Rubber accelerator particle diameter monitoring method and system based on optic nerves
The visual neural-based method and system address the limitations of existing detection methods by using a visual neural network with triangular convolution kernels to achieve precise and stable rubber accelerator particle diameter measurement.
Patent Information
- Application Number
- CN202411561665.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-05
- Publication Date
- 2024-12-03
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Existing methods for detecting rubber accelerator particle diameter, such as optical microscopy, laser particle size analysis, and image processing, lack accuracy, automation, and stability, particularly in large-scale applications.
A visual neural-based method and system using a visual neural network model with triangular convolution kernels to process target images of rubber accelerator particles, extracting features to determine particle diameters accurately.
Enables large-scale, high-precision, and stable detection of rubber accelerator particle diameters, improving detection robustness and accuracy by matching the convolution scale to the particle size.
Smart Images

Figure CN119067973A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of image processing, and in particular to a method and system for monitoring rubber accelerator particle diameter based on visual nerves. Background Art
[0002] With the improvement of industrial automation level and increasingly stringent product quality requirements, the quality control of rubber products has become more and more important. As one of the indispensable additives in the production process of rubber products, the uniformity and stability of the particle diameter of rubber accelerator directly affect the physical properties and service life of rubber products. Therefore, accurate detection of rubber accelerator particle diameter has become a key link to ensure product quality.
[0003] The uniformity of the particle diameter of rubber accelerator is crucial to the processing performance, physical properties and quality stability of rubber products. If the particle diameter of rubber accelerator is unevenly distributed, it may cause uneven vulcanization of rubber products during the vulcanization process, thus affecting the performance indicators of the products such as strength, elasticity and wear resistance. In addition, the size of the particle diameter will also affect the processing technology and production efficiency of rubber products. Therefore, accurate detection of the particle diameter of rubber accelerator is of great significance to ensure the quality of rubber products.
[0004] At present, the detection methods of rubber accelerator particle diameter mainly include optical microscopy, laser particle size analysis and image processing. Optical microscopy is a traditional particle diameter detection method. The morphology and size of rubber accelerator particles are observed under a microscope, and the diameter of the particles is calculated based on the projected area or length of the particles. This method is simple to operate and low in cost, but has low accuracy and cannot achieve automated detection. Laser particle size analysis is a non-contact detection technology. Rubber accelerator particles are irradiated with a laser beam, and particle size analysis is performed using the scattering principle of the laser by the particles. This method has the advantages of fast detection speed, high accuracy and good repeatability, but the equipment cost is relatively high. Image processing is to measure the size of rubber accelerator particles through digital image processing technology. This method obtains particle images through an image acquisition device, and then uses image processing software to identify and measure the size of the particles. This method has the advantage of high automation, but the detection stability is not high.
[0005] Therefore, how to achieve large-scale, high-precision and stable rubber accelerator particle diameter is a current research issue. Summary of the invention
[0006] The embodiment of the present invention provides a method and system for monitoring the particle diameter of rubber accelerator based on visual nerves, which are used to monitor the particle diameter of rubber accelerator in large quantities with high precision and stability.
[0007] In order to achieve the above object, the present invention adopts the following technical scheme: In a first aspect, a visual neural-based rubber accelerator particle diameter monitoring method is provided, which is applied to an electronic device. The method comprises: the electronic device acquires a target image, where the image contains a plurality of rubber accelerator particles, and the plurality of rubber accelerator particles contain a plurality of batches of rubber accelerator particles; the electronic device processes the target image through a visual neural network model to obtain the diameter of each of the plurality of rubber accelerator particles, wherein the convolution scale of the visual neural network model matches the size of the rubber accelerator particles in the image.
[0008] Optionally, the electronic device processes the target image through a visual neural network model to obtain the diameters of multiple rubber accelerator particles, including: the electronic device performs a convolution operation on the target image through at least one convolution kernel in a convolution layer of the visual neural network model to obtain a convolution feature vector of the target image, and at least one convolution kernel is a triangular convolution kernel; the electronic device convolves the feature vector of the target image through a feature processing layer of the visual neural network model to obtain the diameters of multiple rubber accelerator particles, wherein the feature processing layer includes, in order of processing: a ReLU excitation layer, a pooling layer, and a fully connected layer.
[0009] Optionally, at least one convolution kernel includes a first convolution kernel, and the first convolution kernel is a triangular convolution kernel including a first convolution kernel of , K is an integer greater than or equal to 5, -2 means decreasing by 2, That is to say, the size of the first row in the first convolution kernel is K, and the size of each row in the first convolution kernel decreases by 2 until the size of the last row in the first convolution kernel is 1 or 2; the size of the horizontal movement step size and the vertical movement step size of the first convolution kernel are both K and greater than or equal to 1.
[0010] Optionally, the step length of the lateral movement of the first convolution kernel is x1, the step length of the longitudinal movement of the first convolution kernel is y1, and the electronic device performs a convolution operation on the target image through at least one convolution kernel in the convolution layer of the visual neural network model to obtain a convolution feature vector of the target image, including: the electronic device performs a convolution operation on the target image through the first convolution kernel in the convolution layer of the visual neural network model with the step lengths of x1 and y1 respectively to obtain the convolution feature vector of the target image; Wherein, the target image is an image with a resolution of M*N, M and N are integers greater than K, M is the number of rows, and N is the number of columns; during the convolution operation, the first convolution kernel performs one convolution from the position of the 0th row and the 0th column in the target image, and then the first convolution kernel moves horizontally in sequence with a step length of x1 and performs convolution in sequence until it moves to the position of the 0th row and the Nth column in the target image and performs one convolution, the first convolution kernel moves vertically in sequence with a step length of y1 to the position of the y1th row and the Nth column in the target image and performs one convolution, and then the first convolution kernel moves horizontally in sequence with a step length of x1 and performs convolution in sequence until it moves to the position of the y1th row and the 0th column in the target image and performs one convolution, and so on, until the convolution is completed, when the convolution is completed, the first convolution kernel moves to the position of the M-J1th row and the Nth column in the target image, and J1 is the number of rows contained in the first convolution kernel; The convolution scale of the visual neural network model matches the size of the rubber accelerator particles in the target image, which means that the distance between the repeatedly convolved feature vectors in the target image is the diameter of the rubber accelerator particles in the target image.
[0011] Optionally, at least one convolution kernel includes a second convolution kernel, and the second convolution kernel is a triangular convolution kernel including a second convolution kernel of , K is an integer greater than or equal to 5, ++2 means adding 2 successively, That is to say, the size of the first row in the second convolution kernel is 1 or 2, and the size of each row in the second convolution kernel is increased by 2 until the size of the last row in the second convolution kernel is K; the size of the horizontal movement step size and the vertical movement step size of the second convolution kernel are both K and greater than or equal to 1.
[0012] Optionally, the step length of the lateral movement of the second convolution kernel is x2, the step length of the longitudinal movement of the second convolution kernel is y2, and the electronic device performs a convolution operation on the target image through at least one convolution kernel in the convolution layer of the visual neural network model to obtain a convolution feature vector of the target image, including: the electronic device performs a convolution operation on the target image through the second convolution kernel in the convolution layer of the visual neural network model with the step lengths of x2 and y2 respectively to obtain the convolution feature vector of the target image; Wherein, the target image is an image with a resolution of M*N, M and N are integers greater than K, M is the number of rows, and N is the number of columns; during the convolution operation, the second convolution kernel performs one convolution from the position located at the J2th row and the 0th column in the target image, J2 is the number of rows contained in the second convolution kernel, and then the second convolution kernel moves horizontally in sequence with a step length of x2 and performs convolution in sequence until it moves to the position located at the J2th row and the Nth column in the target image and performs one convolution, and the second convolution kernel moves vertically in sequence with a step length of y2 to the position located at the y2th+J2th row and the Nth column in the target image and performs one convolution, and then the second convolution kernel moves horizontally in sequence with a step length of x2 and performs convolution in sequence until it moves to the position located at the y2th+J2th row and the Nth column in the target image and performs one convolution, and so on, until the convolution is completed, when the convolution is completed, the second convolution kernel moves to the position located at the Mth row and the Nth column in the target image; The convolution scale of the visual neural network model matches the size of the rubber accelerator particles in the target image, which means that the distance between the repeatedly convolved feature vectors in the target image is the diameter of the rubber accelerator particles in the target image.
