Conveyor carrier roller running state detection system based on image analysis

Abstract

The invention discloses a conveyor carrier roller operation state detection system based on image analysis, particularly relates to the technical field of target detection, and comprises an image acquisition module, an image recognition module, a feature extraction module, a state evaluation module and a sampling frequency optimization module. The image acquisition module acquires a carrier roller operation image in real time; the image recognition module performs preliminary analysis on the image and recognizes whether an early fault sign exists or not; the feature extraction module extracts a feature group I and a feature group II when detecting a fault sign; the state evaluation module evaluates the running state of the carrier roller based on the extracted feature group; the sampling frequency optimization module optimizes the sampling frequency according to the basic sampling frequency, the rotation consistency index and the appearance damage index in a low-quality operation state, so as to realize accurate detection; according to the method, early signs of faults in the running process of the carrier roller can be captured, and through multi-dimensional state evaluation and dynamic sampling frequency adjustment, the situation that the conveying cost is increased due to fault expansion is effectively avoided.

Description

Technical Field

[0001] The present invention relates to the technical field of target detection, and more specifically, to a conveyor roller running state detection system based on image analysis. Background Art

[0002] Conveyors are a type of material transportation equipment widely used in industrial production, and are widely used in mining, metallurgy, ports, electricity, chemical industry and other fields. As an important part of conveyors, rollers play a key role in supporting the weight of conveyor belts and materials and reducing the running resistance of conveyor belts. Their running status directly affects the stability and service life of conveyors. However, during long-term use, rollers are prone to wear, damage, deviation and other problems due to high loads, harsh environments and lack of effective maintenance, which in turn leads to abnormal operation of the conveyor system.

[0003] Manual inspection is a traditional way to detect the status of conveyors. On-site personnel conduct visual inspections and manual measurements of rollers to determine their operating status. However, inspections require a lot of manpower and time, making it difficult to achieve comprehensive monitoring of large-scale conveyor systems. Manual inspections are usually based on regular inspections, which make it difficult to detect early signs of failure in a timely manner. In addition, the detection results rely on the experience and judgment of the inspectors, which results in large errors. Moreover, the lag of passive maintenance often leads to problems not being discovered in a timely manner, which may cause serious consequences such as conveyor belt deviation, shutdown, and even equipment damage, affecting production efficiency and safety.

[0004] Therefore, a conveyor roller operating status detection system based on image analysis is proposed to improve the efficiency of conveyor roller operating status detection. Summary of the invention

[0005] To achieve the above object, the present invention provides the following technical solutions: A conveyor roller running state detection system based on image analysis, including an image acquisition module, an image recognition module, a feature extraction module, and a state evaluation module; The image acquisition module is used to collect the running images of conveyor rollers in real time; The image recognition module is used to perform preliminary analysis on the running images of the conveyor rollers acquired by the image acquisition module, and identify whether there are early signs of running failures during the running of the conveyor rollers; The feature extraction module is used to extract features from the conveyor roller operation image when there are early signs of operation failure during the operation of the conveyor roller, and then obtain feature group 1 and feature group 2; The state evaluation module is used to evaluate the operating state of the conveyor roller based on feature group one and feature group two.

[0006] In a preferred embodiment, feature extraction is performed on the conveyor roller running image, and then feature group one and feature group two are obtained, which means: extracting the roller rotation trajectory feature data and the appearance damage feature data, and then obtaining feature group one composed of the roller rotation trajectory feature data, and feature group two composed of the appearance damage feature data.

[0007] In a preferred embodiment, the state evaluation module is used to evaluate the running state of the conveyor roller based on the feature group 1 and the feature group 2, which means: The roller rotation consistency index is generated based on feature group one, and the roller appearance damage index is generated based on feature group two. Then, the roller operation status is divided into a high-quality operation status or a low-quality operation status through fuzzy logic.

[0008] In a preferred embodiment, the present invention further comprises a sampling frequency optimization module for optimizing the sampling frequency according to the basic sampling frequency, the roller rotation consistency index and the roller appearance damage index under low-quality operation conditions.

[0009] In a preferred embodiment, the preliminary analysis operation refers to: In a fixed time window, the roller rotation speed data corresponding to each time point in the conveyor is obtained to obtain the roller rotation speed data time series, and then the absolute value of the corresponding data difference between two adjacent time points is calculated, and then the average value and standard deviation of all absolute values ​​are calculated.

