A method and system for detecting and controlling mold frame accuracy with rapid positioning

Abstract

An embodiment of the present invention provides a method and system for detecting and controlling the accuracy of a fast-positioning mold frame, comprising: obtaining a first front initial image and a second front initial image of an upper mold frame and a lower mold frame; performing super-resolution processing on the local image of the guide post and the local image of the guide sleeve through a constructed double-layer super-resolution processing model, positioning the local image of the high-resolution guide post and the local image of the high-resolution guide sleeve to obtain guide post features and guide sleeve features; performing depth-separable convolution and depth-convolution operations on the guide post features and the guide sleeve features to perform positioning reconstruction, calculating the positioning accuracy of the mold frame through the final guide post positioning map and the final guide sleeve positioning map, separating the local image of the positioning device, and then obtaining the final positioning device image by optimizing multiple positioning of the image, and then identifying the size and accuracy of the positioning device, thereby saving precision detection costs, reducing detection time, and improving detection efficiency.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a method for detecting and controlling the accuracy of a fast-positioning mold frame, a detection and control system for the accuracy of a fast-positioning mold frame, a computer device, and a storage medium. Background Art

[0002] The mold frame is the supporting component of the mold during the processing. The mold is installed on the mold frame, and the upper and lower combinations are used to enable processing inside the mold. The mold frame also plays a positioning role in the mold processing process, so high precision requirements are required. The current frame precision is generally tested by multiple inspection tools to detect the verticality, parallelism, concentricity, etc. between the two templates. The relative precision detection is difficult, the detection time is long, the detection efficiency is low, and the detection cost is relatively high. Summary of the Invention

[0003] In view of the above problems, embodiments of the present invention are proposed to provide a method for detecting and controlling the accuracy of a fast-positioning mold frame, a detection and control system for the accuracy of a fast-positioning mold frame, a computer device, and a storage medium that overcome the above problems or at least partially solve the above problems.

[0004] In order to solve the above problems, an embodiment of the present invention discloses a method for detecting and controlling the accuracy of a mold frame for rapid positioning, comprising:

[0005] Acquire a first front initial image and a second front initial image of the upper mold frame and the lower mold frame;

[0006] Extracting a partial image of the guide post within a preset range of the guide post of the upper mold frame from the first front initial image;

[0007] Extracting a partial image of the guide sleeve of the lower mold frame within a preset range of the guide sleeve from the second front initial image;

[0008] The guide post local image and the guide sleeve local image are subjected to super-resolution processing by a constructed double-layer super-resolution processing model to obtain a high-resolution guide post local image and a high-resolution guide sleeve local image;

[0009] Positioning the high-resolution guide post local image and the high-resolution guide sleeve local image to obtain guide post features and guide sleeve features;

[0010] Performing depthwise separable convolution and depthwise convolution operations on the guide post features and the guide sleeve features to perform positioning reconstruction, thereby obtaining a final guide post positioning map and a final guide sleeve positioning map;

[0011] The positioning accuracy of the mold base is calculated based on the final guide column positioning diagram and the final guide sleeve positioning diagram.

[0012] Preferably, extracting a partial image of the guide column within a preset range of the guide column of the upper mold frame from the first front initial image includes:

[0013] The first front initial image is input into the deep backbone network, the final convolution layer of the deep backbone network is adapted using the dilated convolution, and then the maximum pooling operation is performed to obtain a local image of the guide post.

[0014] Preferably, the super-resolution processing of the guide post local image and the guide sleeve local image by a constructed double-layer super-resolution processing model to obtain a high-resolution guide post local image and a high-resolution guide sleeve local image includes:

[0015] A guide post super-resolution processing module is constructed through a plurality of first convolutional layers, first bicubic interpolation layers, first deconvolutional layers, and second bicubic interpolation layers;

[0016] A guide super-resolution processing module is constructed through multiple second convolution layers, third bicubic interpolation layers, second deconvolution layers and fourth bicubic interpolation layers;

[0017] The guide post super-resolution processing module and the guide sleeve super-resolution processing module are constructed into a double-layer super-resolution processing model;

[0018] Inputting the guide post local image into a guide post super-resolution processing module, sequentially passing through a plurality of first convolution layers, a first bicubic interpolation layer, a first deconvolution layer, and a second bicubic interpolation layer, to obtain an output high-resolution guide post local image;

[0019] The guide sleeve partial image is input into the guide sleeve partial image, and sequentially passes through multiple second convolution layers, third bicubic interpolation layers, second deconvolution layers and fourth bicubic interpolation layers to obtain an output high-resolution guide sleeve partial image.

[0020] Preferably, the positioning of the high-resolution guide post local image and the high-resolution guide sleeve local image to obtain the guide post features and the guide sleeve features includes:

[0021] The high-resolution guide post partial image and the high-resolution guide sleeve partial image are subjected to the CBAM attention mechanism to adjust the weight parameters between the feature maps, thereby extracting a plurality of guide post image sub-features and a plurality of guide sleeve image sub-features;

[0022] Performing feature fusion on the multiple guide post image sub-features and the multiple guide sleeve image sub-features to obtain fused guide post image sub-features and fused guide sleeve image sub-features;

[0023] Inputting the fused guide post image sub-feature and the fused guide sleeve image sub-feature into a prediction layer with a set loss function for first positioning, thereby obtaining a positioning guide post image sub-feature and a positioning guide sleeve image sub-feature;

[0024] The edge features and geometric features of the positioning guide post image sub-features and the positioning guide sleeve image sub-features are extracted, and a second positioning is performed based on the edge features and geometric features to obtain the guide post features and the guide sleeve features.

[0025] Preferably, performing depthwise separable convolution and depthwise convolution operations on the guide post features and the guide sleeve features to perform positioning reconstruction to obtain a final guide post positioning map and a final guide sleeve positioning map includes:

[0026] Converting the guide post feature and the guide sleeve feature into a multi-channel vector;

[0027] Performing a spatial convolution operation on the features of each channel in the multi-channel vector and a convolution kernel to obtain a feature map, and converting the feature map into a recombined vector;

[0028] The reorganized vector is subjected to a depthwise convolution operation and a residual connection operation to obtain a final guide post positioning map and a final guide sleeve positioning map.