[0013] Optionally, at least one convolution kernel includes a first convolution kernel and a second convolution kernel; the first convolution kernel is a triangular convolution kernel including the first convolution kernel is , K is an integer greater than or equal to 5, -2 means decreasing by 2, That is, the size of the first row in the first convolution kernel is K, and the size of each row in the first convolution kernel decreases by 2 until the size of the last row in the first convolution kernel is 1 or 2; the size of the horizontal movement step and the vertical movement step of the first convolution kernel are both K and greater than or equal to 1; the second convolution kernel is a triangular convolution kernel including the second convolution kernel is , K is an integer greater than or equal to 5, ++2 means adding 2 successively, That is to say, the size of the first row in the second convolution kernel is 1 or 2, and the size of each row in the second convolution kernel is increased by 2 until the size of the last row in the second convolution kernel is K; the size of the horizontal movement step size and the vertical movement step size of the second convolution kernel are both K and greater than or equal to 1.
[0014] Optionally, the step size of the lateral movement of the first convolution kernel is x1, the step size of the longitudinal movement of the first convolution kernel is y1, the step size of the lateral movement of the second convolution kernel is x2, and the step size of the longitudinal movement of the second convolution kernel is y2. The electronic device performs a convolution operation on the target image through at least one convolution kernel in the convolution layer of the visual neural network model to obtain a convolution feature vector of the target image, including: the electronic device performs a convolution operation on the target image through the first convolution kernel in the convolution layer of the visual neural network model with a step size of x1 and y1, and simultaneously performs a convolution operation on the target image through the second convolution kernel in the convolution layer of the visual neural network model with a step size of x2 and y2 to obtain a convolution feature vector of the target image; wherein the target image is an image with a resolution of M*N, M and N are integers greater than K, M is the number of rows, and N is the number of columns; During the convolution operation: the first convolution kernel performs one convolution from the position of the 0th row and 0th column in the target image, and at the same time, the second convolution kernel moves to the position of the MJth row and Nth column in the target image, where J1 is the number of rows contained in the first convolution kernel; then, the first convolution kernel moves horizontally with a step size of x1 and performs convolutions in sequence until it moves to the position of the 0th row and Nth column in the target image and performs one convolution, and at the same time, the second convolution kernel moves horizontally with a step size of x2 and performs convolutions in sequence until it moves to the position of the J2th row and Nth column in the target image and performs one convolution; then, the first convolution kernel moves horizontally with a step size of y2 and performs convolutions in sequence until it moves to the position of the J2th row and Nth column in the target image and performs one convolution; 1 moves vertically to the position of the y1th row and Nth column in the target image and performs one convolution. At the same time, the second convolution kernel moves vertically to the position of the y2th + J2th row and Nth column in the target image with a step size of y2 and performs one convolution. After that, the first convolution kernel moves horizontally in sequence with a step size of x1 and performs convolutions in sequence until it moves to the position of the y1th row and 0th column in the target image and performs one convolution. At the same time, the second convolution kernel moves horizontally in sequence with a step size of x2 and performs convolutions in sequence until it moves to the position of the y2th + J2th row and Nth column in the target image and performs one convolution. And so on until the convolution is completed. When the convolution is completed, the first convolution kernel moves to the position at the M-J1th row and the Nth column in the target image, where J1 is the number of rows included in the first convolution kernel, and the second convolution kernel moves to the position at the Mth row and the Nth column in the target image.
[0015] The convolution scale of the visual neural network model matches the size of the rubber accelerator particles in the target image, which means that the distance between the repeatedly convolved feature vectors in the target image is the diameter of the rubber accelerator particles in the target image.
[0016] Optionally, the method further includes: the electronic device determines whether there is a batch with quality problems among the multiple batches of rubber accelerator particles by analyzing the diameters of each of the multiple rubber accelerator particles.
[0017] In a second aspect, a visual neural-based rubber accelerator particle diameter monitoring system is provided, the system includes an electronic device, and the system is configured as follows: the electronic device acquires a target image, the image includes multiple rubber accelerator particles, and the multiple rubber accelerator particles include multiple batches of rubber accelerator particles; the electronic device processes the target image through a visual neural network model to obtain the diameters of the multiple rubber accelerator particles output by the visual neural network model, wherein the convolution scale of the visual neural network model matches the size of the rubber accelerator particles in the image.
[0018] Optionally, the electronic device processes the target image through a visual neural network model to obtain the diameters of multiple rubber accelerator particles, including: the electronic device performs a convolution operation on the target image through at least one convolution kernel in a convolution layer of the visual neural network model to obtain a convolution feature vector of the target image, and at least one convolution kernel is a triangular convolution kernel; the electronic device convolves the feature vector of the target image through a feature processing layer of the visual neural network model to obtain the diameters of multiple rubber accelerator particles, wherein the feature processing layer includes, in order of processing: a ReLU excitation layer, a pooling layer, and a fully connected layer.
[0019] Optionally, at least one convolution kernel includes a first convolution kernel, and the first convolution kernel is a triangular convolution kernel including a first convolution kernel of , K is an integer greater than or equal to 5, -2 means decreasing by 2, That is to say, the size of the first row in the first convolution kernel is K, and the size of each row in the first convolution kernel decreases by 2 until the size of the last row in the first convolution kernel is 1 or 2; the size of the horizontal movement step size and the vertical movement step size of the first convolution kernel are both K and greater than or equal to 1.
[0020] Optionally, the step length of the lateral movement of the first convolution kernel is x1, the step length of the longitudinal movement of the first convolution kernel is y1, and the electronic device performs a convolution operation on the target image through at least one convolution kernel in the convolution layer of the visual neural network model to obtain a convolution feature vector of the target image, including: the electronic device performs a convolution operation on the target image through the first convolution kernel in the convolution layer of the visual neural network model with the step lengths of x1 and y1 respectively to obtain the convolution feature vector of the target image; Wherein, the target image is an image with a resolution of M*N, M and N are integers greater than K, M is the number of rows, and N is the number of columns; during the convolution operation, the first convolution kernel performs one convolution from the position of the 0th row and the 0th column in the target image, and then the first convolution kernel moves horizontally in sequence with a step length of x1 and performs convolution in sequence until it moves to the position of the 0th row and the Nth column in the target image and performs one convolution, the first convolution kernel moves vertically in sequence with a step length of y1 to the position of the y1th row and the Nth column in the target image and performs one convolution, and then the first convolution kernel moves horizontally in sequence with a step length of x1 and performs convolution in sequence until it moves to the position of the y1th row and the 0th column in the target image and performs one convolution, and so on, until the convolution is completed, when the convolution is completed, the first convolution kernel moves to the position of the M-J1th row and the Nth column in the target image, and J1 is the number of rows contained in the first convolution kernel; The convolution scale of the visual neural network model matches the size of the rubber accelerator particles in the target image, which means that the distance between the repeatedly convolved feature vectors in the target image is the diameter of the rubber accelerator particles in the target image.
[0021] Optionally, at least one convolution kernel includes a second convolution kernel, and the second convolution kernel is a triangular convolution kernel including a second convolution kernel of , K is an integer greater than or equal to 5, ++2 means adding 2 successively, That is to say, the size of the first row in the second convolution kernel is 1 or 2, and the size of each row in the second convolution kernel is increased by 2 until the size of the last row in the second convolution kernel is K; the size of the horizontal movement step size and the vertical movement step size of the second convolution kernel are both K and greater than or equal to 1.
