Excavator bucket tooth detection method and system and electronic equipment

Abstract

The invention provides an excavator bucket tooth detection method and system and electronic equipment, and relates to the field of engineering machinery operation and maintaining.According to the method, data fusion is conducted on three types of multi-mode sensors, so that bucket tooth service life detection is achieved, the detection limitation of a single sensor is relieved, and the detection efficiency is improved. And the relation curve obtained through data fusion processing can be used for carrying out preventive detection and maintenance on the bucket tooth, and the reliability of the bucket tooth under the complex working condition is improved.

Description

Technical Field

[0001] The present invention relates to the field of operation and maintenance of construction machinery, and in particular, to a method, a system and an electronic device for detecting the bucket teeth of an excavator. Background Art

[0002] In the detection of the bucket teeth of an excavator, the traditional method mainly relies on manual visual inspection, which is prone to missed detection and cannot be monitored in real time; although relevant sensors are used for automatic detection in the prior art, most of them use a single type of sensor, but are greatly interfered by environmental noise and have a high false alarm rate. In addition, in the prior art, the life of the bucket teeth cannot be predicted through historical data, which is prone to unplanned shutdowns and it is difficult to achieve preventive detection and maintenance, thus affecting the reliability of the bucket tooth detection. Summary of the Invention

[0003] In view of this, the purpose of the present invention is to provide a method, a system and an electronic device for detecting the bucket teeth of an excavator. This method fuses the data of three types of multi-modal sensors to detect the life of the bucket teeth, alleviates the detection limitations of a single sensor, and can use the relationship curve obtained by data fusion processing to perform preventive detection and maintenance on the bucket teeth, improving the reliability of the bucket teeth under complex working conditions.

[0004] In a first aspect, an embodiment of the present invention provides a method for detecting the bucket teeth of an excavator, the method comprising: A data acquisition step: respectively acquiring visual data, vibration data and stress data corresponding to the bucket teeth in the excavator based on a visual sensor, a vibration sensor and a stress sensor provided on the excavator; A feature extraction step: respectively extracting visual features, vibration features and stress features corresponding to the bucket teeth according to the spatial feature information of the visual data, the time-frequency feature information of the vibration data, and the energy spectrum feature information corresponding to the stress data; A feature fusion step: using the feature fusion result of the visual features, vibration features and stress features to determine a relationship curve corresponding to the bucket teeth including the stress distribution entropy value and the life value; A detection output step: obtaining the stress distribution entropy value and the life value corresponding to the number of working cycles of the bucket teeth according to the relationship curve, and using the stress distribution entropy value and the life value to determine the detection result of the bucket teeth.

[0005] Optionally, the data acquisition step includes: Obtaining a visual unit installed on the side of the bucket of the excavator, and using the visual sensor provided in the visual unit to acquire visual data corresponding to the bucket teeth at a first sampling rate; Obtaining a vibration sensing array installed at the bucket of the excavator, and using the vibration sensor provided in the vibration sensing array to acquire vibration data corresponding to the bucket teeth at a second sampling rate; Obtain the stress monitoring unit installed in the root of the bucket tooth, and use the strain sensor in the stress monitoring unit to collect the stress data corresponding to the bucket tooth at the third sampling rate.

[0006] Optionally, the feature extraction step includes: Calculate the spatial feature vector of the visual data, and use the temporal change feature corresponding to the spatial feature vector to determine the spatial feature information corresponding to the visual data; extract the defect area feature included in the bucket tooth according to the spatial feature information, and determine the visual feature based on the defect area feature; Determine the time-frequency feature of the impact signal included in the vibration data, use the time-frequency feature of the impact signal to determine the time-frequency feature information corresponding to the vibration data, and use the time-frequency feature information and the time-domain data corresponding to the kurtosis and pulse factor in the vibration data to determine the vibration feature; Use wavelet packet decomposition to calculate the energy spectrum feature corresponding to the stress data, use the energy spectrum feature to determine the energy spectrum feature information, and use the energy spectrum feature information and the statistics corresponding to the mean square deviation value and peak factor in the stress data to determine the stress feature.

[0007] Optionally, the feature fusion step includes: Respectively determine the first weight coefficient, the second weight coefficient and the third weight coefficient based on the abnormal data included in the visual feature, the vibration feature and the stress feature; Use the first weight coefficient, the second weight coefficient and the third weight coefficient to perform cascade splicing and weighted fusion calculation on the visual feature, the vibration feature and the stress feature to obtain the feature fusion result corresponding to the bucket tooth; Determine the stress distribution entropy value corresponding to the bucket tooth according to the feature fusion result, and determine the defect probability and defect type corresponding to the bucket tooth based on the stress distribution entropy value; Use the defect probability and defect type to determine the life value of the bucket tooth, and use the relationship curve between the life value and the stress distribution entropy value.

[0008] Optionally, respectively determining the first weight coefficient, the second weight coefficient and the third weight coefficient based on the abnormal data included in the visual feature, the vibration feature and the stress feature includes: Use the visual feature to determine the defect area feature included in the bucket tooth, and determine the first weight coefficient corresponding to the visual feature based on the defect area feature; Use the vibration feature to determine the abnormal impact feature included in the bucket tooth, and determine the second weight coefficient corresponding to the vibration feature based on the abnormal impact feature; Use the stress feature to determine the stress offset feature included in the bucket tooth, and determine the third weight coefficient corresponding to the stress feature based on the stress offset feature.

[0009] Optionally, the detection output step includes: Real-time obtain the number of working cycles of the bucket tooth; Using the three-stage piecewise curve included in the relationship curve, determine the stress distribution entropy value corresponding to the number of working cycles, and use the two-stage piecewise curve included in the relationship curve to determine the life value corresponding to the number of working cycles; Determine the health risk level of the bucket tooth according to the stress distribution entropy value and the life value, and use the health risk level to determine the detection result.

[0010] Optionally, the three-stage piecewise curve is: ; Wherein, is the normalized entropy value corresponding to the stress distribution entropy value; is the number of working cycles; is the initial growth coefficient; is the exponential growth factor; is the platform constant, used to characterize the entropy value benchmark of the second stage; is the fine-tuning coefficient; is the decay exponent; is the critical number of working cycles; is the quadratic growth coefficient; is the quadratic growth exponent; The two-stage piecewise curve is: ; Wherein, is the remaining life ratio corresponding to the life value; is the linear damage rate coefficient; is the critical remaining life ratio; is the exponential decay coefficient; is the power-law damage acceleration coefficient; is the power exponent.

[0011] Optionally, after the data acquisition step, the method further includes a preprocessing step: After downsampling and histogram equalization of the visual data using the preset first parameter, obtain the first time-series data corresponding to the visual data; After filtering the vibration data using the second parameter corresponding to the preset filter, obtain the second time-series data corresponding to the vibration data; After performing Kalman filtering on the stress data using the preset third parameter, obtain the third time-series data corresponding to the stress data; Perform complementary calculation on the sampling rates corresponding to the first time-series data, the second time-series data, and the third time-series data, and use the complementary calculation results to update the visual data, the vibration data, and the stress data respectively.

