Cable fixing method, system and device based on machine vision

Abstract

The invention provides a cable fixing method, system and device based on machine vision, and the method comprises the steps: obtaining cable fixing parameters in a cable fixing process, and collecting a cable fixing image; performing content identification on the cable fixing image by adopting machine vision, performing identification to obtain a cable gap parameter, and performing fixed cable damage analysis to obtain a first cable damage parameter; environment parameters are collected, cable shaking damage analysis is carried out in combination with the cable gap parameters, cable shaking parameters are obtained, shaking cable damage analysis is carried out, and second cable damage parameters are obtained; according to the first cable damage parameter and the second cable damage parameter, calculating to obtain a cable damage parameter, combining the cable shaking parameter, calculating to obtain a cable fixing score, adjusting and optimizing the cable fixing parameter, and obtaining an optimal cable fixing result. The technical problem of low cable fixing quality caused by an extensive cable fixing method in the prior art is solved.

Description

Technical Field

[0001] The present invention relates to the field of machine vision, and in particular to a cable fixing method, system and device for machine vision. Background Art

[0002] The quality of cable fixing directly determines the performance of the cable. Traditional cable fixing methods lack quantitative fixing parameters and rely entirely on manual experience and judgment, resulting in obvious deviations in 40% of installations: when the fixing is too tight, the insulation layer will be subjected to abnormal pressure exceeding 25N / mm², causing damage to the insulation layer on the cable surface, thereby affecting the cable performance; when the fixing is too loose, the cable will be shaken by wind, causing the insulation layer on the cable surface to wear, thereby affecting the cable performance. Therefore, there is a technical problem in the prior art that the extensive cable fixing method leads to low cable fixing quality. Summary of the invention

[0003] The present invention aims to solve the technical problem that the extensive cable fixing method in the prior art leads to low cable fixing quality, and provides a machine vision cable fixing method, system and device.

[0004] The technical solution of the present invention to solve the above technical problems is as follows:

[0005] In a first aspect, the present invention provides a machine vision cable fixing method, comprising:

[0006] During the process of cable fixing, cable fixing parameters are obtained and cable fixing images are collected;

[0007] Using machine vision to perform content recognition on the cable fixed image, identify and obtain cable gap parameters, perform fixed cable damage analysis, and obtain first cable damage parameters;

[0008] Collecting environmental parameters, performing cable shaking damage analysis in combination with the cable gap parameters to obtain cable shaking parameters, and performing shaking cable damage analysis to obtain second cable damage parameters;

[0009] The cable damage parameter is calculated based on the first cable damage parameter and the second cable damage parameter. The cable fixation score is calculated based on the cable sway parameter. The cable fixation parameters are adjusted and optimized to obtain the optimal cable fixation result.

[0010] In a second aspect, the present invention provides a machine vision system, comprising:

[0011] The data acquisition module is used to obtain the cable fixing parameters and the cable fixing images during the cable fixing process;

[0012] A fixed damage analysis module, used to use machine vision to perform content recognition on the cable fixed image, identify and obtain cable gap parameters, perform fixed cable damage analysis, and obtain first cable damage parameters;

[0013] A sway damage analysis module, used to collect environmental parameters, perform cable sway damage analysis in combination with the cable gap parameters, obtain cable sway parameters, and perform sway cable damage classification to obtain second cable damage parameters;

[0014] The optimization output module is used to calculate the cable damage parameter based on the first cable damage parameter and the second cable damage parameter, calculate the cable fixation score in combination with the cable shaking parameter, adjust and optimize the cable fixation parameters, and obtain the optimal cable fixation result.

[0015] In a third aspect, the present invention provides a machine vision cable fixing device, which includes: a memory for storing a computer software program; a processor for reading and executing the computer software program, thereby implementing a machine vision cable fixing method.

[0016] The beneficial effects of the present invention are:

[0017] The present application firstly obtains the cable fixing parameters and collects the cable fixing images during the cable fixing process, thereby providing reliable data support for the subsequent optimization of the cable fixing parameters; secondly, the machine vision is used to perform content recognition on the cable fixing images, and the cable gap parameters are identified and obtained, and a fixed cable damage analysis is performed to obtain the first cable damage parameters, and a correlation between the cable gap parameters and the insulation layer damage is established; then, the environmental parameters are collected, and a cable shaking damage analysis is performed in combination with the cable gap parameters to obtain the cable shaking parameters, and a shaking cable damage analysis is performed to obtain the second cable damage parameters, and a correlation between the environmental wind parameters and the insulation layer damage is established; finally, the cable damage parameters are calculated based on the first cable damage parameters and the second cable damage parameters, and the cable fixing score is calculated in combination with the cable shaking parameters, and the cable fixing parameters are adjusted and optimized to obtain the optimal cable fixing result.

[0018] Through the above technical solution, the present application first collects the cable fixing parameters and corresponding images during the cable fixing process, and then obtains the first cable damage parameter and the second cable damage parameter based on machine vision technology, establishes the relationship between the cable gap parameter, the environmental wind dynamic parameter and the insulation layer damage, and finally comprehensively considers the damage of the cable insulation layer and the influence of the stability of the electrical connection on the cable fixing quality, and continuously optimizes the cable fixing parameters until the optimal cable fixing parameters are output. In this way, the cable fixing quality is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 A schematic flow chart of a machine vision cable fixing method provided by the present invention;

[0020] Figure 2 A schematic structural diagram of a machine vision cable fixing system provided by the present invention;

[0021] Figure 3 A schematic structural diagram of a machine vision cable fixing device provided by the present invention.