[0022] Optionally, the step length of the lateral movement of the second convolution kernel is x2, the step length of the longitudinal movement of the second convolution kernel is y2, and the electronic device performs a convolution operation on the target image through at least one convolution kernel in the convolution layer of the visual neural network model to obtain a convolution feature vector of the target image, including: the electronic device performs a convolution operation on the target image through the second convolution kernel in the convolution layer of the visual neural network model with the step lengths of x2 and y2 respectively to obtain the convolution feature vector of the target image; Wherein, the target image is an image with a resolution of M*N, M and N are integers greater than K, M is the number of rows, and N is the number of columns; during the convolution operation, the second convolution kernel performs one convolution from the position located at the J2th row and the 0th column in the target image, J2 is the number of rows contained in the second convolution kernel, and then the second convolution kernel moves horizontally in sequence with a step length of x2 and performs convolution in sequence until it moves to the position located at the J2th row and the Nth column in the target image and performs one convolution, and the second convolution kernel moves vertically in sequence with a step length of y2 to the position located at the y2th+J2th row and the Nth column in the target image and performs one convolution, and then the second convolution kernel moves horizontally in sequence with a step length of x2 and performs convolution in sequence until it moves to the position located at the y2th+J2th row and the Nth column in the target image and performs one convolution, and so on, until the convolution is completed, when the convolution is completed, the second convolution kernel moves to the position located at the Mth row and the Nth column in the target image; The convolution scale of the visual neural network model matches the size of the rubber accelerator particles in the target image, which means that the distance between the repeatedly convolved feature vectors in the target image is the diameter of the rubber accelerator particles in the target image.
[0023] Optionally, at least one convolution kernel includes a first convolution kernel and a second convolution kernel; the first convolution kernel is a triangular convolution kernel including the first convolution kernel is , K is an integer greater than or equal to 5, -2 means decreasing by 2, That is, the size of the first row in the first convolution kernel is K, and the size of each row in the first convolution kernel decreases by 2 until the size of the last row in the first convolution kernel is 1 or 2; the size of the horizontal movement step and the vertical movement step of the first convolution kernel are both K and greater than or equal to 1; the second convolution kernel is a triangular convolution kernel including the second convolution kernel is , K is an integer greater than or equal to 5, ++2 means adding 2 successively, That is to say, the size of the first row in the second convolution kernel is 1 or 2, and the size of each row in the second convolution kernel is increased by 2 until the size of the last row in the second convolution kernel is K; the size of the horizontal movement step size and the vertical movement step size of the second convolution kernel are both K and greater than or equal to 1.
[0024] Optionally, the step size of the lateral movement of the first convolution kernel is x1, the step size of the longitudinal movement of the first convolution kernel is y1, the step size of the lateral movement of the second convolution kernel is x2, and the step size of the longitudinal movement of the second convolution kernel is y2. The electronic device performs a convolution operation on the target image through at least one convolution kernel in the convolution layer of the visual neural network model to obtain a convolution feature vector of the target image, including: the electronic device performs a convolution operation on the target image through the first convolution kernel in the convolution layer of the visual neural network model with a step size of x1 and y1, and simultaneously performs a convolution operation on the target image through the second convolution kernel in the convolution layer of the visual neural network model with a step size of x2 and y2 to obtain a convolution feature vector of the target image; wherein the target image is an image with a resolution of M*N, M and N are integers greater than K, M is the number of rows, and N is the number of columns; During the convolution operation: the first convolution kernel performs one convolution from the position of the 0th row and 0th column in the target image, and at the same time, the second convolution kernel moves to the position of the MJth row and Nth column in the target image, where J1 is the number of rows contained in the first convolution kernel; then, the first convolution kernel moves horizontally with a step size of x1 and performs convolutions in sequence until it moves to the position of the 0th row and Nth column in the target image and performs one convolution, and at the same time, the second convolution kernel moves horizontally with a step size of x2 and performs convolutions in sequence until it moves to the position of the J2th row and Nth column in the target image and performs one convolution; then, the first convolution kernel moves horizontally with a step size of y2 and performs convolutions in sequence until it moves to the position of the J2th row and Nth column in the target image and performs one convolution; 1 moves vertically to the position of the y1th row and Nth column in the target image and performs one convolution. At the same time, the second convolution kernel moves vertically to the position of the y2th + J2th row and Nth column in the target image with a step size of y2 and performs one convolution. After that, the first convolution kernel moves horizontally in sequence with a step size of x1 and performs convolutions in sequence until it moves to the position of the y1th row and 0th column in the target image and performs one convolution. At the same time, the second convolution kernel moves horizontally in sequence with a step size of x2 and performs convolutions in sequence until it moves to the position of the y2th + J2th row and Nth column in the target image and performs one convolution. And so on until the convolution is completed. When the convolution is completed, the first convolution kernel moves to the position at the M-J1th row and the Nth column in the target image, where J1 is the number of rows included in the first convolution kernel, and the second convolution kernel moves to the position at the Mth row and the Nth column in the target image.
[0025] The convolution scale of the visual neural network model matches the size of the rubber accelerator particles in the target image, which means that the distance between the repeatedly convolved feature vectors in the target image is the diameter of the rubber accelerator particles in the target image.
[0026] Optionally, the system is further configured such that: the electronic device determines whether there is a batch with quality problems among the multiple batches of rubber accelerator particles by analyzing the diameters of each of the multiple rubber accelerator particles.
[0027] In a third aspect, a computer-readable storage medium is provided, comprising: a computer program or instructions; when the computer program or instructions are executed on a computer, the computer is caused to execute the method described in the first aspect.
[0028] In summary, the above method and system have the following technical effects: The electronic device can acquire a target image containing multiple batches of rubber accelerator particles and process the target image through a visual neural network model. Since the convolution scale of the visual neural network model matches the size of the rubber accelerator particles in the image, better feature extraction can be performed during convolution, and the diameters of multiple rubber accelerator particles can be accurately obtained, so as to achieve large-scale, high-precision and stable rubber accelerator particle diameters. BRIEF DESCRIPTION OF THE DRAWINGS
[0029] Figure 1 A schematic diagram of a flow chart of a method for monitoring rubber accelerator particle diameter based on visual nerves provided in an embodiment of the present invention; Figure 2 A schematic diagram of a target image in a rubber accelerator particle diameter monitoring method based on visual nerves provided in an embodiment of the present invention; Figure 3 A schematic diagram of convolution in a rubber accelerator particle diameter monitoring method based on visual nerves provided in an embodiment of the present invention; Figure 4 A schematic diagram of feature extraction in a rubber accelerator particle diameter monitoring method based on visual nerves provided in an embodiment of the present invention; Figure 5 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION
[0030] The technical solution of the present invention will be described below in conjunction with the accompanying drawings.
[0031] The present invention will present various aspects, embodiments or features around a system that may include multiple devices, components, modules, etc. It should be understood and appreciated that each system may include additional devices, components, modules, etc., and / or may not include all of the devices, components, modules, etc. discussed in conjunction with the drawings. In addition, combinations of these schemes may also be used.
[0032] In addition, in the embodiments of the present invention, words such as "exemplary" and "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design described as "exemplary" in the present invention should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the word "exemplary" is used to present concepts in a specific way.
[0033] In the embodiments of the present invention, "information", "signal", "message", "channel" and "signaling" can sometimes be used interchangeably. It should be noted that when the distinction between them is not emphasized, the meanings to be expressed are matched. "of", "corresponding, relevant" and "corresponding" can sometimes be used interchangeably. It should be noted that when the distinction between them is not emphasized, the meanings to be expressed are matched. In addition, the " / " mentioned in the present invention can be used to express an "or" relationship.
[0034] The network architecture and business scenarios described in the embodiments of the present invention are intended to more clearly illustrate the technical solutions of the embodiments of the present invention, and do not constitute a limitation on the technical solutions provided by the embodiments of the present invention. A person of ordinary skill in the art can appreciate that with the evolution of the network architecture and the emergence of new business scenarios, the technical solutions provided by the embodiments of the present invention are equally applicable to similar technical problems.
