Efficient communication method and system for industrial robots based on edge computing

Abstract

This invention belongs to the field of image data processing technology, specifically relating to an efficient communication method and system for industrial robots based on edge computing. The method includes: determining all narrow-edge pixels in each frame of an image; determining the edge intensity factor of each frame; determining the structured edge factor of each frame; determining the optimization coefficient of the effective frame filtering threshold for each frame, thus obtaining the optimized effective frame filtering threshold for each frame; and judging each frame to achieve effective frame filtering. This invention optimizes the effective frame filtering threshold for each frame by analyzing the gradient magnitude and gradient direction characteristics of the narrow-edge pixels in each frame, overcoming the limitations of a fixed effective frame filtering threshold, and filtering effective frames to improve the timeliness and robustness of industrial robot communication.

Description

Technical Field

[0001] This invention relates to the field of image data processing technology. More specifically, this invention relates to an efficient communication method and system for industrial robots based on edge computing. Background Technology

[0002] With the advancement of intelligent manufacturing, industrial robots have been widely used in automated production lines. In scenarios such as visual servoing, precision assembly, and defect detection, industrial robots are typically equipped with industrial cameras to acquire real-time images of the site and transmit the image data to edge computing nodes or cloud servers for in-depth processing and analysis. Edge computing architecture, by providing computing, storage, and network resources on the side closer to the data source, is a key technical architecture for achieving low-latency and high-reliability communication in industrial settings. Since valid frames of an image typically have clear and sharp workpiece contours, their edges exhibit narrow step changes at the pixel level, providing accurate geometric features and positioning references for backend vision algorithms. Invalid frames, on the other hand, are usually composed of motion blur caused by high-speed robot movement or defocusing due to focusing failure. Their edges exhibit wide, smooth transitions, making them unsuitable for high-precision feature extraction. Transmitting data containing a large number of invalid frames to the cloud not only consumes uplink bandwidth resources and increases the server's unproductive computing load, but more seriously, the blurred edge features interfere with feature matching and target recognition in backend vision algorithms, causing robot control commands to drift. Therefore, in an edge computing architecture, uploading all image data in its entirety can easily lead to network congestion and high latency, and also contain a large amount of redundant information.

[0003] Therefore, it is necessary to perform validity screening on images at the edge. This invention uses an algorithm based on edge width histogram to determine image quality, retains valid frames, and removes invalid frames, thereby reducing the waste of communication and computing resources and improving communication efficiency.

[0004] However, existing edge width histogram-based algorithms for valid frame selection have limitations. These algorithms typically use fixed thresholds to determine whether an image is a valid frame. This static standard is difficult to adapt to the dynamic working conditions of industrial robots and cannot distinguish between the physical contours of real workpieces and the effects of thermal noise and high-frequency textures in cluttered backgrounds in actual working conditions. If images containing thermal noise and high-frequency textures are indiscriminately classified as valid frames for transmission, it can easily lead to data loss or redundant transmission at critical moments. It can also interfere with feature extraction in backend vision algorithms and reduce communication efficiency. Summary of the Invention

[0005] To address the technical problem that existing edge width histogram-based algorithms are susceptible to thermal noise and high-frequency textures when filtering effective frames, leading to bandwidth waste or loss of critical data, this invention provides solutions in the following aspects.

[0006] In a first aspect, the present invention provides an efficient communication method for industrial robots based on edge computing, comprising: acquiring visual images and motion speed data of an industrial robot, performing preprocessing, and determining all narrow edge pixels of each frame image; determining an edge intensity factor of each frame image based on the average gradient magnitude of each narrow edge pixel and the average gradient magnitude of all pixels in each frame image; determining a structured edge factor of each frame image based on the edge intensity factor, combined with the information entropy value and range of the gradient direction values ​​of all narrow edge pixels in each frame image; determining an optimization coefficient for an effective frame filtering threshold of each frame image based on the structured edge factor and the motion speed of the industrial robot corresponding to each frame image; optimizing a fixed effective frame filtering threshold using the optimization coefficient to obtain an optimized effective frame filtering threshold for each frame image; and using the optimized effective frame filtering threshold to determine each frame image, thereby achieving the filtering of effective frames of the image.