[0012] In a second aspect, the present invention provides an excavator bucket tooth detection system, and the system includes: A data acquisition module, configured to respectively acquire visual data, vibration data, and stress data corresponding to the bucket teeth in the excavator based on a visual sensor, a vibration sensor, and a stress sensor set on the excavator; A feature extraction module, configured to respectively extract visual features, vibration features, and stress features corresponding to the bucket teeth according to the spatial feature information of the visual data, the time-frequency feature information of the vibration data, and the energy spectrum feature information corresponding to the stress data; A feature fusion module, configured to determine a relationship curve including the stress distribution entropy value and the life value corresponding to the bucket teeth by using the feature fusion result of the visual features, the vibration features, and the stress features; A detection output module, configured to obtain the stress distribution entropy value and the life value corresponding to the working cycle number of the bucket teeth according to the relationship curve, and determine the detection result of the bucket teeth by using the stress distribution entropy value and the life value.

[0013] In a third aspect, an embodiment of the present invention further provides an electronic device, including a processor and a memory, where the memory stores computer executable instructions that can be executed by the processor, and the processor executes the computer executable instructions to implement the steps of the excavator bucket tooth detection method provided in the first aspect.

[0014] In a fourth aspect, an embodiment of the present invention further provides a storage medium, where the storage medium stores computer executable instructions, and when the computer executable instructions are called and executed by a processor, the computer executable instructions cause the processor to implement the steps of the excavator bucket tooth detection method provided in the first aspect.

[0015] An excavator bucket tooth detection method, system, and electronic device provided by an embodiment of the present invention, in the process of detecting the excavator bucket teeth, first acquire visual data, vibration data, and stress data corresponding to the bucket teeth in the excavator respectively based on a visual sensor, a vibration sensor, and a stress sensor set on the excavator, so as to complete data acquisition; then respectively extract visual features, vibration features, and stress features corresponding to the bucket teeth according to the spatial feature information of the visual data, the time-frequency feature information of the vibration data, and the energy spectrum feature information corresponding to the stress data, so as to complete the feature extraction process; subsequently determine a relationship curve including the stress distribution entropy value and the life value corresponding to the bucket teeth by using the feature fusion result of the visual features, the vibration features, and the stress features, to implement the feature fusion process; finally obtain the stress distribution entropy value and the life value corresponding to the working cycle number of the bucket teeth according to the relationship curve, and determine the detection result of the bucket teeth by using the stress distribution entropy value and the life value, so as to obtain the detection output process of the excavator bucket teeth. This method performs data fusion on three types of multi-modal sensors to realize the detection of the bucket tooth life, alleviates the detection limitations of a single sensor, and can perform preventive detection and maintenance on the bucket teeth by using the relationship curve obtained by data fusion processing, improving the reliability of the bucket teeth under complex working conditions.

[0016] Other features and advantages of the present invention will be set forth in the following description, and in part will be obvious from the description, or may be learned by practice of the present invention. The objectives and other advantages of the present invention are realized and attained by the structure particularly pointed out in the specification, claims as well as the drawings.

[0017] In order to make the above objects, features and advantages of the present invention more comprehensible, the following specific embodiments are given by way of example and in conjunction with the accompanying drawings, and are described in detail as follows. Brief Description of the Drawings

[0018] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required for the description of the specific embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0019] Figure 1 A flowchart of a method for detecting an excavator bucket tooth provided by an embodiment of the present invention; Figure 2 A flowchart of step S101 in a method for detecting an excavator bucket tooth provided by an embodiment of the present invention; Figure 3 A flowchart of step S102 in a method for detecting an excavator bucket tooth provided by an embodiment of the present invention; Figure 4 A flowchart of step S103 in a method for detecting an excavator bucket tooth provided by an embodiment of the present invention; Figure 5 A flowchart of step S401 in a method for detecting an excavator bucket tooth provided by an embodiment of the present invention; Figure 6 A flowchart of step S104 in a method for detecting an excavator bucket tooth provided by an embodiment of the present invention; Figure 7 A flowchart of a method for detecting an excavator bucket tooth provided by an embodiment of the present invention when a preprocessing step is included; Figure 8 A flowchart of another method for detecting an excavator bucket tooth provided by an embodiment of the present invention; Figure 9 A schematic diagram of sensor deployment in a method for detecting an excavator bucket tooth provided by an embodiment of the present invention; Figure 10 A relationship curve graph of the stress distribution entropy value and the bucket tooth life in a method for detecting an excavator bucket tooth provided by an embodiment of the present invention; Figure 11 A schematic structural diagram of a system for detecting an excavator bucket tooth provided by an embodiment of the present invention; Figure 12 A structural schematic diagram of an electronic device provided by an embodiment of the present invention.

[0020] Icon: 1110 - Data acquisition module; 1120 - Feature extraction module; 1130 - Feature fusion module; 1140 - Detection output module; 1 - First vision sensor; 2 - Second vision sensor; 3 - Vibration sensor; 4 - Stress sensor; 5 - Edge computing module; 6 - Decision-making and early warning module; 101 - Processor; 102 - Memory; 103 - Bus; 104 - Communication interface. Specific implementation manners

[0021] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below in conjunction with the embodiments. Apparently, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0022] In the detection of the bucket teeth of an excavator, the traditional method mainly relies on manual visual inspection, which is prone to missed detection and cannot be monitored in real time; although relevant sensors are used for automated detection in the prior art, most use a single type of sensor, but are greatly affected by environmental noise and have a high false alarm rate (such as misidentifying a rock impact as a bucket tooth defect). In addition, in the prior art, the bucket tooth life cannot be predicted through historical data during the bucket tooth detection, which is prone to unplanned shutdowns and it is difficult to achieve preventive detection and maintenance, thus affecting the reliability of the bucket tooth detection. Based on this, the embodiments of the present invention provide a method, system, and electronic device for detecting the bucket teeth of an excavator. This method fuses the data of three types of multi-modal sensors to detect the bucket tooth life, alleviates the detection limitations of a single sensor, and can perform preventive detection and maintenance on the bucket teeth by using the relationship curve obtained through data fusion processing, improving the reliability of the bucket teeth under complex working conditions.

[0023] To facilitate the understanding of this embodiment, first, a method for detecting the bucket teeth of an excavator disclosed in the embodiments of the present invention will be introduced in detail, as Figure 1 shown, this method includes: Data acquisition step S101: Based on the vision sensor, vibration sensor, and stress sensor set on the excavator, respectively collect the vision data, vibration data, and stress data corresponding to the bucket teeth in the excavator.

[0024] A high-resolution industrial camera is reasonably set near the bucket tooth of the excavator as a visual sensor. The camera should have a sufficient frame rate to ensure that it can capture the dynamic changes of the bucket tooth during operation. When collecting visual data, not only the overall appearance image of the bucket tooth should be obtained, but also key parts of the bucket tooth, such as the tooth tip and tooth body, should be focused on for shooting. To improve the accuracy of the data, multiple visual sensors at different angles can be set to shoot the bucket tooth from multiple perspectives, so as to comprehensively obtain visual information such as the shape, wear degree, and surface cracks of the bucket tooth. At the same time, the collected image data should be preprocessed, including operations such as denoising, grayscale conversion, and contrast enhancement, to improve the effect of subsequent feature extraction.