[0022] In the accompanying drawings, the components represented by the reference numerals are as follows:

[0023] Data acquisition module 11, fixed damage analysis module 12, shaking damage analysis module 13, optimization output module 14, cable fixing device 200, memory 210, processor 220, computer program 211. DETAILED DESCRIPTION

[0024] 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 those skilled in the art without creative work are within the scope of protection of the present invention.

[0025] In the description of the present invention, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0026] In the description of the present invention, the term "for example" is used to mean "used as an example, illustration or explanation". Any embodiment described as "for example" in the present invention is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is given to enable any technician in the field to implement and use the present invention. In the following description, details are listed for the purpose of explanation. It should be understood that a person of ordinary skill in the art can recognize that the present invention can be implemented without using these specific details. In other examples, well-known structures and processes will not be 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 is consistent with the widest scope consistent with the principles and features disclosed in the present invention.

[0027] Embodiment 1, as Figure 1As shown, an embodiment of the present invention provides a machine vision cable fixing method, comprising:

[0028] S10: in the process of cable fixing, obtaining cable fixing parameters and collecting cable fixing images;

[0029] Traditional cable fixing operations are mainly completed by manually operating fixing equipment such as cable clamps and cable ties. However, due to the lack of standardized fixing parameters (such as the travel distance of the bolt rotation in the cable clamp, the gap distance between the cable tie and the cable, etc.), the traditional method completely relies on manual skills and experience, resulting in low cable fixing quality.

[0030] In view of the above problems, the present application obtains cable fixing parameters during the process of fixing the cable, and obtains corresponding cable fixing images based on the cable fixing parameters, so as to provide necessary data support for subsequent optimization of the fixing parameters.

[0031] Specifically, step S10 in the method includes:

[0032] In the process of cable fixing, obtaining cable fixing parameters, wherein the cable fixing parameters include fixing dimensions;

[0033] An image after the cable is fixed according to the cable fixing parameters is collected to obtain a cable fixing image.

[0034] In the embodiment of the present application, the cable fixing parameters in the cable fixing process are first obtained, wherein the cable fixing parameters include fixing dimensions. Specifically, the cable is mainly fixed by fixing devices such as cable clamps and cable ties, and the fixing dimensions in the cable fixing process (such as the travel distance of the bolt rotation in the cable clamp, the gap distance between the cable tie and the cable, etc.) are collected as cable fixing parameters. For example, when the cable is fixed by a cable clamp, the travel distance of the bolt rotation in the cable clamp during the fixing process is collected as 4.5 mm, which is used as the fixing parameter of the cable.

[0035] Secondly, an image after the cable is fixed according to the cable fixing parameters is collected to obtain a cable fixing image. Exemplarily, the cable fixing image corresponding to the fixing parameters is captured by a 20-megapixel industrial camera (such as Basler ace).

[0036] In summary, the present application first collects the cable fixing parameters during the cable fixing process, and then collects the cable fixing images corresponding to the cable fixing parameters, so as to provide reliable data support for the subsequent optimization of the cable fixing parameters.

[0037] S20: using machine vision to perform content recognition on the cable fixed image, identify and obtain cable gap parameters, perform fixed cable damage analysis, and obtain first cable damage parameters;

[0038] During the cable fixing process, if the cable is fixed too tightly, the cable insulation layer may be damaged. For example, when the pressure applied by the fixing equipment exceeds the yield strength of the material, it will cause local thinning of the insulation layer. During long-term operation, a radial crack network will form at the location where mechanical stress is concentrated, thereby affecting the cable performance and shortening the cable service life.

[0039] In response to the above problems, the present application is based on machine vision, and establishes an association between cable fixing parameters and cable damage by analyzing the relationship between cable gap parameters and cable insulation layer damage size. The cable gap parameters are then input into the first cable damage classifier to obtain the first cable damage parameters as output.

[0040] Specifically, step S20 in the method includes:

[0041] Preprocessing the cable fixing image;

[0042] The preprocessed cable fixing image is input into a pretrained cable fixing identifier to identify and output cable gap parameters, wherein the cable gap parameters include the gap distance after the cable is fixed. The cable fixing identifier is constructed using a convolutional neural network and is trained using a sample cable fixing image set and a sample cable gap parameter set.

[0043] In the embodiment of the present application, the collected cable fixing image is first preprocessed. For example, Gaussian filtering and median filtering are used to eliminate image noise; adaptive illumination equalization is performed through the CLAHE algorithm to solve the problem of uneven illumination; the cable contour features are enhanced through Canny edge detection and Sobel operator; key areas are extracted through morphological operations, etc. The preprocessed image is geometrically standardized and normalized, and finally converted into a tensor format suitable for convolutional neural network input. In this way, high-quality input data is provided for the subsequent cable gap parameter recognition output.