[0035] For example, Figure 1 A schematic flow chart of a method for monitoring rubber accelerator particle diameter based on visual nerves provided in an embodiment of the present invention. The method can be applied to electronic devices.
[0036] like Figure 1 As shown, the process of the rubber accelerator particle diameter monitoring method based on visual nerve is as follows: S101, the electronic device acquires a target image.
[0037] The target image includes a plurality of rubber accelerator particles, and the plurality of rubber accelerator particles include a plurality of batches of rubber accelerator particles. Figure 2 As shown, multiple rubber accelerator particles can be arranged in a matrix on the support plate, and the image acquisition device can shoot from above to obtain the target image, and then send the target image to the electronic device. Among them, for each batch of rubber accelerator particles, they can be concentrated together and can also be arranged in a matrix. For example, a batch contains 9 rubber accelerator particles, which can be arranged in a 3*3 matrix, such as Figure 2 As shown, four batches of rubber accelerator particles are included so that the electronic device can perform diameter detection on multiple batches at one time.
[0038] S102, the electronic device processes the target image through a visual neural network model to obtain the diameters of the plurality of rubber accelerator particles.
[0039] The convolution scale of the visual neural network model matches the size of the rubber accelerator particles in the image.
[0040] Specific: Step 1: The electronic device performs a convolution operation on the target image through at least one convolution kernel in the convolution layer of the visual neural network model to obtain a convolution feature vector of the target image.
[0041] Wherein, at least one convolution kernel is a triangular convolution kernel. The visual neural network model can be a convolutional neural network (CNN).
[0042] Method 1: At least one convolution kernel includes a first convolution kernel, and the first convolution kernel is a triangular convolution kernel including a first convolution kernel of , K is an integer greater than or equal to 5, -2 means decreasing by 2, That is, it means: the size of the first row in the first convolution kernel is K, and the size of each row in the first convolution kernel decreases by 2 in sequence, until the size of the last row in the first convolution kernel is 1 or 2; the size of the step size of the horizontal movement and the step size of the vertical movement of the first convolution kernel are both K and greater than or equal to 1. For example, in one example, the size of the first row of the first convolution kernel is 8, indicating that the features of 8 pixels can be extracted, the size of the second row of the first convolution kernel is 6, indicating that the features of 6 pixels can be extracted, the size of the third row of the first convolution kernel is 4, indicating that the features of 4 pixels can be extracted, and the size of the fourth row of the first convolution kernel is 2, indicating that the features of 2 pixels can be extracted.
[0043] On this basis, if the step length of the lateral movement of the first convolution kernel is set to x1, and the step length of the longitudinal movement of the first convolution kernel is set to y1, the electronic device performs a convolution operation on the target image through at least one convolution kernel in the convolution layer of the visual neural network model to obtain a convolution feature vector of the target image, including: The electronic device performs a convolution operation on the target image through the first convolution kernel in the convolution layer of the visual neural network model with the step sizes of x1 and y1 respectively, to obtain a convolution feature vector of the target image; The target image is an image with a resolution of M*N, where M and N are integers greater than K, M is the number of rows, and N is the number of columns. During the convolution operation, the first convolution kernel performs one convolution from the position at the 0th row and 0th column in the target image, such as Figure 3As shown in (a) in the above example, the left end of the first row of the first convolution kernel is located at the position of the 0th row and the 0th column in the target image. After that, the first convolution kernel moves horizontally in accordance with the step length x1 and performs convolution in sequence, such as x1=2, and the next position can also be Figure 3 As shown in (a) in the figure, it moves to the position at the 0th row and the Nth column in the target image and performs one convolution, as shown in Figure 3 As shown in (a) in the figure, the right end of the first row of the first convolution kernel is located at the position of the 0th row and the Nth column in the target image. The first convolution kernel moves vertically to the position of the y1th row and the Nth column in the target image with a step length of y1 and performs a convolution, such as y1=2. Specifically, it can also be as follows Figure 3 As shown in (a) in the figure, the right end of the first convolution kernel in the first row is located at the position of the second row and the Nth column in the target image. After that, the first convolution kernel moves horizontally in sequence with a step length of x1 and performs convolution in sequence until it moves to the position of the y1th row and the 0th column in the target image and performs one convolution. Specifically, it can also be as follows Figure 3 As shown in (a), specifically, the left end of the first row of the first convolution kernel is located at the position of the second row and the 0th column in the target image. Then, the same process is repeated until the convolution is completed. When the convolution is completed, the first convolution kernel moves to the position of the M-J1th row and the Nth column in the target image, where J1 is the number of rows included in the first convolution kernel. The convolution scale of the visual neural network model matches the size of the rubber accelerator particles in the target image, which means that the distance between the repeatedly convolved feature vectors in the target image is the diameter of the rubber accelerator particles in the target image. Figure 4 As shown, since the shape of the rubber accelerator particles in the target image is a circle, during the convolution process, due to the inverted triangle structure of the first convolution kernel (or a structure similar to an inverted triangle, or even a structure similar to an inverted trapezoid, that is, this structure is not an inverted triangle or an inverted trapezoid, but is similar to this structure as a whole, such as an external shape that can be an inverted triangle or an inverted trapezoid), it can fit the shape of the circle, so that the features that can be repeatedly extracted by the first convolution kernel at different positions can include more of the edge area of the circle, and the position change of the first convolution kernel can also extract features at two positions in the rubber accelerator particles that can be used to determine the diameter, such as Figure 4 The features in region 1 and region 2 in the image are taken as the diameter. The distance between the parts of the rubber accelerator particles in region 1 and region 2 is the diameter. In other words, by setting the first convolution kernel to a triangular structure, more features for calculating the diameter can be extracted, making the robustness and accuracy higher.
[0044] It can be understood that even if some pixels at the edge of the target image are not convolved, since the pixels at these locations usually do not contain images of rubber accelerator particles, they do not affect subsequent processing, such as Figure 3 As shown in (a) in .
[0045] Method 2: At least one convolution kernel includes a second convolution kernel, and the second convolution kernel is a triangular convolution kernel including a second convolution kernel of , K is an integer greater than or equal to 5, ++2 means adding 2 successively, That is, it means: the size of the first row in the second convolution kernel is 1 or 2, and the size of each row in the second convolution kernel is increased by 2 in sequence until the size of the last row in the second convolution kernel is K; the size of the step size of the lateral movement and the step size of the longitudinal movement of the second convolution kernel are both K and greater than or equal to 1. For example, in one example, the size of the first row of the second convolution kernel is 2, indicating that the features of 2 pixels can be extracted, the size of the second row of the second convolution kernel is 4, indicating that the features of 4 pixels can be extracted, the size of the third row of the second convolution kernel is 6, indicating that the features of 6 pixels can be extracted, and the size of the fourth row of the second convolution kernel is 8, indicating that the features of 8 pixels can be extracted.
