CN114972349A - Carrier roller running state detection method and system based on image processing - Google Patents

Abstract

The invention discloses a method and a system for detecting the running state of a carrier roller based on image processing, and relates to the field of artificial intelligence. The method mainly comprises the following steps: acquiring a multi-frame gray image of the carrier roller to be detected at each moment in the running process of the conveyor belt; overlapping edge vibration areas in adjacent multi-frame gray scale images at each moment to obtain an edge vibration image corresponding to each moment; carrying out non-blind deconvolution on the edge vibration image at each moment to obtain a clear vibration image at each moment; and respectively obtaining the similarity between the edge vibration image and the clear vibration image corresponding to each moment so as to determine whether the carrier roller to be tested is an abnormal carrier roller. According to the embodiment of the invention, the front images of the carrier rollers are processed, so that the state detection result of each carrier roller can be respectively obtained in the operation process of the conveying device provided with the carrier rollers, and the state detection precision of the carrier rollers is improved.

Description

Method and system for detection of idler running state based on image processing

technical field

The present application relates to the field of artificial intelligence, and in particular to a method and system for detecting the running state of an idler based on image processing.

Background technique

Belt conveyor is a large-scale coal mine transportation equipment, which has the advantages of simple structure, large transportation volume, low construction cost and maintenance cost. However, due to the complex working environment of the mine and the operation state of the belt conveyor, many safety accidents have occurred. Among them, the idler is one of the most important parts of the belt conveyor. The main function of the idler is to support the conveyor belt and the conveyed belt. During the operation of the belt conveyor, the idler may not be in close contact with the conveyor belt due to improper installation or defects on the surface of the idler, which will cause the idler to vibrate up and down and further shorten the idler. The service life of the medium bearing is harmful to the service life of the conveyor belt.

In the prior art, the audio frequency at the idler is mainly analyzed to determine whether there is any abnormality in the idler. However, during the operation of the idler, there are sounds during the material transmission process, and parts other than the idler on the belt conveyor. The interference of various sounds, such as sound, makes it difficult to collect the audio frequency of a single idler during operation, so that it is difficult to obtain an accurate state detection result for the idler.

SUMMARY OF THE INVENTION

In view of the above technical problems, the present invention provides a method and system for detecting the running state of an idler based on image processing. As for the state detection result of each idler, compared with the state detection of the idler by audio in the prior art, the state detection accuracy of the idler can be improved.

In a first aspect, an embodiment of the present invention proposes a method for detecting the running state of an idler based on image processing, including:

Acquire multiple frontal images of the roller to be tested at each moment during the operation of the conveyor belt, and perform grayscale to obtain multiple frames of grayscale images corresponding to each moment.

The edge vibration regions in the adjacent multi-frame grayscale images at each moment are superimposed to obtain the edge vibration images corresponding to each moment, and the edge vibration regions are regions located at predetermined heights of the upper and lower boundaries of the grayscale images.

Use variational Bayes to obtain the convolution kernel corresponding to the edge vibration image at each moment, and perform a non-blind deconvolution and deblurring method on the edge vibration image at each moment according to the convolution kernel, and obtain the deblurred image at each moment. Clear vibrating image.

Obtain the similarity between the edge vibration image corresponding to each moment and the clear vibration image respectively, and obtain the proportion of the similarity less than or equal to the similarity threshold among all the similarities corresponding to all the moments within the preset duration. , in the case that the ratio is greater than the ratio threshold, determine the idler to be tested as an abnormal idler.

Further, in the method for detecting the running state of the idler based on image processing, the method further includes updating the similarity corresponding to each moment, including:

Obtain the long-run and low-gray-level advantage of the gray-scale run-length matrix of any frame of grayscale image at each moment in the movement direction, and use it as the first eigenvalue at each moment, and the movement direction is the roller to be tested. The movement direction of the corresponding direction in the grayscale image;

Obtain the long-run and low-gray-level advantages of the gray-scale run-length matrix of any frame of grayscale image at each moment in each direction other than the motion direction, and set the long-run low-level advantages of the grayscale run-length matrix in all directions other than the motion direction. The mean value of the gray-level advantage is used as the second eigenvalue at each moment; the directions other than the motion direction are directions other than the directions corresponding to the motion directions in 0°, 45°, 90° and 135°;

Taking the division operation result of the second eigenvalue at each moment and the first eigenvalue as the third eigenvalue at each moment;

The result of multiplying the similarity corresponding to each moment and the third eigenvalue corresponding to each moment is taken as the updated similarity corresponding to each moment.