[0044] Secondly, the preprocessed cable fixing image is input into the trained cable fixing identifier, and the gap distance after the cable is fixed is output as the cable gap parameter. Among them, the cable gap parameter is a quantifiable parameter such as the maximum distance of the gap between the cable and the fixing device after fixing. The cable fixing identifier is constructed by using a convolutional neural network (CNN), and then trained by using a sample cable fixing image set and a sample cable gap parameter set. Furthermore, the construction and training process of the cable fixing identifier is as follows: 1. Collect training data sets: collect preprocessed cable fixing images as sample cable fixing image sets, manually annotate the cable gap parameters of the fixed images (annotation tools such as LabelImg), as sample cable gap parameter sets, and then divide the data sets into training sets, validation sets, and test sets according to the ratio of 7.5:1.5:1.5. 2. Construct a model: The cable fixing identifier is constructed by using a multi-level convolutional neural network (CNN), which includes 5 convolutional layers, 3 pooling layers, and 2 fully connected layers. ReLU activation function and batch normalization are used to achieve efficient feature extraction. 3. Model training: The transfer learning method is used, with the pre-trained ResNet50 as the basic model. A data set containing 5,000 labeled samples is used for fine-tuning training. The model parameters are optimized through the Adam optimizer and the cosine annealing learning rate scheduling strategy, and finally a recognition accuracy of 95% is achieved on the test set.

[0045] For example, 5,000 cable fixing images are collected, denoised, enhanced and standardized as sample cable fixing image sets, and the gap parameters of each image are manually labeled through the LabelImg tool as sample cable gap parameter sets. The sample cable fixing image set and the sample cable gap parameter set are divided into training set, validation set and test set according to the ratio of 7.5:1.5:1.5. Then the cable fixing identifier built based on convolutional neural network (CNN) is trained and optimized until the output result reaches 95% recognition accuracy on the test set, which is regarded as model convergence. After that, the preprocessed cable fixing image is input into the trained cable fixing identifier, and the maximum distance (such as 2mm) between the cable and the fixing device after fixing can be output as the cable gap parameter.

[0046] Furthermore, the “performing a fixed cable damage analysis to obtain a first cable damage parameter” includes:

[0047] According to the cable fixing data in the historical time, a set of sample cable gap parameters is collected, and the size of the cable insulation layer damage under different sample cable gap parameters is collected and marked as the sample first cable damage parameter, so as to obtain the sample first cable damage parameter set;

[0048] Constructing an index relationship between the sample cable clearance parameter set and the sample first cable damage parameter set to obtain a first cable damage classifier;

[0049] The cable gap parameter is input into the first cable damage classifier, and the index classification output is used to obtain the first cable damage parameter.

[0050] In the embodiment of the present application, firstly, according to the cable fixing data in the historical time, a set of sample cable gap parameters (such as the maximum distance of the gap between the cable and the fixing device after fixing, mm) is collected, and the size of the cable insulation layer damage under different sample cable gap parameters (such as the length direction size of the insulation layer damage, the thickness direction size of the insulation layer damage, mm) is collected, and marked as the sample first cable damage parameter, and the sample first cable damage parameter set is obtained. Exemplarily, a certain cable is fixed with a cable tie, and the maximum distance of the gap between the cable and the cable tie is collected by a digital micrometer, which is 2.1mm, 1.9mm, 1.6mm, 2mm, and 1.3mm, respectively, as the sample cable gap parameter set, and then the thickness direction size of the cable insulation layer damage corresponding to different gap parameters is collected by an ultrasonic thickness gauge, which is 0.8mm, 0.76mm, 0.6mm, 0.77mm, and 0.4mm, respectively, and marked as the sample first cable damage parameter, and the sample first cable damage parameter set is obtained.

[0051] Secondly, construct an index relationship between the sample cable gap parameter set and the sample first cable damage parameter set to obtain the first cable damage classifier. Exemplarily, establish a one-to-one index relationship between the aforementioned sample cable gap parameter set [2.1mm, 1.9mm, 1.6mm, 2mm, 1.3mm] and the sample first cable damage parameter set [0.8mm, 0.76mm, 0.6mm, 0.77mm, 0.4mm]. Furthermore, the construction and training process of the first cable damage classifier can be achieved through the following technical paths: 1. Prepare a training data set: prepare 5000 groups of corresponding sample cable gap parameter sets and sample first cable damage parameter sets, and divide them into training set, validation set, and test set in a ratio of 7.5:1.5:1.5. 2. Model construction: The mixed expert model (MoE) architecture is adopted, which mainly includes feature extraction module, spatial attention module, and decision fusion module. Among them, the feature extraction module uses 1D-CNN to process time series signals (sampling rate 1kHz) and cooperates with the GRU network to capture long-term dependencies; the spatial attention module strengthens key position features through the CBAM mechanism; the decision fusion module integrates the dual model outputs of XGBoost (max_depth=6) and LightGBM (num_leaves=31). 3. Model training: The training process adopts a curriculum learning strategy. First, the model is trained through the training set, and then the difficult samples are expanded through the adversarial generative network (GAN) for fine-tuning. The Bayesian hyperparameter search is introduced in the model optimization stage. Finally, the accuracy rate of 95% is achieved on the test set, which is regarded as model convergence.

[0052] Finally, the cable gap parameter is input into the trained first cable damage classifier, and the index classification output is used to obtain the first cable damage parameter. For example, the cable gap parameter of 2.2 mm is input into the trained first cable damage classifier, and the output is used to obtain the first cable damage parameter of 0.32 mm.

[0053] In summary, compared with the prior art, this application uses machine vision to perform content recognition on cable fixed images, identifies and obtains cable gap parameters, and uses the first cable damage classifier to perform fixed cable damage analysis on different cable gap parameters to obtain the first cable damage parameter. In this way, the association between the cable gap parameter and the insulation layer damage is established, providing reliable data support for the subsequent optimization of fixed parameters.

