Self-adaptive control system and method for propeller machining

Abstract

The invention provides a self-adaptive control system and method for propeller machining, and relates to the field of propeller machining, and the method comprises the steps: collecting a real-time surface image in a propeller polishing process, and carrying out the noise reduction and feature extraction processing of the real-time surface image through a data analysis technology based on computer vision, and finally, the polishing position and the polishing time are controlled in a self-adaptive mode based on the propeller surface polishing condition. In this way, precise grinding of the surface of the propeller can be achieved, and the grinding quality can be improved.

Description

Technical Field

[0001] The present application relates to the field of propeller processing, and more specifically, to an adaptive control system and method for propeller processing. Background Art

[0002] During the manufacturing process, after the initial processing such as casting and milling, the propeller surface often has problems such as unevenness, defects or burrs. These problems not only affect the appearance quality of the propeller, but may also cause unnecessary resistance to the water flow during use, reducing the efficiency of the propeller. Through grinding, these surface defects can be removed, making the propeller surface smoother, thereby reducing resistance and improving propulsion efficiency.

[0003] The traditional propeller grinding process often relies on simple mechanized equipment. Mechanized equipment usually performs grinding according to a preset program and cannot adjust the grinding parameters in real time according to the actual surface conditions of the propeller. Some parts that have already been polished to a qualified level may continue to be polished, while some parts that need to be polished are not polished enough. This will lead to problems such as over-polishing and under-polishing, which seriously affects the quality of polishing.

[0004] Therefore, an adaptive control scheme for propeller machining is needed. Summary of the invention

[0005] In order to solve the above technical problems, the present application is proposed. The embodiments of the present application provide an adaptive control system and method for propeller processing.

[0006] According to one aspect of the present application, there is provided an adaptive control system for propeller processing, comprising: Propeller grinding image acquisition module, used to obtain real-time surface images of the propeller during the grinding process; A propeller grinding image denoising module, used for performing denoising on the real-time surface image to obtain a denoised propeller surface image; A propeller grinding image content coding module, used for performing image content coding on the denoised propeller surface image to obtain a propeller surface grinding condition feature vector; The propeller grinding processing parameter result generation module is used to decode and obtain the grinding position data and grinding time data that need fine grinding based on the propeller surface grinding condition feature vector.

[0007] According to another aspect of the present application, there is provided an adaptive control method for propeller machining, comprising: Get real-time surface images of the propeller during the grinding process; Performing noise reduction processing on the real-time surface image to obtain a noise-reduced propeller surface image; Performing image content encoding on the denoised propeller surface image to obtain a propeller surface polishing condition feature vector; Based on the propeller surface polishing condition feature vector, the polishing position data and polishing time data requiring fine polishing are decoded and obtained.

[0008] Compared with the prior art, the adaptive control system and method for propeller processing provided by the present application collects real-time surface images during the propeller grinding process, and uses computer vision-based data analysis technology to reduce noise and extract features from the real-time surface images, thereby understanding the propeller surface grinding conditions, and finally adaptively controlling the grinding position and grinding time based on the propeller surface grinding conditions. In this way, accurate grinding of the propeller surface can be achieved, which is conducive to improving the grinding quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0010] Figure 1 It is a system block diagram of an adaptive control system for propeller processing according to an embodiment of the present application.

[0011] Figure 2 It is a block diagram of a propeller grinding image content encoding module in an adaptive control system for propeller processing according to an embodiment of the present application.

[0012] Figure 3 It is a block diagram of a propeller surface image local feature extraction unit in an adaptive control system for propeller processing according to an embodiment of the present application.

[0013] Figure 4 It is a block diagram of a global feature extraction unit of a propeller surface image in an adaptive control system for propeller processing according to an embodiment of the present application.

[0014] Figure 5 It is a block diagram of a propeller grinding processing parameter result generation module in an adaptive control system for propeller processing according to an embodiment of the present application.

[0015] Figure 6 It is a flow chart of an adaptive control method for propeller machining according to an embodiment of the present application. DETAILED DESCRIPTION

[0016] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0017] In modern shipbuilding, optimizing propeller performance is the key to improving propulsion efficiency and reducing energy consumption. During the manufacturing process, after the initial processing such as casting and milling, the propeller surface often has problems such as unevenness, defects or burrs. These problems not only affect the appearance quality of the propeller, but may also cause unnecessary resistance to the water flow during use, reducing the efficiency of the propeller. Through grinding, these surface defects can be removed, making the propeller surface smoother, thereby reducing resistance and improving propulsion efficiency.

