A method and system for adjusting and optimizing sandblasting parameters

Abstract

The application discloses a sand blasting parameter adjustment optimization method and system, relates to the field of sand blasting parameter control, and intelligently adjusts key parameters of a sand blasting robot, such as roughness, sand blasting diameter, sand blasting angle and material type, by combining an advanced support vector machine algorithm and a Gaussian kernel processing, significantly improves sand blasting efficiency and adaptability, simultaneously reduces manual intervention, ensures higher operation precision and consistency, and through the application, the sand blasting robot can automatically optimize operation parameters when processing various complex surfaces, so that the automation degree of operation and the ability to adapt to different operation environments can be greatly improved while ensuring a high-quality surface treatment effect.

Description

Technical Field

[0001] This invention relates to the field of sandblasting parameter control, specifically a method and system for adjusting and optimizing sandblasting parameters. Background Technology

[0002] An industrial robot is a machine device capable of automatically performing various industrial tasks. It typically consists of a front-end mechanical structure and a back-end control system. Industrial robots perform precise and efficient operations on production lines, replacing or assisting human labor in repetitive, tedious, or dangerous tasks. The applications of industrial robots are very broad, including automobile manufacturing, electronics assembly, metal processing, and logistics warehousing. They can perform various tasks such as welding, assembly, handling, spraying, grinding, and quality inspection. Through IoT technology and data analysis, industrial robots can achieve highly flexible movements and precise operations, improving equipment intelligence, production efficiency, quality, and safety. Sandblasting robots, specifically designed for automated sandblasting operations, typically consist of a high-precision robotic arm, sandblasting equipment (including a spray gun and abrasive storage unit), a control system, and sensors. They can perform precise and uniform sandblasting on various material surfaces, thereby improving work efficiency and quality while reducing labor costs and safety risks.

[0003] In existing sandblasting robot technology, operators pre-program the parameters of the sandblasting robot according to work requirements, such as sandblasting intensity, nozzle movement speed and path. These parameters are usually set manually before operation, based on experience or standard operating procedures. In addition, during operation, operators may need to monitor and manually adjust the robot's parameters in real time to adapt to specific working conditions or surface conditions.

[0004] Although existing sandblasting robot technology has made some progress in the field of automated surface treatment, it still has some limitations:

[0005] (1) The parameter setting method that relies on preset programs and manual adjustments lacks flexibility and is difficult to adapt to complex and ever-changing working environments;

[0006] (2) Manually adjusting parameters is time-consuming and inefficient, and is easily affected by operator fatigue and subjective judgment, which may lead to inconsistent sandblasting results;

[0007] (3) The lack of advanced data processing and intelligent decision-making capabilities makes it difficult for robots to automatically optimize and adjust according to complex working conditions and surface characteristics.

[0008] In conclusion, although existing sandblasting robot technologies offer automated solutions in some aspects, there is still room for improvement in terms of flexibility, efficiency, accuracy, and intelligence. SUMMARY

[0009] The present application provides a sandblasting parameter adjustment optimization method and system, which aims to make decision assistance through intelligent algorithm to select the most suitable working mode for the current conditions.

[0010] The sandblasting parameter adjustment optimization method comprises the following steps:

[0011] S1. Obtain surface feature data through high-precision surface scanning technology, wherein the surface feature data at least includes roughness, sandblasting diameter and sandblasting angle, and normalize the surface feature data;

[0012] S2. Determine the relationship between the surface feature data and the sandblasting intensity, and classify the surface feature data according to the linear / nonlinear relationship;

[0013] S3. For the data of nonlinear relationship, perform kernel function conversion according to the SVM classification algorithm of radial basis function, and for the data of linear relationship, directly use SVM in the original feature space for training;

[0014] S4. Use the trained SVM model to predict the best parameters of sandblasting operation according to the new working scene.

[0015] Further, in the step S3, the data of nonlinear relationship specifically comprises the following steps:

[0016] S301. Select the SVM model of radial basis function, and define the kernel function: K(x i ,x j )=exp(-γ||x i -x j || 2 ); wherein, the K(x i ,x j ) represents the value of the kernel function, the exp(-γ||x i -x j || 2 ) represents the conversion of the Euclidean distance in the original feature space into the similarity measure in the feature space through the exponential function, the γ represents the kernel function, and the x i and x j represent the sample points in the data set;

[0017] S302. Calculate the kernel function value according to the defined kernel function for each surface feature data feature with nonlinear relationship;

[0018] S303. Determine the best regularization parameter and each surface feature data feature with nonlinear relationship through cross-validation.