[0054] S30: collecting environmental parameters, performing cable shaking damage analysis in combination with the cable gap parameters to obtain cable shaking parameters, and performing shaking cable damage analysis to obtain second cable damage parameters;

[0055] During the cable fixing process, if the fixing is too loose, the cable will swing constantly under the influence of wind force, resulting in friction, which will further damage the cable insulation layer. Specifically, the continuous reciprocating movement between the cable surface and the fixing device will cause abrasive wear of the insulation layer, affecting the cable performance and shortening the service life of the cable.

[0056] To address the above problems, this application collects the pneumatic parameters in the cable fixing environment as environmental parameters, then inputs the environmental parameters into a cable sway predictor constructed based on a feedforward neural network to output cable sway parameters, and then inputs the cable sway parameters into a second cable damage classifier to output and obtain second cable damage parameters.

[0057] Specifically, step S30 in the method includes:

[0058] Collect the pneumatic parameters in the cable fixing environment as environmental parameters;

[0059] Input the cable gap parameters and environmental parameters into a pre-trained cable sway predictor to predict and output cable sway parameters. Among them, the cable sway parameters include the sway distance of the cable per unit time, and the cable sway predictor is constructed based on a feedforward neural network and trained using a sample cable gap parameter set, a sample environmental parameter set, and a sample cable sway parameter set.

[0060] In the embodiment of this application, first, the pneumatic parameters in the cable fixing environment are collected as environmental parameters. Among them, the environmental parameters such as the average environmental wind force, wind direction, eddy current, etc. are used as the main environmental factors affecting cable sway damage. Exemplarily, the average wind force in the cable fixing environment is collected as 1.8 m / s as the environmental parameter.

[0061] Secondly, the cable clearance parameters and environmental parameters are input into the trained cable sway predictor, and the cable sway distance per unit time is output as the cable sway parameter. The cable sway parameter is the cable sway distance per unit time, such as mm / s, cm / min. The cable sway predictor is constructed using a feedforward neural network, and then trained using a sample cable clearance parameter set, a sample environmental parameter set, and a sample cable sway parameter set. Furthermore, the construction and training process of the cable sway predictor is as follows: 1. Collect training data sets: collect wind parameters under different wind conditions (such as the average wind force in the environment, m / s) as sample environmental parameter sets, then collect cable gap data under different fixed states under corresponding wind conditions (such as the maximum distance of the gap between the cable and the fixed equipment after fixing, mm) as sample cable gap parameter sets, and finally collect cable sway distance under corresponding wind conditions and corresponding cable gap distance conditions (such as the actual measured cable sway distance per unit time, mm / s) as sample cable sway parameter sets, and then divide the data set into training set, validation set, and test set according to the ratio of 7.5:1.5:1.5. 2. Model construction: The cable sway predictor is constructed using a feedforward neural network with physical constraints. The input layer receives a 15-dimensional feature vector, contains 3 hidden layers of 128 nodes (Dropout rate 0.3) and a linear output layer, and is optimized by a mixed loss function (L1+L2+physical constraint term). 3. Model training: The training process adopts a transfer learning strategy, first pre-training based on the sample training set, then fine-tuning with the validation set data, and finally achieving a recognition accuracy of 95% on the test set.

[0062] Exemplarily, the average wind force under different environments (such as 2.5m / s, 3.0m / s, 3.5m / s, 4.0m / s, 4.5m / s) is collected as a sample environmental parameter set, and then the cable gap data of different fixed states under the corresponding environmental parameters are collected (such as 2.5m / s environmental parameters, cable gap data: 2.1mm, 1.9mm, 1.6mm, 2mm, 1.3mm) as a sample cable gap parameter set, and finally the cable sway distance per unit time under environmental parameter conditions and corresponding cable gap distance conditions (such as 2.5m / s environmental parameters, cable gap 2.1mm conditions, the sway distance per unit time is 1.6mm / s) as a sample cable sway parameter set, and the data set is divided into a training set, a validation set, and a test set according to a ratio of 7.5:1.5:1.5. Then, the cable sway predictor is trained and optimized until the output data reaches a recognition accuracy of 95% on the test set, which is considered to be model convergence. Finally, the cable gap parameters (such as 2.1mm) and environmental parameters (such as 2.5m / s) are input into the trained cable sway predictor, and the cable sway distance per unit time (such as 1.6mm / s) can be output as the cable sway parameters.

[0063] Furthermore, the “performing a shaking cable damage analysis to obtain a second cable damage parameter” includes:

[0064] According to the cable monitoring data in the historical time, a set of sample cable shaking parameters is collected, and the size of the cable insulation layer damage within a preset time range under different sample cable shaking parameters is collected, marked as the sample second cable damage parameter, and the sample second cable damage parameter set is obtained;

[0065] Constructing an index relationship between the sample cable sway parameter set and the sample second cable damage parameter set to obtain a second cable damage classifier;

[0066] The cable shaking parameters are input into the second cable damage classifier, and the index classification output is used to obtain the second cable damage parameters.