Further, in the method for detecting the running state of the idler based on image processing, the method further includes obtaining the moving direction of the idler to be measured according to the grayscale image, and obtaining the moving direction of the idler to be measured according to the grayscale image includes:

The direction with the smallest difference between the gray value of each pixel in the grayscale image and the pixel in its eight neighborhoods is taken as the characteristic direction of each pixel in the grayscale image;

The characteristic direction with an included angle of 180 degrees is taken as the same characteristic direction, and the characteristic direction with the largest intermediate frequency among all characteristic directions of all pixel points in the grayscale image is taken as the moving direction of the roller to be tested.

Further, in the method for detecting the running state of the idler based on image processing, before the front image of the idler to be measured is grayed to obtain a grayscale image, the method further includes:

Collect an image including the front surface of the roller to be tested and perform image segmentation to obtain the front image of the roller to be tested, and the pixel value of the pixels other than the roller to be tested in the front image is 0.

Further, in the method for detecting the running state of the idler based on image processing, image segmentation is performed on the image including the front surface of the idler to be tested to obtain the frontal image of the idler to be tested through DNN.

Further, in the method for detecting the running state of the idler based on image processing, before acquiring multiple frames of frontal images of the idler to be tested at each moment, the method further includes determining the idler to be measured, wherein the determination of the idler to be measured is determined. The process includes:

During the operation of the conveyor belt, the audio frequency of each idler is collected respectively, and the idler whose difference between the audio frequency at the idler and the standard idler is greater than the difference threshold is determined as the idler to be tested.

Further, in the method for detecting the running state of the idler based on image processing, the similarity between the edge vibration image corresponding to each moment and the clear vibration image is the structural similarity SSIM.

In a second aspect, an embodiment of the present invention provides an image processing-based system for detecting the running state of an idler, including: a memory and a processor, where the processor executes a computer program stored in the memory to implement the embodiments of the present invention A method for detecting the running state of idlers based on image processing.

The present invention provides a method and system for detecting the running state of an idler roller based on image processing. Compared with the prior art, the beneficial effect of the embodiment of the present invention is: by processing the front image of the idler roller, it can be During the operation of the roller conveying device, the state detection result of each idler is obtained separately, which can improve the state detection accuracy of the idler compared to the state detection of the idler by audio in the prior art.

Description of drawings

In order to explain the embodiments of the present invention or the technical solutions in the prior art more clearly, the following briefly introduces the accompanying drawings that need to be used in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only These are some embodiments of the present invention, and for those of ordinary skill in the art, other drawings can also be obtained from these drawings without any creative effort.

1 is a schematic flowchart of a method for detecting the running state of an idler based on image processing provided by an embodiment of the present invention;

2 is a schematic diagram of the relative position of the front surface of the idler and the moving direction of the conveyor belt provided by an embodiment of the present invention;

FIG. 3 is a schematic flowchart of updating the similarity corresponding to each moment according to an embodiment of the present invention.

Detailed ways

In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention, but not to limit the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

In the following description, for the purpose of illustration rather than limitation, specific details such as a specific system structure and technology are set forth in order to provide a thorough understanding of the embodiments of the present application. However, it will be apparent to those skilled in the art that the present application may be practiced in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, devices, circuits, and methods are omitted so as not to obscure the description of the present application with unnecessary detail.

The terms "first" and "second" are only used for descriptive purposes, and should not be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined with "first" and "second" may explicitly or implicitly include one or more of the features; in the description of this embodiment, unless otherwise specified, the meaning of "multiple" are two or more.

An embodiment of the present invention provides a method for detecting the running state of an idler based on image processing, as shown in FIG. 1 , including:

Step S100 , acquiring multiple frames of front images of the roller to be tested at each moment during the running process of the conveyor belt, and performing grayscale to obtain multiple frames of grayscale images corresponding to each moment.

