Image recognition result optimization method, system and equipment based on industrial Internet of Things, and medium

Abstract

The invention discloses an image recognition result optimization method, system and device based on the industrial Internet of Things and a medium, and relates to the technical field of image data processing, and the method comprises the steps: obtaining a surface image, and extracting an edge contour image in the surface image; generating a first image block according to the edge contour image, performing enhancement processing on the first image block and generating a second image block, calculating a first definition index and a second definition index of the first image block and the second image block, obtaining a fluctuation index, and obtaining an optimization degree index; if the fluctuation index exceeds a first threshold value, obtaining a to-be-processed first image block and performing optimization processing on the to-be-processed first image block; and if the fluctuation index does not exceed a first threshold value and the optimization degree index exceeds a second threshold value, obtaining a to-be-processed first image block and carrying out optimization processing on the to-be-processed first image block. The method has the advantages of being high in optimization precision, saving computing power resources and being high in optimization efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of industrial Internet of Things data processing, and in particular to an industrial Internet of Things-based image recognition result optimization method, system, device and medium. Background Art

[0002] With the rapid development of Industrial Internet of Things (IIoT), image recognition technology has become a key force in promoting the intelligent and efficient industrial production. It can not only realize real-time monitoring and precise control of the production process, but also greatly improve product quality and production efficiency.

[0003] However, in complex industrial environments, image acquisition faces many challenges, such as uneven lighting, shadow interference, equipment jitter, etc. These factors seriously affect the image quality, especially the edge contour part in the surface image, whose details and clarity are easily damaged. The edge contour part is the basis for identifying key information such as the shape, size, and position of the object. Once the edge contour image quality is damaged, it will seriously affect the accuracy of subsequent image analysis and recognition.

[0004] For example, in the field of precision manufacturing, slight edge deviations may cause the product to be unqualified; in the quality inspection process, blurred contours may cause defects to be misjudged or missed.

[0005] The existing image recognition processing methods adopt a "one-size-fits-all" strategy, that is, uniformly adjusting the brightness, contrast enhancement or denoising the entire image. Although this method can improve the overall appearance of the image to a certain extent, it ignores the particularities of different areas in the image, especially the edge contours. At the same time, this approach not only wastes a lot of computing resources, but also often fails to fully optimize the edge contours, making it difficult to meet the requirements of industrial detection for high precision and high reliability. Summary of the invention

[0006] In view of the defects in the prior art, the present invention provides an image recognition result optimization method, system, device and medium based on industrial Internet of Things.

[0007] An image recognition result optimization method based on industrial Internet of Things comprises: obtaining a surface image of an image generation node in the industrial Internet of Things, and extracting an edge contour image in the surface image; performing denoising and segmentation processing on the edge contour image and generating a plurality of continuous first image blocks, and performing enhancement processing on the plurality of first image blocks based on an enhancement model and generating a plurality of second image blocks, wherein the first image block and the second image block each include a plurality of pixels; calculating a first clarity index and a second clarity index of the first image block and the second image block based on an evaluation model, and obtaining a plurality of fluctuation indexes of the plurality of continuous first image blocks based on the plurality of first clarity indexes, and obtaining an optimization index based on the first clarity index and the second clarity index; judging whether the fluctuation index exceeds a first threshold value, and if the fluctuation index exceeds the first threshold value, obtaining a first image block to be processed based on the fluctuation index and optimizing the first image block to be processed; judging whether the optimization index exceeds a second threshold value if the fluctuation index does not exceed the first threshold value, and if the optimization index exceeds the second threshold value, obtaining the first image block to be processed based on the optimization index and optimizing the first image block to be processed.

[0008] Furthermore, denoising and segmenting the edge contour image and generating a plurality of continuous first image blocks include: denoising the edge contour image using a filter and generating a denoised image; performing continuity detection on the denoised image and obtaining a segmentation window based on the denoised image after the detection; segmenting the denoised image in sequence according to the segmentation window and generating a plurality of continuous first image blocks.

[0009] Further, based on the enhancement model, the enhancement processing is performed on the plurality of first image blocks respectively and the enhancement models in the plurality of second image blocks are generated as follows: ;in, is the b-th second image block, is the bth first image block, is the mixing coefficient, is the neighborhood window, is the coordinate in the neighborhood window The pixel point at Represented as the b-th second image block in Separated The pixel value of the pixel point, is the number of pixels in the neighborhood window.

[0010] Further, the evaluation model in calculating the first definition index of the first image block based on the evaluation model is expressed as: The evaluation model in which the second definition index of the second image block is calculated based on the evaluation model is expressed as: ;in, is the first definition index of the b-th first image block, is the second definition index of the b-th second image block, is the gradient value of the k-th pixel in the b-th first image block in the x direction, is the gradient value of the k-th pixel in the b-th first image block in the y direction, is the gradient value of the k-th pixel in the b-th second image block in the x direction, is the gradient value of the k-th pixel in the b-th second image block in the y direction, is the number of pixels in the first image block or the second image block.

[0011] Further, multiple fluctuation indices of multiple continuous first image blocks are obtained respectively according to the multiple first clarity indices and are expressed as: ;in, For the The fluctuation index of the first image block, For the The number of adjacent first image blocks of a first image block, For the A first definition indicator of a first image block.