[0067] In the embodiment of the present application, firstly, according to the cable monitoring data in the historical time, the sample cable shaking parameter set (the shaking distance of the cable in the unit time, mm / s, cm / min) is collected, and the size of the cable insulation layer damage within the preset time range under different sample cable shaking parameters (such as the length direction size of the insulation layer damage, the thickness direction size of the insulation layer damage, mm) is collected, and marked as the sample second cable damage parameter, and the sample second cable damage parameter set is obtained. Among them, the preset time is a time range determined by technicians in this field according to actual conditions, such as 1 month, 6 months, 12 months, etc. For example, a certain cable is laid outdoors, and its shaking parameters (1.6mm / s, 1.4mm / s, 1.2mm / s, 1.0mm / s, 0.8mm / s) are collected, and then the thickness direction size of the insulation layer damage under different cable shaking parameters (such as 0.21mm, 0.19mm, 0.16mm, 0.15mm, 0.12mm) is collected, marked as the sample second cable damage parameter, and the sample second cable damage parameter set is obtained.

[0068] Secondly, construct an index relationship between the sample cable sway parameter set and the sample second cable damage parameter set to obtain the second cable damage classifier. Exemplarily, a one-to-one corresponding index relationship is established between the aforementioned sample cable sway parameter set [1.6mm / s, 1.4mm / s, 1.2mm / s, 1.0mm / s, 0.8mm / s] and the sample second cable damage parameter set [0.21mm, 0.19mm, 0.16mm, 0.15mm, 0.12mm]. Furthermore, the construction and training process of the second cable damage classifier can be achieved through the following technical paths: 1. Prepare a training data set: prepare a sample cable sway parameter set and a sample second cable damage parameter set, and divide them into a training set, a validation set, and a test set in a ratio of 7.5:1.5:1.5. 2. Model construction: Adopting the dual-track design concept of "physical mechanism driven + data driven", high-precision damage classification is achieved through a three-layer cascaded LSTM-Attention hybrid neural network. The model input layer receives the feature vector (including the shaking parameters), extracts the timing features through the bidirectional LSTM layer, dynamically focuses on the key vibration period through the self-attention mechanism, and finally outputs the result (second cable damage parameter) through the global maximum pooling and regularized fully connected layer. 3. Model training: The training process adopts a curriculum learning strategy. First, the model is trained with the training set, and then the difficult samples are expanded and fine-tuned through the adversarial generative network (GAN). Bayesian hyperparameter search is introduced in the model optimization stage, and finally an accuracy of 95% is achieved on the test set, which is considered as model convergence.

[0069] Finally, input the cable shaking parameters into the second cable damage classifier, and obtain the second cable damage parameters through index classification output. Exemplarily, input the cable shaking parameter of 1.6 mm / s into the trained second cable damage classifier, and the output second cable damage parameter is 0.21 mm.

[0070] In summary, compared with the prior art, the present application analyzes the cable shaking damage by collecting environmental parameters and combining the cable clearance parameters, obtains the cable shaking parameters, and analyzes the damage of the shaking cable to obtain the second cable damage parameters. In this way, the correlation between the environmental wind parameters and the insulation layer damage condition is established, providing reliable data support for the optimization of subsequent fixed parameters.

[0071] S40: Calculate the cable damage parameters based on the first cable damage parameters and the second cable damage parameters, calculate the cable fixing score in combination with the cable shaking parameters, adjust and optimize the cable fixing parameters, and obtain the optimal cable fixing result.

[0072] The foregoing first cable damage parameters and second cable damage parameters can reflect the relationship between the cable fixing parameters and the insulation layer damage condition. By adjusting the cable fixing parameters, the damage condition of the insulation layer can be improved, and the cable fixing quality can be enhanced. Further, in the natural environment, in addition to being affected by the damage condition of the surface insulation layer, the stability of the electrical connection is also an important influencing factor. Specifically, the cable shaking with the wind will affect the stability of the electrical connection, thereby affecting the cable fixing quality.

[0073] To address the above problems, the present application, based on the foregoing first cable damage parameters and second cable damage parameters, simultaneously considers the influence of the damage condition of the cable insulation layer and the stability of the electrical connection on its fixing quality, and continuously optimizes the cable fixing parameters by calculating the cable fixing score until the optimal cable fixing parameters are output.

[0074] Specifically, step S40 in the method includes:

[0075] Calculate the cable damage parameters based on the first cable damage parameters and the second cable damage parameters;

[0076] Calculate the cable fixing score based on the cable damage parameters and the cable shaking parameters, as follows:

[0077] ;

[0078] where FIHG is the cable fixing score, and are weights, is the cable damage parameter, The preset size of the cable insulation layer, D is the cable sway parameter, To preset the cable shaking parameters;

[0079] The cable fixing parameters are adjusted to obtain adjusted cable fixing parameters, and analysis and calculation are performed to obtain a cable fixing score;

[0080] Iteratively adjust and optimize the cable fixing parameters until convergence, output the optimal cable fixing parameters with the largest cable fixing score, fix the cable, and obtain the optimal cable fixing result.

[0081] In the embodiment of the present application, the cable damage parameter is first calculated based on the aforementioned first cable damage parameter and second cable damage parameter, wherein the cable damage parameter = first cable damage parameter + second cable damage parameter. For example, the first cable damage parameter is 0.13 mm, the second cable damage parameter is 0.09 mm, then the cable damage parameter = 0.13 mm + 0.09 mm = 0.22 mm.