[0029] Preferably, the calculation of the positioning accuracy of the mold base by using the final guide column positioning diagram and the final guide sleeve positioning diagram includes:

[0030] Mark out the dimensional data of the final guide column positioning diagram and the final guide sleeve positioning diagram respectively;

[0031] The dimensional tolerance accuracy data of the guide column and guide sleeve of the mold base are calculated based on the dimensional data.

[0032] The embodiment of the present invention discloses a rapid positioning mold frame accuracy detection and control system, comprising:

[0033] An initial image acquisition module is used to acquire a first front initial image and a second front initial image of the upper mold frame and the lower mold frame;

[0034] A first extraction module is configured to extract a partial image of the guide post within a preset range of the guide post of the upper mold frame from the first front initial image;

[0035] A second extraction module is used to extract a partial image of the guide sleeve of the lower mold frame within a preset range of the guide sleeve from the second front initial image;

[0036] a super-resolution processing module for performing super-resolution processing on the guide post local image and the guide sleeve local image through a constructed double-layer super-resolution processing model to obtain a high-resolution guide post local image and a high-resolution guide sleeve local image;

[0037] A positioning module, configured to perform positioning based on the high-resolution guide post local image and the high-resolution guide sleeve local image to obtain features of the guide post and the guide sleeve;

[0038] A positioning reconstruction module is used to perform depthwise separable convolution and depthwise convolution operations on the guide post features and the guide sleeve features to perform positioning reconstruction, thereby obtaining a final guide post positioning map and a final guide sleeve positioning map;

[0039] A calculation module is used to calculate the positioning accuracy of the mold base through the final guide column positioning diagram and the final guide sleeve positioning diagram.

[0040] Preferably, the first extraction module includes:

[0041] The acquisition submodule is used to input the first front initial image into the deep backbone network, use the void convolution to adapt the final convolution layer of the deep backbone network, and then perform the maximum pooling operation to obtain the local image of the guide column.

[0042] An embodiment of the present invention further discloses a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the above-mentioned steps of detecting and controlling the accuracy of the rapid positioning mold frame when executing the computer program.

[0043] The embodiment of the present invention further discloses a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the computer program implements the above-mentioned steps of detecting and controlling the accuracy of the mold frame for rapid positioning.

[0044] The embodiments of the present invention include the following advantages:

[0045] In an embodiment of the present invention, the method includes acquiring a first front initial image and a second front initial image of an upper mold frame and a lower mold frame; extracting a local image of a guide post within a preset range of the guide post of the upper mold frame from the first front initial image; extracting a local image of a guide sleeve within a preset range of the guide sleeve of the lower mold frame from the second front initial image; performing super-resolution processing on the local image of the guide post and the local image of the guide sleeve through a constructed double-layer super-resolution processing model to obtain a high-resolution local image of the guide post and a high-resolution local image of the guide sleeve; positioning the high-resolution local image of the guide post and the high-resolution local image of the guide sleeve to obtain guide post features and Guide sleeve features; performing depth-separable convolution and depth-convolution operations on the guide post features and the guide sleeve features for positioning reconstruction to obtain a final guide post positioning map and a final guide sleeve positioning map; calculating the positioning accuracy of the mold frame through the final guide post positioning map and the final guide sleeve positioning map, separating the local image of the positioning device, and then obtaining an optimized image by improving the resolution of the local image, achieving high image recovery, and providing a good foundation for image feature positioning; and then obtaining the final positioning device image through multiple positioning of the optimized image, and then identifying the size and accuracy of the positioning device, saving precision detection costs, reducing detection time, and improving detection efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0047] Figure 1 This is a flowchart of a method for detecting and controlling the accuracy of a mold frame for rapid positioning according to an embodiment of the present invention;

[0048] Figure 2 1 is a schematic diagram of an upper mold frame according to an embodiment of the present invention;

[0049] Figure 3 1 is a schematic diagram of a lower mold frame according to an embodiment of the present invention;

[0050] Figure 4 1 is a schematic structural diagram of a double-layer super-resolution processing model according to an embodiment of the present invention;

[0051] Figure 5 This is a structural block diagram of an embodiment of a rapid positioning mold frame accuracy detection and control system according to an embodiment of the present invention;

[0052] Figure 6 The diagram is an internal structural diagram of a computer device according to an embodiment. DETAILED DESCRIPTION

[0053] In order to make the technical problems, technical solutions and beneficial effects solved by the embodiments of the present invention more clearly understood, the embodiments of the present invention are further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0054] In an embodiment of the present invention, the image of the positioning device on the mold frame is subjected to dimensional accuracy recognition. Specifically, the local image of the positioning device is first separated, and then the optimized image is obtained by improving the resolution of the local image, thereby achieving high image recovery and providing a good basis for image feature positioning. Then, multiple positioning of the optimized image is performed, including the first positioning and fine positioning, positioning reconstruction, etc., to obtain the final positioning device image, and then the size and accuracy of the positioning device are identified, thereby saving the cost of precision detection, reducing detection time, and improving detection efficiency.

[0055] Reference Figure 1 , shows a flowchart of a method for detecting and controlling the accuracy of a mold frame for rapid positioning according to an embodiment of the present invention, which may specifically include the following steps:

[0056] Step 101: Acquire a first front initial image and a second front initial image of the upper mold frame and the lower mold frame;

[0057] In the embodiment of the present invention, the mold frame includes an upper mold frame and a lower mold frame. Figure 2 and Figure 3 As shown, the components in the upper mold frame include guide columns, and the components in the lower mold frame include guide sleeves. Of course, the upper mold frame and the lower mold frame can also include specific components, such as mold bases, molds, etc., and the embodiments of the present invention do not impose too many restrictions on this.

[0058] In the embodiment of the present invention, firstly, a first front initial image of the upper mold frame and a second front initial image of the lower mold frame are acquired;

[0059] The detection and control method in the embodiment of the present invention can be applied to terminals, and the terminals can be but are not limited to various personal computers, laptops, smart phones, tablet computers and portable wearable devices. The embodiment of the present invention does not limit the specific type of the terminal. The operating system of the terminal may include Android, Harmony OS, IOS, WindowsPhone, Windows, etc. The present invention does not impose too many restrictions on this. The terminal is connected to a camera, and the camera is used to obtain a first front initial image and a second front initial image. The camera can be provided with a fill light, and the image is acquired when the fill light is turned on to obtain a 1080p or 720p image. The shooting parameters of the camera can be set as follows: aperture: F8, shutter speed: 1 / 250s, ISO: 150. The above shooting parameters are only an example of the embodiment of the present invention. Other shooting parameters can also be set, and the embodiment of the present invention does not impose too many restrictions on this.