[0018] The traditional propeller grinding process often relies on simple mechanized equipment, which usually performs grinding according to preset programs and cannot adjust the grinding parameters in real time according to the actual surface conditions of the propeller. This may result in some parts that have been polished to a qualified level continuing to be polished, while some parts that need to be polished are not polished enough, which will lead to problems such as over-polishing and under-polishing, seriously affecting the quality of polishing. Therefore, an adaptive control solution for propeller processing is expected.

[0019] In recent years, deep learning and neural networks have been widely used in computer vision, natural language processing, text signal processing and other fields. In addition, deep learning and neural networks have also shown a level close to or even beyond that of humans in image classification, object detection, semantic segmentation, text translation and other fields. The development of deep learning and neural networks provides new solutions and solutions for the adaptive control of propeller processing.

[0020] Figure 1 FIG. 1 is a system block diagram of an adaptive control system for propeller machining according to an embodiment of the present application. Figure 1 As shown, in the adaptive control system 100 for propeller processing, it includes: a propeller grinding processing image acquisition module 110, which is used to obtain a real-time surface image of the propeller in the grinding process; a propeller grinding processing image denoising module 120, which is used to perform denoising on the real-time surface image to obtain a denoised propeller surface image; a propeller grinding processing image content encoding module 130, which is used to perform image content encoding on the denoised propeller surface image to obtain a propeller surface grinding condition feature vector; a propeller grinding processing parameter result generating module 140, which is used to decode, based on the propeller surface grinding condition feature vector, to obtain grinding position data and grinding time data that require fine grinding.

[0021] In an embodiment of the present application, the propeller grinding processing image acquisition module 110 is used to obtain a real-time surface image of the propeller during the grinding process. It should be understood that the real-time surface image of the propeller during the grinding process can intuitively display the unevenness of the surface during the propeller grinding process, such as protrusions, depressions, etc., and this information is crucial for determining which areas require fine grinding. In addition to the grinding position, the information in the real-time surface image of the propeller during the grinding process can also be used to determine the grinding time. Specifically, the defects and unevenness in different areas are different, so the required grinding time is also different. By analyzing the feature information in the image, the grinding time required for each area can be estimated. In general, through image processing technology, real-time monitoring of the propeller surface grinding condition can be achieved, and precise control of grinding can be achieved based on the propeller surface condition, thereby improving the grinding quality to ensure the overall performance of the propeller.

[0022] In an embodiment of the present application, the propeller grinding image denoising module 120 is used to perform denoising on the real-time surface image to obtain the denoised propeller surface image. Specifically, in an embodiment of the present application, the propeller grinding image denoising module is used to: pass the real-time surface image through a propeller surface image denoiser to obtain the denoised propeller surface image. It should be understood that in the propeller grinding process, the propeller surface image acquired in real time may be affected by a variety of factors, such as light, equipment noise, dust, etc., which will cause the quality of the collected image to deteriorate and make it difficult to accurately reflect the true situation of the propeller surface. In order to effectively remove these interference factors and improve the clarity and accuracy of the image, it is necessary to perform denoising on the real-time surface image, that is, to input the real-time surface image into the propeller surface image denoiser. It should be noted here that the essence of the propeller surface image denoiser is an image denoiser based on an automatic codec. The automatic codec is mainly divided into an encoder part and a decoder part. The encoder is responsible for compressing the input real-time surface image into a low-dimensional feature representation. These features contain the main information of the image, but remove noise and redundant information; the decoder reconstructs a high-quality denoised propeller surface image based on these low-dimensional features, while removing noise. In particular, in an embodiment of the present application, one possible implementation method of passing the real-time surface image through a propeller surface image denoiser to obtain the denoised propeller surface image may be: inputting the real-time surface image into an encoder of the propeller surface image denoiser, wherein the encoder uses a convolutional layer to perform explicit spatial encoding on the real-time surface image to obtain propeller image features; inputting the propeller image features into a decoder of the propeller surface image denoiser, wherein the decoder uses a deconvolutional layer to perform deconvolution processing on the propeller image features to obtain the denoised propeller surface image.