[0007] This invention analyzes the gradient magnitude of all narrow-edge pixels in each frame of an image to determine the edge intensity factor of each frame, reflecting the contrast significance of all narrow-edge pixels in each frame. Furthermore, by deeply analyzing the statistical characteristics of the gradient directions of all narrow-edge pixels in each frame, it determines the structured edge factor of each frame, eliminating interference from high-frequency textures in cluttered backgrounds. Finally, by analyzing the industrial robot's motion speed corresponding to each frame, it determines the optimization coefficient of the effective frame filtering threshold for each frame and the optimized effective frame filtering threshold, thereby achieving the identification and filtering of effective frame images and enhancing the communication efficiency and adaptability of industrial robots in complex actual working conditions.

[0008] Preferably, the edge intensity factor satisfies the expression: In the formula, For the index of each frame of image, For the first Edge intensity factor of frame image For the first The total number of pixels at the narrow edge of a frame image. For the first The index of each narrow edge pixel in the frame image. For the first The first frame of the image The gradient magnitude of a narrow-edge pixel. For the first The average gradient magnitude of all pixels in the frame image. To prevent tiny constants with a denominator of zero, It is the maximum-minimum normalization function.

[0009] This invention achieves edge intensity factor evaluation by constructing the ratio of the average gradient magnitude of narrow edge pixels to the average gradient magnitude of all pixels, eliminating low-contrast thermal noise, and removing the influence of thermal noise on effective frame selection, thus providing a reliable basis for subsequent optimization calculation of the effective frame selection threshold.

[0010] Preferably, the method for obtaining the information entropy value of the gradient direction value of all narrow edge pixels in each frame image is as follows: the gradient direction from 0 degrees to 360 degrees is divided into several gradient direction intervals, the proportion of narrow edge pixels falling in each gradient direction interval to the total number of narrow edge pixels is counted, the proportion of narrow edge pixels in each gradient direction interval to the total number of narrow edge pixels is input into the information entropy calculation expression, and the result of the information entropy calculation expression is used as the information entropy value of the gradient direction value of all narrow edge pixels in each frame image.

[0011] Preferably, the structured edge factor satisfies the expression: In the formula, For the index of each frame of image, For the first Structured edge factors of frame images For the first Edge intensity factor of frame image For the first The information entropy value of the gradient direction of all narrow-edge pixels in a frame image. For the first The theoretical maximum value of the information entropy of the gradient direction values ​​of all narrow-edge pixels in a frame image. For the first The range of gradient direction values ​​for all narrow-edge pixels in a frame image. It is a natural exponential function.

[0012] This invention achieves the evaluation of structured edge factors by constructing a composite function that includes an edge intensity factor, an information entropy penalty term for gradient direction values, and a range penalty term for gradient direction values. By utilizing the difference in gradient direction consistency between the high-frequency textures of the real workpiece contour and the cluttered background, the influence of the high-frequency textures of the cluttered background on the selection of effective frames is effectively eliminated, ensuring that the selected effective frames contain the geometric contours of the real workpiece that are conducive to machine vision processing.

[0013] Preferably, the optimization coefficients of the effective frame filtering threshold satisfy the expression: In the formula, For the index of each frame of image, For the first The optimization coefficients for the effective frame filtering threshold of the frame image. For the first Structured edge factors of frame images For the first The numerical value of the industrial robot's motion speed data corresponding to the time the frame image was captured. The maximum speed value set for industrial robots It is a natural exponential function. It is the maximum-minimum normalization function.

[0014] This invention constructs a composite function that includes a structured edge factor and an industrial robot motion speed penalty term, thereby evaluating the optimization coefficient of the effective frame screening threshold. When the industrial robot's high-speed movement easily generates motion blur, the optimization coefficient of the effective frame screening threshold is increased to ensure that images with motion blur are not uploaded as effective frames. At the same time, it provides a more realistic reference for optimizing the effective frame screening threshold.