During the operation of the conveyor belt, the conveyor belt and the materials above the conveyor belt need to be supported by the idler roller. However, when the idler roller has defects such as pits on the surface, the idler roller and the conveyor belt cannot be fully contacted, resulting in the idler roller in the up and down direction. Vibration and further shorten the service life of the bearings in the idler. Therefore, it is necessary to find out the abnormal vibration in the idler in time.

Figure 2 is a schematic diagram of the relative position of the front surface of the idler and the moving direction of the conveyor belt in the embodiment of the present invention. Multiple frames of front images of the idler to be tested during the running process of the conveyor belt can be collected from the front of the idler. The collected multiple frames The front image is an RGB image. RGB is a color standard. Various colors are obtained by changing the three color channels of red (R), green (G), and blue (B) and superimposing them on each other. , RGB is the color representing the three channels of red, green and blue.

Grayscale is performed on the multi-frame frontal images collected at each moment, respectively, to obtain multi-frame grayscale images corresponding to each moment, wherein the grayscale process can adopt maximum value grayscale or average value grayscale.

Optionally, before graying the front image of the roller to be tested to obtain a grayscale image, collect an image containing the front of the roller to be tested, and perform image segmentation on the image containing the front of the roller to be tested, The front image of the roller to be tested is obtained, wherein the pixel value of the pixels other than the roller to be tested in the front image obtained after segmentation is 0. In this way, the adverse effects of the pixels other than the roller on the detection result can be removed.

The image segmentation process of the front containing the roller to be tested can be realized by DNN (Deep Neural Networks, deep neural network), and the training process of the DNN can include the following steps: the pixels in the image containing the front of the roller are processed. Manual labeling, dividing the pixels into the idler class and the background class other than the idler, labeling the pixels of the background class as 0, and using the labeled images to train the DNN. At the same time, during the training process The training process is supervised using a cross-entropy loss function.

Optionally, before acquiring multiple frames of front images of the rollers to be tested at each moment, the rollers to be tested may also be determined from all the rollers, which may specifically include: collecting each roller separately during the operation of the conveyor belt. and the difference between the audio frequency at the idler and the standard idler is greater than the difference threshold as the idler to be tested. It should be noted that the idler under the load conditions of different goods and the moving speed, so The corresponding audio frequency will be in different states. When the difference between the audio frequency of the idler and the standard audio corresponding to the current state is greater than the difference threshold, the idler is more likely to be a faulty idler, and it can be determined as the one to be tested. The idler, in this way, can realize the preliminary screening of the idler.

Wherein, for the degree of difference between the audios, the amplitude of the audio can be determined, the amplitude is the difference between the maximum value and the minimum value of the audio, and the absolute value of the difference between the amplitudes between different audios can be determined. , as the degree of difference between different audios.

Step S200, superimpose the edge vibration areas in the adjacent multi-frame grayscale images at each moment to obtain the edge vibration images corresponding to each moment, and the edge vibration areas are located at the predetermined height of the upper and lower boundaries of the grayscale image. area.

The reason for superimposing the edge vibration areas in the adjacent multi-frame grayscale images at each moment in the embodiment of the present invention is that the edge vibration area of a single frame cannot effectively reflect the up and down vibration at the same moment. , since the upper and lower boundary parts of the grayscale image on the front side of the idler can effectively reflect the upper and lower vibrations, the edge vibration area in the embodiment of the present invention is an area located at a predetermined height of the upper and lower boundaries of the grayscale image.

Step S300, use variational Bayes to obtain the convolution kernel corresponding to the edge vibration image at each moment, and perform a non-blind deconvolution and deblurring method on the edge vibration image at each moment according to the convolution kernel, and obtain each time after deblurring. A clear vibrating image of a moment.

In the embodiment of the present invention, a de-motion blurring algorithm is used to obtain a clear vibration image corresponding to the edge vibration image. Since the larger the vibration amplitude is, the more blurred the image is, and the similarity before and after the de-motion blur is smaller. Therefore, In the embodiment of the present invention, in the subsequent steps, the vibration intensity of the roller to be tested is reflected by comparing the similarity between the images before and after the edge vibration image is deblurred.

Step S400, obtain the similarity between the edge vibration image corresponding to each moment and the clear vibration image respectively, and obtain all the similarities corresponding to all the moments in the preset duration, the similarity less than or equal to the similarity threshold. If the ratio is greater than the ratio threshold, the idler to be tested is determined to be an abnormal idler.

