Steel seal position identification method and system

Abstract

According to the steel seal position identification method and system provided by the invention, the steel seal features of the preprocessed steel seal images with different depths are acquired, the ink concentration model is constructed and trained, the steel seal features are substituted into the model, and the ink concentrations corresponding to the preprocessed steel seal images with different depths are calculated; constructing a background suppression model through the ink concentration; and identifying steel seal positions in the preprocessed steel seal images with different depths by combining an edge detection algorithm and morphological operation. Compared with the prior art, the background suppression model is set to effectively recognize the position of the steel seal in combination with an edge detection algorithm and morphological operation in consideration that the influence of the color shade of the steel seal image caused by the ink concentration on the recognition of the position of the steel seal is large, so that the manual interference is reduced, and the recognition accuracy and the recognition speed are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of tobacco production, and particularly to a method and system for identifying the position of a steel seal. Background Art

[0002] The cigarette making and tipping unit is equipped with an on-line cigarette paper printing system, which can instantaneously print characters and patterns during production, and can also adopt single-color marks or double-color marks according to product design, simply referred to as single steel seal or double steel seal.

[0003] During the production process of cigarette rods, the accurate alignment of the cigarette rod steel seal is a technical challenge commonly faced in the industry. Although the ZJ119 type cigarette making and tipping unit is recognized for its high production speed and stable cigarette rod quality, there is still room for improvement in the accuracy of the steel seal. With the increasing requirements of each cigarette factory for product quality and the pursuit of "exquisite" products, higher requirements are put forward for the steel seal accuracy of the cigarette making and tipping unit.

[0004] Currently, there are many factors causing the steel seal to deviate, such as the difference between cold machine operation and long-term operation, different tension levels of the cloth belt pulley, creep gradually generated by the cigarette gun cloth belt due to long-term heating and tension, the influence of tobacco density, clearance caused by wear of relevant couplings, the stretchability of the cigarette paper tape, the paper tape tension balance mechanism, and the mechanical assembly accuracy of the brand box parts. The position of the steel seal after deviation needs to be immediately and accurately identified in order to be corrected in time as a benchmark, thereby avoiding production accidents.

[0005] During the existing steel seal position recognition process, it is easily affected by the change in image color shade, resulting in a decrease in recognition accuracy and speed, and more manual intervention is required.

[0006] Therefore, it is an urgent problem for those skilled in the art to provide a method and system for identifying the position of a steel seal to solve the above problems. Summary of the Invention

[0007] The purpose of the present invention is to provide a method for identifying the position of a steel seal, which has clear logic, is safe, effective, reliable and easy to operate, can reduce the influence of color shade on the steel seal position recognition, reduce manual intervention, and improve the recognition accuracy and speed.

[0008] Based on the above purpose, the technical solution provided by the present invention is as follows: A method for identifying the position of a steel seal, comprising the following steps: Obtain the steel seal features of the preprocessed steel seal images with different shades; Substitute the steel seal features of the preprocessed steel seal images with different shades into the trained ink concentration model, and calculate and obtain the ink concentration corresponding to the preprocessed steel seal images with different shades; Construct a background suppression model based on the ink concentrations corresponding to the pre-processed steel seal images with different shades; Identify the positions of the steel seals in the pre-processed steel seal images with different shades according to the background suppression model, edge detection algorithm, and morphological operations.

[0009] Preferably, the trained ink concentration model is determined through the following steps: Obtain the steel seal features in the image training samples with different ink concentrations after pre-processing; Construct an initial ink concentration model based on a support vector machine; Iteratively train the initial ink concentration model according to the steel seal features in the image training samples with different ink concentrations after pre-processing until the model converges to obtain the trained ink concentration model.

[0010] Preferably, the obtaining of the steel seal features in the image training samples with different ink concentrations after pre-processing includes the following steps: Adjust the ink concentration to obtain multiple image training samples with gradually changing shades and containing steel seals; Pre-process multiple image training samples with gradually changing shades and containing steel seals to obtain the image training samples with different ink concentrations after pre-processing; Extract the steel seal features in the image training samples with different ink concentrations after pre-processing according to the histogram of oriented gradients.

[0011] Preferably, the pre-processing of multiple image training samples with gradually changing shades and containing steel seals to obtain the image training samples with different ink concentrations after pre-processing includes the following steps: Grayscale multiple image training samples with gradually changing shades and containing steel seals to obtain the grayscale histogram of the image training samples; Perform image segmentation, denoising, image enhancement, filtering, and binarization on the grayscale histogram of the image training samples in sequence to obtain the image training samples with different ink concentrations after pre-processing.