[0019] Furthermore, in step S303, the specific steps for determining the optimal regularization parameter through cross-validation are as follows:

[0020] S3031. Define the regularization parameter C and the kernel function γ;

[0021] S3032. For each group (C, γ), use k-fold cross-validation to evaluate the model performance;

[0022] S3033. Select the C,γ combination that has the best average performance in cross-validation.

[0023] Furthermore, in step S302, the specific steps for calculating the kernel function value based on the defined kernel function are as follows:

[0024] For roughness: K 粗糙度 (x i ,x j )=exp(-γ 粗糙度 (x i,粗糙度 -x j,粗糙度 ) 2 ); where K 粗糙度 (x i ,x j ) represents the sample x in the feature dimension of roughness. i and x j The similarity, the γ 粗糙度 The x represents the width parameter of the Gaussian kernel function in the roughness dimension. i,粗糙度 and x j,粗糙度 This represents the value in terms of roughness characteristics.

[0025] Furthermore, in step S3, for data with linear relationships, the specific steps for training directly using SVM in the original feature space are as follows:

[0026] Training using a linear SVM: Where ω represents the weight vector, b represents the bias term, C represents the regularization parameter, and ξ represents the weight vector. i Denotes the slack variable, the min ω,b This represents the process of minimizing the objective function.

[0027] Furthermore, a sandblasting parameter adjustment and optimization system, which is based on any one of the sandblasting parameter adjustment and optimization methods described above, includes:

[0028] The data acquisition module is used to acquire surface feature data through high-precision surface scanning technology. The surface feature data includes at least roughness, blasting diameter and blasting angle, and the surface feature data is normalized.

[0029] The data classification module is used to determine the relationship between surface feature data and sandblasting intensity, and to classify the surface feature data according to linear / nonlinear relationships;

[0030] In the model training module, for non-linear data, kernel function transformation is performed using the SVM classification algorithm based on radial basis function; for linear data, SVM is directly used for training in the original feature space.

[0031] The parameter tuning module is used to predict the optimal parameters for sandblasting operations based on new work scenarios using a trained SVM model.

[0032] Furthermore, the model training module also includes:

[0033] The nonlinear relation data processing unit is used to select an SVM model with radial basis functions to train nonlinear relation data; the linear relation data processing unit is used to train linear relation data using a linear SVM.

[0034] Furthermore, the nonlinear relational data processing unit also includes:

[0035] The kernel function defines sub-units used to select the radial basis function SVM model, and defines the kernel function: K(x i ,x j )=exp(-γ||x i -x j || 2 ); where K(x) i ,x j ) represents the value of the kernel function, where exp(-γ||x) i -x j || 2 ) represents the transformation of the Euclidean distance in the original feature space into a similarity metric in the feature space using an exponential function, where γ represents the kernel function, and x i and x j This represents the sample points in the dataset;

[0036] The kernel function calculation subunit is used to calculate the kernel function value according to the defined kernel function for each surface feature data feature with a nonlinear relationship.

[0037] The cross-validation tuning subunit is used to determine the optimal regularization parameters and surface feature data features with nonlinear relationships through cross-validation.

[0038] Furthermore, in the kernel function calculation subunit, the specific process for calculating the kernel function value according to the defined kernel function is as follows: For roughness: K 粗糙度 (x i ,x j )=exp(-γ粗糙度 (x i,粗糙度 -x j,粗糙度 ) 2 ); where K 粗糙度 (x i ,x j ) represents the sample x in the feature dimension of roughness. i and x j The similarity, the γ 粗糙度 The x represents the width parameter of the Gaussian kernel function in the roughness dimension. i,粗糙度 and x j,粗糙度 This represents the value in terms of roughness characteristics.

[0039] Furthermore, in the linear relationship data processing unit, the specific process for training the linear relationship data using a linear SVM is as follows: training using a linear SVM: Where ω represents the weight vector, b represents the bias term, C represents the regularization parameter, and ξ represents the weight vector. i Denotes the slack variable, the min ω,b This represents the process of minimizing the objective function.