[0010] In a preferred embodiment, identifying whether there is an early sign of an operating failure during the operation of the conveyor roller refers to: The average value of all absolute values ​​is compared with the preset threshold value one, and the standard deviation of all absolute values ​​is compared with the preset threshold value two. If the average value of all absolute values ​​is less than or equal to the preset threshold value one and the standard deviation of all absolute values ​​is less than or equal to the preset threshold value two, a normal signal is generated. If the average value of all absolute values ​​is less than or equal to the preset threshold value one and the standard deviation of all absolute values ​​is less than or equal to the preset threshold value two, an abnormal signal is generated. When the abnormal signal is generated, it indicates that there are early signs of operating failure during the operation of the conveyor rollers.

[0011] In a preferred embodiment, the logic for obtaining the roller rotation consistency index is: Get the rotation trajectory point sequence P of the roller: ; is the center point position of the trajectory of the nth collected running image; Calculate the offset of the center point of the track in each running image: ; Indicates the offset of the center point of the trajectory of the i-th collected running image. is the center point position of the trajectory of the i-th collected running image, is the average position of the center point of the trajectory of all collected running images; Based on the offset of the center point of the trajectory, the variance of the trajectory fluctuation is calculated, that is, the offset of the center point of all trajectories Variance ; Calculate the overall offset of the roller rotation track, that is, the offset of the center point position of all tracks The mean ; Calculate the standard deviation of trajectory fluctuations, that is, the offset of the center point position of all trajectories Standard Deviation ; Substitute into the roller rotation consistency index calculation formula: ; represents the trajectory fluctuation influence term, represents the trajectory deviation influence term, Represents the combined term of positive and negative deviations, Indicates the roller rotation consistency index; ; ; ; , , are all preset non-zero trade-off coefficients, represents the preset benchmark variance, Indicates the preset deviation threshold.

[0012] In a preferred embodiment, the logic for obtaining the roller appearance damage index is: Obtain and calculate the ratio of the damaged area detected in the ith running image to the total surface area of ​​the roller to obtain the damaged area ratio ; Obtain and calculate the ratio of the crack density detected in the i-th running image to the preset reference crack density to obtain the crack density ratio ; Get the wear depth detected in the i-th running image , and then calculate the wear depth effect : ; The proportion of damaged area , crack density ratio , Wear depth influence Perform weighted summation to calculate the roller appearance damage index: ; n represents the total number of running images, w1, w2, w3 are all preset non-zero scale coefficients, and the sum of w1, w2, w3 is one, Indicates the appearance damage index of the roller.

[0013] In a preferred embodiment, the sampling frequency optimization module optimizes the sampling frequency by: ; represents the basic sampling frequency, Represents the optimized sampling frequency.

[0014] Technical effects and advantages of the present invention: Through image acquisition and image analysis, the present invention can obtain multi-dimensional information on the rotation consistency and appearance damage of the roller in real time, comprehensively reflect the operating status of the roller, and adopt fuzzy logic algorithm to achieve accurate evaluation of the operating status of the roller, effectively reducing misjudgment and missed judgment. The present invention can capture early signs of failure during the operation of the roller, such as abnormal fluctuations in rotation speed or increased surface damage, thereby achieving early warning of failure. By dynamically adjusting the sampling frequency, the system can quickly respond to abnormal conditions and ensure timely detection of potential failures.

[0015] The present invention monitors through non-contact image acquisition, does not require the installation of a large number of sensors, reduces the installation and maintenance costs of the equipment, avoids downtime and maintenance costs caused by the expansion of faults through early fault detection and accurate status assessment, and extends the service life of the roller and conveyor system. The present invention is suitable for harsh industrial environments such as high temperature, high humidity, and high dust, has strong environmental adaptability, and can work stably under a variety of complex working conditions. The non-contact monitoring method avoids the danger of on-site inspections and improves the safety and reliability of the system.