[0012] Furthermore, obtaining the first image block to be processed according to the fluctuation index and optimizing the first image block to be processed includes: screening out the first image block whose fluctuation index exceeds the first threshold, and using the screened out first image block as the first image block to be processed; replacing the first image block to be processed with the corresponding second image block in the edge contour image.

[0013] Furthermore, obtaining the first image block to be processed according to the optimization index and optimizing the first image block to be processed includes: screening out the first image block whose optimization index exceeds the second threshold, and using the screened out first image block as the first image block to be processed; replacing the first image block to be processed with the corresponding second image block in the edge contour image.

[0014] A system for optimizing image recognition results based on the industrial Internet of Things is also provided. The system includes a service platform, a management platform, and a sensor network platform that are sequentially connected in communication. The management platform includes: an image acquisition module, which is used to acquire a surface image of an image generation node in the industrial Internet of Things and extract an edge contour image in the surface image; an image processing module, which is used to perform denoising and segmentation processing on the edge contour image and generate a plurality of continuous first image blocks, and to perform enhancement processing on the plurality of first image blocks based on an enhancement model and generate a plurality of second image blocks, wherein the first image blocks and the second image blocks each include a plurality of pixel points; a calculation management module, which is used to calculate the first clear image blocks and the second image blocks respectively based on an evaluation model. a first judgment and optimization module, for judging whether the fluctuation index exceeds a first threshold value, if the fluctuation index exceeds the first threshold value, obtaining the first image block to be processed according to the fluctuation index and performing optimization processing on the first image block to be processed; a second judgment and optimization module, for judging whether the optimization index exceeds a second threshold value when the fluctuation index does not exceed the first threshold value, if the optimization index exceeds the second threshold value, obtaining the first image block to be processed according to the optimization index and performing optimization processing on the first image block to be processed.

[0015] An electronic device is also provided, comprising: a memory on which a computer program is stored; and a processor for executing the computer program in the memory to implement the above-mentioned image recognition result optimization method based on the industrial Internet of Things.

[0016] A non-temporary computer-readable storage medium is also provided, on which a computer program is stored, and when the program is executed by a processor, the above-mentioned image recognition result optimization method based on the industrial Internet of Things is implemented.

[0017] The beneficial effects of the present invention are embodied in: In the entire image recognition result optimization method based on industrial Internet of Things, by accurately extracting the surface image and edge contour image of the image generation node in the industrial Internet of Things, a solid foundation is provided for subsequent image enhancement, quality evaluation and optimization, ensuring the accuracy and effectiveness of the processing object; further, the denoising segmentation technology is used to effectively remove the noise interference in the image, and the image is divided into multiple small blocks that are easy to manage for subsequent processing; further, the image block is initially improved by using the enhancement processing technology, the main purpose of which is to make subsequent optimization judgments more accurately, and at the same time, the clarity and contrast of the image block are improved to a certain extent, providing a better starting point for possible optimization processing; more importantly, by calculating the clarity index through the evaluation model, and obtaining the fluctuation index and optimization index, the scheme can accurately evaluate the quality of the image block and its changes, providing a scientific judgment basis for the optimization processing; finally, according to the comparison results of the indicators, the scheme can select the image blocks that need to be optimized in a targeted manner, and use the results of the previous enhancement processing to assist in further optimization, which not only avoids unnecessary waste of resources, but also ensures the accuracy and efficiency of the optimization processing, and avoids the problems of resource waste and insufficient precision caused by the "one-size-fits-all" strategy. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following is a brief introduction to the drawings required for the specific embodiments or the description of the prior art. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn according to the actual scale.

[0019] Figure 1 A schematic diagram of the steps of the method for optimizing image recognition results based on the industrial Internet of Things of the present invention; Figure 2 It is a schematic diagram of step S2 in the image recognition result optimization method based on industrial Internet of Things of the present invention; Figure 3 It is a schematic diagram of step S4 in the image recognition result optimization method based on industrial Internet of Things of the present invention; Figure 4 It is a schematic diagram of step S5 in the image recognition result optimization method based on industrial Internet of Things of the present invention; Figure 5 It is a schematic diagram of the composition of the image recognition result optimization system based on the industrial Internet of Things of the present invention; Figure 6 It is a schematic diagram of the composition of the optimized industrial Internet of Things involved in the present invention; Figure 7 The present invention is a block diagram of an electronic device according to an embodiment of the present invention.

[0020] Reference numerals: 700 - electronic device, 701 - processor, 702 - memory, 703 - multimedia component, 704 - I / O interface, 705 - communication component. DETAILED DESCRIPTION

[0021] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0022] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0023] It should be noted that similar reference numerals and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further defined and explained in the subsequent drawings. In addition, the terms "first", "second", etc. are only used to distinguish the description and cannot be understood as indicating or implying relative importance.