[0023] In the embodiment of the present application, the propeller grinding image content encoding module 130 is used to perform image content encoding on the denoised propeller surface image to obtain a propeller surface grinding condition feature vector. Specifically, Figure 2 FIG. 1 is a block diagram of a propeller grinding image content encoding module in an adaptive control system for propeller machining according to an embodiment of the present application. Figure 2 As shown, the propeller grinding processing image content encoding module 130 includes: a propeller surface image local feature extraction unit 131, used to perform image local feature extraction on the denoised propeller surface image to obtain a propeller surface local grinding condition feature vector; a propeller surface image global feature extraction unit 132, used to perform image global feature extraction on the denoised propeller surface image to obtain a propeller surface global grinding condition feature vector; a propeller surface feature fusion unit 133, used to perform smoothing adjustment on the propeller surface local grinding condition feature vector and the propeller surface global grinding condition feature vector based on the feature manifold corresponding to the visual mapping to obtain the propeller surface grinding condition feature vector.

[0024] More specifically, Figure 3 FIG. 1 is a block diagram of a propeller surface image local feature extraction unit in an adaptive control system for propeller processing according to an embodiment of the present application. Figure 3 As shown, the propeller surface image local feature extraction unit 131 includes: a propeller surface local polishing condition feature generation subunit 1311, used to pass the denoised propeller surface image through a propeller surface local feature encoder to obtain a propeller surface local polishing condition feature map; a propeller surface local polishing condition feature dimensionality reduction subunit 1312, used to perform maximum pooling on each feature matrix of the propeller surface local polishing condition feature map along the channel dimension to obtain the propeller surface local polishing condition feature vector.

[0025] Considering that the denoised propeller surface image mainly contains information such as the real protrusions and depressions on the propeller surface, this information is the basis for determining the subsequent polishing-related parameters. In order to extract the propeller surface condition features from the denoised surface image, and then help to more accurately judge the propeller's polishing needs, based on this, in the technical solution of the present application, the denoised propeller surface image is passed through a propeller surface local feature encoder to obtain a propeller surface local polishing condition feature map. Here, the propeller surface local feature encoder is a convolutional neural network model using a two-dimensional convolution kernel. It is usually composed of multiple convolutional layers, pooling layers, fully connected layers, etc., and can automatically learn the feature representation in the denoised propeller surface image. Due to the advantages of automatic learning features, local perception and weight sharing, translation invariance, scalability and flexibility, it can efficiently process propeller surface images and extract useful feature information.

[0026] Specifically, in an embodiment of the present application, a possible implementation method of passing the denoised propeller surface image through a propeller surface local feature encoder to obtain a propeller surface local polishing condition feature map may be: using each layer of the propeller surface local feature encoder to perform the following on the input data in the forward transfer of the layer: using the convolution units of each layer of the propeller surface local feature encoder to perform convolution processing based on a two-dimensional convolution kernel on the input data to obtain a convolution feature map; using the pooling units of each layer of the propeller surface local feature encoder to perform pooling processing based on a local feature matrix on the convolution feature map to obtain a pooling feature map; using the activation units of each layer of the propeller surface local feature encoder to perform nonlinear activation on the feature values ​​of each position in the pooling feature map to obtain an activation feature map; wherein the output of the last layer of the propeller surface local feature encoder is the propeller surface local polishing condition feature map, and the input of the first layer of the propeller surface local feature encoder is the denoised propeller surface image.

[0027] Considering that the original propeller surface local polishing condition feature map obtained through feature extraction contains a large amount of feature information, directly processing these features will consume a lot of computing resources and time. In order to significantly reduce the dimension of the feature map, thereby reducing the amount of calculation for subsequent processing, the technical solution of the present application needs to perform maximum pooling on each feature matrix along the channel dimension of the propeller surface local polishing condition feature map to obtain the propeller surface local polishing condition feature vector. Through the maximum pooling operation, the most significant features in the feature map can be retained while removing unimportant redundant information.