[0082] Secondly, according to the cable damage parameters and the cable shaking parameters, the cable fixing score is calculated by the following formula:

[0083] ;

[0084] Among them, FIHG is the cable fixing score. The higher the fixing score, the higher the cable fixing quality. 1 and w 2 is the weight (w 1 +w 2 = 1), and w is determined by technicians in this field according to actual conditions. 1 and w 2 The initial value of w 1 =0.6, w 2 =0.4. C s is the cable damage parameter, C s The smaller it is (indicating less damage to the cable insulation layer), the greater the cable fixation score. y The preset size of the cable insulation layer, that is, the initial thickness of the cable insulation layer. For example, the standard insulation thickness of the medium voltage cable is 4.5mm. D is the cable sway parameter. The smaller D is (the smaller the sway, the more stable the electrical connection), the greater the cable fixation score. D y is the preset cable shaking parameter, that is, the maximum allowable cable shaking parameter. For example, if the maximum allowable cable shaking distance per unit time is 0.5 mm / s, then D y Set to 0.5mm / s.

[0085] For example, w 1 and w 2 0.6, 0.4, Cs 0.53mm, C y is 4.5mm, D is 0.3mm / s, D y is 0.5mm / s, substitute it into the above formula to calculate:

[0086] =0.557+0.089=0.645.

[0087] For example, w 1 and w 2 They are 0.6 and 0.4 respectively, Cs is 0.8mm, Cy is 4.5mm, D is 0.45mm / s, and Dy is 0.5mm / s. Substitute them into the above formula for calculation:

[0088] =0.525+0.018=0.543. It can be seen that when the cable damage parameter and the cable sway parameter decrease, the cable fixation score increases.

[0089] Thirdly, the cable fixing parameters are adjusted to obtain the adjusted cable fixing parameters, and a new cable fixing score is calculated. Specifically, adjusting the cable fixing parameters will result in a cable damage parameter C s The cable shaking parameter D changes, which in turn changes the cable fixing score FIHG. For example, the initial value of the cable fixing parameter (such as the travel distance of the bolt rotation in the cable clamp) is 4.5 mm, and the calculated FIHG=0.73 is obtained. Now the cable fixing parameter is increased by 0.2 mm, and the new FIHG=0.82 is recalculated.

[0090] Finally, the cable fixing parameters are iteratively adjusted and optimized until convergence, and the optimal cable fixing parameters with the largest cable fixing score are output, and the cable is fixed to obtain the optimal cable fixing result. Among them, the iterative optimization process can use an improved genetic algorithm, and each round of iteration adjusts the fixing parameters (such as the travel distance of the bolt rotation in the cable clamp ±0.5mm, the gap distance between the cable tie and the cable ±0.2mm), and calculates the new FIHG value until the FIHG value obtained by 5 consecutive iterations is increased by less than 1%, or the maximum number of iterations (such as 100 times) is reached, the iteration stops, and the optimal cable fixing parameters with the largest cable fixing score are output, and then the cable is fixed with this optimal value to obtain the optimal cable fixing result.

[0091] In summary, compared with the prior art, this application simultaneously considers the impact of the damage of the cable insulation layer and the stability of the electrical connection on the cable fixing quality, calculates the cable fixing score, and then continuously optimizes the cable fixing parameters until the optimal cable fixing parameters are output. In this way, the optimal cable fixing result is obtained.

[0092] In summary, the embodiments of the present application have at least the following technical effects:

[0093] Compared with the prior art, the present application first collects the cable fixing parameters during the cable fixing process, and then collects the cable fixing images corresponding to the cable fixing parameters, so as to provide reliable data support for the subsequent optimization of the cable fixing parameters.

[0094] Secondly, the machine vision is used to identify the content of the cable fixed image, identify the cable gap parameters, and use the first cable damage classifier to perform fixed cable damage analysis on different cable gap parameters to obtain the first cable damage parameter. In this way, the association between the cable gap parameter and the insulation layer damage is established.

[0095] Next, by collecting environmental parameters and combining the cable gap parameters to perform cable shaking damage analysis, cable shaking parameters are obtained, and shaking cable damage classification is performed to obtain second cable damage parameters. In this way, the association between environmental wind parameters and insulation layer damage is established.

[0096] Finally, the effects of the damage to the cable insulation layer and the stability of the electrical connection on the cable fixing quality were considered at the same time. The cable fixing score was calculated, and then the cable fixing parameters were continuously optimized until the optimal cable fixing parameters were output. In this way, the optimal cable fixing parameters were obtained.

[0097] Through the above technical solution, the present application first collects the cable fixing parameters and corresponding images during the cable fixing process, and then obtains the first cable damage parameter and the second cable damage parameter based on machine vision technology, establishes the relationship between the cable gap parameter, the environmental wind dynamic parameter and the insulation layer damage, and finally comprehensively considers the influence of the cable insulation layer damage and the stability of the electrical connection on the cable fixing quality, and continuously optimizes the cable fixing parameters until the optimal cable fixing parameters are output. In this way, the cable fixing quality is improved.

[0098] Embodiment 2, as Figure 2 As shown, based on the same inventive concept of a machine vision cable fixing method provided in Embodiment 1, an embodiment of the present invention further provides a machine vision system, including:

[0099] The data acquisition module 11 is used to obtain cable fixing parameters and capture cable fixing images during the process of cable fixing;

[0100] A fixed damage analysis module 12 is used to use machine vision to perform content recognition on the cable fixed image, identify and obtain cable gap parameters, perform fixed cable damage analysis, and obtain first cable damage parameters;

[0101] The shaking damage analysis module 13 is used to collect environmental parameters, perform cable shaking damage analysis in combination with the cable gap parameters, obtain cable shaking parameters, and classify shaking cable damage to obtain second cable damage parameters;

[0102] The optimization output module 14 is used to calculate the cable damage parameter according to the first cable damage parameter and the second cable damage parameter, calculate the cable fixing score in combination with the cable shaking parameter, adjust and optimize the cable fixing parameters, and obtain the optimal cable fixing result.