[0060] The first front initial image refers to an initial image perpendicular to the guide column direction of the upper mold frame, and the second front initial image refers to an initial image perpendicular to the guide sleeve direction of the lower mold frame.

[0061] Step 102: extracting a partial image of the guide post within a preset range of the guide post of the upper mold frame from the first front initial image;

[0062] In the embodiment of the present invention, first, a partial image of the predetermined area including the guide post is extracted from the initial front image. For example, the partial image of the guide post is a partial image of the guide post extending outward by 0.5 cm in radius.

[0063] Specifically, in an embodiment of the present invention, extracting a partial image of the guide post within a preset range of the guide post of the upper mold frame from the first front initial image includes:

[0064] The first front initial image is input into the deep backbone network, and the final convolution layer of the deep backbone network is adapted using the dilated convolution, followed by a 2x2 maximum pooling operation to obtain a local image of the guide post.

[0065] Specifically, the deep backbone network can be a deep convolutional neural network, which is constructed by stacking multiple smaller convolutional layers, fully connected layers, and pooling layers, which can reduce the size of the image and retain the most significant features. The training method of the deep backbone network can be to pre-input the historical first front initial image, the historical guide column partial image, the historical second front initial image, and the historical guide sleeve partial image as training samples for training respectively to obtain a trained deep backbone network, and then input the new first front initial image into the trained deep backbone network to extract the output guide column partial image and guide sleeve partial image;

[0066] Among them, dilated convolution adaptation can be added to the final convolution layer to reduce the information loss rate. Dilated convolution introduces an expansion rate to enable the convolution kernel to obtain a larger receptive field, thereby improving the accuracy and timeliness of image feature extraction.

[0067] Step 103: extracting a partial image of the guide sleeve within a preset range of the guide sleeve of the lower mold base from the second front initial image;

[0068] Further applied to the embodiment of the present invention, a partial image of the guide sleeve within a preset range of the guide sleeve of the lower mold frame can be extracted from the second front initial image. First, a partial image of the guide sleeve within the preset area is extracted from the front initial image. For example, the partial image of the guide sleeve is a partial image extending outward from the radius of the guide sleeve by 0.5 cm.

[0069] Similarly, extracting a partial image of the guide sleeve within a preset range of the guide sleeve of the lower mold frame from the second front initial image includes:

[0070] The second front initial image is input into the deep backbone network, and the final convolution layer of the deep backbone network is adapted using the dilated convolution, and then the maximum pooling operation is performed to obtain the local image of the guide sleeve. The specific implementation steps are the same as the previous step 102, and the embodiment of the present invention will not be elaborated on this.

[0071] Step 104: Super-resolution processing is performed on the guide post local image and the guide sleeve local image using the constructed double-layer super-resolution processing model to obtain a high-resolution guide post local image and a high-resolution guide sleeve local image;

[0072] After obtaining the local image of the guide post and the local image of the guide sleeve, the local image of the guide post and the local image of the guide sleeve are respectively input into the two submodules of the constructed double-layer super-resolution processing model for super-resolution processing to improve the quality of the extracted image.

[0073] In actual application in the embodiment of the present invention, the guide post local image and the guide sleeve local image are subjected to super-resolution processing by a constructed double-layer super-resolution processing model to obtain a high-resolution guide post local image and a high-resolution guide sleeve local image, including:

[0074] A guide post super-resolution processing module is constructed through a plurality of first convolutional layers, first bicubic interpolation layers, first deconvolutional layers, and second bicubic interpolation layers;

[0075] A guide super-resolution processing module is constructed through multiple second convolution layers, third bicubic interpolation layers, second deconvolution layers and fourth bicubic interpolation layers;

[0076] The guide post super-resolution processing module and the guide sleeve super-resolution processing module are constructed into a double-layer super-resolution processing model;

[0077] like Figure 4 As shown, the guide post super-resolution processing module is used to process the guide post local image, while the guide sleeve super-resolution processing module is used to process the guide sleeve local image; the guide post local image and the guide sleeve local image are processed simultaneously to achieve parallel processing of image data and improve image data processing efficiency;

[0078] Specifically, the first convolutional layer includes six 3×3 convolution kernels, and the first bicubic interpolation layer includes a data module that uses 16 neighboring pixels around the target pixel for weighted calculation. The function of the first bicubic interpolation layer is to improve the ability to restore high-frequency details of the image, and the first deconvolution layer includes six 3×3 deconvolution kernels to achieve image restoration and reconstruction. Finally, a second bicubic interpolation layer is passed to further balance data complexity and image pixels, thereby improving the image restoration details. It should be noted that the second convolutional layer, the third bicubic interpolation layer, the second deconvolution layer, and the fourth bicubic interpolation layer are arranged corresponding to the first convolution layer, the first bicubic interpolation layer, the first deconvolution layer, and the second bicubic interpolation layer, and this embodiment of the present invention will not be described in detail.

[0079] Inputting the guide post local image into a guide post super-resolution processing module, sequentially passing through a plurality of first convolution layers, a first bicubic interpolation layer, a first deconvolution layer, and a second bicubic interpolation layer, to obtain an output high-resolution guide post local image;

[0080] For example, a local image of a guide post is converted into a feature vector, and the feature vector is processed through multiple first convolution layers, first bicubic interpolation layers, first deconvolution layers, and second bicubic interpolation layers to obtain a high-resolution local image of the guide post.

[0081] On the other hand, the guide sleeve local image is input into the guide sleeve local image, and can also pass through multiple second convolution layers, third bicubic interpolation layers, second deconvolution layers and fourth bicubic interpolation layers in sequence to obtain an output high-resolution guide sleeve local image.