[0028] More specifically, Figure 4 FIG. 1 is a block diagram of a propeller surface image global feature extraction unit in an adaptive control system for propeller processing according to an embodiment of the present application. Figure 4 As shown, the propeller surface image global feature extraction unit 132 includes: a propeller surface global polishing condition feature generation subunit 1321, used to pass the denoised propeller surface image through a propeller surface global feature encoder to obtain a propeller surface global polishing condition feature map; a propeller surface global polishing condition feature dimensionality reduction subunit 1322, used to perform maximum value pooling on each feature matrix of the propeller surface global polishing condition feature map along the channel dimension to obtain the propeller surface global polishing condition feature vector.

[0029] Accordingly, it is considered that although the denoised propeller surface image has removed noise, it still contains a lot of information. This information includes both local detail features and global overall features. In order to capture the global features in the image and help understand the overall polishing of the propeller surface, the technical solution of the present application needs to pass the denoised propeller surface image through a propeller surface global feature encoder to obtain a feature map of the global polishing of the propeller surface. The propeller surface global feature encoder in the present application is a convolutional neural network model that uses a global transformation matrix as a filter. By introducing a global transformation matrix, the convolutional neural network model can globally transform the denoised propeller surface image input to the model, thereby extracting global information from the image. Compared with traditional convolutional neural networks, the model using a global transformation matrix can better capture the global features in the image, and thus can more accurately reflect the overall polishing of the propeller surface.

[0030] Similarly, considering that the global polishing condition feature map of the propeller surface usually has a high dimension, in order to significantly reduce the dimension of the feature, thereby helping to reduce the amount of calculation and storage resources required for subsequent processing, the technical solution of the present application needs to perform maximum pooling on each feature matrix along the channel dimension of the global polishing condition feature map of the propeller surface to obtain a global polishing condition feature vector of the propeller surface. Maximum pooling can retain the most significant eigenvalues ​​in each channel, which usually correspond to the key polishing information of the propeller surface. In this way, the most representative features can be extracted from the global feature map, thereby providing strong support for subsequent fusion and decoding analysis.

[0031] The local polishing condition feature vector of the propeller surface mainly captures the detailed features of the propeller surface, while the global polishing condition feature vector of the propeller surface reflects the overall features of the propeller surface. In order to comprehensively reflect the polishing condition of the propeller surface, including both detailed information and overall information, it is necessary to fuse the local polishing condition feature vector of the propeller surface and the global polishing condition feature vector of the propeller surface to obtain the propeller surface polishing condition feature vector in this application.

[0032] In particular, it is considered that the local polishing condition feature vector of the propeller surface mainly captures the detailed changes of the propeller surface, while the global polishing condition feature vector of the propeller surface focuses on the overall situation of the propeller surface. Since the local polishing condition feature vector and the global polishing condition feature vector focus on different aspects, their expressions in the feature space will naturally be quite different. And in the process of extracting the local and global polishing condition feature vectors, the feature map is subjected to maximum pooling operations respectively. Due to the differences in spatial distribution and importance between local features and global features, this pooling operation will further increase the difference between the two, leading to feature heterogeneity. This feature heterogeneity will cause the feature manifold and feature space to be incompatible, resulting in the fused vector being unable to have the feature manifold smoothness, thereby affecting the certainty of the fused feature expression. Based on this, in the technical solution of the present application, the local polishing condition feature vector of the propeller surface and the global polishing condition feature vector of the propeller surface are adjusted to smooth the feature manifold corresponding to the visual mapping to obtain the propeller surface polishing condition feature vector.

[0033] Specifically, in an embodiment of the present application, the local polishing condition feature vector of the propeller surface and the global polishing condition feature vector of the propeller surface are smoothly adjusted based on the characteristic manifold corresponding to the visual mapping to obtain the propeller surface polishing condition feature vector, including: constructing a propeller surface polishing condition distance topological matrix between the local polishing condition feature vector of the propeller surface and the global polishing condition feature vector of the propeller surface; constructing a propeller surface polishing condition eigenvalue granularity association matrix between the local polishing condition feature vector of the propeller surface and the global polishing condition feature vector of the propeller surface; and The surface polishing condition eigenvalue granularity association matrix is ​​topologically modulated to obtain a propeller surface polishing condition topological modulation association matrix; the propeller surface polishing condition topological modulation association matrix is ​​used as the perspective co-projection space, and the propeller surface local polishing condition eigenvector and the propeller surface global polishing condition eigenvector are projected into the perspective co-projection space to obtain a perspective-modulated propeller surface local polishing condition eigenvector and a perspective-modulated propeller surface global polishing condition eigenvector; the perspective-modulated propeller surface local polishing condition eigenvector and the perspective-modulated propeller surface global polishing condition eigenvector are fused to obtain the propeller surface polishing condition eigenvector.