[0024] like Figure 1 As shown, a method for optimizing image recognition results based on industrial Internet of Things is provided, including: S1, obtaining a surface image of an image generation node in the industrial Internet of Things, and extracting an edge contour image from the surface image; S2, performing denoising and segmentation processing on the edge contour image to generate a plurality of continuous first image blocks, and performing enhancement processing on the plurality of first image blocks based on the enhancement model to generate a plurality of second image blocks, wherein the first image blocks and the second image blocks each include a plurality of pixel points; S3, respectively calculating a first clarity index and a second clarity index of the first image block and the second image block based on the evaluation model, and respectively obtaining a plurality of fluctuation indexes of a plurality of continuous first image blocks according to the plurality of first clarity indexes, and obtaining an optimization index according to the first clarity index and the second clarity index; S4, determining whether the fluctuation index exceeds a first threshold value, and if the fluctuation index exceeds the first threshold value, obtaining a first image block to be processed according to the fluctuation index and performing optimization processing on the first image block to be processed; S5. If the fluctuation index does not exceed the first threshold, determine whether the optimization index exceeds the second threshold. If the optimization index exceeds the second threshold, obtain the first image block to be processed according to the optimization index and optimize the first image block to be processed.

[0025] In this embodiment, it should be noted that in S1, the image generation nodes of the industrial Internet of Things environment may include various sensors, cameras or other image acquisition devices, which are deployed in production lines, equipment or working environments to capture real-time surface images, which may contain important information about products, equipment status or production processes. The extraction of edge contour images is the core of S1. The goal is to accurately identify the edge contour area of ​​the object from the surface image; first, the edge area in the image is identified by edge detection, that is, the edge detection algorithm (such as Canny, Sobel, Prewitt, etc., classic algorithms for edge detection) is applied; then, based on edge detection, the adjacent edge areas are connected by region tracking algorithms (such as Freeman chain code, Hough transform, etc.) to form a complete edge contour image of sufficient size. In summary, S1 is the basis for subsequent image enhancement, quality assessment and optimization.

[0026] In S2, first, the edge contour image is processed for noise. Since image acquisition in industrial environments is often disturbed by various noises, such as equipment vibration, electromagnetic interference, etc., these noises will affect the image quality and thus affect the accurate segmentation of the edge contour image. Therefore, the edge contour image needs to be denoised. There are many denoising methods, such as median filtering, Gaussian filtering, bilateral filtering, etc. These methods can effectively remove the noise in the edge contour image while retaining the details of the edge contour image. Then, the denoised edge contour image is segmented. This is because the image needs to be processed more efficiently. Usually, the image is segmented into multiple smaller areas or image blocks for processing. In S2, we segment the edge contour image into multiple continuous first image blocks. The size of these image blocks can be adjusted according to actual needs to ensure a balance between processing efficiency and image quality. Segmentation helps to simplify complex problems into smaller problems that are easier to manage, thereby improving the efficiency and accuracy of the entire processing flow.

[0027] Furthermore, in the S2 stage, each first image block after segmentation is enhanced using an enhancement model; the goal of the enhancement model is to improve key indicators such as clarity and contrast of the image block, making the edge contour more prominent and accurate. The enhancement process inputs each first image block into the enhancement model, and the model outputs an enhanced image block, namely the second image block. The quality of the multiple second image blocks generated in the end will be improved, that is, a higher quality edge contour image is obtained.

[0028] In S3, an evaluation model is first required to calculate the clarity index of the image block. The goal of the evaluation model is to accurately measure the clarity of the image block and output a first clarity index and a second clarity index for subsequent quality evaluation and optimization. The first clarity index reflects the clarity level of the image block in its original state. By comparing the first clarity indexes of different first image blocks, we can understand the clarity distribution of different areas in the image. The second clarity index reflects the clarity level of the image block after enhancement processing. By comparing the first clarity index and the second clarity index, we can evaluate the impact of enhancement processing on the clarity of the image block.

[0029] Furthermore, in order to measure the degree of change in clarity between image blocks, a fluctuation index is calculated based on the first clarity indicators of multiple consecutive first image blocks. The fluctuation index can reflect the unevenness of clarity in the image, that is, the clarity of some areas is significantly higher or lower than that of other areas. By calculating the fluctuation index, areas in the image with large changes in clarity can be identified. These areas are the targets of subsequent optimization processing, providing strong support for subsequent optimization processing.

[0030] Furthermore, in order to evaluate the effect of enhancement processing on improving the overall clarity of the image, an optimization index is calculated based on the first clarity index and the second clarity index. The optimization index can reflect the degree of improvement of the image clarity caused by the enhancement processing. By comparing the optimization index, it can be determined whether the enhancement processing is necessary, thereby providing strong support for subsequent optimization processing.

[0031] In S4, a first threshold is first set. This threshold can be adjusted according to the actual application scenario and image quality requirements. The first threshold is a reference value for judging whether the fluctuation index is too large. When the fluctuation index exceeds the first threshold, it means that there are areas in the image with large changes in clarity. These areas may be caused by uneven lighting, device jitter, changes in shooting angles, etc.

[0032] For each fluctuation index of the first image block, we compare it with the first threshold. If the fluctuation index exceeds the first threshold, the first image block is marked as the first image block to be processed; if the fluctuation index does not exceed the first threshold, it is considered that the clarity change of the image block is within an acceptable range and no further optimization processing is required.

[0033] Furthermore, based on the judgment results of the fluctuation index, all the first image blocks to be processed can be obtained. These image blocks are areas with large changes in clarity and are also the focus of subsequent optimization processing. For the first image blocks to be processed, a variety of optimization processing strategies can be adopted, such as local enhancement, filtering processing, contrast adjustment, etc.; specific optimization processing strategies can be selected according to the actual application scenarios and image quality requirements. The goal of the optimization processing is to improve the clarity of the image block so that it is more consistent with the clarity of the surrounding areas.

