which is a schematic diagram of an embodiment of an electronic device provided by an embodiment of the present invention. As Figure 3As shown in the figure, an embodiment of the present invention provides an electronic device 500, which includes a memory 510, a processor 520, and a first computer program 511 stored on the memory 510 and executable on the processor 520. When the processor 520 executes the first computer program 511, it implements an image recognition and cropping control method in the production of a power cord as described in Embodiment 1.

[0082] It should be noted that in the above embodiments, the descriptions of each embodiment have their own emphases. For parts not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0083] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects.

[0084] The present invention is described with reference to the flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to the embodiments of the present invention. It should be understood that each flow and / or block in the flowchart and / or block diagram can be implemented by computer program instructions, and the combination of the flows and / or blocks in the flowchart and / or block diagram can also be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded computer, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate a device for implementing the functions specified in one Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0085] These computer program instructions can also be loaded onto a computer or other programmable data processing devices, so that a series of operation steps are executed on the computer or other programmable devices to generate a computer-implemented process. Thus, the instructions executed on the computer or other programmable devices provide steps for implementing the functions specified in one Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0086] Although the preferred embodiments of the present invention have been described, those skilled in the art can make additional changes and modifications to these embodiments once they know the basic creative concepts.

[0087] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the present invention and its equivalent technologies, the present invention is also intended to include these modifications and variations.

Claims

1. An image recognition and cutting control method in the production of power cords, characterized in that, Including: Receiving, from a user side, a preset cutting path of a tool, a wire model and wire specifications, a cutting error threshold, a zero-tension length of a power cord, and a desired cutting length; Constrained by the wire model and wire specifications, retrieving a number of stretching ratios and a number of cutting errors, and based on the zero-tension length of the power cord, counting a target stretching length that meets the cutting error threshold; Configuring a cutting plane position according to the target stretching length and in combination with the desired cutting length; Before performing a power cord cutting operation, receiving a real-time image of the power cord, performing recognition, and obtaining a monitored position of the first end of the power cord and a monitored position of the second end of the power cord; When the distance between the monitored position of the first end of the power cord and the monitored position of the second end of the power cord meets the target stretching length, and the preset cutting path of the tool belongs to the cutting plane position, making a non-abnormality mark on the real-time image of the power cord.

2. The image recognition and cutting control method in the production of a power cord according to claim 1, characterized in that, Before performing a power cord cutting operation, receiving a real-time image of the power cord, performing recognition, and obtaining a monitored position of the first end of the power cord and a monitored position of the second end of the power cord, including: Before performing a power cord cutting operation, fixing the first end of the power cord to a first clamping member and fixing the other end of the power cord to a second clamping member; Processing the real-time image of the power cord through a clamping member identifier to obtain a real-time position of the first clamping member, and setting it as the monitored position of the first end of the power cord; Processing the real-time image of the power cord through a clamping member identifier to obtain a real-time position of the second clamping member, and setting it as the monitored position of the second end of the power cord.

3. The image recognition and cutting control method in the production of a power cord according to claim 2, characterized in that, Processing the real-time image of the power cord through a clamping member identifier to obtain a real-time position of the first clamping member, and setting it as the monitored position of the first end of the power cord, including: Collecting a dataset of operation image records and a dataset of clamping member spatial position identifiers of a preset model clamping member at a cutting station where the power cord is produced; Training a clamping member identifier with the dataset of clamping member spatial position identifiers as supervision and the dataset of operation image records as input.

4. The image recognition and cutting control method in the production of a power cord according to claim 3, characterized in that, Training a clamping member identifier with the dataset of clamping member spatial position identifiers as supervision and the dataset of operation image records as input, including: Any set of operation image record data in the dataset of operation image records includes operation image record data above the station and image clamping member identifier record data above the station, operation image record data on the first side of the station and image clamping member identifier record data on the first side of the station, and operation image record data on the second side of the station and image clamping member identifier record data on the second side of the station; Training a convolutional neural network with the image clamping member identifier record data above the station as supervision and the operation image record data above the station as input to obtain a first clamping member identifier branch; Training a convolutional neural network with the image clamping member identifier record data on the first side of the station as supervision and the operation image record data on the first side of the station as input to obtain a second clamping member identifier branch; Training a convolutional neural network with the image clamping member identifier record data on the second side of the station as supervision and the operation image record data on the second side of the station as input to obtain a third clamping member identifier branch; Merge the output layers of the first clamping part identification branch, the second clamping part identification branch, and the third clamping part identification branch with the three input nodes of the fully connected layer respectively to obtain the clamping part identifier architecture; Use the clamping part spatial position identification data set as supervision and the operation image recording data set as input to train the clamping part identifier architecture to obtain the clamping part identifier.