Since the larger the vibration amplitude of the idler is, the greater the degree of blur caused to the edge vibration area, the smaller the similarity between the clear vibration image obtained after deblurring and the edge vibration image. Therefore, in the embodiment of the present invention, by determining the edge vibration The similarity between the vibration image and its corresponding clear vibration image can reflect the vibration amplitude of the idler at the edge.

As a feasible implementation manner, the process of obtaining the similarity between the edge vibration image and the clear vibration image in the embodiment of the present invention includes:

是边缘振动图像中像素点的灰度均值，

是清晰振动图像中像素点的灰度均值，

是边缘振动图像中像素点的灰度值的方差，

是清晰振动图像中像素点的灰度值的方差，

是边缘振动图像与清晰振动图像中像素点的灰度值的协方差，

，

，其中，

为预设第一数值，

为预设第二数值，作为一个示例，本发明实施例中

，

。In the formula,

is the gray mean of the pixels in the edge vibration image,

is the gray mean of the pixels in the clear vibration image,

is the variance of the gray value of the pixel point in the edge vibration image,

is the variance of the gray value of the pixel point in the clear vibration image,

is the covariance of the gray value of the pixel points in the edge vibration image and the clear vibration image,

,

,in,

is the default first value,

is a preset second value. As an example, in this embodiment of the present invention,

,

.

As another feasible implementation, in the embodiment of the present invention, the similarity between the edge vibration image and the clear vibration image is calculated by using the structural similarity SSIM. It should be noted that the structural similarity SSIM (structural similarity index) is a kind of An index to measure the similarity of two images. From the perspective of image composition, it defines the structure information as independent of brightness and contrast, reflecting the properties of the object structure in the scene, and models the distortion as three different brightness, contrast and structure. A combination of factors, where the mean is an estimate of brightness, the standard deviation is an estimate of contrast, and the covariance is a measure of structural similarity.

As another feasible implementation manner, in the embodiment of the present invention, the similarity between the edge vibration image and the clear vibration image is calculated by using the Pearson correlation coefficient. It should be noted that, in statistics, the Pearson correlation coefficient, also known as the Pearson correlation coefficient Pearson product-moment correlation coefficients (PPMCCs or PCCs for short), which can be applied to calculate the similarity between two images.

In the case where the proportion of the similarity less than or equal to the similarity threshold is greater than the proportion threshold among all the similarities corresponding to all the moments within the preset time period, the roller to be tested can be determined as an abnormal roller; as an example , in this embodiment of the present invention, the similarity threshold is 0.7, and the implementer may determine the similarity threshold according to actual needs.

It should be noted that when the ratio of the similarity less than or equal to the similarity threshold among all the similarities within the preset time period is greater than the proportional threshold, it means that the vibration of the idler exceeds the allowable range, and the idler to be tested can be determined to exist. Abnormal idler.

Optionally, after it is determined that there is an abnormal idler, the abnormal idler can be replaced. At the same time, for the replacement process of the abnormal idler, the idler replacement device can be used to temporarily replace the abnormal idler to be replaced to complete the support. In the work of the conveyor belt, after the replacement of the idler rollers is completed, the idler roller replacement device is removed, and the replaced idler rollers are used to continue supporting the conveyor belt.

Optionally, in this embodiment of the present invention, the similarity corresponding to each moment may also be updated, as shown in FIG. 3 , which may specifically include the following steps:

Step S110: Obtain the long-run and low-gray-level advantage of the gray-scale run-length matrix of the grayscale run-length matrix in the motion direction of any frame of grayscale images at each moment, and use it as the first eigenvalue at each moment, and the motion direction is to be Measure the direction of movement of the idler in the corresponding direction in the grayscale image.

During the operation of the conveyor belt, the movement speed of the idler will also affect the vibration of the idler. At the same time, under different movement speeds of the idler with the same texture, the texture presented in the front image is different, mainly as follows: In the front image of the roller with faster moving speed, there will be more elongated connected domains along the moving direction.

At the same time, this feature can be well reflected by using the long-run grayscale of the grayscale run-length matrix. Therefore, in the embodiment of the present invention, the grayscale run-length matrix in the motion direction of any frame of grayscale image at each moment is obtained. The long-run low-gray-level advantage of , and take it as the first eigenvalue at each moment.