[0082] Step 105, positioning the high-resolution guide post partial image and the high-resolution guide sleeve partial image to obtain guide post features and guide sleeve features;

[0083] Further applied to the embodiment of the present invention, after obtaining the high-resolution partial image of the guide post and the high-resolution partial image of the guide sleeve, the guide post and the guide sleeve can be positioned based on the high-resolution partial image of the guide post and the high-resolution partial image of the guide sleeve;

[0084] Specifically, positioning the high-resolution guide post local image and the high-resolution guide sleeve local image to obtain the guide post features and the guide sleeve features includes:

[0085] The high-resolution guide post partial image and the high-resolution guide sleeve partial image are subjected to the CBAM (Convolutional Block Attention Module) attention mechanism to adjust the weight parameters between the feature maps, thereby extracting multiple guide post image sub-features and multiple guide sleeve image sub-features;

[0086] Among them, the CBAM attention mechanism includes a channel attention module and a spatial attention module. The CBAM attention mechanism is used to adjust the weight parameters between feature maps, and then the local features and global features are fused through a 3×3 convolutional layer, a Transformer block and a 1×1 convolutional layer to obtain multiple guide post image sub-features and multiple guide sleeve image sub-features.

[0087] Performing feature fusion on the multiple guide post image sub-features and the multiple guide sleeve image sub-features to obtain fused guide post image sub-features and fused guide sleeve image sub-features;

[0088] Furthermore, in the embodiment of the present invention, the feature fusion method is global pooling, and the multiple guide post image sub-features are subjected to global average pooling operations to obtain fused guide post image sub-features;

[0089] In the embodiment of the present invention, global pooling reduces the spatial dimension of the feature map to a single value by averaging or maximizing the value, which can be used to retain global features and reduce the number of parameters.

[0090] Inputting the fused guide post image sub-feature and the fused guide sleeve image sub-feature into a prediction layer with a set loss function for first positioning, thereby obtaining a positioning guide post image sub-feature and a positioning guide sleeve image sub-feature;

[0091] Furthermore, the fused guide post image sub-feature is input into the prediction layer with the loss function set to perform the first positioning, thereby obtaining the positioning guide post image sub-feature;

[0092] The fused guide sleeve image sub-feature is input into the prediction layer with the loss function set to perform the first positioning, thereby obtaining the positioning guide sleeve image sub-feature;

[0093] Specifically applied to the embodiment of the present invention, the loss function is as follows:

[0094] ;in, represents the loss function, Indicates the distance between the two center points of the fused guide column image, Indicates the distance between the two center points of the fused guide sleeve image, , , respectively represent the center points of the fused guide column image prediction frame and the sample frame, , , respectively represent the center points of the fusion guide image prediction frame and the sample frame, represents the diameter of the enclosing circle, Represents the similarity of the edge aspect ratios of the fused guide sleeve image and the fused guide sleeve image sub-features; where, ;in, Indicates the width of the fused guide pillar image, Indicates the height of the fused guide post image, Indicates the width of the fused guide sleeve image, Indicates the height of the fused guide sleeve image. In the embodiment of the present invention, a loss function is provided for predicting the guide post image and the guide sleeve image, and the data of the guide post and the guide sleeve are used to compare with each other to improve the prediction accuracy.

[0095] Furthermore, edge features and geometric features of the positioning guide post image sub-features and the positioning guide sleeve image sub-features are extracted, and a second positioning is performed based on the edge features and geometric features to obtain guide post features and guide sleeve features.

[0096] Taking the edge features and geometric features as the starting point, a 20×20 square detection operator is set in each of the four directions, and the mean of the detection area with the edge features and geometric features as the starting point is calculated. If it is within the predicted grayscale threshold range, it means that the detected area is already at the edge. Then the center coordinates of the rectangle are calculated as the center of the circle for the second positioning, and the area composed of the edge and the center of the circle is obtained, that is, the guide column feature and guide sleeve feature can be obtained.

[0097] Step 106, performing depthwise separable convolution and depthwise convolution operations on the guide post features and the guide sleeve features to perform positioning reconstruction, thereby obtaining a final guide post positioning map and a final guide sleeve positioning map;

[0098] After obtaining the guide post features and guide sleeve features, they need to be subjected to depthwise separable convolution and depthwise convolution operations for positioning and reconstruction to obtain the final guide post positioning map and the final guide sleeve positioning map.

[0099] Specifically, in an embodiment of the present invention, the positioning reconstruction is performed by performing depthwise separable convolution and depthwise convolution operations on the guide post features and the guide sleeve features to obtain the final guide post positioning map and the final guide sleeve positioning map, including:

[0100] Converting the guide post feature and the guide sleeve feature into a multi-channel vector;

[0101] Performing a spatial convolution operation on the features of each channel in the multi-channel vector and a convolution kernel to obtain a feature map, and converting the feature map into a recombined vector;

[0102] The reorganized vector is subjected to a depthwise convolution operation and a residual connection operation to obtain a final guide post positioning map and a final guide sleeve positioning map.

[0103] Step 107 : Calculate the positioning accuracy of the mold base using the final guide post positioning diagram and the final guide sleeve positioning diagram.

[0104] In an embodiment of the present invention, the positioning accuracy of the mold frame is calculated using the final guide post positioning diagram and the final guide sleeve positioning diagram, including: marking the dimensional data of the final guide post positioning diagram and the final guide sleeve positioning diagram respectively; calculating the dimensional tolerance accuracy data of the guide post and the guide sleeve of the mold frame using the dimensional data, the dimensional tolerance accuracy data of the guide post diameter can be 0.003~0.004mm or 0.004~0.005mm, etc., and the dimensional tolerance accuracy data of the guide sleeve diameter can be 0.004~0.005 mm or 0.005~0.007 mm, etc.