[0012] Preferably, the performing of image segmentation, denoising, image enhancement, filtering, and binarization on the grayscale histogram of the image training samples in sequence is specifically as follows: Perform image segmentation on the grayscale histogram of the image training samples by the threshold segmentation method; Perform denoising through Gaussian filtering or median filtering; Equalize the grayscale histogram of the image training samples to enhance the contrast of the image; Highlight the edge information in the image through Laplacian filtering or directional filtering; Convert it into a binary image.

[0013] Preferably, identifying the position of the steel seal in the pre - processed steel seal images with different depths and shades according to the background suppression model, edge detection algorithm, and morphological operation includes the following steps: Determine the image area containing the position of the steel seal in the pre - processed steel seal image according to the background suppression model; Identify the edge contour of the steel seal in the image area through the Canny edge detection algorithm; Refine the edge contour through morphological operations to determine the position of the steel seal in the pre - processed steel seal images with different depths and shades.

[0014] A steel seal position recognition system includes: A steel seal feature acquisition module, configured to acquire the steel seal features of the pre - processed steel seal images with different depths and shades; An ink concentration module, configured to substitute the steel seal features of the pre - processed steel seal images with different depths and shades into the trained ink concentration model to calculate and obtain the ink concentrations corresponding to the pre - processed steel seal images with different depths and shades; A background suppression model construction module, configured to construct a background suppression model according to the ink concentrations corresponding to the pre - processed steel seal images with different depths and shades; A steel seal position recognition module, configured to identify the position of the steel seal in the pre - processed steel seal images with different depths and shades according to the background suppression model, edge detection algorithm, and morphological operation.

[0015] Preferably, it further includes: an ink concentration model construction module; The ink concentration model construction module is configured to acquire the steel seal features in the image training samples with different pre - processed ink concentrations; The ink concentration model construction module is further configured to construct an initial ink concentration model based on the support vector machine; The ink concentration model construction module is further configured to iteratively train the initial ink concentration model according to the steel seal features in the image training samples with different pre - processed ink concentrations until the model converges to obtain the trained ink concentration model.

[0016] The steel seal position recognition method provided by the present invention is to acquire the steel seal features of the pre - processed steel seal images with different depths and shades, construct and train an ink concentration model, substitute the steel seal features into the model to calculate and obtain the ink concentrations corresponding to the pre - processed steel seal images with different depths and shades, construct a background suppression model through the ink concentrations; and combine the edge detection algorithm and morphological operation to identify the position of the steel seal in the pre - processed steel seal images with different depths and shades.

[0017] Compared with the prior art, considering that the color shade of the embossed image due to the ink concentration has a great influence on the recognition of the embossed position, a background suppression model is set up, and the edge detection algorithm and morphological operations are combined to effectively recognize the embossed position, reducing manual interference and improving the recognition accuracy and speed.

[0018] The present invention also provides a system for recognizing the position of an embossed seal. Since it belongs to the same technical concept as this method and solves the same technical problems, it should have the same beneficial effects, which will not be elaborated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0020] Figure 1 It is a flowchart of a method for recognizing the position of an embossed seal provided by an embodiment of the present invention; Figure 2 It is a flowchart of determining the trained ink concentration model provided by an embodiment of the present invention; Figure 3 It is a flowchart of step A1 provided by an embodiment of the present invention; Figure 4 It is a flowchart of step B2 provided by an embodiment of the present invention; Figure 5 It is a flowchart of step S4 provided by an embodiment of the present invention; Figure 6 It is a schematic structural diagram of a system for recognizing the position of an embossed seal provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0021] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0022] The embodiments of the present invention are written in a progressive manner.

[0023] The embodiments of the present invention provide a method and a system for recognizing the position of an embossed seal. The main technical problem to be solved is that in the prior art, during the recognition of the position of an embossed seal, it is easily affected by the change of the color shade of the image, resulting in a decrease in the recognition accuracy and speed, and more manual interference is required.