[0015] Preferably, the optimized effective frame filtering threshold satisfies the expression: In the formula, For the index of each frame of image, For the first The optimized effective frame filtering threshold for frame images. A fixed threshold for selecting valid frames. For the first The optimization coefficients for the effective frame filtering threshold of the frame image. It is the maximum-minimum normalization function.

[0016] This invention optimizes a fixed effective frame filtering threshold by optimizing the effective frame filtering threshold coefficient, thereby achieving adaptive effective frame filtering threshold. When affected by thermal noise and high-frequency texture, the effective frame filtering threshold is automatically increased to avoid incorrect image transmission, thus improving the efficiency and robustness of industrial robot communication.

[0017] Preferably, the determination of each frame image includes: for each frame image, in response to its edge intensity factor being greater than its optimized effective frame filtering threshold, determining it as a valid frame image and transmitting it to the edge computing node for further processing and analysis; for each frame image, in response to its edge intensity factor not being greater than its optimized effective frame filtering threshold, determining it as an invalid frame image and discarding it in local memory.

[0018] Preferably, acquiring the visual images and motion speed data of the industrial robot includes: acquiring visual images of the industrial robot using an industrial camera deployed on the industrial robot, and acquiring motion speed data of the industrial robot using a speed sensor deployed on the industrial robot.

[0019] Preferably, determining all narrow edge pixels in each frame of an image includes: for each pixel in each frame of an image, determining it as a narrow edge pixel in response to its edge width not being greater than a narrow edge judgment threshold; and traversing all pixels in each frame of an image to determine all narrow edge pixels in each frame of an image.

[0020] Secondly, the present invention provides an efficient communication system for industrial robots based on edge computing, including a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned efficient communication method for industrial robots based on edge computing is implemented.

[0021] By adopting the above technical solution, the above-mentioned efficient communication method for industrial robots based on edge computing is generated into a computer program and stored in a memory for loading and execution by a processor. This allows for the creation of terminal devices based on the memory and processor, making them convenient to use.

[0022] The beneficial effects of this invention are as follows: By introducing an effective frame screening threshold optimization mechanism based on gradient magnitude features and gradient direction features, this invention solves the problem that existing algorithms based on edge width histograms are difficult to accurately screen out the physical contours of real workpieces when performing effective frame screening.

[0023] This invention establishes the intrinsic mapping relationship between the edge intensity factor and the physical contour of the real workpiece by analyzing the gradient magnitude and average gradient magnitude of narrow edge pixels. On this basis, it further integrates the information entropy value and range of the gradient direction value, and tightly couples the index reflecting the difference in gradient magnitude with the index reflecting the distribution of gradient direction, thereby realizing the identification of thermal noise and the influence of high-frequency texture.

[0024] This invention calculates a structured edge factor matching the gradient features of narrow-edge pixels for each image frame. By adaptively optimizing a fixed effective frame selection threshold based on the motion speed characteristics of an industrial robot, it achieves accurate matching between the effective frame selection threshold, the gradient features of narrow-edge pixels, and the robot's motion speed. For images affected by thermal noise and high-frequency textures, the threshold is automatically increased to prevent false selection. For images containing the physical contours of real workpieces, the threshold is maintained for complete retention. Ultimately, this invention improves the accuracy and anti-interference capability of edge computing-based industrial robot communication, enhances the robustness of effective frame image selection, and enables timely and accurate identification of effective frames, providing a solid technical guarantee for communication optimization and robot performance improvement. Attached Figure Description

[0025] Figure 1 This is a flowchart illustrating the efficient communication method for industrial robots based on edge computing in this invention; Figure 2 This is a comparison chart showing the results of effective frame filtering using a fixed effective frame filtering threshold in the existing technology and using an optimized effective frame filtering threshold in this invention. Detailed Implementation

[0026] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0027] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0028] This invention discloses an efficient communication method for industrial robots based on edge computing, referring to... Figure 1 The process includes steps S001 to S005, specifically: S001: Acquire the visual image of the industrial robot and the motion speed data of the industrial robot, perform preprocessing, and determine all narrow edge pixels of each frame image.