[0105] In an embodiment of the present invention, the method includes acquiring a first front initial image and a second front initial image of an upper mold frame and a lower mold frame; extracting a local image of a guide post within a preset range of the guide post of the upper mold frame from the first front initial image; extracting a local image of a guide sleeve within a preset range of the guide sleeve of the lower mold frame from the second front initial image; performing super-resolution processing on the local image of the guide post and the local image of the guide sleeve through a constructed double-layer super-resolution processing model to obtain a high-resolution local image of the guide post and a high-resolution local image of the guide sleeve; positioning the high-resolution local image of the guide post and the high-resolution local image of the guide sleeve to obtain guide post features and Guide sleeve features; performing depth-separable convolution and depth-convolution operations on the guide post features and the guide sleeve features for positioning reconstruction to obtain a final guide post positioning map and a final guide sleeve positioning map; calculating the positioning accuracy of the mold frame through the final guide post positioning map and the final guide sleeve positioning map, separating the local image of the positioning device, and then obtaining an optimized image by improving the resolution of the local image, achieving high image recovery, and providing a good foundation for image feature positioning; and then obtaining the final positioning device image through multiple positioning of the optimized image, and then identifying the size and accuracy of the positioning device, saving precision detection costs, reducing detection time, and improving detection efficiency.

[0106] It should be noted that for the sake of simplicity, the method embodiments are described as a series of actions. However, those skilled in the art should be aware that the embodiments of the present invention are not limited by the order of the actions described, because according to the embodiments of the present invention, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in this specification are all preferred embodiments, and the actions involved are not necessarily required by the embodiments of the present invention.

[0107] Reference Figure 5 , shows a structural block diagram of an embodiment of a rapid positioning mold frame accuracy detection and control system according to an embodiment of the present invention, which may specifically include the following modules:

[0108] The initial image acquisition module 301 is used to acquire a first front initial image and a second front initial image of the upper mold frame and the lower mold frame;

[0109] A first extraction module 302 is configured to extract a partial image of the guide post within a preset range of the guide post of the upper mold frame from the first front initial image;

[0110] A second extraction module 303 is configured to extract a partial image of the guide sleeve within a preset range of the guide sleeve of the lower mold frame from the second front initial image;

[0111] The super-resolution processing module 304 is configured to perform super-resolution processing on the guide post local image and the guide sleeve local image by using the constructed double-layer super-resolution processing model to obtain a high-resolution guide post local image and a high-resolution guide sleeve local image;

[0112] A positioning module 305 is used to perform positioning based on the high-resolution guide post partial image and the high-resolution guide sleeve partial image to obtain guide post features and guide sleeve features;

[0113] A positioning reconstruction module 306 is configured to perform depthwise separable convolution and depthwise convolution operations on the guide post features and the guide sleeve features to perform positioning reconstruction, thereby obtaining a final guide post positioning map and a final guide sleeve positioning map.

[0114] The calculation module 307 is used to calculate the positioning accuracy of the mold base according to the final guide column positioning diagram and the final guide sleeve positioning diagram.

[0115] Preferably, the first extraction module includes:

[0116] The acquisition submodule is used to input the first front initial image into the deep backbone network, use the void convolution to adapt the final convolution layer of the deep backbone network, and then perform the maximum pooling operation to obtain the local image of the guide column.

[0117] Preferably, the super-resolution processing module includes:

[0118] A first construction submodule is used to construct a guide post super-resolution processing module through a plurality of first convolution layers, first bicubic interpolation layers, first deconvolution layers and second bicubic interpolation layers;

[0119] A second construction submodule is used to construct a guide sleeve super-resolution processing module through multiple second convolution layers, third bicubic interpolation layers, second deconvolution layers and fourth bicubic interpolation layers;

[0120] A third construction submodule is used to construct the guide post super-resolution processing module and the guide sleeve super-resolution processing module into a double-layer super-resolution processing model;

[0121] A first input submodule is configured to input the guide post local image into a guide post super-resolution processing module, and sequentially pass the image through a plurality of first convolution layers, a first bicubic interpolation layer, a first deconvolution layer, and a second bicubic interpolation layer to obtain an output high-resolution guide post local image;

[0122] The second input submodule is used to input the guide sleeve partial image into the guide sleeve partial image, and sequentially pass through multiple second convolution layers, third bicubic interpolation layers, second deconvolution layers and fourth bicubic interpolation layers to obtain an output high-resolution guide sleeve partial image.

[0123] Preferably, the positioning module includes:

[0124] an extraction submodule for adjusting weight parameters between feature maps of the high-resolution guide post local image and the high-resolution guide sleeve local image through the CBAM attention mechanism to extract multiple guide post image sub-features and multiple guide sleeve image sub-features; a feature fusion submodule for fusing the multiple guide post image sub-features and the multiple guide sleeve image sub-features to obtain fused guide post image sub-features and fused guide sleeve image sub-features;

[0125] A first positioning submodule, configured to input the fused guide post image sub-feature and the fused guide sleeve image sub-feature into a prediction layer having a loss function set therein for first positioning, thereby obtaining a positioning guide post image sub-feature and a positioning guide sleeve image sub-feature;

[0126] The second positioning submodule is used to extract the edge features and geometric features of the positioning guide post image subfeatures and the positioning guide sleeve image subfeatures, and perform second positioning based on the edge features and geometric features to obtain the guide post features and guide sleeve features.

[0127] Preferably, the positioning reconstruction module includes:

[0128] A conversion submodule, configured to convert the guide post feature and the guide sleeve feature into a multi-channel vector;

[0129] A spatial convolution operation submodule, configured to perform a spatial convolution operation on the features of each channel in the multi-channel vector with a convolution kernel to obtain a feature map, and convert the feature map into a recombined vector;

[0130] The connection operation submodule is used to perform a depth convolution operation and a residual connection operation on the reorganized vector to obtain a final guide column positioning map and a final guide sleeve positioning map.

[0131] Preferably, the calculation module includes:

[0132] The marking submodule is used to mark the dimensional data of the final guide column positioning diagram and the final guide sleeve positioning diagram respectively;

[0133] The calculation submodule is used to calculate the dimensional tolerance accuracy data of the guide column and guide sleeve of the mold base based on the dimensional data.

[0134] As for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0135] The specific definitions of the rapid positioning mold frame accuracy detection and control system can be found in the definitions of the rapid positioning mold frame accuracy detection and control method described above and will not be repeated here. Each module in the rapid positioning mold frame accuracy detection and control system described above can be implemented in whole or in part through software, hardware, or a combination thereof. Each of the above modules can be embedded in or independent of a processor in a computer device in hardware form, or can be stored in a computer device memory in software form, so that the processor can call and execute the corresponding operations of each of the above modules.