[0024] As Figure 1 shown, a method for identifying the position of a steel seal includes the following steps: S1. Obtain the steel seal features of the preprocessed steel seal images with different shades; S2. Substitute the steel seal features of the preprocessed steel seal images with different shades into the trained ink concentration model, and calculate and obtain the ink concentrations corresponding to the preprocessed steel seal images with different shades; S3. Construct a background suppression model according to the ink concentrations corresponding to the preprocessed steel seal images with different shades; S4. Identify the position of the steel seal in the preprocessed steel seal images with different shades according to the background suppression model, edge detection algorithm, and morphological operations.

[0025] In step S1, by adjusting the ink concentration and setting a steel seal position detector at a specific position, multiple steel seal images with different ink concentrations are collected, and the steel seal features are extracted after preprocessing; In step S2, the extracted steel seal features are input into the trained ink concentration model to obtain the ink concentrations corresponding to the preprocessed steel seal images with different shades; In step S3, gradient matching is implemented according to the ink concentration corresponding to each steel seal image to construct a background suppression model; In step S4, according to the constructed background suppression model, combined with the edge detection algorithm and morphological operations, the position of the steel seal in the preprocessed steel seal images with different shades is identified.

[0026] As Figure 2 shown, preferably, the trained ink concentration model is determined through the following steps: A1. Obtain the steel seal features in the image training samples with different ink concentrations after preprocessing; A2. Construct an initial ink concentration model based on a support vector machine; A3. Iteratively train the initial ink concentration model according to the steel seal features in the image training samples with different ink concentrations after preprocessing until the model converges to obtain the trained ink concentration model.

[0027] In steps A1 to A3, the trained ink concentration model is obtained by obtaining the steel seal features in the image training samples with different ink concentrations after preprocessing in the manner of step S1; constructing an initial ink concentration model based on a support vector machine; inputting the steel seal features in the image training samples into the initial ink concentration model for iterative training, and continuously adjusting and optimizing parameters during the training process, such as selecting a suitable kernel function and adjusting the penalty parameter C, until the model converges to obtain the trained ink concentration model; It should be noted that the Support Vector Machine (SVM) is a class of generalized linear classifiers for binary classification of data in a supervised learning manner. Its decision boundary is the maximum-margin hyperplane solved for the learning samples. SVM uses the hinge loss function to calculate the empirical risk and adds a regularization term to the solution system to optimize the structural risk. It is a classifier with sparsity and robustness. SVM can perform non-linear classification through the kernel method and is one of the common kernel learning methods.

[0028] As Figure 3 shown, preferably, step A1 includes the following steps: B1. Adjust the ink concentration to obtain multiple image training samples with gradually changing shades of a steel seal. B2. Preprocess the multiple image training samples with gradually changing shades of a steel seal to obtain preprocessed image training samples with different ink concentrations. B3. Extract the steel seal features from the preprocessed image training samples with different ink concentrations according to the histogram of oriented gradients.

[0029] In steps B1 to B3, by adjusting the ink concentration, the color of the steel seal presented in the image training samples has gradually changing shades, multiple image training samples with gradually changing shades of a steel seal are obtained, the training samples are preprocessed, and the steel seal features in the preprocessed image samples are extracted by means of the histogram of oriented gradients.

[0030] It should be noted that the histogram of oriented gradient (HOG for short) is a feature descriptor used in the fields of computer vision and image processing for object detection. This technique is used to calculate the statistical value of the direction information of local image gradients. Its principle is specifically as follows: In an image, the appearance and shape of local objects can be well described by the direction density distribution of gradients or edges. The specific implementation method is: First, divide the image into small connected regions, which we call cell units. Then collect the direction histograms of the gradients or edges of each pixel point in the cell units. Finally, combine these histograms to form a feature descriptor. To improve performance, we can also perform contrast normalization on these local histograms in a larger range of the image (which we call an interval or block). The method used is: First, calculate the density of each histogram in this interval (block), and then normalize each cell unit in the interval according to this density. After this normalization, better effects can be obtained for illumination changes and shadows.

[0031] As Figure 4 shown, preferably, step B2 includes the following steps: C1. Grayscale multiple training image samples of gradually changing shades with embossing to obtain the grayscale histogram of the training image samples; C2. Successively perform image segmentation, denoising, image enhancement, filtering, and binarization on the grayscale histogram of the training image samples to obtain the preprocessed training image samples with different ink concentrations.

[0032] In steps C1 to C2, the training image samples are grayscaled to obtain the grayscale histogram; and successively perform operations such as image segmentation, denoising, image enhancement, filtering, and binarization to obtain the preprocessed training image samples with different ink concentrations.