[0034] In S5, if the fluctuation index of the first image block does not exceed the set first threshold, it means that the change in clarity of the image block is relatively uniform and there is no significant fluctuation. Then a second threshold is set, which is set based on the optimization index and is used to determine whether the optimization process has achieved the expected improvement effect and whether the optimization is necessary. The selection of the second threshold should take into account factors such as the actual application scenario, image quality requirements, and the cost of the optimization process.

[0035] Then the optimization index is compared with the second threshold. If the optimization index exceeds the second threshold, it means that the enhancement processing has a significant improvement effect on the clarity of the image block, so it is necessary to optimize, and then these first image blocks are marked as first image blocks to be processed; if the optimization index does not exceed the second threshold, it is considered that the original first image block is clearer, or the optimization cost-effectiveness is low, the current enhancement processing is no longer necessary, and no further optimization is required.

[0036] For the first image block marked as to be processed, we can use an optimization processing strategy similar to the S4 stage for further processing, selecting a variety of optimization processing strategies such as local enhancement, contrast adjustment, and filtering processing. The goal is to further improve the clarity of the image block and make it closer to the ideal image quality.

[0037] In summary, in the entire image recognition result optimization method based on industrial Internet of Things, by accurately extracting the surface image and edge contour image of the image generation node in the industrial Internet of Things, a solid foundation is provided for subsequent image enhancement, quality evaluation and optimization, ensuring the accuracy and effectiveness of the processing object; further, the denoising segmentation technology is used to effectively remove the noise interference in the image, and the image is divided into multiple small blocks that are easy to manage for subsequent processing; further, the image block is initially improved by using the enhancement processing technology, the main purpose of which is to make subsequent optimization judgments more accurately, and at the same time, the clarity and contrast of the image block are improved to a certain extent, providing a better starting point for possible optimization processing; more importantly, by calculating the clarity index through the evaluation model, and obtaining the fluctuation index and optimization index, the scheme can accurately evaluate the quality of the image block and its changes, providing a scientific judgment basis for the optimization processing; finally, according to the comparison results of the indicators, the scheme can select the image blocks that need to be optimized in a targeted manner, and use the results of the previous enhancement processing to assist in further optimization, which not only avoids unnecessary waste of resources, but also ensures the accuracy and efficiency of the optimization processing, and avoids the problems of resource waste and insufficient precision caused by the "one-size-fits-all" strategy.

[0038] like Figure 2 As shown, in one embodiment, performing denoising and segmentation processing on the edge contour image and generating a plurality of continuous first image blocks in S2 includes: S21, performing denoising processing on the edge contour image using a filter and generating a denoised image; S22, performing continuity detection on the denoised image, and obtaining a segmentation window according to the denoised image after detection; S23, sequentially segmenting the denoised image according to the segmentation windows and generating a plurality of continuous first image blocks.

[0039] In this embodiment, it should be noted that in S21, in the process of optimizing the image recognition results, the quality of the edge contour image is crucial. Therefore, it is necessary to perform denoising on the edge contour image before the segmentation process. In the denoising process, the selection of the filter is the key. Commonly used filters include mean filtering, Gaussian filtering, median filtering, bilateral filtering, etc. These filters have their own advantages and disadvantages and are suitable for different noise types and image characteristics. The specific steps of denoising are: according to the characteristics of the edge contour image and the type of noise, select the appropriate filter type and parameters. For example, if Gaussian filtering is selected, the size and standard deviation of the Gaussian kernel need to be determined; apply the selected filter to the edge contour image to generate a denoised image. This step can usually be implemented by functions in an image processing library (such as OpenCV, MATLAB image processing toolbox, etc.); evaluate the denoised image to check whether the noise is effectively removed while retaining the edge and detail information of the image. If the denoising effect is not ideal, it may be necessary to adjust the filter parameters or try other types of filters.

[0040] In S22, after denoising, the quality of the edge contour image is improved, but there may still be some discontinuous areas or edges in the image. These discontinuous parts may affect the subsequent image segmentation and optimization processing effects. Therefore, before image segmentation, it is necessary to perform continuity detection on the denoised image. Continuity detection is usually based on the similarity and connectivity between image pixels for analysis. Commonly used methods include gradient-based methods, region growing-based methods, morphology-based methods, etc. For example, by processing the image using morphological operations (such as dilation, erosion, opening operations, closing operations, etc.), the shape and structure in the image can be identified. By setting certain morphological parameters, the image can be divided into different regions and the continuous parts therein can be identified.

[0041] Furthermore, after the continuity detection is completed, a segmentation window can be obtained according to the detection result. The segmentation window refers to a rectangular frame that divides the image into multiple continuous regions or edges. The size and position of these rectangular frames can be determined according to the continuity detection result. Specifically, the size and number of the segmentation windows are determined according to the continuity detection result and the requirements of subsequent processing. Generally, the size of the segmentation window should be sufficient to include a complete region or edge in the image, and the number is weighed according to the complexity of the image and the efficiency of subsequent processing.