[0103] Wherein, the data acquisition module 11 is specifically used for:

[0104] In the process of cable fixing, obtaining cable fixing parameters, wherein the cable fixing parameters include fixing dimensions;

[0105] An image after the cable is fixed according to the cable fixing parameters is collected to obtain a cable fixing image.

[0106] Wherein, the fixed damage analysis module 12 is specifically used for:

[0107] Preprocessing the cable fixing image;

[0108] The preprocessed cable fixing image is input into a pretrained cable fixing identifier to identify and output cable gap parameters, wherein the cable gap parameters include the gap distance after the cable is fixed. The cable fixing identifier is constructed using a convolutional neural network and is trained using a sample cable fixing image set and a sample cable gap parameter set.

[0109] Furthermore, the “performing a fixed cable damage analysis to obtain a first cable damage parameter” includes:

[0110] According to the cable fixing data in the historical time, a set of sample cable gap parameters is collected, and the size of the cable insulation layer damage under different sample cable gap parameters is collected and marked as the sample first cable damage parameter, so as to obtain the sample first cable damage parameter set;

[0111] Constructing an index relationship between the sample cable clearance parameter set and the sample first cable damage parameter set to obtain a first cable damage classifier;

[0112] The cable gap parameter is input into the first cable damage classifier, and the index classification output is used to obtain the first cable damage parameter.

[0113] The shaking damage analysis module 13 is specifically used for:

[0114] Collect wind parameters in the cable fixing environment as environmental parameters;

[0115] The cable clearance parameters and environmental parameters are input into a pre-trained cable sway predictor to predict and output cable sway parameters, wherein the cable sway parameters include the sway distance of the cable in unit time. The cable sway predictor is constructed based on a feedforward neural network and is trained using a sample cable clearance parameter set, a sample environmental parameter set and a sample cable sway parameter set.

[0116] Furthermore, the “classifying the shaking cable damage and obtaining the second cable damage parameter” includes:

[0117] According to the cable monitoring data in the historical time, a set of sample cable shaking parameters is collected, and the size of the cable insulation layer damage within a preset time range under different sample cable shaking parameters is collected, marked as the sample second cable damage parameter, and the sample second cable damage parameter set is obtained;

[0118] Constructing an index relationship between the sample cable sway parameter set and the sample second cable damage parameter set to obtain a second cable damage classifier;

[0119] The cable shaking parameters are input into the second cable damage classifier, and the index classification output is used to obtain the second cable damage parameters.

[0120] The optimization output module 14 is specifically used for:

[0121] Calculating and obtaining a cable damage parameter according to the first cable damage parameter and the second cable damage parameter;

[0122] According to the cable damage parameters and cable shaking parameters, the cable fixation score is calculated as follows:

[0123] ;

[0124] Among them, FIHG is the cable fixing score, and is the weight, is the cable damage parameter, The preset size of the cable insulation layer, D is the cable sway parameter, To preset the cable shaking parameters;

[0125] The cable fixing parameters are adjusted to obtain adjusted cable fixing parameters, and analysis and calculation are performed to obtain a cable fixing score;

[0126] Iteratively adjust and optimize the cable fixing parameters until convergence, output the optimal cable fixing parameters with the largest cable fixing score, fix the cable, and obtain the optimal cable fixing result.

[0127] In summary, the embodiments of the present application have at least the following technical effects:

[0128] The data acquisition module collects the cable fixing parameters during the cable fixing process, and then collects the cable fixing images corresponding to the cable fixing parameters, providing reliable data support for the subsequent optimization of the cable fixing parameters; the fixed damage analysis module uses the first cable damage classifier to perform fixed cable damage analysis on different cable gap parameters, obtains the first cable damage parameter, and establishes the association between the cable gap parameter and the insulation layer damage, providing reliable data support for the subsequent optimization of the fixed parameters; the sway damage analysis module collects environmental parameters and combines the cable gap parameters to perform cable sway damage analysis, obtains the cable sway parameters, and classifies the sway cable damage to obtain the second cable damage parameter, establishes the association between the environmental wind parameters and the insulation layer damage, and provides reliable data support for the subsequent optimization of the fixed parameters; the optimization output module simultaneously considers the impact of the damage of the cable insulation layer and the stability of the electrical connection on its performance, calculates the cable fixing score, and then continuously optimizes the cable fixing parameters until the optimal cable fixing parameters are output. In this way, the optimal cable fixing parameters are obtained and the cable fixing quality is improved.

[0129] like Figure 3 As shown, an embodiment of the present invention provides a machine vision cable fixing device, which includes: a memory 210 for storing a computer software program 211; a processor 220 for reading and executing the computer software program 211, thereby implementing a machine vision cable fixing method.

[0130] It should be noted that in the above embodiments, the description of each embodiment has its own emphasis, and for parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0131] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0132] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded computer, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0133] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0134] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0135] Although preferred embodiments of the present invention have been described, additional changes and modifications may occur to these embodiments once those skilled in the art understand the basic inventive concepts.

[0136] 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 belong to the scope of the present invention and its equivalent technologies, the present invention is also intended to include these changes and variations.