[0040] The beneficial effects of the invention are:

[0041] (1) This invention combines advanced support vector machine (SVM) algorithm and Gaussian kernel processing to intelligently adjust key parameters of sandblasting robot such as roughness, sandblasting diameter, sandblasting angle and material type, which significantly improves sandblasting efficiency and adaptability, while reducing manual intervention and ensuring higher operation accuracy and consistency.

[0042] (2) Through this invention, the sandblasting robot can automatically optimize operating parameters when processing various complex surfaces, thereby ensuring high-quality surface treatment while greatly improving the automation level of the operation and the ability to adapt to different working environments. Attached Figure Description

[0043] Figure 1 A flowchart illustrating a method for adjusting and optimizing sandblasting parameters according to an embodiment of the present invention;

[0044] Figure 2 This is a schematic diagram of the structure of a terminal device for adjusting and optimizing sandblasting parameters according to an embodiment of the present invention;

[0045] Figure 3 This is a schematic diagram of the product structure of a sandblasting parameter adjustment and optimization method proposed in an embodiment of the present invention. Detailed Implementation

[0046] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings, but the scope of protection of the present invention is not limited to the following description.

[0047] For the purposes of the present application, the technical solutions and advantages will be clearer. The present application will be further described in detail in conjunction with the drawings and examples. It should be understood that the specific examples described herein are only intended to explain the present application and are not intended to limit the present application. That is, the described examples are only a part of the examples of the present application, but not all the examples. The components of the embodiments of the present application generally described and shown in the drawings herein can be arranged and designed in various different configurations.

[0048] Therefore, the detailed description of the embodiments of the present application provided below in the drawings is not intended to limit the scope of the claimed application, but only represents selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application. It should be noted that the relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between the entities or operations.

[0049] Moreover, the term "comprise", "include" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such a process, method, article or device. Without more limitations, the element defined by the statement "comprises a" does not exclude the presence of additional identical elements in the process, method, article or device comprising the element.

[0050] The features and properties of the present application are further described in detail below in conjunction with the examples.

[0051] Among them, a sand blasting parameter adjustment optimization method, such as Figure 1 , comprising the following steps:

[0052] S1. Obtain surface feature data by high-precision surface scanning technology, the surface feature data at least including roughness, sand blasting diameter and sand blasting angle, and normalize the surface feature data;

[0053] S2. Determine the relationship between the surface feature data and the sand blasting intensity, and classify the surface feature data according to the linear / nonlinear relationship;

[0054] S3. For data with nonlinear relationship, kernel function conversion is performed according to SVM classification algorithm of radial basis function, and for data with linear relationship, SVM is directly used in the original feature space for training;

[0055] S4. Using the trained SVM model to predict the optimal parameters of the sandblasting operation according to the new job scenario.

[0056] Specifically, for the nonlinear relationship data described in the embodiment, the corresponding surface feature data is roughness data, and for the linear relationship data, the corresponding surface feature data is sandblasting angle and sandblasting diameter. The specific implementation process of the embodiment is to comprehensively consider roughness (processed by Gaussian kernel), sandblasting diameter and sandblasting angle (processed by linear kernel) through the SVM model to predict the most suitable sandblasting parameters. For example, for a region with high roughness, the model may recommend stronger sandblasting intensity and finer sandblasting diameter.

[0057] As a preferred specific implementation of the above embodiment, the specific process of normalizing the surface feature data is as follows: for each feature x i By: normalization to reduce the deviation between different orders of magnitude of features, wherein the μ i represents the mean of the i-th feature, and the σ i represents the standard deviation of the i-th feature.

[0058] Further, in step S3, the nonlinear relationship data specifically includes the following steps:

[0059] S301. Select the SVM model of the radial basis function and define the kernel function: K(x i ,x j ) = exp(-γ‖x i -x j ‖ 2 ); wherein the K(x i ,x j ) represents the value of the kernel function, the exp(-γ‖x i -x j ‖ 2 ) represents the conversion of the Euclidean distance in the original feature space to the similarity measure in the feature space by the exponential function, the γ represents the kernel function, and the x i and x j represent sample points in the data set.

[0060] S302. For each surface feature data feature with a nonlinear relationship, calculate the kernel function value according to the defined kernel function.

[0061] S303. Determine the optimal regularization parameter and each surface feature data feature with a nonlinear relationship by cross-validation.