[0016] The present invention combines image recognition, feature extraction and fuzzy logic evaluation to realize intelligent monitoring and decision support of roller operation status. It can automatically optimize sampling frequency and detection strategy according to different operation status, and further improve monitoring efficiency. Through accurate monitoring and early warning of roller operation status, the downtime of conveyor failure is effectively reduced, and the overall operation efficiency and productivity of conveyor system are improved. By optimizing sampling frequency and data processing method, the collection and storage requirements of invalid data are reduced, and energy consumption is reduced, which is in line with the development trend of green intelligence. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to facilitate understanding by those skilled in the art, the present invention is further described below in conjunction with the accompanying drawings; Figure 1 The schematic diagram of a conveyor roller operating status detection system based on image analysis in the present invention. DETAILED DESCRIPTION

[0018] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. 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.

[0019] Reference Figure 1 The following embodiments are obtained: Example 1

[0020] A conveyor roller running state detection system based on image analysis comprises an image acquisition module, an image recognition module, a feature extraction module and a state evaluation module; the image acquisition module, the image recognition module, the feature extraction module and the state evaluation module are communicatively connected with each other.

[0021] The image acquisition module is used to collect the running images of the conveyor rollers in real time; the image acquisition module is the input end of the system and is used to collect image data of the conveyor rollers in real time during their operation. Through high-definition camera equipment, the running status of the rollers is captured, including the rotation trajectory, appearance surface and overall movement, providing the original data basis for subsequent image analysis and feature extraction. Through continuous sampling, complete running status information is obtained to ensure that the collected images have sufficient resolution and clarity to support subsequent analysis, adapt to complex industrial environments such as high dust and high humidity, and provide stable data.

[0022] The image recognition module is used to perform preliminary analysis on the operating images of the conveyor rollers collected by the image acquisition module, and identify whether there are early signs of operating failures during the operation of the conveyor rollers; the image recognition module is responsible for preliminary processing and analysis of the original images obtained by the image acquisition module, and quickly identify whether there are early signs of operating failures. Preliminary processing refers to obtaining processed images that are conducive to subsequent analysis, calculation, and data acquisition through preset processing strategies in the existing technology, which will not be elaborated here. When the image recognition module discovers potential faults at an early stage, it will promptly trigger subsequent feature extraction and status evaluation to improve safety.

[0023] The feature extraction module is used to extract features from the conveyor roller operation image when there are early signs of operation failure during the operation of the conveyor roller, and then obtain feature group 1 and feature group 2; when the image recognition module detects signs of failure, the feature extraction module is responsible for in-depth processing of the roller operation image and extracting key feature parameters (such as rotation trajectory stability, surface damage, etc.). These features are divided into feature group 1 and feature group 2 for subsequent status evaluation, extracting core features that can directly reflect the roller operation status from the image, and classifying the features according to different dimensions (such as rotation status and appearance status) to form structured data that is easy to evaluate, providing a scientific and quantitative evaluation basis for the status evaluation module.

[0024] The status assessment module is used to assess the operating status of conveyor rollers based on feature group one and feature group two. The status assessment module uses an assessment algorithm based on feature group one and feature group two to conduct a comprehensive analysis of the operating status of the rollers to determine whether they are in a high-quality operating state or are at risk of low-quality operation.

[0025] Extracting features from the conveyor roller running image and then obtaining feature group 1 and feature group 2 means: extracting the roller rotation trajectory feature data and appearance damage feature data, and then obtaining feature group 1 composed of the roller rotation trajectory feature data and feature group 2 composed of appearance damage feature data. The state assessment module is used to assess the conveyor roller running state based on feature group 1 and feature group 2, which means: The roller rotation consistency index is generated based on feature group one, and the roller appearance damage index is generated based on feature group two. Then, the roller operation status is divided into a high-quality operation status or a low-quality operation status through fuzzy logic.

[0026] In the operating status evaluation based on the roller rotation consistency index and the roller appearance damage index, fuzzy logic uses membership functions and fuzzy rules to comprehensively analyze the two and output the operating status of the roller, dividing it into high-quality operating status or low-quality operating status.

[0027] Input variables: Roller rotation consistency index (input one), which reflects the smoothness of roller rotation, and the value range is 0 to 1. The larger the value, the smoother the rotation. Roller appearance damage index (input two), which reflects the degree of damage on the roller surface, and the value range is 0 to 1. The larger the value, the more serious the damage.