[0046] On this basis, if the step length of the lateral movement of the second convolution kernel is set to x2, and the step length of the longitudinal movement of the second convolution kernel is set to y2, the electronic device performs a convolution operation on the target image through at least one convolution kernel in the convolution layer of the visual neural network model to obtain a convolution feature vector of the target image, including: The electronic device performs a convolution operation on the target image through the second convolution kernel in the convolution layer of the visual neural network model with the step sizes of x2 and y2 respectively, to obtain a convolution feature vector of the target image; Among them, the target image is an image with a resolution of M*N, M and N are integers greater than K, M is the number of rows, and N is the number of columns; during the convolution operation, the second convolution kernel performs one convolution from the position located at the J2th row and the 0th column in the target image, and J2 is the number of rows contained in the second convolution kernel, such as Figure 3 As shown in (b) in the above example, the left end of the 4th row of the second convolution kernel is located at the 4th row and 0th column in the target image. After that, the second convolution kernel moves horizontally in sequence with a step size of x2 and performs convolution in sequence, such as x2=2. The next position can also be Figure 3 As shown in (b) in the figure, it moves to the position of the J2th row and the Nth column in the target image and performs one convolution. The second convolution kernel moves vertically with a step size of y2 to the position of the y2th + J2th row and the Nth column in the target image and performs one convolution, such as y1=2. Specifically, it can also be as follows Figure 3As shown in (b) in FIG. 1 , the right end of the 4th row of the second convolution kernel is located at the position of the 6th row and the Nth column in the target image. After that, the second convolution kernel moves horizontally in sequence with a step length of x2 and performs convolution in sequence until it moves to the position of the y2th + J2th row and the Nth column in the target image and performs one convolution. Specifically, it can also be as follows: Figure 3 As shown in (b), specifically, the left end of the 4th row of the first convolution kernel is located at the position of the 6th row and the 0th column in the target image. Then, the same process is repeated until the convolution is completed. When the convolution is completed, the second convolution kernel moves to the position of the Mth row and the Nth column in the target image. Among them, the convolution scale of the visual neural network model matches the size of the rubber accelerator particles in the target image, which means that the distance between the repeatedly convolved feature vectors in the target image is the diameter of the rubber accelerator particles in the target image. In other words, the second convolution kernel is set to a triangular structure opposite to the above-mentioned method 1 (or a structure similar to a triangle, or even a structure similar to a trapezoid, that is, this structure is not a triangle or a trapezoid, but is similar to this structure as a whole, such as an external shape that can be a triangle or a trapezoid) can also achieve a similar effect, and can extract more features for calculating the diameter, so that both robustness and accuracy are higher.
[0047] Method 3: The at least one convolution kernel includes a first convolution kernel and a second convolution kernel.
[0048] The first convolution kernel is a triangular convolution kernel. The first convolution kernel is , K is an integer greater than or equal to 5, -2 means decreasing by 2, That is to say, the size of the first row in the first convolution kernel is K, and the size of each row in the first convolution kernel decreases by 2 until the size of the last row in the first convolution kernel is 1 or 2; the size of the horizontal movement step size and the vertical movement step size of the first convolution kernel are both K and greater than or equal to 1.
[0049] The second convolution kernel is a triangular convolution kernel. The second convolution kernel is , K is an integer greater than or equal to 5, ++2 means adding 2 successively, That is to say, the size of the first row in the second convolution kernel is 1 or 2, and the size of each row in the second convolution kernel is increased by 2 until the size of the last row in the second convolution kernel is K; the size of the horizontal movement step size and the vertical movement step size of the second convolution kernel are both K and greater than or equal to 1.
[0050] On this basis, the step length of the lateral movement of the first convolution kernel is set to x1, the step length of the longitudinal movement of the first convolution kernel is set to y1, the step length of the lateral movement of the second convolution kernel is set to x2, and the step length of the longitudinal movement of the second convolution kernel is set to y2. The electronic device performs a convolution operation on the target image through at least one convolution kernel in the convolution layer of the visual neural network model to obtain a convolution feature vector of the target image, including: The electronic device performs a convolution operation on the target image through a first convolution kernel in a convolution layer of the visual neural network model with step sizes of x1 and y1, and simultaneously performs a convolution operation on the target image through a second convolution kernel in a convolution layer of the visual neural network model with step sizes of x2 and y2, to obtain a convolution feature vector of the target image; Wherein, the target image is an image with a resolution of M*N, M and N are integers greater than K, M is the number of rows, and N is the number of columns; During the convolution operation: the first convolution kernel performs one convolution from the position at the 0th row and 0th column in the target image, and at the same time, the second convolution kernel moves to the position at the MJth row and Nth column in the target image, where J1 is the number of rows contained in the first convolution kernel; then, the first convolution kernel moves horizontally with a step size of x1 and performs convolutions in sequence, until it moves to the position at the 0th row and Nth column in the target image and performs one convolution, and at the same time, the second convolution kernel moves horizontally with a step size of x2 and performs convolutions in sequence, until it moves to the position at the J2th row and Nth column in the target image and performs one convolution; then, the first convolution kernel moves vertically with a step size of y1 to the position at the y1th row and Nth column in the target image and performs one convolution, and at the same time, the second convolution kernel moves vertically with a step size of y2 and performs convolutions in sequence, until it moves to the position at the J2th row and Nth column in the target image and performs one convolution; y2 moves vertically to the position of the y2+J2th row and the Nth column in the target image and performs one convolution; then, the first convolution kernel moves horizontally in sequence with a step size of x1 and performs convolution in sequence until it moves to the position of the y1th row and the 0th column in the target image and performs one convolution, and at the same time, the second convolution kernel moves horizontally in sequence with a step size of x2 and performs convolution in sequence until it moves to the position of the y2+J2th row and the Nth column in the target image and performs one convolution, and so on until the convolution is completed; when the convolution is completed, the first convolution kernel moves to the position of the M-J1th row and the Nth column in the target image, J1 is the number of rows contained in the first convolution kernel, and the second convolution kernel moves to the position of the Mth row and the Nth column in the target image.
[0051] The convolution scale of the visual neural network model matches the size of the rubber accelerator particles in the target image, which means that the distance between the repeatedly convolved feature vectors in the target image is the diameter of the rubber accelerator particles in the target image.
[0052] That is to say, the above-mentioned method 1 and method 2 can be combined with embodiments synchronously, and two convolution kernels of opposite shapes are set during the convolution process, and move synchronously, that is, the first convolution kernel and the second convolution kernel both move synchronously from the initial position with the same step size, and when convolution is performed at each position, the features extracted by the first convolution kernel and the second convolution kernel are used as the features extracted by this convolution. In this case, compared with methods 1 and 2, method 3 can also extract more features for calculating the diameter, making the robustness and accuracy higher.
[0053] Step 2: The electronic device convolves the feature vector of the target image through the feature processing layer of the visual neural network model to obtain the diameters of the multiple rubber accelerator particles.
[0054] The feature processing layer includes, in order of processing: a ReLU excitation layer, a pooling layer, and a fully connected layer, which may be a prior art, and the embodiment of the present invention does not limit this. The diameter of each of the plurality of rubber accelerator particles may be the diameter of each rubber accelerator particle marked in the target image, so that the electronic device can determine the diameter of each rubber accelerator particle in each batch.
[0055] Optionally, after S102, the method further includes: The electronic device determines whether there is a batch of rubber accelerator particles with quality problems among the multiple batches of rubber accelerator particles by analyzing the diameters of the multiple rubber accelerator particles. The electronic device can determine the number of rubber accelerator particles with diameters greater than an upper limit or less than a lower limit in each batch of rubber accelerator particles according to the batches in the target image, and if the number exceeds a number threshold, the electronic device determines that the batch is a batch with quality problems.
[0056] In summary: the electronic device can acquire a target image containing multiple batches of rubber accelerator particles, and process the target image through a visual neural network model. Since the convolution scale of the visual neural network model matches the size of the rubber accelerator particles in the image, better feature extraction can be performed during convolution, and the diameters of multiple rubber accelerator particles can be accurately obtained, so as to achieve large-scale, high-precision and stable rubber accelerator particle diameters.
[0057] Combination of the above Figure 1 The rubber accelerator particle diameter monitoring method based on visual nerve provided by the embodiment of the present invention is described in detail. The following is a rubber accelerator particle diameter monitoring system based on visual nerve for executing the rubber accelerator particle diameter monitoring method based on visual nerve provided by the embodiment of the present invention.
[0058] The system includes an electronic device, which is configured as follows: the electronic device acquires a target image, the image contains multiple rubber accelerator particles, and the multiple rubber accelerator particles contain multiple batches of rubber accelerator particles; the electronic device processes the target image through a visual neural network model to obtain the diameters of the multiple rubber accelerator particles output by the visual neural network model, wherein the convolution scale of the visual neural network model matches the size of the rubber accelerator particles in the image.