[0034] More specifically, in the embodiment of the present application, another possible implementation method of performing smooth adjustment on the local polishing condition feature vector of the propeller surface and the global polishing condition feature vector of the propeller surface based on the characteristic manifold corresponding to the visual mapping to obtain the propeller surface polishing condition feature vector may be: processing the local polishing condition feature vector of the propeller surface and the global polishing condition feature vector of the propeller surface according to the following formula to obtain the propeller surface polishing condition feature vector; wherein the formula is: ;in, represents the characteristic vector of the local polishing condition of the propeller surface, represents the feature vector of the global polishing condition of the propeller surface, is the propeller surface polishing condition distance topological matrix between the propeller surface local polishing condition feature vector and the propeller surface global polishing condition feature vector, that is, , The surface polishing condition of the propeller is the distance from the topological matrix The eigenvalues ​​of the positions, The first step of calculating the characteristic vector of the local polishing condition of the propeller surface is represented by The eigenvalue of the position and the eigenvector of the global polishing condition of the propeller surface The distance of the eigenvalue of the position, and and are column vectors, The distance topology matrix representing the polishing condition of the propeller surface, represents the transpose of a vector, represents matrix multiplication, The eigenvalue granularity correlation matrix representing the polishing condition of the propeller surface, represents the natural exponential function, The topological modulation correlation matrix representing the polishing condition of the propeller surface is: represents the feature vector of the local polishing condition of the propeller surface modulated by the viewing angle, represents the feature vector of the global polishing condition of the propeller surface modulated by the viewing angle, and represents the weighted hyperparameter, A feature vector representing the polishing condition of the propeller surface.

[0035] That is, in the technical solution of the present application, the feature vector of the local polishing condition of the propeller surface and the feature vector of the global polishing condition of the propeller surface are smoothly adjusted based on the feature manifold corresponding to the visual mapping, and firstly, a propeller surface polishing condition distance topological matrix between the feature vector of the local polishing condition of the propeller surface and the feature vector of the global polishing condition of the propeller surface is constructed. This step utilizes the idea of ​​graph theory, taking the eigenvalues ​​of each position in the feature vector of the local polishing condition of the propeller surface and the feature vector of the global polishing condition of the propeller surface as nodes, and using the distance metric function to determine the low-dimensional embedding expression of the edges between nodes, so as to construct a feature expression that can quantify the topological correlation relationship between the feature vector of the local polishing condition of the propeller surface and the feature vector of the global polishing condition of the propeller surface. This step is crucial to understanding the spatial distribution of data points and their mutual connections, and also provides basic structural information for subsequent steps.

[0036] Next, a propeller surface polishing condition characteristic value granularity association matrix is ​​constructed between the propeller surface local polishing condition characteristic vector and the propeller surface global polishing condition characteristic vector. This step focuses on exploring the correlation between the internal attributes of the characteristic vectors. In a specific example, the product between the propeller surface local polishing condition characteristic vector and the transposed vector of the propeller surface global polishing condition characteristic vector is calculated to obtain the propeller surface polishing condition characteristic value granularity association matrix. Furthermore, the propeller surface polishing condition characteristic value granularity association matrix is ​​topologically modulated based on the propeller surface polishing condition distance topological matrix to obtain the propeller surface polishing condition topological modulation association matrix. That is, the characteristic value granularity association information of the propeller surface local polishing condition characteristic vector and the propeller surface global polishing condition characteristic vector are driven to wander on the distance topological map to construct a low-dimensional modulated perspective co-projection space domain for mapping the propeller surface local polishing condition characteristic vector and the propeller surface global polishing condition characteristic vector to a continuous high-dimensional regression space attribute.