It should be noted that the grayscale run-length matrix of an image reflects the comprehensive information of the grayscale of the image, such as direction, adjacent interval, and variation range, which is one of the basis for the local pattern of the analyzed image and its arrangement rules. The grayscale run-length matrix can realize the statistics of the number of consecutive occurrences of the same grayscale value in the same direction in an image. On an image, consecutive pixels in a certain direction have the same grayscale value, and the grayscale run-length matrix is The texture features are obtained by statistical distribution of these pixel points.

Specifically, the process of obtaining the long-run gray-level advantage of the gray-level co-occurrence matrix of the gray-level image includes:

为灰度图像在

方向的灰度共生矩阵的长游程第灰度优势，

为灰度图像中第

个灰度级的像素点沿着

方向的游程为

的频数，

为灰度游程矩阵中最大游程，

为灰度图像中的灰度级数，游程指的是连续出现的次数，同时，

为0度、45度、90度和135度中的任意一种。In the formula,

for grayscale images in

The gray-scale advantage of the long-run gray co-occurrence matrix of the direction,

for the grayscale image

grayscale pixels along the

The run in the direction is

frequency,

is the maximum run length in the grayscale run length matrix,

is the number of grayscale levels in the grayscale image, and the run length refers to the number of consecutive occurrences. At the same time,

Any of 0 degrees, 45 degrees, 90 degrees and 135 degrees.

Optionally, the moving direction of the roller to be tested can also be obtained according to the grayscale image, which may specifically include: taking the direction with the smallest difference between the grayscale values of each pixel in the grayscale image and the pixels in its eight neighborhoods as the grayscale respectively. The characteristic direction of each pixel point in the grayscale image; the characteristic direction with an included angle of 180 degrees is regarded as the same characteristic direction, and the characteristic direction with the largest frequency among all characteristic directions of all pixel points in the grayscale image is used as the characteristic direction of the roller to be tested. direction of movement.

Step S120: Obtain the long-run and low-gray-level advantages of the gray-scale run-length matrix of any frame of grayscale image at each moment in each direction other than the motion direction, and use the grayscale run-length matrix of each direction other than the motion direction. The mean value of the advantages of long run and low gray level is used as the second eigenvalue at each moment; the directions other than the motion direction are directions other than the directions corresponding to the motion directions in 0°, 45°, 90° and 135°.

Step S130, taking the division operation result of the second eigenvalue at each moment and the first eigenvalue as the third eigenvalue at each moment; taking the similarity corresponding to each moment and the third characteristic corresponding to each moment The result of the product of the values is used as the updated similarity corresponding to each moment.

In this way, the obtained updated similarity can include the influence of the vibration caused by the movement speed, so as to obtain a more accurate detection result of the idler state.

Based on the same inventive concept as the above method, this embodiment also provides an image processing-based idler running state detection system. In this embodiment, the image processing-based idler running state detection system includes a memory and a processor. The processor executes the computer program stored in the memory to realize the detection of the state of the idler as described in the embodiment of the method for detecting the running state of the idler based on image processing.

Since the method for detecting the state of the idler has already been described in the embodiment of the method for detecting the running state of the idler based on image processing, it will not be repeated here.

To sum up, the present invention provides a method and system for detecting the running state of the idler based on image processing. By processing the front image of the idler, it is possible to obtain the detection results respectively during the operation of the conveying device equipped with the idler. As for the state detection result of each idler, compared with the state detection of the idler by audio in the prior art, the state detection accuracy of the idler can be improved.

Words such as "including", "comprising", "having", etc. referred to herein are open-ended words meaning "including but not limited to" and are used interchangeably therewith. As used herein, the words "or" and "and" refer to and are used interchangeably with the word "and/or" unless the context clearly dictates otherwise. As used herein, the word "such as" refers to and is used interchangeably with the phrase "such as but not limited to".

It should also be pointed out that in the method and system of the present invention, each component or each step can be decomposed and/or recombined. These disaggregations and/or recombinations should be considered equivalents of the present disclosure.

The above-mentioned embodiments are only examples for clear description, and do not limit the protection scope of the present invention. For those of ordinary skill in the art, changes or modifications in other different forms can also be made on the basis of the above description, and it is not necessary and impossible to list all the implementations here. All the same or similar designs as the present invention fall within the protection scope of the present invention.