[0033] It should be noted that grayscaling is the process of converting a color image into a grayscale image, aiming to simplify image data and improve the operation speed. In a color image, the color of each pixel is determined by three components: R (red), G (green), and B (blue), and the value range of each component is between 0 and 255. Therefore, there are 16,777,216 color variations for a pixel point in a color image. A grayscale image is a special color image with the same R, G, and B components, and the variation range of each pixel point is only 0 to 255, greatly reducing the amount of data. There are various grayscaling methods, including: the maximum value method: using the maximum value of the brightness of the three components in the color image as the grayscale value of the grayscale image; the average value method: averaging the brightness of the three components in the color image to obtain a grayscale value; the weighted average method: weighted averaging the three components with different weights according to importance and other indicators. Since the human eye is most sensitive to green and least sensitive to blue, a more reasonable grayscale image can be obtained by performing weighted averaging on the RGB three components according to the following formula. Grayscaling is very important in image recognition because it can simplify image data, reduce computational complexity, and thus improve the operation speed and processing efficiency.

[0034] It should also be noted that image preprocessing is an operation on the image at the lowest abstraction level and is called "preprocessing". Both the input and output are luminance images. The purpose of preprocessing is to suppress unwanted deformations or enhance certain image features important for subsequent processing. Preprocessing does not increase the amount of information in the image. Generally, it reduces the amount of information, and geometric transformation of the image is also regarded as a "preprocessing method".

[0035] Preferably, step C2 is specifically: Segment the grayscale histogram of the training image samples by the threshold segmentation method; Denoise by Gaussian filtering or median filtering; Equalize the grayscale histogram of the image training samples to enhance the image contrast; Highlight the edge information in the image through Laplacian filtering or directional filtering; Convert it into a binary image.

[0036] In the actual application process, the threshold segmentation method is a region-based image segmentation technology. Its basic principle is to divide the pixel points of the image into several categories. The threshold segmentation method divides the pixel points of the image into several categories by setting different feature thresholds. The commonly used features include the grayscale or color features directly from the original image, as well as the features transformed from the original grayscale or color values.

[0037] Gaussian filtering is a linear smoothing filter, suitable for eliminating Gaussian noise and widely used in the denoising process of image processing. Generally speaking, Gaussian filtering is a process of weighted averaging of the entire image. The value of each pixel point is obtained by weighted averaging of itself and other pixel values in the neighborhood. The specific operation of Gaussian filtering is as follows: scan each pixel in the image with a template (or convolution, mask), and replace the value of the pixel at the center of the template with the weighted average grayscale value of the pixels in the neighborhood determined by the template; the median filtering method is a non-linear smoothing technology. It sets the grayscale value of each pixel point to the median of all pixel point grayscale values in a certain neighborhood window of this point. Median filtering is a non-linear signal processing technology based on sorting statistics theory that can effectively suppress noise. The basic principle of median filtering is to replace the value of a point in a digital image or digital sequence with the median of the values of each point in a neighborhood of this point, making the surrounding pixel values close to the true values, thereby eliminating isolated noise points. The method is to use a two-dimensional sliding template of a certain structure, sort the pixels in the template according to the pixel values, and generate a two-dimensional data sequence that increases (or decreases) monotonically.

[0038] Laplacian filtering is a classic linear filtering method, commonly used in image and signal processing. In a 3D model, Laplacian filtering can smooth the model surface and remove high-frequency noise. This filter is achieved by calculating the curvature of the model surface at each vertex. Curvature is an index describing the surface change rate, with a lower value for flat areas and a higher value for sharp edges and details.

[0039] Directional filtering selectively enhances the gradient features in specific directions. The sum of the elements of the directional filtering transformation kernel is "0". In the output image, the value of the area where the pixel gray values are exactly the same is "0" after filtering, and the area where the gray values change violently appears as a brighter edge.

[0040] Such as Figure 5 shown, preferably, step S4 includes the following steps: D1. Determine the image area containing the position of the steel seal in the preprocessed steel seal image according to the background suppression model; D2. Identify the edge contour of the steel seal in the image area through the Canny edge detection algorithm; D3. Refine the edge contour through morphological operations to determine the position of the steel seal in the preprocessed steel seal images with different shades.

[0041] In steps D1 to D3, the area where the steel seal is located is highlighted through the background suppression model, the edge contour of the steel seal in this area is identified by combining the Canny edge detection algorithm, and further the edge contour is refined through morphological operations, so as to determine the position of the steel seal.