[0042] In S23, the segmentation is based on the previously acquired segmentation windows. The sizes and positions of these windows are determined according to the results of the continuity detection. Therefore, during segmentation, we only need to cut the image according to the boundaries of these windows to obtain multiple continuous first image blocks.

[0043] The first image block generated will serve as input for subsequent image enhancement, quality assessment and optimization processing. Since these image blocks have undergone denoising and continuity detection and are accurately cut according to the segmentation window, they have high continuity, which provides a good foundation for the subsequent indicator generation, especially the generation of fluctuation indicators, and helps to improve the accuracy and reliability of optimization judgments.

[0044] In one embodiment, the enhancement model is expressed as: ;in, is the b-th second image block, is the bth first image block, is the mixing coefficient, is the neighborhood window, is the coordinate in the neighborhood window The pixel point at It is represented as the bth second image block adjacent to (x, y) The pixel value of the pixel point, is the number of pixels in the neighborhood window.

[0045] In this embodiment, it should be noted that represents the b-th second image block (i.e., the enhanced image block), and is therefore also the pixel value of the second image block at the coordinate (x, y). represents the bth first image block (i.e., the original image block), and is therefore also the pixel value of the first image block at the coordinate (x, y). is a mixing coefficient used to adjust the weight of the original image block and the neighborhood pixel information in the enhanced result. Its value range is usually between [0, 1]. L is a neighborhood window, which defines the range of surrounding pixels to be considered. It can be a 2x2, 3x3 or 2x3 window. is the number of pixels in the neighborhood window, for example, for a 2x3 window, N=6.

[0046] This part is the smoothing enhancement process, which aims to improve the clarity of unclear areas in the image. It means that the value of the enhanced image at each pixel position (x, y) is the average value of all pixel values ​​within L of the original image at that pixel position (x, y). This makes the enhancement process more localized and can better adapt to the characteristics of different areas in the image. It is particularly suitable for processing images with complex textures or edges. is the original pixel value, so the entire enhancement model can both enhance the unclear areas in the image (through the weighted average of the neighborhood pixels) and enhance the details of the image (by retaining some of the original pixel information). This dual effect significantly improves the image enhancement processing effect.

[0047] In one embodiment, the evaluation model in calculating the first definition index of the first image block based on the evaluation model is expressed as: ; The evaluation model in calculating the second definition index of the second image block based on the evaluation model is expressed as: ;in, is the first definition index of the b-th first image block, is the second definition index of the b-th second image block, is the gradient value of the k-th pixel in the b-th first image block in the x direction, is the gradient value of the k-th pixel in the b-th first image block in the y direction, is the gradient value of the k-th pixel in the b-th second image block in the x direction, is the gradient value of the k-th pixel in the b-th second image block in the y direction, is the number of pixels in the first image block or the second image block.

[0048] In this embodiment, it should be noted that the image clarity is usually related to the details and edge information in the image. The gradient is an important indicator to measure the rate of change of pixel values ​​in the image, which can reflect the edge and texture information of the image. Therefore, by calculating the gradient value of the image block in the x and y directions, it can be used to evaluate the clarity of the image block. and represent the gradient values ​​of the k-th pixel in the first image block in the x and y directions, respectively, and and represents the gradient value of the corresponding pixel in the second image block (i.e. the enhanced image block), which reflects the degree of change of the image block in the horizontal and vertical directions. In order to obtain a comprehensive clarity index, the evaluation model calculates the gradient amplitude of each pixel (i.e. and ) are averaged to obtain the clarity index of the entire image block (i.e. and ), the larger the clarity index is, the more dramatic the change in pixel values ​​in the image block is, that is, the richer the edge and texture information of the image block is, and therefore the higher the clarity is. In summary, by calculating the clarity index of the image block, the evaluation model can objectively and quantitatively evaluate the clarity of the image block. This is of great significance for image quality evaluation and image enhancement effect evaluation in image processing and computer vision tasks.

[0049] It should also be noted that the gradient value here is usually obtained by applying the gradient operator to the image. The gradient operator is a mathematical tool used to calculate the rate of change of a function at a certain point. In a two-dimensional image, the gradient is a vector containing components in the x-direction and the y-direction, which can be expressed as (Right now and )and (Right now and ). Furthermore, commonly used gradient operators include Sobel operator, Prewitt operator and Roberts operator. These operators perform convolution operations with the image through the convolution kernel to calculate the gradient value of each pixel in the x direction and the y direction.

[0050] In one embodiment, a plurality of fluctuation indices of a plurality of consecutive first image blocks are obtained respectively according to a plurality of first clarity indices and are expressed as: ;in, For the The fluctuation index of the first image block, For the The number of adjacent first image blocks of a first image block, For the A first definition indicator of a first image block.

[0051] In this embodiment, it should be noted that It represents the fluctuation index of the bth first image block. The whole expression measures the degree of change in clarity between the image block and its adjacent image blocks. (the first definition index of the b-th image block) and The difference between (the first clarity index of the g-th image block) and taking the absolute value can quantify the difference in clarity between adjacent image blocks. It represents the number of adjacent image blocks of the b-th first image block, which determines how many adjacent image blocks are considered when calculating the fluctuation index. Introduce the formula and average the differences (i.e. divide by ), which can smooth the impact of the number of adjacent image blocks on the fluctuation index, making the evaluation result more stable and reliable. Fluctuation index It reflects the fluctuation of clarity between adjacent image blocks in the image. If the clarity difference between adjacent image blocks is large, the fluctuation index will be higher; conversely, if the clarity between adjacent image blocks is relatively consistent, the fluctuation index will be lower. Therefore, the fluctuation index can be used as an important basis for evaluating the overall stability and consistency of the image.