[0028] Membership function design: Fuzzy membership functions are designed for the input variables respectively, and they are divided into three fuzzy sets: Fuzzy set of roller rotation consistency index: High consistency: The membership degree gradually increases as the consistency index approaches 1. Medium consistency: The membership degree is the largest near the middle value (about 0.5). Low consistency: The membership degree increases as the consistency index approaches 0.

[0029] Fuzzy set of roller appearance damage index: Slight damage: The membership gradually increases as the damage index approaches 0. Moderate damage: The membership is the largest near the middle value (about 0.5). Severe damage: The membership gradually increases as the damage index approaches 1.

[0030] Based on the membership of the two input variables, a fuzzy rule table is formulated to infer the operating status of the roller. An example of the rule is shown in Table 1 below: Table 1. Fuzzy rule design table Roller rotation consistency index Roller appearance damage index Roller running status High consistency Minor damage High quality operation status High consistency Moderate damage High quality operation status High consistency Severe injury Low quality operation status Moderate consistency Minor damage High quality operation status Moderate consistency Moderate damage Low quality operation status Moderate consistency Severe injury Low quality operation status Low consistency Minor damage Low quality operation status Low consistency Moderate damage Low quality operation status Low consistency Severe injury Low quality operation status Fuzzy reasoning process: Fuzzy processing: According to the membership function, the specific values ​​of the input variables (roller rotation consistency index and roller appearance damage index) are mapped to the corresponding fuzzy membership.

[0031] Fuzzy rule matching: According to the fuzzy membership of the input variables, the fuzzy rules that meet the conditions are activated, and the fuzzy set of operating status is generated according to the weight of the rules.

[0032] Membership synthesis: The membership of all activated rules is synthesized to generate the fuzzy result of the roller running status.

[0033] Defuzzification processing: Defuzzify the fuzzy operating state through weighted average method or centroid method, and output the final operating state result.

[0034] Output results: According to the results of fuzzy reasoning, the system outputs the following operating status: High-quality operation status: indicates that the roller rotates smoothly, has slight appearance damage, and has no obvious hidden dangers of failure.

[0035] Low-quality operation status: Indicates that the roller rotates unsteadily, has serious appearance damage, has obvious faults, and needs maintenance or replacement.

[0036] The present invention also includes a sampling frequency optimization module, which is used to optimize the sampling frequency according to the basic sampling frequency, the roller rotation consistency index and the roller appearance damage index under low-quality operating conditions.

[0037] The preliminary analysis operation refers to: obtaining the roller rotation speed data corresponding to each time point in the conveyor within a fixed time window, obtaining the roller rotation speed data time series, and then calculating the absolute value of the corresponding data difference between two adjacent time points, and then calculating the average and standard deviation of all absolute values. Identifying whether there are early signs of operating failures during the operation of the conveyor rollers refers to: comparing the average of all absolute values ​​with a preset threshold value one, and comparing the standard deviation of all absolute values ​​with a preset threshold value two. If the average of all absolute values ​​is less than or equal to the preset threshold value one and the standard deviation of all absolute values ​​is less than or equal to the preset threshold value two, a normal signal is generated. If the average of all absolute values ​​is less than or equal to the preset threshold value one and the standard deviation of all absolute values ​​is less than or equal to the preset threshold value two, an abnormal signal is generated. When the abnormal signal is generated, it indicates that there are early signs of operating failures during the operation of the conveyor rollers.

[0038] By obtaining the conveyor roller rotation speed data in a fixed time window and calculating the average and standard deviation of the absolute value of the difference between adjacent time points in the time series, a preliminary quantitative analysis of the roller's operating status can be performed. By performing statistical characteristic analysis on small fluctuations in the rotation speed, it is possible to quickly identify whether the operating status is stable. Focusing on the average and standard deviation directly reflects the overall change trend and fluctuation amplitude of the roller operation, which is suitable for rapid detection of early faults. Using simple mathematical operations (average and standard deviation) to obtain preliminary judgment results is convenient for real-time analysis at industrial sites and meets the requirements of continuous operation of conveyors.