[0059] Optionally, the electronic device processes the target image through a visual neural network model to obtain the diameters of multiple rubber accelerator particles, including: the electronic device performs a convolution operation on the target image through at least one convolution kernel in a convolution layer of the visual neural network model to obtain a convolution feature vector of the target image, and at least one convolution kernel is a triangular convolution kernel; the electronic device convolves the feature vector of the target image through a feature processing layer of the visual neural network model to obtain the diameters of multiple rubber accelerator particles, wherein the feature processing layer includes, in order of processing: a ReLU excitation layer, a pooling layer, and a fully connected layer.
[0060] Optionally, at least one convolution kernel includes a first convolution kernel, and the first convolution kernel is a triangular convolution kernel including a first convolution kernel of , K is an integer greater than or equal to 5, -2 means decreasing by 2, That is to say, the size of the first row in the first convolution kernel is K, and the size of each row in the first convolution kernel decreases by 2 until the size of the last row in the first convolution kernel is 1 or 2; the size of the horizontal movement step size and the vertical movement step size of the first convolution kernel are both K and greater than or equal to 1.
[0061] Optionally, the step length of the lateral movement of the first convolution kernel is x1, the step length of the longitudinal movement of the first convolution kernel is y1, and the electronic device performs a convolution operation on the target image through at least one convolution kernel in the convolution layer of the visual neural network model to obtain a convolution feature vector of the target image, including: the electronic device performs a convolution operation on the target image through the first convolution kernel in the convolution layer of the visual neural network model with the step lengths of x1 and y1 respectively to obtain the convolution feature vector of the target image; Wherein, the target image is an image with a resolution of M*N, M and N are integers greater than K, M is the number of rows, and N is the number of columns; during the convolution operation, the first convolution kernel performs one convolution from the position of the 0th row and the 0th column in the target image, and then the first convolution kernel moves horizontally in sequence with a step length of x1 and performs convolution in sequence until it moves to the position of the 0th row and the Nth column in the target image and performs one convolution, the first convolution kernel moves vertically in sequence with a step length of y1 to the position of the y1th row and the Nth column in the target image and performs one convolution, and then the first convolution kernel moves horizontally in sequence with a step length of x1 and performs convolution in sequence until it moves to the position of the y1th row and the 0th column in the target image and performs one convolution, and so on, until the convolution is completed, when the convolution is completed, the first convolution kernel moves to the position of the M-J1th row and the Nth column in the target image, and J1 is the number of rows contained in the first convolution kernel; The convolution scale of the visual neural network model matches the size of the rubber accelerator particles in the target image, which means that the distance between the repeatedly convolved feature vectors in the target image is the diameter of the rubber accelerator particles in the target image.
[0062] Optionally, at least one convolution kernel includes a second convolution kernel, and the second convolution kernel is a triangular convolution kernel including a second convolution kernel of , K is an integer greater than or equal to 5, ++2 means adding 2 successively, That is to say, the size of the first row in the second convolution kernel is 1 or 2, and the size of each row in the second convolution kernel is increased by 2 until the size of the last row in the second convolution kernel is K; the size of the horizontal movement step size and the vertical movement step size of the second convolution kernel are both K and greater than or equal to 1.
[0063] Optionally, the step length of the lateral movement of the second convolution kernel is x2, the step length of the longitudinal movement of the second convolution kernel is y2, and the electronic device performs a convolution operation on the target image through at least one convolution kernel in the convolution layer of the visual neural network model to obtain a convolution feature vector of the target image, including: the electronic device performs a convolution operation on the target image through the second convolution kernel in the convolution layer of the visual neural network model with the step lengths of x2 and y2 respectively to obtain the convolution feature vector of the target image; Wherein, the target image is an image with a resolution of M*N, M and N are integers greater than K, M is the number of rows, and N is the number of columns; during the convolution operation, the second convolution kernel performs one convolution from the position located at the J2th row and the 0th column in the target image, J2 is the number of rows contained in the second convolution kernel, and then the second convolution kernel moves horizontally in sequence with a step length of x2 and performs convolution in sequence until it moves to the position located at the J2th row and the Nth column in the target image and performs one convolution, and the second convolution kernel moves vertically in sequence with a step length of y2 to the position located at the y2th+J2th row and the Nth column in the target image and performs one convolution, and then the second convolution kernel moves horizontally in sequence with a step length of x2 and performs convolution in sequence until it moves to the position located at the y2th+J2th row and the Nth column in the target image and performs one convolution, and so on, until the convolution is completed, when the convolution is completed, the second convolution kernel moves to the position located at the Mth row and the Nth column in the target image; The convolution scale of the visual neural network model matches the size of the rubber accelerator particles in the target image, which means that the distance between the repeatedly convolved feature vectors in the target image is the diameter of the rubber accelerator particles in the target image.
[0064] Optionally, at least one convolution kernel includes a first convolution kernel and a second convolution kernel; the first convolution kernel is a triangular convolution kernel including the first convolution kernel is , K is an integer greater than or equal to 5, -2 means decreasing by 2, That is, the size of the first row in the first convolution kernel is K, and the size of each row in the first convolution kernel decreases by 2 until the size of the last row in the first convolution kernel is 1 or 2; the size of the horizontal movement step and the vertical movement step of the first convolution kernel are both K and greater than or equal to 1; the second convolution kernel is a triangular convolution kernel including the second convolution kernel is , K is an integer greater than or equal to 5, ++2 means adding 2 successively, That is to say, the size of the first row in the second convolution kernel is 1 or 2, and the size of each row in the second convolution kernel is increased by 2 until the size of the last row in the second convolution kernel is K; the size of the horizontal movement step size and the vertical movement step size of the second convolution kernel are both K and greater than or equal to 1.
[0065] Optionally, the step size of the lateral movement of the first convolution kernel is x1, the step size of the longitudinal movement of the first convolution kernel is y1, the step size of the lateral movement of the second convolution kernel is x2, and the step size of the longitudinal movement of the second convolution kernel is y2. The electronic device performs a convolution operation on the target image through at least one convolution kernel in the convolution layer of the visual neural network model to obtain a convolution feature vector of the target image, including: the electronic device performs a convolution operation on the target image through the first convolution kernel in the convolution layer of the visual neural network model with a step size of x1 and y1, and simultaneously performs a convolution operation on the target image through the second convolution kernel in the convolution layer of the visual neural network model with a step size of x2 and y2 to obtain a convolution feature vector of the target image; wherein the target image is an image with a resolution of M*N, M and N are integers greater than K, M is the number of rows, and N is the number of columns; During the convolution operation: the first convolution kernel performs one convolution from the position of the 0th row and 0th column in the target image, and at the same time, the second convolution kernel moves to the position of the MJth row and Nth column in the target image, where J1 is the number of rows contained in the first convolution kernel; then, the first convolution kernel moves horizontally with a step size of x1 and performs convolutions in sequence until it moves to the position of the 0th row and Nth column in the target image and performs one convolution, and at the same time, the second convolution kernel moves horizontally with a step size of x2 and performs convolutions in sequence until it moves to the position of the J2th row and Nth column in the target image and performs one convolution; then, the first convolution kernel moves horizontally with a step size of y2 and performs convolutions in sequence until it moves to the position of the J2th row and Nth column in the target image and performs one convolution; 1 moves vertically to the position of the y1th row and Nth column in the target image and performs one convolution. At the same time, the second convolution kernel moves vertically to the position of the y2th + J2th row and Nth column in the target image with a step size of y2 and performs one convolution. After that, the first convolution kernel moves horizontally in sequence with a step size of x1 and performs convolutions in sequence until it moves to the position of the y1th row and 0th column in the target image and performs one convolution. At the same time, the second convolution kernel moves horizontally in sequence with a step size of x2 and performs convolutions in sequence until it moves to the position of the y2th + J2th row and Nth column in the target image and performs one convolution. And so on until the convolution is completed. When the convolution is completed, the first convolution kernel moves to the position at the M-J1th row and the Nth column in the target image, where J1 is the number of rows included in the first convolution kernel, and the second convolution kernel moves to the position at the Mth row and the Nth column in the target image.