[0062] Further, in step S303, the specific steps for determining the optimal regularization parameter by cross-validation are as follows:

[0063] S3031. Define the regularization parameter C and the kernel function γ;

[0064] S3032. For each group (C, γ), use k-fold cross-validation to evaluate the model performance;

[0065] S3033. Select the C,γ combination that has the best average performance in cross-validation.

[0066] Specifically, for the above embodiments, a parameter tuning process is proposed to find the optimal parameter combination so that the model can perform best on unseen data. The process is as follows: Define a series of C and γ values ​​to be tested; use grid search and k-fold cross-validation to test each set of parameters; find a set of C and γ that performs best in cross-validation; use these parameters to retrain the SVM model on all available training data; and validate its performance on the test set. It should be noted that the regularization parameter C controls the model's tolerance for misclassification. A larger C value reduces the model's tolerance for misclassification, forcing the model to classify each sample more strictly, leading to overfitting on training data but poor generalization ability on unseen data; a smaller C value allows for more misclassification, increasing the model's generalization ability, but may also lead to an overly simplistic model that cannot capture complex patterns in the data. A suitable C value is selected through experimentation and cross-validation. By trying different C values, the optimal C value on the validation dataset is found, thus balancing performance on training data and generalization ability.

[0067] Furthermore, in step S302, the specific steps for calculating the kernel function value based on the defined kernel function are as follows:

[0068] For roughness: K 粗糙度 (x i ,x j )=exp(-γ 粗糙度 (x i,粗糙度 -x j,粗糙度 ) 2 ); where K 粗糙度 (x i ,x j ) represents the sample x in the feature dimension of roughness. i and x j The similarity, the γ 粗糙度 The x represents the width parameter of the Gaussian kernel function in the roughness dimension. i,粗糙度 and x j,粗糙度 This represents the value in terms of roughness characteristics.

[0069] Furthermore, in step S3, for data with linear relationships, the specific steps for training directly using SVM in the original feature space are as follows:

[0070] Training by linear SVM: wherein, ω represents a weight vector, b represents a bias term, C represents a regularization parameter, ξ i represents a slack variable, min ω,b represents a process of minimizing an objective function.

[0071] Further, a sandblasting parameter adjustment optimization system is implemented based on any one of the above-described sandblasting parameter adjustment optimization methods, comprising:

[0072] A data acquisition module is configured to acquire surface feature data by high-precision surface scanning technology, and the surface feature data at least includes roughness, sandblasting diameter and sandblasting angle, and the surface feature data is normalized.

[0073] A data classification module is configured to judge the relationship between the surface feature data and the sandblasting intensity, and classify the surface feature data according to linear / nonlinear relationship.

[0074] A model training module is configured to perform kernel function conversion according to the SVM classification algorithm of the radial basis function for the data of nonlinear relationship, and directly use SVM to train in the original feature space for the data of linear relationship.

[0075] A parameter optimization module is configured to use the trained SVM model to predict the best parameters of sandblasting operation according to a new job scene.

[0076] Further, the model training module further comprises:

[0077] A nonlinear relationship data processing unit is configured to select the SVM model of the radial basis function to train the data of nonlinear relationship; and a linear relationship data processing unit is configured to train the data of linear relationship by linear SVM.

[0078] Further, the nonlinear relationship data processing unit further comprises:

[0079] A kernel function definition subunit is configured to select the SVM model of the radial basis function, and define the kernel function: K(x i ,x j )=exp(-γ||x i -x j || 2 ); wherein, K(x i ,x j ) represents the value of the kernel function, exp(-γ||x i -x j || 2) represents the conversion of the Euclidean distance in the original feature space into a similarity measure in the feature space by an exponential function, the γ represents a kernel function, the x i and x j represent sample points in the data set;

[0080] The kernel function calculation sub-unit is configured to calculate the kernel function value according to the defined kernel function for each surface feature data feature with a nonlinear relationship.

[0081] The cross-validation tuning sub-unit is configured to determine the optimal regularization parameter and each surface feature data feature with a nonlinear relationship through cross-validation.

[0082] Further, in the kernel function calculation sub-unit, the specific process of calculating the kernel function value according to the defined kernel function is as follows: for roughness: K 粗糙度 (x i ,x j ) = exp(-γ 粗糙度 (x i,粗糙度 -x j,粗糙度 ) 2 ); wherein, the K 粗糙度 (x i ,x j ) represents the similarity of sample x i and x j in the feature dimension of roughness, the γ 粗糙度 represents the width parameter of the Gaussian kernel function in the roughness dimension, and the x i,粗糙度 and x j,粗糙度 represent the values on the roughness feature.