[0042] It should be noted that for the Canny edge detection algorithm, specifically, the Sobel operator is used to calculate the gradient values of the image in the horizontal and vertical directions, so as to obtain the gradient intensity and direction of each pixel point. This step helps to determine the edge direction in the image. Non-maximum suppression: By comparing the gradient values of the neighboring pixel points of each pixel point in its gradient direction, only the pixel points with the maximum gradient value are retained to eliminate the blurred effect on the edge. Double-threshold processing and edge connection: Set two thresholds, high and low. The pixel points above the high threshold are classified as strong edges, the pixel points below the low threshold are classified as non-edges, and the pixel points between the two are classified as weak edges. The weak edges are connected to the strong edges to form a complete edge.

[0043] Morphological operations are a technique widely used in the fields of image processing and machine vision, especially in the stages of image preprocessing and feature extraction. These operations are based on the shape structure of the image and are mainly applied to binary images, but can also be extended to grayscale images. Morphological operations are based on concepts such as dilation and erosion in set theory, and more complex operations such as opening and closing are derived through the combination of these two basic operations.

[0044] As Figure 6 shown, a steel seal position recognition system includes: A steel seal feature acquisition module for acquiring the steel seal features of preprocessed steel seal images with different shades; An ink concentration module for substituting the steel seal features of preprocessed steel seal images with different shades into the trained ink concentration model to calculate and obtain the ink concentration corresponding to the preprocessed steel seal images with different shades; A background suppression model construction module for constructing a background suppression model according to the ink concentration corresponding to the preprocessed steel seal images with different shades; The steel seal position recognition module is used to recognize the steel seal position in the pre - processed steel seal images with different depths according to the background suppression model, edge detection algorithm and morphological operations.

[0045] Preferably, it further includes: an ink concentration model construction module; The ink concentration model construction module is used to obtain the steel seal features in the image training samples with different ink concentrations after pre - processing; The ink concentration model construction module is also used to construct an initial ink concentration model based on the support vector machine; The ink concentration model construction module is also used to iteratively train the initial ink concentration model according to the steel seal features in the image training samples with different ink concentrations after pre - processing until the model converges, and obtain the trained ink concentration model.

[0046] In the actual application process, the present invention also provides a steel seal position recognition system. In this recognition system, there are a steel seal feature acquisition module, an ink concentration module, a background suppression model construction module, a steel seal position recognition module and an ink concentration model construction module. The functions of each module correspond to each step of the steel seal position recognition method, and have the same beneficial effects, which will not be elaborated here.

[0047] In the embodiments provided in the present application, it should be understood that the disclosed methods and systems can be implemented in other ways. The system embodiments described above are only illustrative. For example, the division of modules is only a logical function division. In actual implementation, there may be other division methods, such as: multiple modules or components can be combined, or can be integrated into another system, or some features can be ignored, or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed with each other can be through some interfaces, and the indirect coupling or communication connection of devices or modules can be electrical, mechanical or other forms.

[0048] In addition, in each embodiment of the present invention, each functional module can be all integrated in one processor, or each module can be separately used as a device alone, or two or more modules can be integrated in one device; each functional module in each embodiment of the present invention can be implemented in the form of hardware, or in the form of a combination of hardware and software functional units.

[0049] Those of ordinary skill in the art will understand that all or part of the steps to implement the above method embodiments can be completed through program instructions and related hardware. The foregoing program instructions can be stored in a computer-readable storage medium. When the program instructions are executed, the steps of the above method embodiments are performed; and the foregoing storage medium includes: various media such as removable storage devices, read-only memory (ROM), magnetic disks, or optical discs that can store program codes.

[0050] It should be understood that in this application, if the terms "system", "device", "unit" and / or "module" are used, they are only a method for distinguishing different components, elements, parts, portions or assemblies at different levels. However, if other words can achieve the same purpose, they can be replaced by other expressions.

[0051] As shown in this application and the claims, unless the context clearly indicates an exception, words such as "a", "an", "one" and / or "the" are not specifically singular and may also include the plural. Generally speaking, the terms "comprising" and "including" only indicate the inclusion of the steps and elements that have been clearly identified, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements. An element defined by the statement "comprising one..." does not exclude the existence of other identical elements in the process, method, commodity or device including the element.