[0025] For the vibration sensor, a high-precision accelerometer is selected as the vibration sensor and firmly installed at a suitable position on the bucket tooth, such as the middle or root of the tooth body, to ensure that the vibration condition of the bucket tooth during operation can be accurately measured. The vibration data collected by the vibration sensor contains information such as the vibration frequency and amplitude of the bucket tooth under different working conditions. During the data collection process, classification collection should be carried out according to the working modes of the excavator (such as excavation, lifting, slewing, etc.) to facilitate the subsequent analysis of the vibration characteristics of the bucket tooth under different working conditions. At the same time, to ensure the stability and reliability of the data, multiple vibration sensors can be set for redundant collection, and the collected data can be filtered to remove high-frequency noise and interference signals.

[0026] For the stress sensor, a strain gauge or piezoresistive stress sensor is used and pasted or installed at the stress concentration parts of the bucket tooth, such as the transition area between the tooth tip and the tooth body, and the tooth root. The stress sensor can measure the magnitude and direction of the stress borne by the bucket tooth during operation in real time. When collecting stress data, the change laws of stress under different working conditions, such as the influence of excavation force and impact force on the stress of the bucket tooth, should be considered. At the same time, to improve the accuracy of the stress data, the sensor can be calibrated regularly, and a suitable signal amplification and acquisition circuit can be used to ensure that the stress data can be accurately obtained.

[0027] Feature extraction step S102: Extract the visual features, vibration features, and stress features corresponding to the bucket tooth according to the spatial feature information of the visual data, the time-frequency feature information of the vibration data, and the energy spectrum feature information corresponding to the stress data respectively.

[0028] The visual feature extraction process is based on the collected visual data, and uses computer vision technology to extract the spatial feature information of the bucket teeth. First, an edge detection algorithm (such as the Canny algorithm) is used to extract the contour edges of the bucket teeth, thereby determining the shape features of the bucket teeth. Then, through image segmentation technology, the bucket teeth are separated from the background, and the surface texture features of the bucket teeth are further analyzed, such as the texture changes in the worn area and the morphology of cracks. In addition, deep learning methods, such as convolutional neural networks (CNNs), can also be used to train the images of the bucket teeth to automatically extract more advanced visual features, such as the wear degree of the bucket teeth, the length and width of the cracks, etc.

[0029] When extracting vibration features, for the collected vibration data, time-frequency analysis methods (such as short-time Fourier transform, wavelet transform, etc.) are used to convert the vibration signals in the time domain into feature information in the time-frequency domain. By analyzing the time-frequency diagram, the vibration energy distribution of the bucket teeth at different frequencies can be obtained, thereby extracting vibration features. For example, features such as the vibration amplitude and the change of frequency components of the bucket teeth at a specific working frequency are determined. At the same time, the statistical features of the vibration data, such as mean, variance, kurtosis, etc., can also be calculated to reflect the stability and abnormal conditions of the vibration.

[0030] According to the collected stress data, its energy spectrum feature information is calculated. First, the Fourier transform is performed on the stress data to convert it into a frequency-domain signal, and then the energy distribution at different frequencies is calculated. By analyzing the energy spectrum, the main frequency components and energy concentration regions of the stress data can be extracted, thereby obtaining stress features. In addition, the correlation coefficient, cross-correlation function, etc. of the stress data can also be calculated to analyze the relationship and variation law between the stresses at different parts.

[0031] Feature fusion step S103: Use the feature fusion results of visual features, vibration features, and stress features to determine the relationship curve corresponding to the bucket teeth that includes the stress distribution entropy value and the life value.

[0032] The feature fusion process needs to adopt appropriate feature fusion methods to fuse the extracted visual features, vibration features, and stress features. Common feature fusion methods include weighted fusion, serial fusion, parallel fusion, etc. When selecting a fusion method, it should be reasonably set according to the characteristics and importance of different features. For example, for visual features that are closely related to the wear degree of the bucket teeth, a higher weight can be assigned; for vibration features and stress features that reflect the working state of the bucket teeth, the weight can be allocated according to their influence degree on the life of the bucket teeth.

[0033] It is worth mentioning that although the visual data, vibration data, and stress data do not have the same dimension, when performing data fusion, they can be fused by combining their corresponding data values with relevant weight values. For example, specific weight values can be set according to different data types, and then the weight values are multiplied by the data of that type to obtain the corresponding product results. Furthermore, by setting different weight values, the product results can be made to fall within a specific range, thus solving the problem of difficult fusion caused by different dimensions.

[0034] After obtaining the fusion features through feature fusion, a relationship model between the fusion features and the stress distribution entropy value and life value of the bucket teeth can be established using machine learning or data mining methods. Regression analysis methods (such as linear regression, non-linear regression, etc.) or neural network methods (such as multi-layer perceptron, radial basis function network, etc.) can be used to train the fusion features, thereby obtaining a relationship curve including the stress distribution entropy value and the life value. When establishing the relationship model, a large amount of historical data should be used for training and verification to ensure the accuracy and reliability of the model.

[0035] Detection output step S104: Obtain the stress distribution entropy value and life value corresponding to the working cycle times of the bucket teeth according to the relationship curve, and use the stress distribution entropy value and life value to determine the detection result of the bucket teeth.

[0036] According to the working records or real-time monitoring system of the excavator, obtain the working cycle times of the bucket teeth. Then, according to the relationship curve, find the stress distribution entropy value and life value corresponding to the working cycle times. During the search process, interpolation methods (such as linear interpolation, spline interpolation, etc.) can be used to process the relationship curve to improve the accuracy of the obtained data.

[0037] According to the obtained stress distribution entropy value and life value, combined with the pre-set thresholds and standards, determine the detection result of the bucket teeth. For example, if the stress distribution entropy value exceeds a certain threshold, it indicates that the stress distribution of the bucket teeth is uneven and there may be potential failure risks; if the life value is less than a certain set value, it means that the remaining life of the bucket teeth is short and needs to be replaced in time. At the same time, the detection results can also be output in the form of intuitive charts or reports to provide a reference basis for the maintenance and management of the excavator.

[0038] Optionally, the data acquisition step S101, as Figure 2 shown, includes: Step S201, obtain the visual unit installed on the side of the bucket of the excavator, and use the visual sensor set in the visual unit to collect the visual data corresponding to the bucket teeth at the first sampling rate.

[0039] The vision unit is an industrial-grade image acquisition unit, which has high resolution, high frame rate, and good environmental adaptability. For example, in a dusty excavation environment, a vision unit with dust and moisture protection functions needs to be selected. The vision unit is installed in the cab area, with the shooting angle facing the bucket directly, and the field of view can completely cover the bucket teeth, minimizing occlusion as much as possible. At the same time, the installation angle needs to be precisely adjusted so that the collected vision data can truly reflect the actual situation of the bucket teeth. The setting of the first sampling rate needs to comprehensively consider the working speed of the excavator and the frequency of changes in the bucket tooth state. If the excavator works at a high speed and the bucket tooth state changes rapidly, a higher sampling rate is required to ensure that the subtle changes in the bucket teeth can be captured. The vision sensor continuously collects the vision data of the bucket teeth at the set first sampling rate. During the collection process, the integrity and accuracy of the data need to be ensured to prevent frame loss or image blurring.

[0040] Step S202: Obtain the vibration sensing array installed at the bucket of the excavator, and use the vibration sensors set in the vibration sensing array to collect the vibration data corresponding to the bucket teeth at the second sampling rate.