[0037] Next, the propeller surface polishing condition topological modulation correlation matrix is ​​used as the perspective co-projection space, and the propeller surface local polishing condition feature vector and the propeller surface global polishing condition feature vector are projected into the perspective co-projection space to obtain the perspective modulated propeller surface local polishing condition feature vector and the perspective modulated propeller surface global polishing condition feature vector. That is, the key to this technical solution is to find an implicit third space with modulation capability, in which as much information of the original high-dimensional data as possible is retained, so that the converted new feature vector not only reduces the dimension, but also maintains the relative position relationship between the original features.

[0038] Finally, the local polishing condition feature vector of the propeller surface modulated by the perspective and the global polishing condition feature vector of the propeller surface modulated by the perspective are fused to obtain the propeller surface polishing condition feature vector. Fusion can be achieved in various forms, ranging from simple arithmetic average to complex ensemble learning algorithms, depending on the application scenario and the goals pursued. The final propeller surface polishing condition feature vector combines the main features of the two sets of input data and provides a more comprehensive data representation, which is not only conducive to improving the performance of machine learning models, but also promotes the ability of cross-modal or multi-source data analysis, thereby significantly enhancing the understanding and utilization efficiency of complex data structures.

[0039] In the embodiment of the present application, the propeller grinding processing parameter result generation module 140 is used to decode and obtain the grinding position data and grinding time data that require fine grinding based on the propeller surface grinding condition feature vector. Specifically, Figure 5 FIG. 1 is a block diagram of a propeller grinding processing parameter result generation module in an adaptive control system for propeller processing according to an embodiment of the present application. Figure 5 As shown, the propeller grinding processing parameter result generation module 140 includes: a grinding position result decoding unit 141, which is used to pass the propeller surface grinding condition feature vector through a grinding position decoder to obtain a grinding position decoding value, and the grinding position decoding value is used to represent the grinding position data that requires fine grinding; a grinding time result decoding unit 142, which is used to pass the propeller surface grinding condition feature vector through a grinding time decoder to obtain a grinding time decoding value, and the grinding time decoding value is used to represent the grinding time data that requires fine grinding.

[0040] Although the propeller surface polishing condition feature vector contains rich propeller surface condition information, this information exists in the form of a vector and cannot be directly used to guide the polishing operation. In order to decode the information in the feature vector into specific polishing position data and realize the mapping of features to positions, the propeller surface polishing condition feature vector is passed through a polishing position decoder in the technical solution of the present application to obtain a polishing position decoding value, and the polishing position decoding value is used to represent the polishing position data that requires fine polishing. The polishing position decoding value directly indicates the position that requires fine polishing. In this way, the polishing equipment can achieve fine polishing at the position that requires polishing, which helps to improve the polishing effect.

[0041] The polishing conditions of the propeller surface are complex and changeable, and the polishing conditions in different areas may be different. In order to ensure that the polishing is neither excessive nor insufficient, the propeller surface polishing condition feature vector needs to be passed through a polishing time decoder to obtain a polishing time decoding value, and the polishing time decoding value is used to represent the polishing time data that requires fine polishing. The polishing time decoder is an algorithm model based on deep learning technology, which is specifically used to analyze the propeller surface polishing condition feature vector and output the polishing time data required for the area that needs to be polished. This decoder is usually trained and optimized with a large amount of training data, and can accurately convert the information in the feature vector into specific polishing time data. Through the polishing time decoder, it can be ensured that the area that requires fine polishing receives appropriate polishing treatment, thereby achieving the expected polishing effect.

[0042] In summary, the adaptive control system 100 for propeller processing based on the embodiment of the present application is explained, which collects real-time surface images in the propeller grinding process, and uses computer vision-based data analysis technology to reduce noise and extract features from the real-time surface images, thereby understanding the propeller surface grinding situation, and finally adaptively controls the grinding position and grinding time based on the propeller surface grinding situation. In this way, accurate grinding of the propeller surface can be achieved, which is conducive to improving the grinding quality.

[0043] As described above, the adaptive control system 100 for propeller processing according to the embodiment of the present application can be implemented in various terminal devices, such as a server for adaptive control of propeller processing. In one example, the adaptive control system 100 for propeller processing according to the embodiment of the present application can be integrated into the terminal device as a software module and / or a hardware module. For example, the adaptive control system 100 for propeller processing can be a software module in the operating system of the terminal device, or can be an application developed for the terminal device; of course, the adaptive control system 100 for propeller processing can also be one of the many hardware modules of the terminal device.