[0066] The convolution scale of the visual neural network model matches the size of the rubber accelerator particles in the target image, which means that the distance between the repeatedly convolved feature vectors in the target image is the diameter of the rubber accelerator particles in the target image.
[0067] Optionally, the system is further configured such that: the electronic device determines whether there is a batch with quality problems among the multiple batches of rubber accelerator particles by analyzing the diameters of each of the multiple rubber accelerator particles.
[0068] Figure 5 The electronic device provided in the embodiment of the present invention is a schematic diagram of the structure of the electronic device. For example, the electronic device may be a terminal device, or a chip (system) or other components or assemblies that can be set in the terminal device. Figure 5 As shown, the electronic device 400 may include a processor 401. Optionally, the electronic device 400 may also include a memory 402 and / or a transceiver 403. The processor 401 is coupled to the memory 402 and the transceiver 403, such as being connected via a communication bus. In addition, the electronic device 400 may also be a chip, such as including the processor 401, in which case the transceiver may be an output and input interface of the chip.
[0069] Combine the following Figure 5 The components of the electronic device 400 are specifically introduced as follows: The processor 401 is the control center of the electronic device 400, and may be a processor or a general term for multiple processing elements. For example, the processor 401 is one or more central processing units (CPUs), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present invention, such as one or more microprocessors (digital signal processors, DSPs), or one or more field programmable gate arrays (FPGAs).
[0070] Optionally, the processor 401 can execute various functions of the electronic device 400 by running or executing the software program stored in the memory 402 and calling the scientific data stored in the memory 402, such as executing the above Figure 1 The visual nerve-based rubber accelerator particle diameter monitoring method shown.
[0071] In a specific implementation, as an embodiment, the processor 401 may include one or more CPUs, such as Figure 5 CPU0 and CPU1 are shown in FIG.
[0072] In a specific implementation, as an embodiment, the electronic device 400 may also include multiple processors. Each of these processors may be a single-core processor (single-CPU) or a multi-core processor (multi-CPU). The processor here may refer to one or more devices, circuits, and / or processing cores for processing scientific data (such as computer programs or instructions).
[0073] The memory 402 is used to store the software program for executing the solution of the present invention, and the execution is controlled by the processor 401. The specific implementation method can refer to the above method embodiment, which will not be repeated here.
[0074] Optionally, the memory 402 may be a read-only memory (ROM) or other types of static storage devices that can store static information and instructions, a random access memory (RAM) or other types of dynamic storage devices that can store information and instructions, or an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical disc, laser disc, optical disc, digital versatile disc, Blu-ray disc, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of instructions or scientific data structures and can be accessed by a computer, but is not limited thereto. The memory 402 may be integrated with the processor 401, or may exist independently and accessed through the interface circuit ( Figure 5 The processor 401 is coupled to the processor 401 (not shown), which is not specifically limited in this embodiment of the present invention.
[0075] The transceiver 403 is used for communication with other electronic devices. For example, if the electronic device 400 is a terminal device, the transceiver 403 can be used to communicate with a network device, or with another terminal device. For another example, if the electronic device 400 is a network device, the transceiver 403 can be used to communicate with a terminal device, or with another network device.
[0076] Optionally, the transceiver 403 may include a receiver and a transmitter ( Figure 5 The receiver is used to implement a receiving function, and the transmitter is used to implement a sending function.
[0077] Optionally, the transceiver 403 may be integrated with the processor 401, or may exist independently and communicate with the electronic device 400 through an interface circuit ( Figure 5 The processor 401 is coupled to the processor 401 (not shown), which is not specifically limited in this embodiment of the present invention.
[0078] Understandably, Figure 5 The structure of the electronic device 400 shown in the figure does not constitute a limitation on the electronic device, and the actual electronic device may include more or less components than those shown in the figure, or combine certain components, or arrange the components differently.
[0079] In addition, the technical effects of the electronic device 400 can refer to the technical effects of the methods described in the above method embodiments, which will not be repeated here.
[0080] It should be understood that the processor in the embodiment of the present invention may be a central processing unit (CPU), and the processor may also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.
[0081] It should also be understood that the memory in the embodiments of the present invention may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memories. Among them, the non-volatile memory may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory may be a random access memory (RAM), which is used as an external cache. By way of example and not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic random access memory (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), and direct rambus RAM (DR RAM).
[0082] The above embodiments can be implemented in whole or in part by software, hardware (such as circuits), firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present invention is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer program or instruction can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer program or instruction can be transmitted from one website, computer, server or scientific data center to another website, computer, server or scientific data center by wired (such as infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer or a scientific data storage device such as a server, scientific data center, etc. that contains one or more available media sets. The available medium can be a magnetic medium (such as a floppy disk, a hard disk, a tape), an optical medium (such as a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.
[0083] It should be understood that the term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. A and B can be singular or plural. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship, but it may also indicate an "and / or" relationship. Please refer to the context for specific understanding.
[0084] In the present invention, "at least one" means one or more, and "plurality" means two or more. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single items or plural items. For example, at least one of a, b, or c can mean: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or multiple.
[0085] It should be understood that in various embodiments of the present invention, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.
[0086] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.
[0087] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.
[0088] In the several embodiments provided by the present invention, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.
[0089] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
[0090] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.
[0091] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, and other media that can store program codes.
[0092] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.
Claims
1. A method for monitoring rubber accelerator particle diameter based on visual nerves, characterized in that: Applied to electronic equipment, the method comprises: The electronic device acquires a target image, wherein the target image includes a plurality of rubber accelerator particles, and the plurality of rubber accelerator particles include a plurality of batches of rubber accelerator particles; The electronic device processes the target image through a visual neural network model to obtain the diameter of each of the plurality of rubber accelerator particles, wherein the convolution scale of the visual neural network model matches the size of the rubber accelerator particles in the image.
2. The method for monitoring rubber accelerator particle diameter based on visual nerve according to claim 1, characterized in that: The electronic device processes the target image through a visual neural network model to obtain the diameter of each of the plurality of rubber accelerator particles, including: The electronic device performs a convolution operation on the target image through at least one convolution kernel in the convolution layer of the visual neural network model to obtain a convolution feature vector of the target image, wherein the at least one convolution kernel is a triangular convolution kernel; The electronic device obtains the diameter of each of the plurality of rubber accelerator particles by convolving the feature vector of the target image through a feature processing layer of a visual neural network model, wherein the feature processing layer includes, in order of processing, a ReLU excitation layer, a pooling layer, and a fully connected layer.
3. The method for monitoring rubber accelerator particle diameter based on visual nerve according to claim 2, characterized in that: The at least one convolution kernel includes a first convolution kernel, and the first convolution kernel is a triangular convolution kernel. The first convolution kernel includes , K is an integer greater than or equal to 5, -2 means decreasing by 2, That is to say, the size of the first row in the first convolution kernel is K, and the size of each row in the first convolution kernel decreases by 2 until the size of the last row in the first convolution kernel is 1 or 2; the size of the horizontal movement step size and the vertical movement step size of the first convolution kernel are both K and greater than or equal to 1.