[0083] Further, in the linear relationship data processing unit, the specific process of training the linear relationship data by the linear SVM is as follows: training by the linear SVM: wherein, the ω represents a weight vector, the b represents a bias term, the C represents a regularization parameter, the ξ i represents a slack variable, and the min ω,b represents the process of minimizing the objective function.

[0084] As a further preferred embodiment, a terminal device for adjusting and optimizing sandblasting parameters is provided, such as Figure 2 The terminal device 200 comprises at least one memory 210, at least one processor 220, and a bus 230 connecting different platform systems.

[0085] The memory 210 can include a readable medium in the form of a volatile memory, such as a random access memory (RAM) 211 and / or a cache memory 212, and can further include a read-only memory (ROM) 213.

[0086] The memory 210 also stores computer programs which can be executed by the processor 220 to enable the processor 220 to perform any of the above-described sandblasting parameter adjustment optimization methods, the specific implementation manners and the achieved technical effects of which are consistent with those described in the embodiments of the above-described methods, and some of the contents will not be described herein. The memory 210 can also include programs / utilities 214 having a set of (at least one) program modules 215, which include but are not limited to: an operating system, one or more application programs, other program modules, and program data, each of which or some combination thereof can include the implementation of a network environment.

[0087] Correspondingly, the processor 220 can execute the above-described computer programs and can execute the programs / utilities 214.

[0088] The bus 230 can represent one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, a graphics acceleration port, a processor or a local bus using any of a variety of bus structures.

[0089] The terminal device 200 can also communicate with one or more external devices 240, such as a keyboard, a pointing device, a Bluetooth device, or the like, one or more devices that enable interaction with the terminal device 200 (for example, a router, a modem, or the like), and / or one or more devices (for example, a router, a modem, or the like) that enable communication with one or more other computing devices. Such communication can be through an input / output (I / O) interface 250. Furthermore, the terminal device 200 can communicate with one or more networks (for example, a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) through a network adapter 260. The network adapter 260 can communicate with other modules of the terminal device 200 through the bus 230. It should be understood that although not shown in the figure, other hardware and / or software modules can be used in conjunction with the terminal device 200, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage platforms, etc.

[0090] As a further preferred embodiment, a computer-readable storage medium for sandblasting parameter adjustment optimization is proposed, and the computer-readable storage medium stores instructions which, when executed by a processor, implement any of the above-described sandblasting parameter adjustment optimization methods. The specific implementation manners and the achieved technical effects of which are consistent with those described in the embodiments of the above-described methods, and some of the contents will not be described herein.

[0091] Figure 3A program product 300 for implementing the above method provided by the embodiment is shown, which can adopt a portable compact disc read-only memory (CD-ROM) and include program codes, and can run on a terminal device, for example, a personal computer. However, the program product 300 of the present embodiment is not limited to this, and in the present embodiment, the readable storage medium can be any tangible medium containing or storing a program, which can be used by or in conjunction with an instruction execution system, apparatus or device. The program product 300 can adopt any combination of one or more readable media. The readable medium can be a readable signal medium or a readable storage medium. The readable storage medium may, for example, be but is not limited to an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus or device, or any combination of the above. More specific examples (non-exhaustive list) of the readable storage medium include an electrical connection having one or more wires, a portable disc, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0092] The computer readable storage medium can include a data signal transported in a baseband or as part of a carrier wave over a communications link, and can be stored on or in one or more types of computer readable media, such as one or more types of removable computer readable media or components thereof, one or more types of internal computer readable media or components thereof, or one or more types of computer readable media combinations thereof. The computer readable storage medium can also include any computer readable medium, or combination thereof, that is non-transitory or that is not a propagated signal. The program code can be transmitted as a carrier wave, or can be embodied in any other volatile or non-volatile storage medium or memory (e.g., a computer memory, a register, or a transitory signal).

[0093] The foregoing is considered as illustrative only of the principles of the application. Further, since numerous modifications and changes will readily occur to those skilled in the art, it is not desired to limit the application to the exact construction and operation described. Accordingly, all such variations are intended to be included within the scope of the present application as defined in the claims below and their equivalents.