[0039] By comparing the mean and standard deviation with the preset thresholds, it is possible to accurately determine whether there are early signs of roller operation failure. Early detection of signs of failure allows maintenance measures to be taken before the problem expands or causes damage to the overall operation of the conveyor, reducing downtime and repair costs. Setting clear thresholds 1 (for the mean) and 2 (for the standard deviation) allows the system to distinguish between normal and abnormal signals with quantitative standards, avoiding the subjectivity of human judgment. Once an abnormal signal is generated, the subsequent feature extraction and status assessment modules can be immediately triggered to ensure that the abnormality is analyzed and processed in a timely manner.

[0040] The logic for obtaining the roller rotation consistency index is: Get the rotation trajectory point sequence P of the roller: ; is the center point position of the trajectory of the nth collected running image; Calculate the offset of the center point of the track in each running image: ; Indicates the offset of the center point of the trajectory of the i-th collected running image. is the center point position of the trajectory of the i-th collected running image, is the average position of the center point of the trajectory of all collected running images; Based on the offset of the center point of the trajectory, the variance of the trajectory fluctuation is calculated, that is, the offset of the center point of all trajectories Variance , quantify the instability of the rotation trajectory, the higher the value, the worse the stability; calculate the overall offset of the roller rotation trajectory, that is, the offset of the position of the center point of all trajectories The mean , judge whether the roller has a tendency to rotate off-center. The larger the value, the more obvious the tendency to rotate off-center. Calculate the standard deviation of the trajectory fluctuation, that is, the offset of the center point position of all trajectories. Standard Deviation ; Substitute into the roller rotation consistency index calculation formula: ; represents the trajectory fluctuation influence term, represents the trajectory deviation influence term, Represents the combined term of positive and negative deviations, Indicates the roller rotation consistency index; ; ; ; , , are all preset non-zero trade-off coefficients, Indicates the preset benchmark variance, which represents the preset variance benchmark value during normal operation. Indicates the preset deviation threshold, which represents the preset maximum allowable deviation value. Introducing nonlinear fluctuations and increasing offset sensitivity, The absolute difference between fluctuation and deviation is measured. In the calculation formula of roller rotation consistency index, the sine function and square sum calculation are used to enhance the response ability to small deviation and fluctuation abnormality, and the weight coefficient is adjusted according to the load, speed or operating environment. , , , adapt to complex working conditions, and consider the impact of trajectory fluctuations and offsets on consistency to ensure the accuracy of the judgment results. Dynamic adjustment to prevent excessive impact on roller rotation consistency index under high fluctuation or high deviation conditions.

[0041] The roller appearance damage index (DAI) measures the degree of damage on the roller surface, taking into account the damage area, crack density, and wear depth. The calculation process uses multiple characteristic parameters and nonlinear transformations to fully reflect the damage of the roller appearance. The logic for obtaining the roller appearance damage index is: Obtain and calculate the ratio of the damaged area detected in the ith running image to the total surface area of ​​the roller to obtain the damaged area ratio The larger the value, the more serious the loss; the proportion of damaged area It can be calculated by the ratio of the number of pixels in the damaged area to the total number of pixels; Obtain and calculate the ratio of the crack density detected in the i-th running image to the preset reference crack density to obtain the crack density ratio , the larger the value, the more serious the loss; the crack density ratio It can be calculated by the ratio of the number of pixels with cracks identified to the total number of pixels; Get the wear depth detected in the i-th running image , and then calculate the wear depth effect : , Indicates the preset wear depth threshold. The wear depth affects The larger the value, the more serious the loss; the wear depth The edge detection and depth estimation algorithms in the prior art are used for calculation, specifically, intensity mapping of grayscale gradients. The specific process is briefly described as follows: The grayscale image of the roller surface is obtained through high-definition camera equipment. The change of grayscale value reflects the texture and morphological characteristics of the roller surface. The grayscale image is processed by edge detection algorithm to extract the edge information of the wear area on the roller surface: the edge is usually manifested as an area with a large change in grayscale value. Edge detection can distinguish the wear boundary from the normal surface area, providing basic data for depth estimation.

[0042] Based on the intensity mapping of grayscale gradient, the depth of the wear area is estimated by using the gradient change of grayscale value: Gradient calculation: Analyze the grayscale value change rate of the wear area. The larger the grayscale gradient, the greater the wear depth. Intensity mapping: Map the grayscale gradient value to the corresponding depth value to form a depth distribution map of the wear area.

