A method and system for full life-cycle health management of mining equipment
By collecting and analyzing point cloud data and operating parameters of mining 3D laser scanners in real time, the health status of the equipment is automatically assessed, solving the problem of lag in periodic verification, realizing precision management throughout the entire life cycle, and improving the availability and safety of the equipment.
Patent Information
- Application Number
- CN202611229411.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-13
- Publication Date
- 2026-11-17
AI Technical Summary
In the existing technology, the health management of mining 3D laser scanners relies on periodic offline calibration, which cannot monitor changes in equipment accuracy in real time. This leads to delayed fault detection, affects measurement reliability and safety, and frequent inspections increase maintenance costs.
By collecting point cloud data, laser emission power, scanning motor drive current, and equipment vibration data of the scanner in real time during the scanning task, the feature cross-sectional contour of the roadway is automatically extracted, the deviation is calculated, and compensation is made in combination with laser emission power and vibration amplitude to generate a full life cycle health index and achieve adaptive management.
It enables real-time accuracy assessment of mining 3D laser scanners, reduces monitoring blind spots, improves equipment availability and safety, and lowers maintenance costs.
Smart Images

Figure CN123118012A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of equipment maintenance, repair, operation and management technology, and in particular to a method and system for full life cycle health management of mining equipment. Background Technology
[0002] A mining 3D laser scanner is a measuring device that rapidly acquires high-precision 3D point cloud data of a mining scene through laser ranging. It is widely used in scenarios such as underground roadway cross-section measurement, goaf deformation monitoring, geological logging, and engineering acceptance. Its ranging and angle measurement accuracy directly determines the reliability of mining measurement results and the safety of engineering decisions. Currently, the health management of mining 3D laser scanners mainly relies on periodically sending the equipment to a metrology institution for offline verification, or manual calibration and comparison on-site using standard parts at fixed intervals. This involves measuring standard targets of known dimensions to verify whether the equipment's accuracy remains within the acceptable range.
[0003] However, this periodic offline verification method has significant shortcomings. During the interval between two calibrations, the equipment's accuracy is completely obscured. If a sudden accuracy failure occurs due to factors such as mechanical vibration, laser aging, or motor wear, on-site measurement personnel cannot detect it during the scanning process and can only passively confirm it during the next periodic verification. This results in a serious lag between fault detection and handling, with maintenance and repair work always lagging behind the occurrence of the fault. At this time, a large amount of point cloud data collected during the failure period may have deviated from the true values, posing an irreversible risk of misjudgment to roadway deformation analysis and mining engineering decisions. In addition, frequent inspections lead to long equipment downtime and high maintenance costs, making it difficult to meet the actual needs of continuous mine production for high availability of measurement equipment. Summary of the Invention
[0004] This application provides a method and system for full life-cycle health management of mining equipment, which upgrades the maintenance and operation management of mining scanning equipment from a passive maintenance mode that relies on periodic inspections to an adaptive management method based on online perception of multiple precision dimensions. This improves the technical problem of delayed anomaly detection caused by the lack of real-time assessment methods in the prior art.
[0005] This application discloses the following technical solution: In a first aspect, this application provides a method for full life-cycle health management of mining equipment, the method comprising: The system acquires raw point cloud data output in real time when the target scanner performs scanning tasks in the mine roadway, and simultaneously collects the laser emission power readback value, scanning motor drive current and equipment housing vibration amplitude at the corresponding time. Based on the original point cloud data, the fitted contours of multiple feature sections of the roadway cross section are extracted, and the contour deviation between the contour of each feature section and the corresponding reference section contour line is calculated to form a set of deviations. Using the laser emission power readback value and the vibration amplitude of the device housing as correction factors, the contour deviation in the deviation set is compensated to obtain the first health factor characterizing the ranging accuracy; Based on the fluctuation variance of the scanning motor drive current and the attenuation rate of the laser emission power readback value, a second health factor characterizing the accuracy of angle measurement is calculated. The first health factor and the second health factor are fused together and compared with the original accuracy parameters calibrated by the target scanner under the baseline state to generate the full life cycle health index of the current scanning task. When the full life cycle health index is lower than the dynamic failure threshold, it is determined that the target scanner has experienced a sudden accuracy failure.