Claims

1. A method of optimizing shot parameter adjustment, characterized by, The method comprises the following steps: S1. Obtain surface feature data by high-precision surface scanning technology, the surface feature data at least including roughness, sandblasting diameter and sandblasting angle, and normalize the surface feature data; S2. Determine the relationship between the surface feature data and the sandblasting intensity, and classify the surface feature data according to linear / nonlinear relationship; S3. For the data of nonlinear relationship, perform kernel function conversion according to the SVM classification algorithm of radial basis function, and for the data of linear relationship, directly train in the original feature space using SVM; S4. Use the trained SVM model to predict the best parameters of sandblasting operation according to a new working scenario. In the step S3, the data of nonlinear relationship specifically comprises the following steps: S301. Select a SVM model of radial basis function, and define a kernel function: ; wherein the represents a value of the kernel function, the represents a conversion of the Euclidean distance in the original feature space into a similarity measure in the feature space by an exponential function, the represents a kernel function, the and represent sample points in a data set; S302. Calculate the kernel function value according to the defined kernel function for each surface feature data feature with nonlinear relationship; S303. Determine the best regularization parameter and each surface feature data feature with nonlinear relationship through cross-validation; For the data of linear relationship, the specific steps of directly training in the original feature space using SVM are as follows: Training by linear SVM: ; where the represents a weight vector, the represents a bias term, the represents a regularization parameter, the represents a relaxation variable, the represents a process of minimizing an objective function.

2. The method of claim 1, wherein, In the step S303, the specific steps of determining the best regularization parameter through cross-validation are as follows: S3031. Define the regularization parameter and kernel functions ; S3032. For each group Model performance is evaluated using k-fold cross-validation; S3033. Selecting the best combination in cross-validation combination.

3. The method of claim 1, wherein, In the step S302, the specific steps of calculating the kernel function value according to the defined kernel function are as follows: For roughness: ; wherein the represents a similarity of the samples and in the feature dimension of the roughness, the represents a width parameter of the Gaussian kernel function in the roughness dimension, the and represent values in the roughness feature.

4. A blasting parameter adjustment optimization system, which is implemented based on the blasting parameter adjustment optimization method according to any one of claims 1-3, characterized in that, Comprise: A data acquisition module is configured to obtain surface feature data by high-precision surface scanning technology, the surface feature data at least including roughness, sandblasting diameter and sandblasting angle, and normalize the surface feature data; A data classification module is configured to determine the relationship between the surface feature data and the sandblasting intensity, and classify the surface feature data according to linear / nonlinear relationship; A model training module is configured to, for the data of nonlinear relationship, perform kernel function conversion according to the SVM classification algorithm of radial basis function, and for the data of linear relationship, directly train in the original feature space using SVM; A parameter optimization module is configured to use the trained SVM model to predict the best parameters of sandblasting operation according to a new working scenario.

5. The blasting parameter adjustment optimization system of claim 4, wherein, The model training module further comprises: A nonlinear relationship data processing unit is configured to select the SVM model of radial basis function to train the data of nonlinear relationship; A linear relationship data processing unit is configured to train the data of linear relationship by linear SVM.

6. The blasting parameter adjustment optimization system of claim 5, wherein, The nonlinear relationship data processing unit further comprises: The kernel function definition subunit is configured to select a SVM model of a radial basis function and define a kernel function: ; wherein the represents a value of the kernel function, and the represents a conversion of an Euclidean distance in an original feature space into a similarity measure in a feature space by an exponential function, and the represents the kernel function, and the and represent sample points in a data set; A kernel function calculation subunit is configured to calculate the kernel function value according to the defined kernel function for each surface feature data feature with nonlinear relationship; A cross-validation optimization subunit is configured to determine the best regularization parameter and each surface feature data feature with nonlinear relationship through cross-validation.

7. The blasting parameter adjustment optimization system of claim 6, wherein, In the kernel function calculation subunit, the specific process of calculating the kernel function value according to the defined kernel function is as follows: For roughness: ; wherein the represents the similarity of the samples and in the feature dimension of roughness, the represents the width parameter of the Gaussian kernel function in the roughness dimension, the and represent the values in the roughness feature.

8. The blasting parameter adjustment optimization system of claim 5, wherein, In the linear relationship data processing unit, the specific process of training the linear relationship data by the linear SVM is as follows: training by the linear SVM ; wherein the represents a weight vector, and the represents a bias term, and the represents a regularization parameter, and the represents a relaxation variable, and the represents a process of minimizing the objective function.

Images