[0043] The following indicators are obtained by statistically analyzing the depth distribution of the wear area: Maximum wear depth: the deepest point in the wear area, used to determine whether the critical value for maintenance or replacement has been reached. Average wear depth: the average of the overall wear degree, used to evaluate the long-term use of the roller. Since the analysis is generally based on data in a short time window, the wear depth The maximum wear depth is used as an indicator.

[0044] It should be noted that: multiple high-definition images of the roller operation process are collected, edge detection and texture analysis are performed on the images, and the extracted features include the damage area detected by the ith operation image, the crack density detected by the ith operation image, and the wear depth detected by the ith operation image. ; The proportion of damaged area , crack density ratio , Wear depth influence Perform weighted summation to calculate the roller appearance damage index: ; n represents the total number of running images, w1, w2, w3 are all preset non-zero scale coefficients, and the sum of w1, w2, w3 is one, It represents the appearance damage index of the roller, taking into account the damage area, crack density and wear depth, and comprehensively assessing the degree of appearance damage. Through nonlinear processing of the wear depth, the response capability to severe damage is enhanced. The proportional coefficients w1, w2 and w3 can be adjusted according to the working conditions to adapt to the sensitive requirements of different damage characteristics. Through the weighted average of multiple image data, the interference of a single image abnormality on the calculation of the roller appearance damage index is reduced.

[0045] Sampling frequency optimization module optimizes the sampling frequency by: ; represents the basic sampling frequency, Indicates the optimized sampling frequency. The sampling frequency in high-quality operation state can be restored to the basic sampling frequency. As the basic adjustment item, : The worse the rotation consistency, the larger the increment; : The more serious the appearance damage, the larger the denominator, which reduces the sensitivity of the adjustment and avoids over-adjustment. is the nonlinear damage response term, : When CRI is low (unsteady), the impact of damage index DAI is more significant. : Logarithmic processing is performed on DAI to amplify the impact of small damage and smooth the adjustment of large damage. The combination of basic adjustment terms and nonlinear damage response terms ensures that the different weights of CRI and DAI on the sampling frequency are considered at the same time, enhancing the dynamic adaptability of the formula.

[0046] When the roller is in abnormal operation or potential failure occurs, increasing the sampling frequency can capture data changes in more detail, generate high-resolution time series, and provide an accurate data basis for subsequent analysis. High sampling frequency can avoid missing important fault precursor information due to long intervals, thereby achieving accurate monitoring; adjusting the sampling frequency according to the roller operation status, appropriately reducing the sampling frequency during normal operation, and increasing the sampling frequency during abnormal operation, so as to make system resource allocation more efficient. By optimizing the sampling frequency, excessive collection of redundant data in a stable state can be avoided, the workload of data processing can be reduced, and the detection efficiency can be improved. When there is no obvious sign of failure, the sampling frequency can be reduced to reduce the amount of data, save storage and computing resources, and reduce the system operation cost. Dynamic optimization of the sampling frequency can reduce the high-intensity operation time of the sensor equipment, reduce its wear and tear, and increase its service life. When the roller operation status changes to low-quality operation, the automatic increase in the sampling frequency enables the system to quickly capture state changes and provide real-time data support for early fault warning and maintenance. Combined with the evaluation results of the roller operation status, the sampling frequency is adjusted intelligently to reflect the adaptive ability of the system.

[0047] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0048] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0049] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0050] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0051] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A conveyor roller operating status detection system based on image analysis, characterized in that: It includes image acquisition module, image recognition module, feature extraction module and state assessment module; The image acquisition module is used to collect the running images of conveyor rollers in real time; The image recognition module is used to perform preliminary analysis on the running images of the conveyor rollers acquired by the image acquisition module, and identify whether there are early signs of running failures during the running of the conveyor rollers; The feature extraction module is used to extract features from the conveyor roller operation image when there are early signs of operation failure during the operation of the conveyor roller, and then obtain feature group 1 and feature group 2; The state evaluation module is used to evaluate the operating state of the conveyor roller based on feature group one and feature group two.