[0006] Secondly, this application provides a full life-cycle health management system for mining equipment, the system comprising: The data acquisition module is used to acquire the raw point cloud data output in real time when the target scanner performs scanning tasks in the mine roadway, and simultaneously acquire the laser emission power readback value, scanning motor drive current and equipment housing vibration amplitude at the corresponding time. The contour deviation module is used to extract the fitted contours of multiple feature sections of the roadway cross section based on the original point cloud data, calculate the contour deviation between the contour of each feature section and the corresponding reference section contour line, and form a set of deviations. The ranging evaluation module is used to compensate for the contour deviation in the deviation set by using the laser emission power readback value and the vibration amplitude of the device housing as correction factors, so as to obtain a first health factor characterizing the ranging accuracy. An angle evaluation module is used to calculate a second health factor characterizing the accuracy of angle measurement based on the fluctuation variance of the scanning motor drive current and the attenuation rate of the laser emission power readback value. The health assessment module is used to integrate the first health factor and the second health factor, and compare them with the original accuracy parameters calibrated by the target scanner under the baseline state to generate the full life cycle health index of the current scanning task. When the full life cycle health index is lower than the dynamic failure threshold, it is determined that the target scanner has experienced a sudden accuracy failure.
[0007] One or more technical solutions provided in this application have at least the following technical effects or advantages: The technical solution of this application provides a method for full life cycle health management of mining equipment. First, by synchronously collecting raw point cloud data, laser emission power readback values, scanning motor drive current, and equipment shell vibration amplitude during the scanning task, a multi-dimensional correlation data foundation is established between the measurement output and the internal operating status of the equipment in terms of time series and spatial location. This provides traceable multi-source synchronous information for subsequent accuracy degradation analysis, and solves the problem that traditional offline verification completely separates accuracy assessment from the actual operation process of the equipment and cannot detect gradual degradation.
[0008] Furthermore, by automatically extracting multiple feature cross-section fitting contours distributed along the roadway from the original point cloud, and performing deviation quantification based on curvature features and robust registration with the reference cross-section contour line, the accuracy degradation problem of large-scale point clouds is transformed into a sequence of deviations arranged by mileage. This achieves continuous quantitative evaluation of the scanning accuracy of the entire roadway, solving the problems of low efficiency and inability to cover the full cross-section accuracy analysis caused by traditional manual selection of a small number of cross-sections for visual comparison.
[0009] Furthermore, by using the laser emission power attenuation rate and the equipment housing vibration interference rate to bidirectionally compensate and correct the contour deviation, the amplification effect of power attenuation and vibration, two types of non-precision factors, on the ranging deviation is eliminated, and the first health factor characterizing the pure ranging accuracy is obtained. This solves the problem that the original deviation is mixed with environmental interference and cannot truly reflect the changes in the equipment's ranging capability.
[0010] Furthermore, by extracting the fluctuation variance of the scanning motor drive current and combining it with the laser emission power attenuation rate for weighted fusion, the dynamic change of angle measurement accuracy was indirectly quantified using the equipment's own operating parameters, thus obtaining a second health factor. This solved the problems of lacking online evaluation methods for angle measurement accuracy and the difficulty in separating the coupling effect of mechanical wear and signal attenuation.
[0011] Finally, by dynamically allocating fusion weights with vibration impact weight coefficients, the dual-dimensional health factors of ranging and angle are synthesized into a full life cycle health index. A dynamic failure threshold is constructed using historical statistics, cumulative usage time, and power attenuation degree for judgment. This achieves a comprehensive assessment of the overall health status of the equipment and adaptive failure early warning, solving the problem that traditional fixed tolerance thresholds cannot take into account both normal aging tolerance and abnormal degradation capture throughout the entire life cycle of the equipment.
[0012] In summary, the technical solution of this application realizes full life-cycle health management of the ranging accuracy and angle accuracy of mining 3D laser scanners through multi-source parameter synchronous acquisition and feature section deviation extraction, power vibration bidirectional compensation and current power weighted fusion, vibration adaptive weighting and dynamic failure threshold determination. Through multi-dimensional data synchronization and echo adaptive filtering, curvature feature registration and robust mismatch elimination, bidirectional correction stripping and weighted fusion quantization, historical statistical benchmark and time power dynamic correction, it effectively improves the technical problems of existing technologies that rely on periodic offline verification for health management, resulting in monitoring blind spots during maintenance intervals, inability to continuously perceive multi-dimensional accuracy degradation during operation, and difficulty in taking into account both normal aging and abnormal failure with fixed thresholds. Attached Figure Description
[0013] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0014] Figure 1 A flowchart illustrating a method for full lifecycle health management of mining equipment provided in this application embodiment; Figure 2 A flowchart illustrating the health fusion and failure determination process of a full lifecycle health management method for mining equipment provided in this application embodiment; Figure 3 This is a schematic diagram of the structure of a mining equipment lifecycle health management system provided in an embodiment of this application.