4. The method for monitoring rubber accelerator particle diameter based on visual nerve according to claim 3, characterized in that: The step length of the lateral movement of the first convolution kernel is x1, the step length of the longitudinal movement of the first convolution kernel is y1, and the electronic device performs a convolution operation on the target image through at least one convolution kernel in the convolution layer of the visual neural network model to obtain a convolution feature vector of the target image, including: The electronic device performs a convolution operation on the target image through the first convolution kernel in the convolution layer of the visual neural network model with the step sizes of x1 and y1 respectively, to obtain a convolution feature vector of the target image; Wherein, the target image is an image with a resolution of M*N, M and N are integers greater than K, M is the number of rows, and N is the number of columns; during the convolution operation, the first convolution kernel performs one convolution from the position of the 0th row and the 0th column in the target image, and then the first convolution kernel is sequentially moved horizontally according to the step size x1 and sequentially convolved until it moves to the position of the 0th row and the Nth column in the target image and performs one convolution, the first convolution kernel is vertically moved to the position of the y1th row and the Nth column in the target image according to the step size y1 and performs one convolution, and then the first convolution kernel is sequentially moved horizontally according to the step size x1 and sequentially convolved until it moves to the position of the y1th row and the 0th column in the target image and performs one convolution, and so on, until the convolution is completed, when the convolution is completed, the first convolution kernel moves to the position of the M-J1th row and the Nth column in the target image, and J1 is the number of rows contained in the first convolution kernel; The convolution scale of the visual neural network model matches the size of the rubber accelerator particles in the target image, which means that the distance between the repeatedly convolved feature vectors in the target image is the diameter of the rubber accelerator particles in the target image.
5. The method for monitoring rubber accelerator particle diameter based on visual nerve according to claim 2, characterized in that: The at least one convolution kernel includes a second convolution kernel, and the second convolution kernel is a triangular convolution kernel. The second convolution kernel includes , K is an integer greater than or equal to 5, ++2 means adding 2 successively, That is to say, the size of the first row in the second convolution kernel is 1 or 2, and the size of each row in the second convolution kernel is increased by 2 until the size of the last row in the second convolution kernel is K; the size of the horizontal movement step size and the vertical movement step size of the second convolution kernel are both K and greater than or equal to 1.
6. The method for monitoring rubber accelerator particle diameter based on visual nerve according to claim 5, characterized in that: The step length of the lateral movement of the second convolution kernel is x2, the step length of the longitudinal movement of the second convolution kernel is y2, and the electronic device performs a convolution operation on the target image through at least one convolution kernel in the convolution layer of the visual neural network model to obtain a convolution feature vector of the target image, including: The electronic device performs a convolution operation on the target image through the second convolution kernel in the convolution layer of the visual neural network model with the step sizes of x2 and y2 respectively, to obtain a convolution feature vector of the target image; Wherein, the target image is an image with a resolution of M*N, M and N are integers greater than K, M is the number of rows, and N is the number of columns; during the convolution operation, the second convolution kernel performs one convolution from the position located at the J2th row and the 0th column in the target image, J2 is the number of rows contained in the second convolution kernel, and then the second convolution kernel moves horizontally in sequence according to the step length x2 and performs convolution in sequence until it moves to the position located at the J2th row and the Nth column in the target image and performs one convolution, the second convolution kernel moves vertically in sequence according to the step length y2 to the position located at the y2th+J2th row and the Nth column in the target image and performs one convolution, and then the second convolution kernel moves horizontally in sequence according to the step length x2 and performs convolution in sequence until it moves to the position located at the y2th+J2th row and the Nth column in the target image and performs one convolution, and so on, until the convolution is completed, when the convolution is completed, the second convolution kernel moves to the position located at the Mth row and the Nth column in the target image; The convolution scale of the visual neural network model matches the size of the rubber accelerator particles in the target image, which means that the distance between the repeatedly convolved feature vectors in the target image is the diameter of the rubber accelerator particles in the target image.
7. The method for monitoring rubber accelerator particle diameter based on visual nerve according to claim 2, characterized in that: The at least one convolution kernel includes a first convolution kernel and a second convolution kernel; The first convolution kernel is a triangular convolution kernel including: , K is an integer greater than or equal to 5, -2 means decreasing by 2, That is, it means that the size of the first row in the first convolution kernel is K, and the size of each row in the first convolution kernel decreases by 2 in sequence until the size of the last row in the first convolution kernel is 1 or 2; the size of the horizontal movement step size and the vertical movement step size of the first convolution kernel are both K and greater than or equal to 1; The second convolution kernel is a triangular convolution kernel including the first convolution kernel being , K is an integer greater than or equal to 5, ++2 means adding 2 successively, That is to say, the size of the first row in the second convolution kernel is 1 or 2, and the size of each row in the second convolution kernel is increased by 2 until the size of the last row in the second convolution kernel is K; the size of the horizontal movement step size and the vertical movement step size of the second convolution kernel are both K and greater than or equal to 1.
8. The method for monitoring rubber accelerator particle diameter based on visual nerve according to claim 7, characterized in that: The step length of the lateral movement of the first convolution kernel is x1, the step length of the longitudinal movement of the first convolution kernel is y1, the step length of the lateral movement of the second convolution kernel is x2, and the step length of the longitudinal movement of the second convolution kernel is y2. The electronic device performs a convolution operation on the target image through at least one convolution kernel in the convolution layer of the visual neural network model to obtain a convolution feature vector of the target image, including: The electronic device performs a convolution operation on the target image through the first convolution kernel in the convolution layer of the visual neural network model with the step sizes of x1 and y1, and simultaneously performs a convolution operation on the target image through the second convolution kernel in the convolution layer of the visual neural network model with the step sizes of x2 and y2, to obtain a convolution feature vector of the target image; Wherein, the target image is an image with a resolution of M*N, M and N are integers greater than K, M is the number of rows, and N is the number of columns; During the convolution operation: the first convolution kernel performs one convolution from the position at the 0th row and the 0th column in the target image, and at the same time, the second convolution kernel moves to the position at the MJth row and the Nth column in the target image, where J1 is the number of rows included in the first convolution kernel; Afterwards, the first convolution kernel sequentially moves horizontally according to the step size x1 and performs convolution sequentially until it moves to the position of the 0th row and the Nth column in the target image and performs one convolution, and at the same time, the second convolution kernel sequentially moves horizontally according to the step size x2 and performs convolution sequentially until it moves to the position of the J2th row and the Nth column in the target image and performs one convolution; Afterwards, the first convolution kernel is longitudinally moved to the position of the y1th row and the Nth column in the target image according to the step size y1 and performs one convolution, and at the same time, the second convolution kernel is longitudinally moved to the position of the y2th row and the Nth column in the target image according to the step size y2 and performs one convolution; Afterwards, the first convolution kernel moves laterally in sequence according to the step size x1 and performs convolution in sequence until it moves to the position of the y1th row and the 0th column in the target image and performs one convolution, and at the same time, the second convolution kernel moves laterally in sequence according to the step size x2 and performs convolution in sequence until it moves to the position of the y2th + J2th row and the Nth column in the target image and performs one convolution, and so on, until the convolution is completed; When the convolution is completed, the first convolution kernel moves to a position located at the M-J1th row and the Nth column in the target image, where J1 is the number of rows included in the first convolution kernel, and the second convolution kernel moves to a position located at the Mth row and the Nth column in the target image; The convolution scale of the visual neural network model matches the size of the rubber accelerator particles in the target image, which means that the distance between the repeatedly convolved feature vectors in the target image is the diameter of the rubber accelerator particles in the target image.
9. A method for monitoring rubber accelerator particle diameter based on visual nerve according to any one of claims 1 to 8, characterized in that: The method further comprises: The electronic device determines whether there is a batch with quality problems among the batches of rubber accelerator particles by analyzing the diameters of the respective batches of rubber accelerator particles.
10. A rubber accelerator particle diameter monitoring system based on visual nerves, characterized in that: The system comprises an electronic device, wherein the system is configured to: The electronic device acquires a target image, wherein the target image includes a plurality of rubber accelerator particles, and the plurality of rubber accelerator particles include a plurality of batches of rubber accelerator particles; The electronic device processes the target image through a visual neural network model to obtain the diameter of each of the multiple rubber accelerator particles output by the visual neural network model, wherein the convolution scale of the visual neural network model matches the size of the rubber accelerator particles in the image.
Citation Information
Patent Citations
Intelligent particle size detection method using multi-scale feature weighting
CN112800995A
Coal dust particle feature extraction method
CN113449743A
Automatic particle size control method and device for granulation production and readable medium
CN114565562A
Adversarial training method based on linear piecewise loss function
CN116012672A
Industrial park environment monitoring method and system
CN117011782A