2. The conveyor roller operating status detection system based on image analysis according to claim 1 is characterized in that: Feature extraction is performed on the conveyor roller running image, and then feature group one and feature group two are obtained, which means: extracting the roller rotation trajectory feature data and appearance damage feature data, and then obtaining feature group one composed of the roller rotation trajectory feature data and feature group two composed of the appearance damage feature data.

3. The conveyor roller operating status detection system based on image analysis according to claim 2 is characterized in that: The status assessment module is used to assess the operating status of conveyor rollers based on feature group 1 and feature group 2. The roller rotation consistency index is generated based on feature group one, and the roller appearance damage index is generated based on feature group two. Then, the roller operation status is divided into a high-quality operation status or a low-quality operation status through fuzzy logic.

4. The conveyor roller operating status detection system based on image analysis according to claim 3 is characterized in that: A sampling frequency optimization module is also included for optimizing the sampling frequency according to a basic sampling frequency, a roller rotation consistency index, and a roller appearance damage index under low-quality operating conditions.

5. The conveyor roller running status detection system based on image analysis according to claim 4 is characterized in that: The initial analysis operation refers to: In a fixed time window, the roller rotation speed data corresponding to each time point in the conveyor is obtained to obtain the roller rotation speed data time series, and then the absolute value of the corresponding data difference between two adjacent time points is calculated, and then the average value and standard deviation of all absolute values ​​are calculated.

6. The conveyor roller operating status detection system based on image analysis according to claim 5 is characterized in that: Identifying early signs of conveyor roller failure during operation means: The average value of all absolute values ​​is compared with the preset threshold value one, and the standard deviation of all absolute values ​​is compared with the preset threshold value two. If the average value of all absolute values ​​is less than or equal to the preset threshold value one and the standard deviation of all absolute values ​​is less than or equal to the preset threshold value two, a normal signal is generated. If the average value of all absolute values ​​is less than or equal to the preset threshold value one and the standard deviation of all absolute values ​​is less than or equal to the preset threshold value two, an abnormal signal is generated. When the abnormal signal is generated, it indicates that there are early signs of operating failure during the operation of the conveyor rollers.

7. The conveyor roller operating status detection system based on image analysis according to claim 5 is characterized in that: The logic for obtaining the roller rotation consistency index is: Get the rotation trajectory point sequence P of the roller: ; is the center point position of the trajectory of the nth collected running image; Calculate the offset of the center point of the track in each running image: ; Indicates the offset of the center point of the trajectory of the i-th collected running image. is the center point position of the trajectory of the i-th collected running image, is the average position of the center point of the trajectory of all collected running images; Based on the offset of the center point of the trajectory, the variance of the trajectory fluctuation is calculated, that is, the offset of the center point of all trajectories Variance ; Calculate the overall offset of the roller rotation track, that is, the offset of the center point position of all tracks The mean ; Calculate the standard deviation of trajectory fluctuations, that is, the offset of the center point position of all trajectories Standard Deviation ; Substitute into the roller rotation consistency index calculation formula: ; represents the trajectory fluctuation influence term, represents the trajectory deviation influence term, Represents the combined term of positive and negative deviations, Indicates the roller rotation consistency index; ； ； ； , , are all preset non-zero trade-off coefficients, represents the preset benchmark variance, Indicates the preset deviation threshold.

8. The conveyor roller running status detection system based on image analysis according to claim 7 is characterized in that: The logic for obtaining the roller appearance damage index is: Obtain and calculate the ratio of the damaged area detected in the ith running image to the total surface area of ​​the roller to obtain the damaged area ratio ; Obtain and calculate the ratio of the crack density detected in the i-th running image to the preset reference crack density to obtain the crack density ratio ; Get the wear depth detected in the i-th running image , and then calculate the wear depth effect : ; Indicates a preset wear depth threshold; The proportion of damaged area , crack density ratio , Wear depth influence Perform weighted summation to calculate the roller appearance damage index: ; n represents the total number of running images, w1, w2, w3 are all preset non-zero scale coefficients, and the sum of w1, w2, w3 is one, Indicates the appearance damage index of the roller.

9. The conveyor roller running status detection system based on image analysis according to claim 8 is characterized in that: Sampling frequency optimization module optimizes the sampling frequency by: ; represents the basic sampling frequency, Represents the optimized sampling frequency.

Images