[0029] Specifically, an industrial camera is deployed on the end effector of the industrial robot to acquire visual images of the robot. A speed sensor is also deployed on the end effector to acquire motion speed data of the robot. The visual images of the industrial robot are converted to grayscale to obtain each frame. The Sobel operator is used to perform a convolution operation on each frame to determine the edge width of each pixel in each frame. For each pixel in each frame, it is determined to be a narrow edge pixel if its edge width is not greater than the narrow edge judgment threshold. All pixels in each frame are traversed to determine all narrow edge pixels in each frame. In this embodiment, the acquisition frequency of the industrial robot visual images and the industrial robot motion speed data is set to 100Hz, and the narrow edge judgment threshold is set to 3. In other embodiments, the implementer can set the acquisition frequency of the industrial robot visual images and the industrial robot motion speed data, as well as the narrow edge judgment threshold, according to the actual implementation situation.

[0030] S002: Determine the edge intensity factor of each frame of the image based on the gradient magnitude of each narrow edge pixel and the average gradient magnitude of all narrow edge pixels.

[0031] It should be noted that existing edge-width histogram-based algorithms for selecting effective frames consider any edge as sharp as it is narrow enough, ignoring the physical properties of the edges. In real industrial settings, insufficient lighting or sensor thermal noise can generate numerous narrow but low-contrast edges. These weak edges contribute little to visual servoing and can affect the effectiveness of image data transmission. The physical contours of real workpieces typically exhibit significant gray-level abrupt changes, displaying high gradient amplitudes. Conversely, contours generated by insufficient lighting or sensor thermal noise, while narrower, have relatively weak gradient amplitudes. Therefore, this invention determines an edge intensity factor for each frame to characterize the contrast significance of narrow edges within the image.

[0032] Specifically, the edge intensity factor satisfies the following expression: In the formula, For the index of each frame of image, For the first Edge intensity factor of frame image For the first The total number of pixels at the narrow edge of a frame image. For the first The index of each narrow edge pixel in the frame image. For the first The first frame of the image The gradient magnitude of a narrow-edge pixel. For the first The average gradient magnitude of all pixels in the frame image. To prevent tiny constants with a denominator of zero, For the maximum-minimum normalization function, in this embodiment, the small constant to prevent the denominator from being zero is set to 0.001. In other embodiments, implementers can set the small constant to prevent the denominator from being zero according to the actual implementation situation.

[0033] In the formula, Reflects the first The average gradient magnitude of all narrow-edge pixels in the frame image; the larger this value, the stronger the gradient. The sharper the narrow edges in a frame image, the higher the contrast, so the first... The larger the edge intensity factor of the frame image, the better.

[0034] S003: Based on the edge intensity factor, combined with the information entropy value and range of the gradient direction values ​​of all narrow edge pixels in each frame of the image, determine the structured edge factor of each frame of the image.

[0035] It should be noted that, because high-frequency textures in cluttered industrial environments often possess high gradient amplitude characteristics similar to the geometric contours of real workpieces, it is difficult to effectively distinguish between the geometric contours of real workpieces and high-frequency texture interference in a single gradient amplitude dimension. This can easily lead to high-frequency textures being mistaken for the geometric contours of real workpieces. Based on the physical properties of edge imaging, the geometric contours of real workpieces exhibit continuity and consistency in gradient direction within a local range, while the gradient directions of high-frequency textures in cluttered backgrounds typically exhibit a chaotic and random distribution. Therefore, this invention determines a structured edge factor for each frame of the image to characterize the structural order and physical realism of high-intensity edges in the image.

[0036] Specifically, the gradient direction from 0 degrees to 360 degrees is divided into several gradient direction intervals. The proportion of narrow edge pixels falling within each gradient direction interval is counted to the total number of narrow edge pixels. Combined with the information entropy calculation expression, the information entropy value of the gradient direction value of all narrow edge pixels in each frame image is calculated.