[0041] Select a vibration sensing array with high sensitivity and a wide frequency response range for the vibration sensing array. For example, piezoelectric vibration sensors have good dynamic response characteristics and can accurately measure the vibration of the bucket teeth under different working conditions. When installing the vibration sensing array at the bucket, the vibration transmission path and vibration characteristics of the bucket teeth need to be considered. Generally, the vibration sensors can be installed at the connection part between the bucket and the bucket teeth, so that the vibration of the bucket teeth can be measured more directly. At the same time, it is necessary to ensure that the sensors are firmly installed to avoid affecting the measurement results due to loosening.

[0042] The setting of the second sampling rate should be based on the frequency range of the bucket tooth vibration. When the bucket tooth vibration frequency is high, a higher sampling rate is required to ensure that the details of the vibration signal can be accurately captured. The vibration sensors collect the vibration data of the bucket teeth at the set second sampling rate. During the collection process, the data needs to be monitored in real time to ensure the stability and reliability of the data.

[0043] Step S203: Obtain the stress monitoring unit embedded at the root of the bucket teeth, and use the strain sensors in the stress monitoring unit to collect the stress data corresponding to the bucket teeth at the third sampling rate.

[0044] Select a stress monitoring unit with high precision and good stability for the stress monitoring unit. For example, foil strain gauges are a commonly used stress measurement element, which has advantages such as high sensitivity and good linearity. When embedding the stress monitoring unit at the root of the bucket teeth, it is necessary to ensure that the strain sensors are in close contact with the bucket teeth, and the bucket teeth should not be damaged during the installation process. Before installation, the surface of the root of the bucket teeth needs to be treated to ensure that the surface is flat and clean; after installation, sealing treatment should be carried out to prevent moisture and dust from entering and affecting the measurement results.

[0045] The setting of the third sampling rate should be based on the frequency and amplitude of the stress change of the bucket tooth. During the working process of the bucket tooth, the stress change is relatively complex, and the sampling rate needs to be reasonably set according to the actual situation. The strain sensor collects the stress data of the bucket tooth at the set third sampling rate. During the collection process, the data should be recorded and stored in real time for subsequent analysis and processing.

[0046] Optionally, the feature extraction step S102, as Figure 3 shown, includes: Step S301, calculating the spatial feature vector of the visual data, determining the spatial feature information corresponding to the visual data by using the temporal change feature corresponding to the spatial feature vector; extracting the defect area feature included in the bucket tooth according to the spatial feature information, and determining the visual feature based on the defect area feature.

[0047] When calculating the spatial feature vector of the visual data, various image features can be selected. For example, for edge features, using the Canny edge detection algorithm can accurately identify the edge of the bucket tooth and outline the general contour of the bucket tooth. For corner features, the Harris corner detection can be used to find the corners with drastic gray level changes in the bucket tooth image, and these corners are often the key positions where the shape of the bucket tooth changes. In terms of texture features, the gray level co-occurrence matrix can quantify the roughness or fineness of the surface texture of the bucket tooth and provide rich information for the spatial feature vector.

[0048] The spatial feature vector changes with time. By using the sliding window method, the feature vector sequence is divided at a certain time interval. Calculate the Euclidean distance between the vectors in adjacent windows. A large distance means that the spatial features of the bucket tooth change significantly. Then, by synthesizing these change situations, such as the concentrated area of feature changes and the change frequency, the spatial feature information of the visual data is determined.

[0049] When extracting the defect area feature, the threshold segmentation algorithm (such as the Otsu algorithm) is used to separate the bucket tooth from the background. For the suspected defect area, morphological operations (such as dilation and erosion) are combined to remove noise interference. Then, features such as the area, perimeter, and shape factor of the defect area are extracted. The area can reflect the size of the defect, the perimeter reflects the length of the defect boundary, and the shape factor can distinguish different-shaped defects.

[0050] During the process of determining the visual feature, the extracted defect area feature and the spatial feature information are combined to form a complete visual feature. For example, the area of the defect area is spliced with features such as the number of corners and texture roughness in the spatial feature vector to form a feature vector that can comprehensively describe the visual condition of the bucket tooth.

[0051] In step S302, determine the time-frequency characteristics of the impact signals contained in the vibration data, use the time-frequency characteristics of the impact signals to determine the time-frequency characteristic information corresponding to the vibration data, and use the time-frequency characteristic information and the time-domain data corresponding to the kurtosis and impulse factor in the vibration data to determine the vibration characteristics.

[0052] When determining the time-frequency characteristics of the impact signals, use the short-time Fourier transform (STFT) or wavelet transform to analyze the vibration data. STFT can expand the signal in the time domain and frequency domain, but the time and frequency resolutions are fixed; the wavelet transform has the characteristic of multi-resolution and is more suitable for analyzing impact signals. Through these transforms, obtain the time-frequency diagram, and extract the characteristics such as the occurrence time, frequency components, and energy distribution of the impact signals from the diagram.

[0053] When determining the time-frequency characteristic information, an energy threshold can be set. When the energy of the transform coefficient exceeds the threshold, it is determined as an impact signal. At the same time, combine conditions such as the frequency range and duration of the impact signal for screening and confirmation. Statistically calculate the information such as the occurrence frequency and average energy of the impact signals to obtain the time-frequency characteristic information of the vibration data.

[0054] When determining the vibration characteristics, it is necessary to calculate the kurtosis and impulse factor of the vibration data. The kurtosis reflects the degree to which the signal peak deviates from the normal distribution, and the impulse factor reflects the ratio relationship between the signal peak and the effective value. Combine the time-frequency characteristic information with the time-domain data such as kurtosis and impulse factor to form the final vibration characteristics.

[0055] In step S303, use wavelet packet decomposition to calculate the energy spectrum characteristics corresponding to the stress data, use the energy spectrum characteristics to determine the energy spectrum characteristic information, and use the energy spectrum characteristic information and the statistics corresponding to the mean square deviation value and peak factor in the stress data to determine the stress characteristics.

[0056] When calculating the energy spectrum characteristics, use wavelet packet decomposition for the stress data and decompose it into different frequency bands. Select appropriate wavelet bases and decomposition levels, for example, determine according to the frequency range of the stress signal and the requirements of analysis accuracy. Calculate the energy in each frequency band, draw the energy spectrum with the frequency band as the horizontal axis and the energy as the vertical axis to obtain the energy spectrum characteristics of the stress data.

[0057] The process of determining the energy spectrum characteristic information is to extract key information from the energy spectrum, such as the frequency band where the energy is concentrated, the standard deviation of the energy distribution, etc. The frequency band where the energy is concentrated may correspond to a specific stress mode of the bucket teeth, and the standard deviation of the energy distribution reflects the degree of dispersion of the energy in each frequency band.

[0058] When determining the stress characteristics, it is necessary to calculate the mean square deviation value and peak factor of the stress data. The mean square deviation reflects the fluctuation magnitude of the stress data, and the peak factor reflects the stress peak situation. Combine the energy spectrum characteristic information with the statistics such as the mean square deviation value and peak factor to form the stress characteristics for comprehensively evaluating the stress state of the bucket teeth.

[0059] Optionally, the feature fusion step S103 is as follows Figure 4 shown and includes: Step S401: Determine the first weight coefficient, the second weight coefficient, and the third weight coefficient respectively based on the abnormal data included in the visual feature, the vibration feature, and the stress feature.