[0136] The detection and control system for the accuracy of mold frames for rapid positioning provided above can be used to execute the detection and control method for the accuracy of mold frames for rapid positioning provided by any of the above embodiments, and has corresponding functions and beneficial effects.

[0137] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Figure 6As shown. The computer device includes a processor, a memory, a network interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a method for detecting and controlling the accuracy of a mold frame with rapid positioning is realized. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad provided on the computer device housing, or an external keyboard, touchpad or mouse, etc.

[0138] Those skilled in the art will understand that Figure 6 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0139] In one embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the following steps are implemented:

[0140] Acquire a first front initial image and a second front initial image of the upper mold frame and the lower mold frame;

[0141] Extracting a partial image of the guide post within a preset range of the guide post of the upper mold frame from the first front initial image;

[0142] Extracting a partial image of the guide sleeve of the lower mold frame within a preset range of the guide sleeve from the second front initial image;

[0143] The local images of the guide post and the guide sleeve are subjected to super-resolution processing by a constructed double-layer super-resolution processing model to obtain high-resolution local images of the guide post and the guide sleeve; positioning is performed on the high-resolution local images of the guide post and the guide sleeve to obtain features of the guide post and the guide sleeve;

[0144] Performing depthwise separable convolution and depthwise convolution operations on the guide post features and the guide sleeve features to perform positioning reconstruction, thereby obtaining a final guide post positioning map and a final guide sleeve positioning map;

[0145] The positioning accuracy of the mold base is calculated based on the final guide column positioning diagram and the final guide sleeve positioning diagram.

[0146] Preferably, extracting a partial image of the guide column within a preset range of the guide column of the upper mold frame from the first front initial image includes:

[0147] The first front initial image is input into the deep backbone network, the final convolution layer of the deep backbone network is adapted using the dilated convolution, and then the maximum pooling operation is performed to obtain a local image of the guide post.

[0148] Preferably, the super-resolution processing of the guide post local image and the guide sleeve local image by a constructed double-layer super-resolution processing model to obtain a high-resolution guide post local image and a high-resolution guide sleeve local image includes:

[0149] A guide post super-resolution processing module is constructed through a plurality of first convolutional layers, first bicubic interpolation layers, first deconvolutional layers, and second bicubic interpolation layers;

[0150] A guide super-resolution processing module is constructed through multiple second convolution layers, third bicubic interpolation layers, second deconvolution layers and fourth bicubic interpolation layers;

[0151] The guide post super-resolution processing module and the guide sleeve super-resolution processing module are constructed into a double-layer super-resolution processing model;

[0152] Inputting the guide post local image into a guide post super-resolution processing module, sequentially passing through a plurality of first convolution layers, a first bicubic interpolation layer, a first deconvolution layer, and a second bicubic interpolation layer, to obtain an output high-resolution guide post local image;

[0153] The guide sleeve partial image is input into the guide sleeve partial image, and sequentially passes through multiple second convolution layers, third bicubic interpolation layers, second deconvolution layers and fourth bicubic interpolation layers to obtain an output high-resolution guide sleeve partial image.

[0154] Preferably, the positioning of the high-resolution guide post local image and the high-resolution guide sleeve local image to obtain the guide post features and the guide sleeve features includes:

[0155] The high-resolution guide post partial image and the high-resolution guide sleeve partial image are subjected to the CBAM attention mechanism to adjust the weight parameters between the feature maps, thereby extracting a plurality of guide post image sub-features and a plurality of guide sleeve image sub-features;

[0156] Performing feature fusion on the multiple guide post image sub-features and the multiple guide sleeve image sub-features to obtain fused guide post image sub-features and fused guide sleeve image sub-features;

[0157] Inputting the fused guide post image sub-feature and the fused guide sleeve image sub-feature into a prediction layer with a set loss function for first positioning, thereby obtaining a positioning guide post image sub-feature and a positioning guide sleeve image sub-feature;

[0158] The edge features and geometric features of the positioning guide post image sub-features and the positioning guide sleeve image sub-features are extracted, and a second positioning is performed based on the edge features and geometric features to obtain the guide post features and the guide sleeve features.

[0159] Preferably, performing depthwise separable convolution and depthwise convolution operations on the guide post features and the guide sleeve features to perform positioning reconstruction to obtain a final guide post positioning map and a final guide sleeve positioning map includes:

[0160] Converting the guide post feature and the guide sleeve feature into a multi-channel vector;

[0161] Performing a spatial convolution operation on the features of each channel in the multi-channel vector and a convolution kernel to obtain a feature map, and converting the feature map into a recombined vector;

[0162] The reorganized vector is subjected to a depthwise convolution operation and a residual connection operation to obtain a final guide post positioning map and a final guide sleeve positioning map.

[0163] Preferably, the calculation of the positioning accuracy of the mold base by using the final guide column positioning diagram and the final guide sleeve positioning diagram includes:

[0164] Mark out the dimensional data of the final guide column positioning diagram and the final guide sleeve positioning diagram respectively;

[0165] The dimensional tolerance accuracy data of the guide column and guide sleeve of the mold base are calculated based on the dimensional data.

[0166] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented: obtaining a first front initial image and a second front initial image of an upper mold frame and a lower mold frame;

[0167] Extracting a partial image of the guide post within a preset range of the guide post of the upper mold frame from the first front initial image;

[0168] Extracting a partial image of the guide sleeve of the lower mold frame within a preset range of the guide sleeve from the second front initial image;

[0169] The guide post local image and the guide sleeve local image are subjected to super-resolution processing by a constructed double-layer super-resolution processing model to obtain a high-resolution guide post local image and a high-resolution guide sleeve local image;

[0170] Positioning the high-resolution guide post local image and the high-resolution guide sleeve local image to obtain guide post features and guide sleeve features;

[0171] Performing depthwise separable convolution and depthwise convolution operations on the guide post features and the guide sleeve features to perform positioning reconstruction, thereby obtaining a final guide post positioning map and a final guide sleeve positioning map;

[0172] The positioning accuracy of the mold base is calculated based on the final guide column positioning diagram and the final guide sleeve positioning diagram.