5. The image recognition and cutting control method in the production of a power cord according to claim 1, characterized in that, With the wire model and wire specification as constraints, retrieve a number of stretching ratios and a number of cutting errors. Based on the zero-tension length of the power cord, count the target stretching lengths that meet the cutting error threshold, including: Based on the number of cutting errors, extract a set of target stretching ratios less than or equal to the cutting error threshold from the number of stretching ratios; Based on the zero-tension length of the power cord, traverse the set of target stretching ratios to find the ratio and obtain the initial set of target stretching lengths; Perform a central tendency analysis on the initial set of target stretching lengths to obtain the stretching length distribution interval, which is set as the target stretching length.

6. The image recognition and cutting control method in the production of a power cord according to claim 5, wherein, With the wire model and wire specification as constraints, retrieve a number of stretching ratios and a number of cutting errors, including: With the wire model and the wire specification as constraints, retrieve a number of wire cutting logs, where any one of the number of wire cutting logs includes a zero-tension length record value of the power cord, a stretching length record value of the power cord, an included angle record value between the cutting section and the preset diameter plane, and a cutting length deviation record value; Traverse the number of wire cutting logs and calculate the ratio of the zero-tension length record value of the power cord to the stretching length record value of the power cord to obtain the number of stretching ratios; Configure a first weight for the normalization parameter of the included angle record value between the cutting section and the preset diameter plane, and configure a second weight for the normalization parameter of the cutting length deviation record value; According to the first weight and the second weight, traverse the number of wire cutting logs and perform a weighted mean calculation on the normalization values of the included angle record value between the cutting section and the preset diameter plane and the cutting length deviation record value to obtain the number of cutting errors.

7. The image recognition and cutting control method in the production of a power cord according to claim 1, characterized in that, When the distance between the first monitoring position of the power cord and the second monitoring position of the power cord meets the target stretching length and the preset cutting path of the tool belongs to the cutting diameter plane position, perform an anomaly-free identification on the real-time image of the power cord, including: When the distance between the first monitoring position of the power cord and the second monitoring position of the power cord meets the target stretching length and the preset cutting path of the tool belongs to the cutting diameter plane position, receive the target peeling lateral distance from the user side, and process the wire model and the wire specification based on the peeling depth calibration database to obtain the target peeling radial distance; Convey the cut power cord to the insulation layer peeling station, and extract the radial tool movement image and the lateral tool movement image from the real-time image of the power cord; Process the radial tool movement image to obtain the radial cutting-in distance, and process the lateral tool movement image to obtain the lateral peeling distance; When the target peeling lateral distance is consistent with the lateral peeling distance, and the radial cutting-in distance is consistent with the target peeling radial distance, an abnormality-free identification is performed on the real-time image of the power cord.

8. An image recognition and cutting control system in the production of power cords, characterized in that, A system for implementing the image recognition and cutting control method in the production of a power cord according to any one of claims 1-7, the system comprising: An information acquisition module, configured to receive, from a user terminal, a preset cutting path of a tool, a wire model and a wire specification, a cutting error threshold, a zero-tension length of the power cord, and a desired cutting length; A data retrieval module, configured to retrieve a plurality of stretching ratios and a plurality of cutting errors with the wire model and the wire specification as constraints, and based on the zero-tension length of the power cord, count a target stretching length that meets the cutting error threshold; A cutting configuration module, configured to configure a cutting radial plane position according to the target stretching length and in combination with the desired cutting length; An image recognition module, configured to receive a real-time image of the power cord before performing a power cord cutting operation, perform recognition, and obtain a monitoring position of a first end of the power cord and a monitoring position of a second end of the power cord; An abnormality identification module, configured to perform an abnormality-free identification on the real-time image of the power cord when the distance between the monitoring position of the first end of the power cord and the monitoring position of the second end of the power cord meets the target stretching length, and the preset cutting path of the tool belongs to the cutting radial plane position; 9. An electronic device, characterized in that, Comprising: A memory, configured to store a computer software program; A processor, configured to read and execute the computer software program, thereby implementing the image recognition and cutting control method in the production of a power cord according to any one of claims 1-7.

Images