[0060] When determining the first, second, and third weight coefficients, the judgment of abnormal data is crucial. For the abnormal data in the visual feature, such as the feature of the bucket tooth defect area exceeding the normal range or the spatial feature information changing violently, it can be identified by setting a threshold. When the area of the defect area in the visual feature suddenly increases to exceed a certain proportion of the historical mean, it is considered abnormal. By counting the proportion of abnormal data in the visual feature and combining expert experience and historical data, the first weight coefficient is determined. If the abnormal proportion is high, it indicates that the visual feature is more important for judging the state of the bucket tooth, and the first weight coefficient can be appropriately increased.

[0061] In terms of the vibration feature, if the time-frequency feature or time-domain data (kurtosis, impulse factor) of the impact signal shows abnormalities, such as the frequency of the impact signal suddenly increasing or the kurtosis value far exceeding the normal range, the abnormality is also judged by the threshold. The second weight coefficient is determined according to the proportion of abnormal data in the vibration feature. If the vibration feature is abnormally frequent, it indicates that the vibration condition has a great impact on the state of the bucket tooth, and the second weight coefficient should be increased.

[0062] In the stress feature, if the energy spectrum feature information or statistic (mean square deviation value, peak factor) shows abnormalities, such as a sudden large increase in the energy of a certain frequency band in the energy spectrum or the mean square deviation value exceeding the normal fluctuation range, the abnormality is judged based on this. The third weight coefficient is determined according to the proportion of abnormal data in the stress feature. If the stress feature is abnormally obvious, it indicates that the stress has a significant impact on the state of the bucket tooth, and the third weight coefficient needs to be increased.

[0063] Step S402: Use the first weight coefficient, the second weight coefficient, and the third weight coefficient to perform cascade splicing and weighted fusion calculation on the visual feature, the vibration feature, and the stress feature to obtain the corresponding feature fusion result of the bucket tooth.

[0064] When performing cascade splicing and weighted fusion calculation, first cascade splice the visual feature, the vibration feature, and the stress feature in sequence to form a longer feature vector. Then, use the first, second, and third weight coefficients to weight the corresponding part features after splicing respectively. Specifically, cascade splicing + attention weighted fusion can be adopted, as shown in the following formula: F fusion =αF vision ⊕βF vibration ⊕γF stress ; where α, β, and γ are the first weight coefficient, the second weight coefficient, and the third weight coefficient respectively; Fvision , F vibration , F stress are visual features, vibration features, and stress features respectively, and F fusion is the result of feature fusion.

[0065] Step S403: Determine the stress distribution entropy value corresponding to the bucket tooth according to the feature fusion result, and determine the defect probability and defect type corresponding to the bucket tooth based on the stress distribution entropy value.

[0066] When determining the stress distribution entropy value according to the feature fusion result, a machine learning model such as a support vector machine (SVM) or a neural network can be used. Taking the feature fusion result as the input, the model outputs the stress distribution entropy value after training. The stress distribution entropy value reflects the degree of disorder of the stress distribution of the bucket tooth. The larger the entropy value, the more uneven the stress distribution.

[0067] To determine the defect probability and defect type of the bucket tooth based on the stress distribution entropy value, a mapping relationship table can be established. Through a large amount of experimental data and historical records, the defect probability and common defect types corresponding to different stress distribution entropy value intervals are statistically analyzed.

[0068] Step S404: Determine the life value of the bucket tooth using the defect probability and defect type, and determine the relationship curve between the life value and the stress distribution entropy value.

[0069] When determining the life value of the bucket tooth using the defect probability and defect type, an empirical formula or a life prediction model can be used. For different defect types, corresponding life decay models are established. For example, for defects of the wear type, the remaining available working hours of the bucket tooth can be calculated as the life value according to the wear degree (quantified by the defect probability) and the wear rate. For defects of the crack type, considering the crack propagation rate and the critical crack size, the remaining number of working cycles before the bucket tooth fails is predicted as the life value.

[0070] After determining the life value, multiple groups of corresponding stress distribution entropy value and life value data are plotted in a coordinate system. Through a fitting algorithm (such as linear fitting, polynomial fitting), a curve that best represents the distribution law of these data points is found, which is the relationship curve between the life value and the stress distribution entropy value. This curve can be used to predict the remaining life of the bucket tooth according to the real-time stress distribution entropy value of the bucket tooth in the future.

[0071] Optionally, step S401 of determining the first weight coefficient, the second weight coefficient, and the third weight coefficient respectively based on the abnormal data included in the visual features, vibration features, and stress features, as Figure 5 shown, includes: Step S501: Use the visual features to determine the defect area features included in the bucket tooth, and determine the first weight coefficient corresponding to the visual features based on the defect area features; Step S502: Determine the abnormal impact characteristics included in the bucket teeth using the vibration characteristics, and determine the second weight coefficient corresponding to the vibration characteristics based on the abnormal impact characteristics; Step S503: Determine the stress offset characteristics included in the bucket teeth using the stress characteristics, and determine the third weight coefficient corresponding to the stress characteristics based on the stress offset characteristics.

[0072] In a specific scenario, when determining the weight coefficient, it can be achieved by constructing a cascaded neural network architecture. For example, in the visual branch of this network architecture, EfficientNet-B4 is used to extract the characteristics of the defective area (cracks, missing area) to determine the first weight coefficient; the vibration branch uses 1D-CNN to analyze the time-frequency characteristics of the impact signal and identify abnormal impact patterns to determine the second weight coefficient; the stress branch uses the LSTM network to predict the stress distribution offset trend to determine the third weight coefficient, which can specifically correspond to the above-mentioned α, β, γ.

[0073] Optionally, the detection output step S104, as Figure 6 shown, includes: Step S601: Obtain the working cycle count of the bucket teeth in real time.

[0074] The real-time acquisition of the working cycle count of the bucket teeth can be achieved by means of sensors installed at key parts of the bucket teeth. For example, a pressure sensor is installed in the bucket teeth. When the bucket teeth perform actions such as excavation, lifting, and unloading, the pressure of the pressure sensor will change periodically. By analyzing the period of the pressure change, the working cycle count can be accurately counted.

[0075] Step S602: Determine the stress distribution entropy value corresponding to the working cycle count using the three-stage piecewise curve included in the relationship curve, and determine the life value corresponding to the working cycle count using the two-stage piecewise curve included in the relationship curve.

[0076] The three-stage piecewise curve and the two-stage piecewise curve in the relationship curve are obtained based on a large amount of experimental data and actual working conditions analysis. Optionally, the three-stage piecewise curve is: ; where, is the normalized entropy value corresponding to the stress distribution entropy value; is the working cycle count; is the initial growth coefficient; is the exponential growth factor; is the platform constant, used to characterize the entropy value benchmark of the second stage; is the fine-tuning coefficient; is the decay exponent; is the critical working cycle count; is the secondary growth coefficient; is the secondary growth exponent; The formula indicates that the initial applied stress is relatively dispersed and unevenly distributed, the entropy value rises rapidly, microcracks appear on the surface or inside of the material in the initial stage, and the stress is rapidly released at the defects, resulting in the entropy value (degree of disorder) increasing in a power-law manner. n is the number of cycles (such as the number of periodic actions like mechanical loads and temperature changes); a1 is the initial growth coefficient, reflecting the sensitivity of the material's initial defects or microstructure to the increase in entropy value; b1 is the exponential growth factor (b1 > 1), controlling the accelerating rising rate of the entropy value.