[0052] like Figure 3 As shown, in one embodiment, obtaining the first image block to be processed according to the fluctuation index and optimizing the first image block to be processed in S4 includes: S41, screening out first image blocks whose fluctuation index exceeds a first threshold, and using the screened out first image blocks as first image blocks to be processed; S42: Replace the first image block to be processed with the corresponding second image block in the edge contour image.

[0053] In this embodiment, it should be noted that, in S41, the fluctuation index of each first image block calculated previously is screened one by one, and the first image blocks whose fluctuation index exceeds the first threshold are screened out; when the fluctuation index of a first image block exceeds this first threshold, it means that there is a large difference between the image block and the surrounding image blocks, which may be caused by noise, blur, uneven lighting or other factors. Therefore, this image block is marked as the first image block to be processed, which requires further optimization processing.

[0054] In S42, a replacement operation is performed on the first image block to be processed screened out in step S41, and it is directly replaced with the second image block, that is, the processing result in the optimization judgment process is utilized, which provides a better starting point for the optimization processing and greatly improves the efficiency of image result optimization.

[0055] like Figure 4 As shown, in one embodiment, obtaining the first image block to be processed according to the optimization index and optimizing the first image block to be processed in S5 includes: S51, screening out first image blocks whose optimization index exceeds a second threshold, and using the screened out first image blocks as first image blocks to be processed; S52: Replace the first image block to be processed with the corresponding second image block in the edge contour image.

[0056] In the present embodiment, it should be noted that, similarly, in S51, the optimization index of each first image block is evaluated, and the optimization index and the second threshold are used to determine whether the optimization processing has achieved the expected improvement effect and whether there is a need for optimization. When the optimization index of a first image block does not exceed the second threshold, it means that although the image block has undergone preliminary processing, its optimization effect is not obvious, and there is no need to waste resources for optimization processing; therefore, the system will mark the first image blocks whose optimization index exceeds the second threshold as first image blocks to be processed.

[0057] S52 is similar to S42. In S52, a replacement operation is performed on the first image block to be processed screened out in step S51, and it is directly replaced with the second image block. That is, the processing results in the optimization judgment process are utilized, which provides a better starting point for the optimization processing and greatly improves the efficiency of image result optimization.

[0058] like Figure 5 As shown, an image recognition result optimization system based on industrial Internet of Things is also provided. The system includes a service platform, a management platform and a sensor network platform which are sequentially connected in communication. The management platform includes: An image acquisition module is used to acquire a surface image of an image generation node in the industrial Internet of Things and extract an edge contour image from the surface image; An image processing module, used to perform denoising and segmentation processing on the edge contour image and generate a plurality of continuous first image blocks, and to perform enhancement processing on the plurality of first image blocks based on the enhancement model and generate a plurality of second image blocks, wherein the first image blocks and the second image blocks each include a plurality of pixel points; A calculation management module, configured to calculate a first clarity index and a second clarity index of the first image block and the second image block respectively based on the evaluation model, and obtain a plurality of fluctuation indexes of a plurality of continuous first image blocks respectively according to the plurality of first clarity indexes, and obtain an optimization index according to the first clarity index and the second clarity index; A first judgment and optimization module, used for judging whether the fluctuation index exceeds a first threshold value, and if the fluctuation index exceeds the first threshold value, obtaining a first image block to be processed according to the fluctuation index and performing optimization processing on the first image block to be processed; The second judgment and optimization module is used to judge whether the optimization index exceeds the second threshold when the fluctuation index does not exceed the first threshold. If the optimization index exceeds the second threshold, the first image block to be processed is obtained according to the optimization index and the first image block to be processed is optimized.

[0059] In this embodiment, it should be noted that, regarding the above-mentioned image recognition result optimization system based on industrial Internet of Things, the specific method of performing the operation has been described in detail in the implementation of the image recognition result optimization method based on industrial Internet of Things, and will not be elaborated here.

[0060] It should also be noted that the entire industrial Internet of Things-based image recognition result optimization system can be applied to the optimized industrial Internet of Things, such as Figure 6 As shown, the optimized industrial Internet of Things includes a user platform, a service platform, a management platform, a sensor network platform and an object platform that establish communication in sequence; The user platform is configured to provide front-end service functions to users; users obtain the required perceived service information through the user platform, process the perceived service information, and convert it into user perceived information; users analyze the user perceived information and make corresponding decisions based on their own wishes, and convert the user perceived information into user control information through the corresponding information system and send it to the service platform, thereby expressing the user's corresponding service demand intention.

[0061] The physical entities of the user platform include various user terminals, such as mobile phones, computers, dedicated terminals, etc., which realize user-side services through integration with user information system software.

[0062] The service platform is configured as an API server or other server used to establish communication between the management platform and the user platform to implement corresponding functions; the physical entity of the service platform includes various servers.