[0037] Specifically, the structured edge factor satisfies the expression: In the formula, For the index of each frame of image, For the first Structured edge factors of frame images For the first Edge intensity factor of frame image For the first The information entropy value of the gradient direction of all narrow-edge pixels in a frame image. For the first The theoretical maximum value of the information entropy of the gradient direction values ​​of all narrow-edge pixels in a frame image. For the first The range of gradient direction values ​​for all narrow-edge pixels in a frame image. It is a natural exponential function.

[0038] In the formula, The larger the value, the more likely it is to be the first. The sharper the narrow edges and the higher the contrast in a frame image, the better. The narrower the edge in the frame image, the greater the probability that it belongs to the geometric contour of the actual workpiece, therefore the... The larger the structured edge factor of the frame image, the better. The larger it is, the more likely it is to be the first The more chaotic the gradient directions of all narrow-edge pixels in a frame image, the more pronounced the gradient direction. The less likely a narrow edge in a frame image is to belong to the geometric contour of the actual workpiece, the more likely it is to be a true workpiece. The narrow edges in a frame image are more likely to contain high-frequency textures in a cluttered background, therefore the first... The smaller the structured edge factor of a frame image. The larger the value, the more likely it is to be the first. The more chaotic the gradient directions of all narrow-edge pixels in a frame image, the greater the reliability, indicating that the first... The narrow edges in a frame image are more likely to contain high-frequency textures in a cluttered background, therefore the first... The smaller the structured edge factor of a frame image.

[0039] S004: Based on the structured edge factor and the industrial robot's movement speed corresponding to each frame of the image, determine the optimization coefficient of the effective frame screening threshold for each frame of the image. Optimize the fixed effective frame screening threshold using the optimization coefficient of the effective frame screening threshold to obtain the optimized effective frame screening threshold for each frame of the image.

[0040] It should be noted that after obtaining the structured edge factor of each frame image, this invention will determine the optimization coefficient of the effective frame screening threshold of each frame image based on the structured edge factor of each frame image and the physical motion state of the industrial robot. Furthermore, the optimization coefficient of the effective frame screening threshold is used to optimize the fixed effective frame screening threshold. Existing algorithms based on edge width histograms usually use a fixed effective frame screening threshold for effective frame screening, which is difficult to adapt to the dynamic working state of the industrial robot. This invention uses the structured edge factor and the industrial robot's motion speed corresponding to each frame image to determine the optimization coefficient of the effective frame screening threshold of each frame image and the optimized effective frame screening threshold of each frame image, making the screening strategy more flexible and in line with actual physical laws.

[0041] Specifically, the optimization coefficients for the effective frame filtering threshold satisfy the expression: In the formula, For the index of each frame of image, For the first The optimization coefficients for the effective frame filtering threshold of the frame image. For the first Structured edge factors of frame images For the first The numerical value of the industrial robot's motion speed data corresponding to the time the frame image was captured. The maximum speed value set for industrial robots It is a natural exponential function. It is the maximum-minimum normalization function.

[0042] In the formula, The smaller the value, the better. The greater the likelihood that a frame image is affected by thermal noise and high-frequency textures in a cluttered background, the more likely the first frame image will be to be affected. The larger the optimization coefficient of the effective frame filtering threshold of the frame image, the more likely it is to ensure that the first frame is selected. The frame image was determined to be an invalid frame. The larger the value, the more likely it is to be the first. The greater the speed of the industrial robot's movement during the capture of the first frame image, the more significant the change in the second frame. The greater the likelihood of motion blur in a frame image, the more likely the 1st frame is to have motion blur. The larger the optimization coefficient of the effective frame filtering threshold of the frame image, the more likely it is to ensure that the first frame is selected. The frame image was determined to be an invalid frame.