[0077] The formula indicates that the stress distribution and wear enter a stable period, the entropy value slows down, and the wear enters a dynamic equilibrium stage. The newly generated cracks and the self-repair of the material (such as the formation of an oxide layer) reach a temporary equilibrium, and the entropy value tends to be stable. If there is , it reflects small fluctuations (such as changes in environmental temperature and humidity). c2 is the platform constant, representing the reference level of the entropy value in the stable wear stage; a2 is the fine-tuning coefficient, usually a2 ≪ a1, used to describe the influence of weak environmental disturbances (such as random vibrations) on the entropy value. b2 is the decay exponent (b2 ≈ 0), when b2 → 0, n b2 → 1, and the formula degenerates into a constant c2.

[0078] The formula indicates that the crack or damage expansion leads to stress distribution, and the crack changes from stable expansion to unstable expansion (such as reaching a critical size), the stress distribution is extremely chaotic, and the entropy value rises in a power-law manner again. The hysteresis effect (the accelerated failure is triggered only after the damage accumulates to a threshold) is reflected by (n - n crit ). Specifically, n crit is the critical number of cycles, corresponding to the starting point of the rapid decline of the remaining life of the material (such as the threshold of the sudden change in the crack growth rate); a3 is the secondary growth coefficient, reflecting the sensitivity of the entropy value in the crack expansion stage; b3 is the secondary growth exponent (b3 > 1), controlling the secondary accelerated rise of the entropy value.

[0079] The two-stage piecewise curve is: ; Among them, is the proportion of the remaining life corresponding to the life value; is the linear damage rate coefficient; is the critical proportion of the remaining life; is the exponential decay coefficient; is the power-law damage acceleration coefficient; is the power exponent.

[0080] The formula represents the first stage, where the damage is approximately linear with the number of cycles. In the initial stage, internal damage in the material (such as microcracks and plastic deformation) accumulates at an approximately constant rate, and the remaining life decreases linearly with the number of cycles. For example, if k1 = 0.0005, the remaining life decreases by 50% after every 1000 cycles, that is, L(1000) = 1 - 0.0005×1000 = 0.5. Specifically, L(n) is the remaining life ratio, representing the percentage of the remaining life of the material or component to the initial life. For example, L(n) = 0.6 means the remaining life is 60% of the initial; n is the number of cycles (the number of times of periodic actions such as mechanical loads and temperature cycles); k1 is the linear damage rate coefficient (k1 > 0 and the value is small), controlling the linear decline rate of the remaining life.

[0081] The formula represents the second stage, where the damage accelerates (exponential or power-law). When the damage accumulates to a critical value (such as a crack reaching a critical size), the material enters the instability failure stage. If k2 = 0.002, then in stage 2, for every additional 100 cycles, the remaining life decays to e -0.002×100 ≈81.9% of the original. If k3 = 0.001 and b4 = 2, then in stage 2, for every additional 100 cycles, the remaining life decreases by 0.001×1002 = 10%. Specifically, L crit is the critical remaining life ratio, that is, the remaining life at the end of stage 1 (for example, L crit ≈40% = 0.4); n crit is the critical number of cycles, that is, the end point of stage 1 (1200 cycles in the figure); k2 is the exponential decay coefficient (k2 > 0), controlling the exponential decline rate of the remaining life; k3 is the power-law damage acceleration coefficient (k3 > 0), controlling the power-law decline amplitude of the remaining life; b4 is the power exponent (b4 > 1), determining the steepness of the curve (the larger the value, the faster the decline).

[0082] According to the working cycle number obtained in real time, search in the three-stage piecewise curve and the two-stage piecewise curve. If the working cycle number exactly corresponds to a known point on the curve, the corresponding stress distribution entropy value and life value can be directly read; if it is between two known points, the linear interpolation method can be used to estimate the corresponding values.

[0083] Step S603, determine the health risk level of the bucket teeth according to the stress distribution entropy value and the life value, and use the health risk level to determine the test result.

[0084] Pre-set different stress distribution entropy value intervals and life value intervals, and associate them with corresponding health risk levels. For example, when the stress distribution entropy value is low and the life value is high, it is determined that the bucket tooth is in a healthy state and the health risk level is low; when the stress distribution entropy value increases and the life value drops to a certain extent, it indicates that the bucket tooth may have minor damage and the health risk level is medium; if the stress distribution entropy value is very high and the life value is extremely low, it means that the bucket tooth faces the risk of serious damage and the health risk level is high.

[0085] According to the determined health risk level, further clarify the detection result. For a low risk level, it can be recommended to continue using the bucket tooth normally, but regular monitoring is required; for a medium risk level, it is recommended to conduct a detailed inspection of the bucket tooth to evaluate whether repair or replacement of some components is needed; for a high risk level, the use of the bucket tooth should be stopped immediately and replaced in a timely manner to avoid safety accidents.

[0086] Optionally, as Figure 7 shown, after the data acquisition step, the method further includes a preprocessing step: Step S701, after downsampling and histogram equalization of the visual data using a preset first parameter, obtain the first time-series data corresponding to the visual data; Step S702, after filtering the vibration data using the second parameter corresponding to a preset filter, obtain the second time-series data corresponding to the vibration data; Step S703, after performing Kalman filtering on the stress data using a preset third parameter, obtain the third time-series data corresponding to the stress data; Step S704, perform a filling calculation on the sampling rates corresponding to the first time-series data, the second time-series data, and the third time-series data, and use the filling calculation results to update the visual data, the vibration data, and the stress data respectively.

[0087] In the preprocessing process, the visual data uses adaptive histogram equalization (CLAHE) to enhance low-light images, segments the bucket tooth profile through Mask R-CNN, and finally realizes downsampling (224×224) and histogram equalization of the visual data. The vibration data extracts the impact characteristics in the 0.5 - 5 kHz frequency band through wavelet transform and filters out the engine noise, specifically using a Butterworth filter for denoising (cutoff frequency 20 - 500 Hz). The stress data eliminates the dynamic load fluctuations based on Kalman filtering and calculates the stress distribution entropy value of the bucket tooth. Finally, a filling calculation is performed on the sampling rates corresponding to the first time-series data, the second time-series data, and the third time-series data. Specifically, the dynamic time warping algorithm DTW is used for the time-series data of the three sensors to compensate for the sampling rate differences.

[0088] As Figure 8The flowchart of another detection method for excavator bucket teeth is as follows. During the feature extraction process, for visual features, ResNet-18 is used to extract spatial features (2048 dimensions), and 3D-CNN is used to extract temporal change features. For vibration features, an LSTM network is used to extract time-frequency features (128 dimensions), and time-domain indicators such as kurtosis and impulse factor are calculated synchronously. For stress features, wavelet packet decomposition (5 layers) is used to obtain energy spectrum features, and statistics such as RMS and peak factor are calculated. In the feature fusion process, cascade splicing + attention weighted fusion is adopted, and it is calculated by the following formula: F fusion =αF vision ⊕βF vibration ⊕γF stress ; where α, β, and γ are the first weight coefficient, the second weight coefficient, and the third weight coefficient respectively; F vision 、F vibration 、F stress are visual features, vibration features, and stress features respectively, and F fusion is the feature fusion result.