[0063] The management platform is configured to perform at least one of equipment operation status monitoring management, data monitoring management, equipment parameter management, and life cycle management; the management platform is an operational coordination platform for the Internet of Things, which may include various management sub-platforms, and different management sub-platforms perform different management services; the physical entities of the management platform include various servers.

[0064] The sensor network platform is configured to perform at least one of network management, command management, device status management, data protocol management, data analysis, data classification, data transmission monitoring, and data transmission security management. The sensor network platform provides data communication, analysis, identification, classification, and other functions, avoiding the direct aggregation of data from various object platforms on the management platform, resulting in redundant data on the management platform and low data processing efficiency; the physical entities of the object platform include various gateways, edge computing devices, etc.

[0065] The object platform is configured to carry out specific production work such as production control, detection, and measurement; the physical entities of the production objects include various production equipment, sensors, etc.

[0066] Figure 7is a block diagram of an electronic device showing an image recognition result optimization method based on the industrial Internet of Things according to an exemplary embodiment. Figure 7 As shown, the electronic device 700 may include: a processor 701 and a memory 702. The electronic device 700 may also include one or more of a multimedia component 703, an I / O interface 704, and a communication component 705.

[0067] The processor 701 is used to control the overall operation of the electronic device 700 to complete all or part of the steps in the above-mentioned method for optimizing image recognition results based on the industrial Internet of Things. The memory 702 is used to store various types of data to support the operation of the electronic device 700, and these data may include, for example, instructions for any application or method used to operate on the electronic device 700, and application-related data, such as contact data, sent and received messages, pictures, audio, video, etc. The memory 702 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (Static Random Access Memory, referred to as SRAM), electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, referred to as EEPROM), erasable programmable read-only memory (Erasable Programmable Read-Only Memory, referred to as EPROM), programmable read-only memory (Programmable Read-Only Memory, referred to as PROM), read-only memory (Read-Only Memory, referred to as ROM), magnetic memory, flash memory, magnetic disk or optical disk. The multimedia component 703 may include a screen and an audio component. The screen may be, for example, a touch screen, and the audio component is used to output and / or input audio signals. For example, the audio component may include a microphone, which is used to receive external audio signals. The received audio signal may be further stored in the memory 702 or sent through the communication component 705. The audio component also includes at least one speaker for outputting audio signals. The I / O interface 704 provides an interface between the processor 701 and other interface modules, and the above-mentioned other interface modules may be keyboards, mice, buttons, etc. These buttons may be virtual buttons or physical buttons. The communication component 705 is used for wired or wireless communication between the electronic device 700 and other devices. Wireless communication, such as Wi-Fi, Bluetooth, Near Field Communication (NFC), 2G, 3G, 4G, NB-IOT, eMTC or other 5G, etc., or a combination of one or more of them, is not limited here. Therefore, the corresponding communication component 705 may include: Wi-Fi module, Bluetooth module, NFC module, etc.

[0068] In an exemplary embodiment, the electronic device 700 can be implemented by one or more application specific integrated circuits (ASIC), digital signal processors (DSP), digital signal processing devices (DSPD), programmable logic devices (PLD), field programmable gate arrays (FPGA), controllers, microcontrollers, microprocessors or other electronic components to execute the above-mentioned image recognition result optimization method based on industrial Internet of Things.

[0069] In another exemplary embodiment, a computer-readable storage medium including program instructions is also provided, and when the program instructions are executed by a processor, the steps of the above-mentioned method for optimizing image recognition results based on the industrial Internet of Things are implemented. For example, the computer-readable storage medium may be the above-mentioned memory 702 including program instructions, and the above-mentioned program instructions may be executed by the processor 701 of the electronic device 700 to complete the above-mentioned method for optimizing image recognition results based on the industrial Internet of Things.

[0070] In another exemplary embodiment, a computer program product is also provided, which includes a computer program that can be executed by a programmable device, and the computer program has a code portion for executing the above-mentioned image recognition result optimization method based on the industrial Internet of Things when executed by the programmable device.

[0071] The preferred embodiments of the present disclosure are described in detail above in conjunction with the accompanying drawings; however, the present disclosure is not limited to the specific details in the above embodiments. Within the technical concept of the present disclosure, a variety of simple modifications can be made to the technical solution of the present disclosure, and these simple modifications all fall within the protection scope of the present disclosure.

[0072] It should also be noted that the various specific technical features described in the above specific embodiments can be combined in any suitable manner without contradiction. In order to avoid unnecessary repetition, the present disclosure will not further describe various possible combinations.

[0073] In addition, various embodiments of the present disclosure may be arbitrarily combined, and as long as they do not violate the concept of the present disclosure, they should also be regarded as the contents disclosed by the present disclosure.

[0074] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. These modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be included in the scope of the claims and specification of the present invention.