Claims (8)

1. A method for detecting the running state of an idler based on image processing, characterized in that, comprising: Acquire multiple frontal images of the roller to be tested at each moment during the operation of the conveyor belt, and perform grayscale to obtain multiple frames of grayscale images corresponding to each moment; The edge vibration regions in the adjacent multi-frame grayscale images at each moment are superimposed to obtain the edge vibration images corresponding to each moment, and the edge vibration regions are regions located at the predetermined heights of the upper and lower boundaries of the grayscale images; Use variational Bayes to obtain the convolution kernel corresponding to the edge vibration image at each moment, and perform a non-blind deconvolution and deblurring method on the edge vibration image at each moment according to the convolution kernel, and obtain the deblurred image at each moment. Clear vibration image; Obtain the similarity between the edge vibration image corresponding to each moment and the clear vibration image respectively, and obtain the proportion of the similarity less than or equal to the similarity threshold among all the similarities corresponding to all the moments within the preset duration. , in the case that the ratio is greater than the ratio threshold, determine the idler to be tested as an abnormal idler. 2 . The method for detecting the running state of an idler based on image processing according to claim 1 , wherein the method further comprises updating the similarity corresponding to each moment, including: 2 . Obtain the long-run and low-gray-level advantage of the gray-scale run-length matrix of any frame of grayscale image at each moment in the movement direction, and use it as the first eigenvalue at each moment, and the movement direction is the roller to be tested. The movement direction of the corresponding direction in the grayscale image; Obtain the long-run and low-gray-level advantages of the gray-scale run-length matrix of any frame of grayscale image at each moment in each direction other than the motion direction, and set the long-run low-level advantages of the grayscale run-length matrix in all directions other than the motion direction. The mean value of the gray-level advantage is used as the second eigenvalue at each moment; the directions other than the motion direction are directions other than the directions corresponding to the motion directions in 0°, 45°, 90° and 135°; Taking the division operation result of the second eigenvalue at each moment and the first eigenvalue as the third eigenvalue at each moment; The result of multiplying the similarity corresponding to each moment and the third eigenvalue corresponding to each moment is taken as the updated similarity corresponding to each moment. 3 . The method for detecting the running state of an idler based on image processing according to claim 2 , wherein the method further comprises obtaining the moving direction of the idler to be measured according to the grayscale image, and obtaining the moving direction of the idler to be measured according to the grayscale image. 4 . Measure the direction of movement of the idler, including: The direction with the smallest difference between the gray value of each pixel in the grayscale image and the pixel in its eight neighborhoods is taken as the characteristic direction of each pixel in the grayscale image; The characteristic direction with an included angle of 180 degrees is taken as the same characteristic direction, and the characteristic direction with the largest intermediate frequency among all characteristic directions of all pixel points in the grayscale image is taken as the moving direction of the roller to be tested. 4. The method for detecting the running state of an idler based on image processing according to claim 1, characterized in that, before the frontal image of the idler to be measured is grayed to obtain a grayscale image, the method further comprises: Collect an image including the front surface of the roller to be tested and perform image segmentation to obtain the front image of the roller to be tested, and the pixel value of the pixels other than the roller to be tested in the front image is 0. 5. The method for detecting the running state of an idler based on image processing according to claim 4, characterized in that, performing image segmentation on the image containing the front of the idler to be tested to obtain the frontal image of the idler to be measured is achieved by DNN . 6 . The method for detecting the running state of an idler based on image processing according to claim 1 , wherein before acquiring multiple frames of front images of the idler to be tested at each moment, the method further comprises determining the idler to be measured. 7 . , wherein, the process of determining the idler to be tested includes: During the operation of the conveyor belt, the audio frequency of each idler is collected separately, and the idler whose difference between the audio frequency at the idler and the standard idler is greater than the threshold of the difference degree is determined as the idler to be tested. 7 . The method for detecting the running state of an idler based on image processing according to claim 1 , wherein the similarity between the edge vibration image corresponding to each moment and the clear vibration image is the structural similarity SSIM. 8 . 8. A system for detecting the running state of idlers based on image processing, comprising: a memory and a processor, wherein the processor executes a computer program stored in the memory to realize any one of claims 1-7 The method for detecting the running state of the idler based on image processing described in item.

Images