[0089] As Figure 9 shown in the schematic diagram of sensor deployment, the multimodal sensing module therein includes the following three modules: The high-dynamic vision unit includes a first vision sensor 1 and a second vision sensor 2, which are anti-shake wide-angle cameras (resolution ≥ 4K, frame rate 60fps) installed on the side of the bucket, integrated with a polarization filter to eliminate reflection interference, and are used to collect bucket tooth images in real time; The vibration sensing array includes a vibration sensor 3. Specifically, triaxial acceleration sensors (sampling rate ≥ 10kHz) are deployed at key positions of the bucket to capture the impact vibration spectrum features of the bucket teeth; The stress monitoring unit includes a stress sensor 4: a micro strain gauge (accuracy ±0.1%) is embedded at the root of the bucket tooth to measure the working load distribution in real time.

[0090] The edge computing module 5 is equipped with an NVIDIA Jetson AGX Orin chip, with a lightweight multi-task deep learning model built-in, supporting real-time inference. At the same time, a 5G communication module is integrated to achieve data cloud synchronization and remote control.

[0091] The decision-making and warning module 6 displays the defect position and severity through an HMI (human-machine interface), triggering an audible and visual alarm.

[0092] Through the attention mechanism (Transformer-based Fusion), the three-modal features are dynamically weighted and fused to output the following results: defect probability (0-1), defect type (crack, fracture, wear), and remaining life prediction (based on the Weibull distribution model). On this basis, asFigure 10 The relationship curve between the entropy value of the stress distribution shown and the bucket tooth life. There are a total of 3 parameters in this coordinate system. The X-axis is the number of working cycles of the bucket tooth, and there are 2 Y-axes, which are the normalized entropy value of the stress and the percentage of the remaining life of the bucket tooth respectively. The normalized entropy value of the stress has 3 stages, namely rapid increase, stable, and rapid increase again. The remaining life has 2 stages, slow decline and rapid decline. The starting point of the rapid decline of the remaining life coincides with the rapid increase again of the normalized entropy value of the stress.

[0093] The curve characteristics are analyzed as follows: In the initial stage (0 - 900 cycles), the entropy value rises rapidly (0.5 → 0.7). In the initial running-in period, the stress distribution is uneven and locally concentrated. In the stable stage (900 - 1800 cycles), the entropy value fluctuates at 0.7 ± 0.05. After using for a period of time, the stress release reaches equilibrium. In the failure stage (>1800 cycles), the entropy value suddenly increases to more than 0.95. Microcracks become macrocracks, resulting in stress concentration. When the remaining life is <20%, it needs to be replaced immediately.

[0094] The hardware implementation example is as follows: The camera is installed on the side bracket of the bucket, with an IP67 protection level, and an internal heating film is used to prevent the mirror from fogging; the vibration sensor is fixed through a magnetic base, supporting quick disassembly and assembly; the edge computing module integrates a power management unit, supporting 24V DC power supply and solar backup power.

[0095] The algorithm implementation example is as follows: The training data is obtained by collecting 100,000 bucket tooth images under different lighting and dust conditions, and the defect types and positions are labeled; the model uses knowledge distillation technology to compress the model (ResNet - 101) to EfficientNet - B4 that can run on edge devices; online learning is carried out through the federated learning framework, and the model is continuously optimized using the on-site data of multiple excavators.

[0096] From the excavator bucket tooth detection method mentioned in the above embodiments, it can be seen that this method fuses the data of three types of multi-modal sensors to realize the detection of the bucket tooth life, alleviates the detection limitations of a single sensor, and can perform preventive detection and maintenance on the bucket tooth using the relationship curve obtained by data fusion processing, improving the reliability of the bucket tooth under complex working conditions.

[0097] Corresponding to the excavator bucket tooth detection method provided in the foregoing embodiments, the embodiment of the present invention provides an excavator bucket tooth detection system, as Figure 11 shown, this system includes: A data acquisition module 1110, configured to respectively acquire visual data, vibration data, and stress data corresponding to the bucket teeth in the excavator based on a visual sensor, a vibration sensor, and a stress sensor set on the excavator; A feature extraction module 1120 is configured to extract visual features, vibration features, and stress features corresponding to the bucket teeth according to the spatial feature information of the visual data, the time-frequency feature information of the vibration data, and the energy spectrum feature information corresponding to the stress data, respectively. A feature fusion module 1130 is configured to determine a relationship curve including the stress distribution entropy value and the life value corresponding to the bucket teeth by using the feature fusion results of the visual features, vibration features, and stress features. A detection output module 1140 is configured to obtain the stress distribution entropy value and the life value corresponding to the number of working cycles of the bucket teeth according to the relationship curve, and determine the detection result of the bucket teeth by using the stress distribution entropy value and the life value.

[0098] As can be seen from the excavator bucket tooth detection system mentioned in the above embodiments, the system performs data fusion on three types of multi-modal sensors to realize the detection of the bucket tooth life, alleviates the detection limitations of a single sensor, and can perform preventive detection and maintenance on the bucket teeth by using the relationship curve obtained through data fusion processing, improving the reliability of the bucket teeth under complex working conditions.

[0099] For the excavator bucket tooth detection system provided by the embodiments of the present invention, its implementation principle and the technical effects generated are the same as those of the foregoing embodiments of the excavator bucket tooth detection method. For the sake of brief description, for the parts not mentioned in the system embodiments, reference may be made to the corresponding content in the foregoing embodiments of the excavator bucket tooth detection method.

[0100] This embodiment further provides an electronic device. The structural schematic diagram of the electronic device is as Figure 12 shown. A processing unit is provided in the excavator device. The processing unit includes a processor 101 and a memory 102. Among them, the memory 102 is used to store one or more computer instructions, and the one or more computer instructions are executed by the processor to implement the steps of the foregoing excavator bucket tooth detection method.

[0101] Figure 12 The processing unit in the electronic device shown further includes a bus 103 and a communication interface 104. The processor 101, the communication interface 104, and the memory 102 are connected through the bus 103.

[0102] Among them, the memory 102 may include a high-speed random access memory (RAM, Random Access Memory), and may also include a non-volatile memory, such as at least one disk memory. The bus 103 may be an ISA bus, a PCI bus, or an EISA bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. For the sake of convenience of representation, Figure 12 only a bidirectional arrow is used in the figure, but it does not mean that there is only one bus or one type of bus.

[0103] The communication interface 104 is used to connect to at least one user terminal and other network units through a network interface, and send the encapsulated IPv4 packet or IPv4 packet to the user terminal through the network interface.

[0104] The processor 101 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the integrated logic circuit in hardware or instructions in software form in the processor 101. The above-mentioned processor 101 may be a general-purpose processor, including a central processing unit (CPU for short), a network processor (NP for short), etc.; it may also be a digital signal processor (DSP for short), an application specific integrated circuit (ASIC for short), a field-programmable gate array (FPGA for short), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components. The storage medium is located in the memory 102, and the processor 101 reads the information in the memory 102 and combines its hardware to complete the steps of the method in the foregoing embodiments.