[0043] Specifically, the optimized effective frame filtering threshold satisfies the expression: In the formula, For the index of each frame of image, For the first The optimized effective frame filtering threshold for frame images. A fixed threshold for selecting valid frames. For the first The optimization coefficients for the effective frame filtering threshold of the frame image. As a maximum-minimum normalization function, in this embodiment, the fixed effective frame filtering threshold is set to 0.9. In other embodiments, implementers can set the fixed effective frame filtering threshold according to the actual implementation situation. For example, when the communication link bandwidth resources are relatively tight or the image clarity standard required to be transmitted to the server is extremely high, the fixed effective frame filtering threshold can be appropriately increased to save network bandwidth to the maximum extent. When industrial robots perform highly dynamic precision operations and have high requirements for the continuity and integrity of visual feedback signals, the fixed effective frame filtering threshold can be appropriately reduced to avoid the omission of key visual information due to over-filtering.

[0044] In the formula, The larger the value, the more likely it is to be the first. The greater the likelihood that a frame image is affected by thermal noise and high-frequency textures in a cluttered background, the greater the likelihood of motion blur. Therefore, the first frame... The larger the optimized effective frame filtering threshold of the frame image, the more likely it is to ensure that the first frame is selected. The frame image was determined to be an invalid frame.

[0045] S005: Use the optimized effective frame filtering threshold to determine the effective frames of each image, thereby achieving the filtering of effective frames of the image.

[0046] Specifically, the effective frame selection of an image includes: for each frame image, if its edge intensity factor is greater than its optimized effective frame selection threshold, it is determined to be an effective frame image and transmitted to the edge computing node for further processing and analysis; for each frame image, if its edge intensity factor is not greater than its optimized effective frame selection threshold, it is determined to be an invalid frame image and discarded in local memory.

[0047] It's important to note that the edge intensity factor, as a core indicator for analyzing the edge sharpness and contrast quality of each image frame, directly reflects the richness and usability of the visual information contained in each frame. When compared with the optimized effective frame selection threshold, a higher edge intensity factor indicates that the image is not only sufficiently clear in contrast but also possesses a high degree of physical realism in its edge structure, sufficient to support the backend visual servoing task. This judgment method eliminates low-quality image data at the source, ensuring that the backend receives high-value effective frames while minimizing the bandwidth consumption of invalid image data, significantly improving the efficiency and robustness of the entire communication process.

[0048] like Figure 2As shown in the figure, this figure is a comparison of the effects of the present invention on optimizing the effective frame screening threshold. The horizontal axis in the figure represents the image frame sequence acquired by the industrial robot's vision, and the vertical axis represents the normalized values ​​of the image feature indicators. As can be seen from the figure, when the industrial robot is in the high-speed movement stage, although the edge intensity factor fluctuates due to motion blur, the corresponding curve of the present invention adaptively increases accordingly, effectively preventing blurred images from being misjudged as effective frames. Particularly significant is that in the stage affected by high-frequency textures, although noise causes the edge intensity factor to be artificially high and exceeds the fixed effective frame screening threshold of the existing technology, the corresponding curve of the present invention can drive the effective frame screening threshold to rise significantly and be significantly higher than the edge intensity factor curve. This result intuitively proves that the present invention can dynamically adjust the judgment criteria according to the working conditions and accurately intercept invalid frames, thereby significantly reducing bandwidth consumption while ensuring communication quality.

[0049] This invention also discloses an efficient communication system for industrial robots based on edge computing, including a processor and a memory. The memory stores computer program instructions, which, when executed by the processor, implement the efficient communication method for industrial robots based on edge computing according to this invention.

[0050] The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the settings and functions of which are known in the art and will not be described in detail here.

Claims

1. A high-efficiency communication method for industrial robots based on edge computing, characterized in that, include: Acquire visual images and motion speed data of the industrial robot, perform preprocessing, and determine all narrow edge pixels in each frame of the image. Based on the average gradient magnitude of each narrow-edge pixel in each frame and the average gradient magnitude of all pixels in each frame, the edge intensity factor of each frame is determined. Based on the edge intensity factor, combined with the information entropy value and range of the gradient direction values ​​of all narrow-edge pixels in each frame, the structured edge factor of each frame is determined. Based on the structured edge factor and the industrial robot motion speed corresponding to each frame, the optimization coefficient of the effective frame filtering threshold of each frame is determined. The fixed effective frame filtering threshold is optimized using the optimization coefficient of the effective frame filtering threshold to obtain the optimized effective frame filtering threshold of each frame. The optimized effective frame filtering threshold is used to determine the effective frames of each image, thereby achieving the filtering of effective frames. The determination of each image includes: for each image, if its edge intensity factor is greater than its optimized effective frame filtering threshold, it is determined to be an effective frame image and transmitted to the edge computing node for further processing and analysis; for each image, if its edge intensity factor is not greater than its optimized effective frame filtering threshold, it is determined to be an invalid frame image and discarded in local memory.