Claims

1. A method for optimizing image recognition results based on industrial Internet of Things, characterized in that: include: Obtain the surface image of the image generation node in the industrial Internet of Things, and extract the edge contour image in the surface image; Performing denoising and segmentation processing on the edge contour image to generate a plurality of continuous first image blocks, and performing enhancement processing on the plurality of first image blocks based on the enhancement model to generate a plurality of second image blocks, wherein the first image blocks and the second image blocks each include a plurality of pixel points; Based on the evaluation model, respectively calculate a first clarity index and a second clarity index of the first image block and the second image block, respectively obtain a plurality of fluctuation indexes of a plurality of consecutive first image blocks according to the plurality of first clarity indexes, and obtain an optimization index according to the first clarity index and the second clarity index; Determine whether the fluctuation index exceeds a first threshold value, and if the fluctuation index exceeds the first threshold value, obtain a first image block to be processed according to the fluctuation index and perform optimization processing on the first image block to be processed; If the fluctuation index does not exceed the first threshold, it is determined whether the optimization index exceeds the second threshold. If the optimization index exceeds the second threshold, the first image block to be processed is obtained according to the optimization index and the first image block to be processed is optimized.

2. The method for optimizing image recognition results based on industrial Internet of Things according to claim 1 is characterized in that: The performing denoising and segmentation processing on the edge contour image and generating a plurality of continuous first image blocks comprises: Using a filter to perform denoising on the edge contour image and generate a denoised image; Perform continuity detection on the denoised image, and obtain a segmentation window based on the denoised image after detection; The denoised image is segmented in sequence according to the segmentation window to generate a plurality of continuous first image blocks.

3. The method for optimizing image recognition results based on industrial Internet of Things according to claim 1, characterized in that: The enhancement model in which the plurality of first image blocks are enhanced respectively based on the enhancement model and the plurality of second image blocks are generated is expressed as: ;in, is the b-th second image block, is the bth first image block, is the mixing coefficient, is the neighborhood window, is the coordinate in the neighborhood window The pixel point at Represented as the b-th second image block in Separated The pixel value of the pixel point, is the number of pixels in the neighborhood window.

4. The method for optimizing image recognition results based on industrial Internet of Things according to claim 3 is characterized in that: The evaluation model in calculating the first definition index of the first image block based on the evaluation model is expressed as: ； The evaluation model in calculating the second definition index of the second image block based on the evaluation model is expressed as: ;in, is the first definition index of the b-th first image block, is the second definition index of the b-th second image block, is the gradient value of the k-th pixel in the b-th first image block in the x direction, is the gradient value of the k-th pixel in the b-th first image block in the y direction, is the gradient value of the k-th pixel in the b-th second image block in the x direction, is the gradient value of the k-th pixel in the b-th second image block in the y direction, is the number of pixels in the first image block or the second image block.

5. The method for optimizing image recognition results based on industrial Internet of Things according to claim 4 is characterized in that: The method of obtaining a plurality of fluctuation indices of a plurality of continuous first image blocks according to a plurality of first definition indices is expressed as: ;in, For the The fluctuation index of the first image block, For the The number of adjacent first image blocks of a first image block, For the A first clarity indicator of a first image block.

6. The method for optimizing image recognition results based on industrial Internet of Things according to claim 1, characterized in that: The obtaining of the first image block to be processed according to the fluctuation index and performing optimization processing on the first image block to be processed comprises: Screening out a first image block whose fluctuation index exceeds a first threshold, and using the screened out first image block as a first image block to be processed; The first image block to be processed is replaced with the corresponding second image block in the edge contour image.

7. The method for optimizing image recognition results based on industrial Internet of Things according to claim 1, characterized in that: The obtaining of the first image block to be processed according to the optimization index and performing optimization processing on the first image block to be processed comprises: Screening out a first image block whose optimization index exceeds a second threshold, and using the screened out first image block as a first image block to be processed; The first image block to be processed is replaced with the corresponding second image block in the edge contour image.

8. An image recognition result optimization system based on industrial Internet of Things, characterized in that: The system includes a service platform, a management platform and a sensor network platform which are connected in sequence, and the management platform includes: An image acquisition module is used to acquire a surface image of an image generation node in the industrial Internet of Things and extract an edge contour image from the surface image; An image processing module, used to perform denoising and segmentation processing on the edge contour image and generate a plurality of continuous first image blocks, and to perform enhancement processing on the plurality of first image blocks based on the enhancement model and generate a plurality of second image blocks, wherein the first image blocks and the second image blocks each include a plurality of pixel points; A calculation management module, configured to calculate a first clarity index and a second clarity index of the first image block and the second image block respectively based on the evaluation model, and obtain a plurality of fluctuation indexes of a plurality of continuous first image blocks respectively according to the plurality of first clarity indexes, and obtain an optimization index according to the first clarity index and the second clarity index; A first judgment and optimization module, used for judging whether the fluctuation index exceeds a first threshold value, and if the fluctuation index exceeds the first threshold value, obtaining a first image block to be processed according to the fluctuation index and performing optimization processing on the first image block to be processed; The second judgment and optimization module is used to judge whether the optimization index exceeds the second threshold when the fluctuation index does not exceed the first threshold. If the optimization index exceeds the second threshold, the first image block to be processed is obtained according to the optimization index and the first image block to be processed is optimized.

9. An electronic device, characterized in that: include: a memory having a computer program stored thereon; A processor, configured to execute the computer program in the memory to implement the method for optimizing image recognition results based on the industrial Internet of Things as described in any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method for optimizing image recognition results based on the industrial Internet of Things described in any one of claims 1 to 7 is implemented.

Images