[0173] Preferably, extracting a partial image of the guide column within a preset range of the guide column of the upper mold frame from the first front initial image includes:

[0174] The first front initial image is input into the deep backbone network, the final convolution layer of the deep backbone network is adapted using the dilated convolution, and then the maximum pooling operation is performed to obtain a local image of the guide post.

[0175] Preferably, the super-resolution processing of the guide post local image and the guide sleeve local image by a constructed double-layer super-resolution processing model to obtain a high-resolution guide post local image and a high-resolution guide sleeve local image includes:

[0176] A guide post super-resolution processing module is constructed through a plurality of first convolutional layers, first bicubic interpolation layers, first deconvolutional layers, and second bicubic interpolation layers;

[0177] A guide super-resolution processing module is constructed through multiple second convolution layers, third bicubic interpolation layers, second deconvolution layers and fourth bicubic interpolation layers;

[0178] The guide post super-resolution processing module and the guide sleeve super-resolution processing module are constructed into a double-layer super-resolution processing model;

[0179] Inputting the guide post local image into a guide post super-resolution processing module, sequentially passing through a plurality of first convolution layers, a first bicubic interpolation layer, a first deconvolution layer, and a second bicubic interpolation layer, to obtain an output high-resolution guide post local image;

[0180] The guide sleeve partial image is input into the guide sleeve partial image, and sequentially passes through multiple second convolution layers, third bicubic interpolation layers, second deconvolution layers and fourth bicubic interpolation layers to obtain an output high-resolution guide sleeve partial image.

[0181] Preferably, the positioning of the high-resolution guide post local image and the high-resolution guide sleeve local image to obtain the guide post features and the guide sleeve features includes:

[0182] The high-resolution guide post partial image and the high-resolution guide sleeve partial image are subjected to the CBAM attention mechanism to adjust the weight parameters between the feature maps, thereby extracting a plurality of guide post image sub-features and a plurality of guide sleeve image sub-features;

[0183] Performing feature fusion on the multiple guide post image sub-features and the multiple guide sleeve image sub-features to obtain fused guide post image sub-features and fused guide sleeve image sub-features;

[0184] Inputting the fused guide post image sub-feature and the fused guide sleeve image sub-feature into a prediction layer with a set loss function for first positioning, thereby obtaining a positioning guide post image sub-feature and a positioning guide sleeve image sub-feature;

[0185] The edge features and geometric features of the positioning guide post image sub-features and the positioning guide sleeve image sub-features are extracted, and a second positioning is performed based on the edge features and geometric features to obtain the guide post features and the guide sleeve features.

[0186] Preferably, performing depthwise separable convolution and depthwise convolution operations on the guide post features and the guide sleeve features to perform positioning reconstruction to obtain a final guide post positioning map and a final guide sleeve positioning map includes:

[0187] Converting the guide post feature and the guide sleeve feature into a multi-channel vector;

[0188] Performing a spatial convolution operation on the features of each channel in the multi-channel vector and a convolution kernel to obtain a feature map, and converting the feature map into a recombined vector;

[0189] The reorganized vector is subjected to a depthwise convolution operation and a residual connection operation to obtain a final guide post positioning map and a final guide sleeve positioning map.

[0190] Preferably, calculating the positioning accuracy of the mold base according to the final guide post positioning diagram and the final guide sleeve positioning diagram comprises: marking the dimension data of the final guide post positioning diagram and the final guide sleeve positioning diagram respectively;

[0191] The dimensional tolerance accuracy data of the guide column and guide sleeve of the mold base are calculated based on the dimensional data.

[0192] Those skilled in the art will understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0193] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0194] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, apparatus, or computer program products. Thus, embodiments of the present invention may take the form of a fully hardware embodiment, a fully software embodiment, or an embodiment combining software and hardware. Furthermore, embodiments of 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 magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0195] The embodiments of the present invention are described with reference to the flowcharts and / or block diagrams of the apparatus, terminal device (system), and computer program product according to the 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 the 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 processor, or other programmable data processing terminal device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing terminal 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.

[0196] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing terminal device to work in a specific manner, so that the instructions stored in the computer readable memory produce a manufactured product including an instruction method, which is implemented in the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0197] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal device so that a series of operating steps are executed on the computer or other programmable terminal device to produce a computer-implemented process, thereby providing instructions for executing on the computer or other programmable terminal device to implement the process. Figure 1 a process or multiple processes and / or boxes Figure 1 The steps for the function specified in one or more boxes.

[0198] Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they become aware of the basic creative concepts. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the embodiments of the present invention.

[0199] Finally, it should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, apparatus, article, or terminal device that includes a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, apparatus, article, or terminal device. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of additional identical elements in the process, apparatus, article, or terminal device that includes the element.

[0200] The above is a detailed introduction to a method for detecting and controlling the accuracy of a rapid positioning mold frame, a detection and control system for the accuracy of a rapid positioning mold frame, a computer device and a storage medium provided by the present invention. Specific examples are used herein to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. At the same time, for general technical personnel in this field, according to the ideas of the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. A method for detecting and controlling the accuracy of a rapid positioning mold frame, characterized in that: include: Acquire a first front initial image and a second front initial image of the upper mold frame and the lower mold frame; Extracting a partial image of the guide post within a preset range of the guide post of the upper mold frame from the first front initial image; Extracting a partial image of the guide sleeve of the lower mold frame within a preset range of the guide sleeve from the second front initial image; The guide post local image and the guide sleeve local image are subjected to super-resolution processing by a constructed double-layer super-resolution processing model to obtain a high-resolution guide post local image and a high-resolution guide sleeve local image; Positioning the high-resolution guide post local image and the high-resolution guide sleeve local image to obtain guide post features and guide sleeve features; Performing depthwise separable convolution and depthwise convolution operations on the guide post features and the guide sleeve features to perform positioning reconstruction, thereby obtaining a final guide post positioning map and a final guide sleeve positioning map; Calculating the positioning accuracy of the mold base through the final guide column positioning diagram and the final guide sleeve positioning diagram; Positioning the high-resolution guide post local image and the high-resolution guide sleeve local image to obtain guide post features and guide sleeve features includes: The high-resolution guide post partial image and the high-resolution guide sleeve partial image are subjected to the CBAM attention mechanism to adjust the weight parameters between the feature maps, thereby extracting a plurality of guide post image sub-features and a plurality of guide sleeve image sub-features; Performing feature fusion on the multiple guide post image sub-features and the multiple guide sleeve image sub-features to obtain fused guide post image sub-features and fused guide sleeve image sub-features; Inputting the fused guide post image sub-feature and the fused guide sleeve image sub-feature into a prediction layer with a set loss function for first positioning, thereby obtaining a positioning guide post image sub-feature and a positioning guide sleeve image sub-feature; Extracting edge features and geometric features of the positioning guide post image sub-features and the positioning guide sleeve image sub-features, performing a second positioning based on the edge features and geometric features to obtain guide post features and guide sleeve features; The positioning reconstruction is performed by performing depthwise separable convolution and depthwise convolution operations on the guide post features and the guide sleeve features to obtain a final guide post positioning map and a final guide sleeve positioning map, including: Converting the guide post feature and the guide sleeve feature into a multi-channel vector; Performing a spatial convolution operation on the features of each channel in the multi-channel vector and a convolution kernel to obtain a feature map, and converting the feature map into a recombined vector; The reorganized vector is subjected to a depthwise convolution operation and a residual connection operation to obtain a final guide post positioning map and a final guide sleeve positioning map.