2. The efficient communication method for industrial robots based on edge computing according to claim 1, characterized in that, The edge strength factor satisfies the expression: In the formula, For the index of each frame of image, For the first Edge intensity factor of frame image For the first The total number of pixels at the narrow edge of a frame image. For the first The index of each narrow edge pixel in the frame image. For the first The first frame of the image The gradient magnitude of a narrow-edge pixel. For the first The average gradient magnitude of all pixels in the frame image. To prevent tiny constants with a denominator of zero, It is the maximum-minimum normalization function.

3. The efficient communication method for industrial robots based on edge computing according to claim 1, characterized in that, The method for obtaining the information entropy value of the gradient direction of all narrow edge pixels in each frame image is as follows: the gradient direction from 0 degrees to 360 degrees is divided into several gradient direction intervals, the proportion of narrow edge pixels falling in each gradient direction interval to the total number of narrow edge pixels is counted, the proportion of narrow edge pixels in each gradient direction interval to the total number of narrow edge pixels is input into the information entropy calculation expression, and the result of the information entropy calculation expression is used as the information entropy value of the gradient direction of all narrow edge pixels in each frame image.

4. The efficient communication method for industrial robots based on edge computing according to claim 1 or 3, characterized in that, The structured edge factor satisfies the expression: In the formula, For the index of each frame of image, For the first Structured edge factors of frame images For the first Edge intensity factor of frame image For the first The information entropy value of the gradient direction of all narrow-edge pixels in a frame image. For the first The theoretical maximum value of the information entropy of the gradient direction values ​​of all narrow-edge pixels in a frame image. For the first The range of gradient direction values ​​for all narrow-edge pixels in a frame image. It is a natural exponential function.

5. The efficient communication method for industrial robots based on edge computing according to claim 1, characterized in that, The optimization coefficients of the effective frame filtering threshold satisfy the expression: In the formula, For the index of each frame of image, For the first The optimization coefficients for the effective frame filtering threshold of the frame image. For the first Structured edge factors of frame images For the first The numerical value of the industrial robot's motion speed data corresponding to the time the frame image was captured. The maximum speed value set for industrial robots It is a natural exponential function. It is the maximum-minimum normalization function.

6. The efficient communication method for industrial robots based on edge computing according to claim 1, characterized in that, The optimized effective frame filtering threshold satisfies the expression: In the formula, For the index of each frame of image, For the first The optimized effective frame filtering threshold for frame images. A fixed threshold for selecting valid frames. For the first The optimization coefficients for the effective frame filtering threshold of the frame image. It is the maximum-minimum normalization function.

7. The efficient communication method for industrial robots based on edge computing according to claim 1, characterized in that, The acquisition of visual images and motion speed data of the industrial robot includes: acquiring visual images of the industrial robot using an industrial camera deployed on the industrial robot, and acquiring motion speed data of the industrial robot using a speed sensor deployed on the industrial robot.

8. The efficient communication method for industrial robots based on edge computing according to claim 1, characterized in that, The step of determining all narrow edge pixels in each frame of an image includes: for each pixel in each frame of an image, determining it as a narrow edge pixel in response to its edge width not being greater than the narrow edge judgment threshold; and traversing all pixels in each frame of an image to determine all narrow edge pixels in each frame of an image.

9. A high-efficiency communication system for industrial robots based on edge computing, characterized in that, include: A processor and a memory, the memory storing computer program instructions that, when executed by the processor, implement the efficient communication method for industrial robots based on edge computing according to any one of claims 1-8.

Images