[0052] If a flowchart is used in this application, the flowchart is used to illustrate the operations performed by the system according to the embodiments of this application. It should be understood that the previous or subsequent operations are not necessarily executed precisely in order. On the contrary, the steps can be executed in reverse order or simultaneously. At the same time, other operations can also be added to these processes, or one or several steps can be removed from these processes.

[0053] The above has introduced in detail a method and system for steel seal position recognition provided by the present invention. The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be obvious to those skilled in the art. The general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to these embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for identifying the position of a steel seal, characterized in that, Including the following steps: Obtain the steel seal features of preprocessed steel seal images with different depths; Substitute the steel seal features of the preprocessed steel seal images with different depths into the trained ink concentration model, and calculate and obtain the ink concentration corresponding to the preprocessed steel seal images with different depths; Construct a background suppression model according to the ink concentration corresponding to the preprocessed steel seal images with different depths; Identify the steel seal positions in the preprocessed steel seal images with different depths according to the background suppression model, edge detection algorithm and morphological operation.

2. The steel seal position recognition method according to claim 1, characterized in that The trained ink concentration model is determined through the following steps: Obtain the steel seal features in the image training samples with different ink concentrations after preprocessing; Construct an initial ink concentration model based on the support vector machine; Iteratively train the initial ink concentration model according to the steel seal features in the image training samples with different ink concentrations after preprocessing until the model converges to obtain the trained ink concentration model.

3. The steel seal position recognition method according to claim 2, wherein The obtaining of the steel seal features in the image training samples with different ink concentrations after preprocessing includes the following steps: Adjust the ink concentration to obtain multiple image training samples with gradually changing depths and containing steel seals; Preprocess multiple image training samples with gradually changing depths and containing steel seals to obtain the image training samples with different ink concentrations after preprocessing; Extract the steel seal features in the image training samples with different ink concentrations after preprocessing according to the histogram of oriented gradients.

4. The steel seal position recognition method according to claim 1, characterized in that The preprocessing of multiple image training samples with gradually changing depths and containing steel seals to obtain the image training samples with different ink concentrations after preprocessing includes the following steps: Grayscale multiple image training samples with gradually changing depths and containing steel seals to obtain the grayscale histogram of the image training samples; Perform image segmentation, denoising, image enhancement, filtering and binarization on the grayscale histogram of the image training samples in sequence to obtain the image training samples with different ink concentrations after preprocessing.

5. The steel seal position recognition method according to claim 4, characterized in that, The performing of image segmentation, denoising, image enhancement, filtering and binarization on the grayscale histogram of the image training samples in sequence is specifically: Perform image segmentation on the grayscale histogram of the image training samples by the threshold segmentation method; Perform denoising through Gaussian filtering or median filtering; Equalize the grayscale histogram of the image training samples to enhance the contrast of the image; Highlight the edge information in the image through Laplacian filtering or directional filtering; Convert it into a binary image.

6. The steel seal position recognition method according to claim 1, characterized in that The identifying of the steel seal positions in the preprocessed steel seal images with different depths according to the background suppression model, edge detection algorithm and morphological operation includes the following steps: Determine the image area containing the steel seal position in the preprocessed steel seal image according to the background suppression model; Identify the edge contour of the steel seal in the image area through the Canny edge detection algorithm; Refine the edge contour through morphological operations to determine the steel seal positions in the preprocessed steel seal images with different depths.

7. A steel seal position recognition system, characterized in that Including: A steel seal feature acquisition module for obtaining the steel seal features of preprocessed steel seal images with different depths; An ink concentration module, which is used to substitute the steel seal features of the preprocessed steel seal images with different shades into the trained ink concentration model to calculate and obtain the ink concentration corresponding to the preprocessed steel seal images with different shades; A background suppression model construction module, which is used to construct a background suppression model according to the ink concentration corresponding to the preprocessed steel seal images with different shades; A steel seal position recognition module, which is used to recognize the steel seal position in the preprocessed steel seal images with different shades according to the background suppression model, edge detection algorithm and morphological operation.

8. The steel seal position recognition system according to claim 7, wherein, It further includes: An ink concentration model construction module; The ink concentration model construction module is used to obtain the steel seal features in the image training samples with different preprocessed ink concentrations; The ink concentration model construction module is further used to construct an initial ink concentration model based on a support vector machine; The ink concentration model construction module is further used to iteratively train the initial ink concentration model according to the steel seal features in the image training samples with different preprocessed ink concentrations until the model converges to obtain the trained ink concentration model.

Images