[0105] If the above functions are implemented in the form of software function units and sold or used as independent products, they can be stored in a non-volatile computer-readable storage medium executable by a processor. Finally, it should be noted that: the above embodiments are only specific implementation manners of the present invention, used to illustrate the technical solutions of the present invention, rather than limiting it. The protection scope of the present invention is not limited thereto. Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: any person skilled in the art within the technical scope disclosed by the present invention can still modify the technical solutions recorded in the foregoing embodiments, or easily think of changes, or perform equivalent replacements on some of the technical features; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be covered by the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. A method for detecting the bucket teeth of an excavator, characterized in that, The method includes: A data acquisition step: acquiring visual data, vibration data, and stress data corresponding to the bucket teeth in the excavator based on a visual sensor, a vibration sensor, and a stress sensor provided in the excavator respectively; A feature extraction step: extracting visual features, vibration features, and stress features corresponding to the bucket teeth according to the spatial feature information of the visual data, the time-frequency feature information of the vibration data, and the energy spectrum feature information corresponding to the stress data respectively; A feature fusion step: determining a relationship curve corresponding to the bucket teeth and including a stress distribution entropy value and a life value by using the feature fusion result of the visual features, the vibration features, and the stress features; A detection output step: obtaining the stress distribution entropy value and the life value corresponding to the number of working cycles of the bucket teeth according to the relationship curve, and determining the detection result of the bucket teeth by using the stress distribution entropy value and the life value.

2. The excavator bucket tooth detection method according to claim 1, wherein The data acquisition step includes: Obtaining a visual unit installed on the side of the bucket in the excavator, and acquiring the visual data corresponding to the bucket teeth by using the visual sensor provided in the visual unit at a first sampling rate; Obtaining a vibration sensing array installed at the bucket in the excavator, and acquiring the vibration data corresponding to the bucket teeth by using the vibration sensor provided in the vibration sensing array at a second sampling rate; Obtaining a stress monitoring unit embedded at the root of the bucket teeth, and acquiring the stress data corresponding to the bucket teeth by using the strain sensor in the stress monitoring unit at a third sampling rate.

3. The excavator bucket tooth detection method according to claim 1, characterized in that, The feature extraction step includes: Calculating a spatial feature vector of the visual data, determining the spatial feature information corresponding to the visual data by using the time series change feature corresponding to the spatial feature vector; extracting the defect area feature included in the bucket teeth according to the spatial feature information, and determining the visual feature based on the defect area feature; Determining the time-frequency feature of the impact signal included in the vibration data, determining the time-frequency feature information corresponding to the vibration data by using the time-frequency feature of the impact signal, and determining the vibration feature by using the time-frequency feature information and the time domain data corresponding to the kurtosis and impulse factor in the vibration data; Calculating the energy spectrum feature corresponding to the stress data by using wavelet packet decomposition, determining the energy spectrum feature information by using the energy spectrum feature, and determining the stress feature by using the energy spectrum feature information and the statistical quantities corresponding to the mean square deviation value and peak factor in the stress data.

4. The excavator bucket tooth detection method according to claim 1, characterized in that, The feature fusion step includes: Respectively determining a first weight coefficient, a second weight coefficient, and a third weight coefficient based on the abnormal data included in the visual features, the vibration features, and the stress features; Performing cascade splicing and weighted fusion calculation on the visual features, the vibration features, and the stress features by using the first weight coefficient, the second weight coefficient, and the third weight coefficient to obtain the feature fusion result corresponding to the bucket teeth; Determining the stress distribution entropy value corresponding to the bucket teeth according to the feature fusion result, and determining the defect probability and defect type corresponding to the bucket teeth based on the stress distribution entropy value. Determine the service life value of the bucket tooth using the defect probability and the defect type, and determine the relationship curve using the service life value and the stress distribution entropy value.

5. The excavator bucket tooth detection method according to claim 4, characterized in that, Based on the abnormal data included in the visual feature, the vibration feature, and the stress feature, determine the first weight coefficient, the second weight coefficient, and the third weight coefficient respectively, including: Use the visual feature to determine the defect area feature included in the bucket tooth, and determine the first weight coefficient corresponding to the visual feature based on the defect area feature; Use the vibration feature to determine the abnormal impact feature included in the bucket tooth, and determine the second weight coefficient corresponding to the vibration feature based on the abnormal impact feature; Use the stress feature to determine the stress offset feature included in the bucket tooth, and determine the third weight coefficient corresponding to the stress feature based on the stress offset feature.

6. The excavator bucket tooth detection method according to claim 1, characterized in that, The detection output step includes: Obtain the working cycle number of the bucket tooth in real time; Use the three-stage piecewise curve included in the relationship curve to determine the stress distribution entropy value corresponding to the working cycle number, and use the two-stage piecewise curve included in the relationship curve to determine the service life value corresponding to the working cycle number; Determine the health risk level of the bucket tooth according to the stress distribution entropy value and the service life value, and use the health risk level to determine the detection result.

7. The excavator bucket tooth detection method according to claim 6, characterized in that, The three-stage piecewise curve is: ； wherein, is the normalized entropy value corresponding to the stress distribution entropy value; is the number of working cycles; is the initial growth coefficient; is the exponential growth factor; is the platform constant, used to characterize the entropy value benchmark in the second stage; is the fine-tuning coefficient; is the decay exponent; is the critical number of working cycles; is the quadratic growth coefficient; is the quadratic growth exponent; The two-stage piecewise curve is: ； wherein, is the remaining life ratio corresponding to the life value; is the linear damage rate coefficient; is the critical remaining life ratio; is the exponential decay coefficient; is the power-law damage acceleration coefficient; is the power exponent.

8. The excavator bucket tooth detection method according to claim 1, wherein After the data acquisition step, the method further includes a preprocessing step: After downsampling and histogram equalization of the visual data using a preset first parameter, obtain the first time-series data corresponding to the visual data; After filtering the vibration data using the second parameter corresponding to a preset filter, obtain the second time-series data corresponding to the vibration data; After performing Kalman filtering on the stress data using a preset third parameter, obtain the third time-series data corresponding to the stress data; Perform a complementary calculation on the sampling rates corresponding to the first time-series data, the second time-series data, and the third time-series data, and use the complementary calculation result to update the visual data, the vibration data, and the stress data respectively.

9. An excavator bucket tooth detection system, characterized in that, The system includes: A data acquisition module for respectively collecting the visual data, the vibration data, and the stress data corresponding to the bucket tooth in the excavator based on a visual sensor, a vibration sensor, and a stress sensor set on the excavator; A feature extraction module for respectively extracting the visual feature, the vibration feature, and the stress feature corresponding to the bucket tooth according to the spatial feature information of the visual data, the time-frequency feature information of the vibration data, and the energy spectrum feature information corresponding to the stress data; A feature fusion module for determining the relationship curve including the stress distribution entropy value and the service life value corresponding to the bucket tooth using the feature fusion result of the visual feature, the vibration feature, and the stress feature; A detection output module for obtaining the stress distribution entropy value and the service life value corresponding to the working cycle number of the bucket tooth according to the relationship curve, and using the stress distribution entropy value and the service life value to determine the detection result of the bucket tooth.

10. An electronic device, characterized in that, It includes a processor and a memory, and the memory stores computer-executable instructions that can be executed by the processor. The processor executes the computer-executable instructions to implement the steps of the excavator bucket tooth detection method according to any one of claims 1 to 8.

Images