2. The method according to claim 1, characterized in that The step of extracting a partial image of the guide column within a preset range of the guide column of the upper mold frame from the first front initial image includes: The first front initial image is input into the deep backbone network, the final convolution layer of the deep backbone network is adapted using the dilated convolution, and then the maximum pooling operation is performed to obtain a local image of the guide post.

3. The method according to claim 1, characterized in that The method of performing super-resolution processing on the guide post local image and the guide sleeve local image by using a constructed double-layer super-resolution processing model to obtain a high-resolution guide post local image and a high-resolution guide sleeve local image includes: A guide post super-resolution processing module is constructed through a plurality of first convolutional layers, first bicubic interpolation layers, first deconvolutional layers, and second bicubic interpolation layers; A guide super-resolution processing module is constructed through multiple second convolution layers, third bicubic interpolation layers, second deconvolution layers and fourth bicubic interpolation layers; The guide post super-resolution processing module and the guide sleeve super-resolution processing module are constructed into a double-layer super-resolution processing model; Inputting the guide post local image into a guide post super-resolution processing module, sequentially passing through a plurality of first convolution layers, a first bicubic interpolation layer, a first deconvolution layer, and a second bicubic interpolation layer, to obtain an output high-resolution guide post local image; The guide sleeve partial image is input into the guide sleeve partial image, and sequentially passes through multiple second convolution layers, third bicubic interpolation layers, second deconvolution layers and fourth bicubic interpolation layers to obtain an output high-resolution guide sleeve partial image.

4. The method according to claim 1, wherein The calculation of the positioning accuracy of the mold base by using the final guide column positioning diagram and the final guide sleeve positioning diagram includes: Mark out the dimensional data of the final guide column positioning diagram and the final guide sleeve positioning diagram respectively; The dimensional tolerance accuracy data of the guide column and guide sleeve of the mold base are calculated based on the dimensional data.

5. A rapid positioning mold frame accuracy detection and control system, characterized in that: include: An initial image acquisition module is used to acquire a first front initial image and a second front initial image of the upper mold frame and the lower mold frame; A first extraction module is configured to extract a partial image of the guide post within a preset range of the guide post of the upper mold frame from the first front initial image; A second extraction module is used to extract a partial image of the guide sleeve of the lower mold frame within a preset range of the guide sleeve from the second front initial image; a super-resolution processing module for performing super-resolution processing on the guide post local image and the guide sleeve local image through a constructed double-layer super-resolution processing model to obtain a high-resolution guide post local image and a high-resolution guide sleeve local image; A positioning module, configured to perform positioning based on the high-resolution local image of the guide post and the high-resolution local image of the guide sleeve to obtain features of the guide post and the guide sleeve; A positioning reconstruction module is used to perform depthwise separable convolution and depthwise convolution operations on the guide post features and the guide sleeve features to perform positioning reconstruction, thereby obtaining a final guide post positioning map and a final guide sleeve positioning map; A calculation module, configured to calculate the positioning accuracy of the mold base according to the final guide column positioning diagram and the final guide sleeve positioning diagram; Positioning the high-resolution guide post local image and the high-resolution guide sleeve local image to obtain guide post features and guide sleeve features includes: The high-resolution guide post partial image and the high-resolution guide sleeve partial image are subjected to the CBAM attention mechanism to adjust the weight parameters between the feature maps, thereby extracting a plurality of guide post image sub-features and a plurality of guide sleeve image sub-features; Performing feature fusion on the multiple guide post image sub-features and the multiple guide sleeve image sub-features to obtain fused guide post image sub-features and fused guide sleeve image sub-features; Inputting the fused guide post image sub-feature and the fused guide sleeve image sub-feature into a prediction layer with a set loss function for first positioning, thereby obtaining a positioning guide post image sub-feature and a positioning guide sleeve image sub-feature; Extracting edge features and geometric features of the positioning guide post image sub-features and the positioning guide sleeve image sub-features, performing a second positioning based on the edge features and geometric features to obtain guide post features and guide sleeve features; The positioning reconstruction is performed by performing depthwise separable convolution and depthwise convolution operations on the guide post features and the guide sleeve features to obtain a final guide post positioning map and a final guide sleeve positioning map, including: Converting the guide post feature and the guide sleeve feature into a multi-channel vector; Performing a spatial convolution operation on the features of each channel in the multi-channel vector and a convolution kernel to obtain a feature map, and converting the feature map into a recombined vector; The reorganized vector is subjected to a depthwise convolution operation and a residual connection operation to obtain a final guide post positioning map and a final guide sleeve positioning map.

6. The system according to claim 5, characterized in that The first extraction module includes: The acquisition submodule is used to input the first front initial image into the deep backbone network, use the void convolution to adapt the final convolution layer of the deep backbone network, and then perform the maximum pooling operation to obtain the local image of the guide column.

7. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method for detecting and controlling the accuracy of a rapid positioning mold frame according to any one of claims 1 to 4 are implemented.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for detecting and controlling the accuracy of a rapid positioning mold frame according to any one of claims 1 to 4 are implemented.

Images