[0044] Alternatively, in another example, the propeller processing adaptive control system 100 and the terminal device may also be separate devices, and the propeller processing adaptive control system 100 may be connected to the terminal device via a wired and / or wireless network and transmit interactive information in accordance with an agreed data format.

[0045] Figure 6 FIG. 1 is a flow chart of an adaptive control method for propeller machining according to an embodiment of the present application. Figure 6As shown, in the adaptive control method for propeller processing, the method includes: S110, acquiring a real-time surface image of the propeller in the grinding process; S120, performing noise reduction processing on the real-time surface image to obtain a noise-reduced propeller surface image; S130, performing image content encoding on the noise-reduced propeller surface image to obtain a propeller surface grinding condition feature vector; S140, based on the propeller surface grinding condition feature vector, decoding to obtain grinding position data and grinding time data that require fine grinding.

[0046] Here, those skilled in the art will appreciate that the specific operations of each step in the above-mentioned propeller processing adaptive control method have been described in the above reference. Figures 1 to 5 The adaptive control system of the propeller processing has been described in detail, and therefore, its repeated description will be omitted.

[0047] In summary, the adaptive control method for propeller processing based on the embodiment of the present application is explained, which collects real-time surface images in the propeller grinding process, and uses computer vision-based data analysis technology to reduce noise and extract features from the real-time surface images, thereby understanding the propeller surface grinding situation, and finally adaptively controlling the grinding position and grinding time based on the propeller surface grinding situation. In this way, accurate grinding of the propeller surface can be achieved, which is conducive to improving the grinding quality.

[0048] Finally, the present application also relates to a computer-readable storage medium for implementing an adaptive control method for propeller processing according to one or more embodiments of the present application. The computer-readable storage medium mentioned herein includes various types of computer storage media, which may be any available medium that can be accessed by a general-purpose or special-purpose computer. For example, a computer-readable storage medium may include RAM, ROM, EPROM, E2PROM, registers, hard disks, removable disks, CD-ROMs or other optical disk storage, disk storage or other magnetic storage devices, or any other temporary or non-temporary medium that can be used to carry or store desired program code units in the form of instructions or data structures and can be accessed by a general-purpose or special-purpose computer, or a general-purpose or special-purpose processor. The description of the computer-readable storage medium according to the present application can refer to the explanation of the adaptive control method for propeller processing according to the present application, and no further description will be given.

[0049] The above is only a preferred embodiment of the present application and does not constitute any form of limitation to the present application. Although the present application has been disclosed as a preferred embodiment as above, it is not intended to limit the present application. Any technician familiar with the profession can make some changes or modifications to equivalent embodiments of the technical contents disclosed above without departing from the scope of the technical solution of the present application. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present application without departing from the content of the technical solution of the present application still falls within the scope of the technical solution of the present application.

Claims

1. An adaptive control system for propeller processing, characterized in that: include: Propeller grinding image acquisition module, used to obtain real-time surface images of the propeller during the grinding process; A propeller grinding image denoising module, used for performing denoising on the real-time surface image to obtain a denoised propeller surface image; A propeller grinding image content coding module, used for performing image content coding on the denoised propeller surface image to obtain a propeller surface grinding condition feature vector; The propeller grinding processing parameter result generation module is used to decode and obtain the grinding position data and grinding time data that need fine grinding based on the propeller surface grinding condition feature vector.

2. The adaptive control system for propeller machining according to claim 1, characterized in that: The propeller grinding image denoising module is used to: pass the real-time surface image through a propeller surface image denoiser to obtain the denoised propeller surface image.

3. The adaptive control system for propeller machining according to claim 2, characterized in that: The propeller grinding image content encoding module includes: A propeller surface image local feature extraction unit, used for performing image local feature extraction on the denoised propeller surface image to obtain a feature vector of the local polishing condition of the propeller surface; A propeller surface image global feature extraction unit, used for performing image global feature extraction on the denoised propeller surface image to obtain a propeller surface global polishing condition feature vector; The propeller surface feature fusion unit is used to perform smoothing adjustment on the local polishing condition feature vector of the propeller surface and the global polishing condition feature vector of the propeller surface based on the feature manifold corresponding to the visual mapping to obtain the propeller surface polishing condition feature vector.