Claims

1. A machine vision cable fixing method, characterized in that: The method comprises: During the process of cable fixing, cable fixing parameters are obtained and cable fixing images are collected; Using machine vision to perform content recognition on the cable fixed image, identify and obtain cable gap parameters, perform fixed cable damage analysis, and obtain first cable damage parameters; Collecting environmental parameters, performing cable shaking damage analysis in combination with the cable gap parameters to obtain cable shaking parameters, and performing shaking cable damage analysis to obtain second cable damage parameters; The cable damage parameter is calculated based on the first cable damage parameter and the second cable damage parameter. The cable fixation score is calculated based on the cable sway parameter. The cable fixation parameters are adjusted and optimized to obtain the optimal cable fixation result.

2. The cable fixing method for machine vision according to claim 1, characterized in that: During the cable fixing process, the cable fixing parameters are obtained and the cable fixing images are collected, including: In the process of cable fixing, obtaining cable fixing parameters, wherein the cable fixing parameters include fixing dimensions; An image after the cable is fixed according to the cable fixing parameters is collected to obtain a cable fixing image.

3. The cable fixing method for machine vision according to claim 1, characterized in that: The cable fixing image is subjected to content recognition by machine vision to obtain cable clearance parameters, including: Preprocessing the cable fixing image; The preprocessed cable fixing image is input into a pretrained cable fixing identifier to identify and output cable gap parameters, wherein the cable gap parameters include the gap distance after the cable is fixed. The cable fixing identifier is constructed using a convolutional neural network and is trained using a sample cable fixing image set and a sample cable gap parameter set.

4. The cable fixing method for machine vision according to claim 1, characterized in that: Perform fixed cable damage analysis to obtain the first cable damage parameters, including: According to the cable fixing data in the historical time, a set of sample cable gap parameters is collected, and the size of the cable insulation layer damage under different sample cable gap parameters is collected and marked as the sample first cable damage parameter, so as to obtain the sample first cable damage parameter set; Constructing an index relationship between the sample cable clearance parameter set and the sample first cable damage parameter set to obtain a first cable damage classifier; The cable gap parameter is input into the first cable damage classifier, and the index classification output is used to obtain the first cable damage parameter.

5. The cable fixing method for machine vision according to claim 1, characterized in that: Collect environmental parameters, combine the cable gap parameters to perform cable sway damage analysis, and obtain cable sway parameters, including: Collect wind parameters in the cable fixing environment as environmental parameters; The cable clearance parameters and environmental parameters are input into a pre-trained cable sway predictor to predict and output cable sway parameters, wherein the cable sway parameters include the sway distance of the cable in unit time. The cable sway predictor is constructed based on a feedforward neural network and is trained using a sample cable clearance parameter set, a sample environmental parameter set and a sample cable sway parameter set.

6. The cable fixing method for machine vision according to claim 1, characterized in that: Perform shaking cable damage classification and obtain the second cable damage parameters, including: According to the cable monitoring data in the historical time, a set of sample cable shaking parameters is collected, and the size of the cable insulation layer damage within a preset time range under different sample cable shaking parameters is collected, marked as the sample second cable damage parameter, and the sample second cable damage parameter set is obtained; Constructing an index relationship between the sample cable sway parameter set and the sample second cable damage parameter set to obtain a second cable damage classifier; The cable shaking parameters are input into the second cable damage classifier, and the index classification output is used to obtain the second cable damage parameters.

7. The cable fixing method for machine vision according to claim 1, characterized in that: The cable damage parameter is calculated based on the first cable damage parameter and the second cable damage parameter, and the cable fixing score is calculated based on the cable shaking parameter, and the cable fixing parameter is adjusted and optimized to obtain the optimal cable fixing result, including: Calculating and obtaining a cable damage parameter according to the first cable damage parameter and the second cable damage parameter; According to the cable damage parameters and cable shaking parameters, the cable fixation score is calculated as follows: ； Among them, FIHG is the cable fixing score, and is the weight, is the cable damage parameter, is the preset size of the cable insulation layer, D is the cable sway parameter, To preset the cable shaking parameters; The cable fixing parameters are adjusted to obtain adjusted cable fixing parameters, and analysis and calculation are performed to obtain a cable fixing score; Iteratively adjust and optimize the cable fixing parameters until convergence, output the optimal cable fixing parameters with the largest cable fixing score, fix the cable, and obtain the optimal cable fixing result.

8. A machine vision cable fixing system, characterized in that: The method for executing any one of claims 1 to 7 comprises: The data acquisition module is used to obtain the cable fixing parameters and the cable fixing images during the cable fixing process; A fixed damage analysis module, used to use machine vision to perform content recognition on the cable fixed image, identify and obtain cable gap parameters, perform fixed cable damage analysis, and obtain first cable damage parameters; A sway damage analysis module, used to collect environmental parameters, perform cable sway damage analysis in combination with the cable gap parameters, obtain cable sway parameters, and perform sway cable damage classification to obtain second cable damage parameters; The optimization output module is used to calculate the cable damage parameter based on the first cable damage parameter and the second cable damage parameter, calculate the cable fixation score in combination with the cable shaking parameter, adjust and optimize the cable fixation parameters, and obtain the optimal cable fixation result.

9. A machine vision cable fixing device, characterized in that: include: Memory for storing computer software programs; A processor is used to read and execute the computer software program, thereby implementing the machine vision cable fixing method described in any one of claims 1-7.

Images