4. The adaptive control system for propeller machining according to claim 3, characterized in that: The propeller surface image local feature extraction unit comprises: A propeller surface local polishing condition feature generating subunit, used for passing the noise-reduced propeller surface image through a propeller surface local feature encoder to obtain a propeller surface local polishing condition feature map; The propeller surface local polishing condition feature dimensionality reduction subunit is used to perform maximum pooling on each feature matrix along the channel dimension of the propeller surface local polishing condition feature map to obtain the propeller surface local polishing condition feature vector.

5. The adaptive control system for propeller machining according to claim 4, characterized in that: The propeller surface image global feature extraction unit comprises: A propeller surface global polishing condition feature generating subunit, used for passing the noise-reduced propeller surface image through a propeller surface global feature encoder to obtain a propeller surface global polishing condition feature map; The propeller surface global polishing condition feature dimensionality reduction subunit is used to perform maximum pooling on each feature matrix along the channel dimension of the propeller surface global polishing condition feature map to obtain the propeller surface global polishing condition feature vector.

6. The adaptive control system for propeller machining according to claim 5, characterized in that: The propeller surface image denoiser is an image denoiser based on an automatic codec, the propeller surface local feature encoder is a convolutional neural network model using a two-dimensional convolution kernel, and the propeller surface global feature encoder is a convolutional neural network model using a global transformation matrix as a filter.

7. The adaptive control system for propeller machining according to claim 6, characterized in that: The propeller surface feature fusion unit is used to: Constructing a propeller surface polishing condition distance topological matrix between the propeller surface local polishing condition feature vector and the propeller surface global polishing condition feature vector; Constructing a propeller surface polishing condition characteristic value granularity correlation matrix between the propeller surface local polishing condition characteristic vector and the propeller surface global polishing condition characteristic vector; Based on the propeller surface polishing condition distance topological matrix, the propeller surface polishing condition eigenvalue granularity association matrix is ​​topologically modulated to obtain a propeller surface polishing condition topological modulation association matrix; Taking the propeller surface polishing condition topological modulation correlation matrix as the perspective co-projection space, projecting the propeller surface local polishing condition feature vector and the propeller surface global polishing condition feature vector into the perspective co-projection space to obtain the perspective-modulated propeller surface local polishing condition feature vector and the perspective-modulated propeller surface global polishing condition feature vector; The local polishing condition feature vector of the view-angle-modulated propeller surface and the global polishing condition feature vector of the view-angle-modulated propeller surface are fused to obtain the propeller surface polishing condition feature vector.

8. The adaptive control system for propeller machining according to claim 7, characterized in that: The propeller grinding processing parameter result generating module includes: A grinding position result decoding unit, used for passing the propeller surface grinding condition feature vector through a grinding position decoder to obtain a grinding position decoding value, wherein the grinding position decoding value is used to represent grinding position data that requires fine grinding; The polishing time result decoding unit is used to pass the propeller surface polishing condition feature vector through a polishing time decoder to obtain a polishing time decoding value, and the polishing time decoding value is used to represent the polishing time data that requires fine polishing.

9. An adaptive control method for propeller processing, characterized in that: include: Get real-time surface images of the propeller during the grinding process; Performing noise reduction processing on the real-time surface image to obtain a noise-reduced propeller surface image; Performing image content encoding on the denoised propeller surface image to obtain a propeller surface polishing condition feature vector; Based on the propeller surface polishing condition feature vector, the polishing position data and polishing time data requiring fine polishing are decoded and obtained.

10. The adaptive control method for propeller machining according to claim 9, characterized in that: Based on the propeller surface polishing condition feature vector, the polishing position data and polishing time data that need fine polishing are decoded and obtained, including: Passing the propeller surface polishing condition feature vector through a polishing position decoder to obtain a polishing position decoding value, wherein the polishing position decoding value is used to represent polishing position data that requires fine polishing; The propeller surface polishing condition feature vector is passed through a polishing time decoder to obtain a polishing time decoding value, and the polishing time decoding value is used to represent polishing time data that requires fine polishing.

Images