[0015] In the attached diagram, Figure 3 The components represented by each number are described as follows: Data acquisition module 11, contour deviation module 12, distance measurement evaluation module 13, angle evaluation module 14, and health evaluation module 15. Detailed Implementation
[0016] This application provides a method and system for full life-cycle health management of mining equipment, which solves the technical problem in the prior art that the maintenance and repair of mining scanning equipment relies on periodic offline verification, and cannot use the equipment's own operating parameters to conduct online assessment of the overall accuracy and trigger maintenance decisions, resulting in uncontrollable risks.
[0017] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application. It should be noted that the numerical values in the embodiments are for illustrative purposes only and do not constitute a limitation on this application.
[0018] Example 1, as shown in the appendix Figure 1 As shown, this application provides a method for full life-cycle health management of mining equipment, the method comprising the following steps: S100: Acquire the raw point cloud data output in real time by the target scanner when it performs scanning tasks in the mine roadway, and simultaneously collect the laser emission power readback value, scanning motor drive current and equipment housing vibration amplitude at the corresponding time. In this embodiment, when a mining 3D laser scanner performs routine scanning tasks in mine roadways, it generates a large amount of point cloud data in real time. Simultaneously, internal operating parameters such as laser emission power, scanning motor drive current, and casing vibration are also continuously changing. Traditional health management methods only focus on accuracy indicators during offline calibration, completely ignoring the multi-source sensor data generated by the equipment itself during operation, which contains early signs of accuracy degradation. The lack of means to simultaneously collect and correlate point cloud quality and operating parameters makes it impossible to perceive changes in the equipment's health during scanning, resulting in accuracy anomalies only being detected during post-calibration, leaving a serious monitoring blind spot during maintenance intervals. Therefore, a method is needed to simultaneously capture measurement outputs and internal equipment status parameters during scanning. However, this method needs to address the problems of varying sampling frequencies, different physical dimensions, and difficulties in time alignment among different sensors.
[0019] Step S100 in the method provided in this application embodiment includes: The target scanner performs a helical scan of the tunnel cross-section at a fixed angular resolution. The echo signals from each scan cycle are timestamped to generate raw point cloud data, where each point cloud contains three-dimensional coordinates and a corresponding echo intensity value. From the laser emitter drive circuit of the target scanner, the laser emission power is extracted at the instant of each laser pulse emission, forming a power sequence corresponding to each scan point. From the horizontal rotary motor drive circuit of the target scanner, the scanning motor drive current is continuously acquired at a sampling rate matching the scanner's line scan cycle, forming a current sequence associated with each scan cross-section. An accelerometer is attached to the optomechanical base of the target scanner, continuously acquiring vibration acceleration values in three orthogonal directions throughout the scan. The square root of the sum of the squares of the vibration acceleration values in the three directions is taken as the maximum value within a sliding window as the vibration amplitude of the equipment housing, and a time synchronization relationship is established with the corresponding scan cross-section. Detailed explanation follows: In this embodiment, the raw point cloud data refers to the three-dimensional spatial coordinate dataset collected and stored at a fixed angular resolution when the target scanner performs a spiral scan. Each point includes the horizontal rotation angle, vertical deflection angle, X-axis coordinates, Y-axis coordinates, Z-axis coordinates calculated from the laser ranging value, and laser echo intensity value. The laser emission power readback value refers to the instantaneous emission power collected from the drive circuit at the moment of each pulse emitted by the laser transmitter, reflecting the output energy changes caused by laser aging or drive circuit attenuation. The scanning motor drive current refers to the real-time operating current of the scanner's horizontal rotation motor, and its fluctuation reflects changes in mechanical condition such as motor bearing wear and transmission mechanism jamming.
[0020] The vibration amplitude of the equipment housing refers to the scalar value obtained by synthesizing the vibration accelerations of the X, Y, and Z axes collected by the triaxial accelerometer in real time and taking the maximum value within a sliding window. It reflects the intensity of external vibrations or abnormal impacts from internal moving mechanisms experienced by the equipment during scanning. Fixed angular resolution refers to the fixed, minute angular increment of horizontal rotation between every two adjacent laser pulses of the scanner, determining the spatial sampling density of the scanning points. Echo intensity refers to the light intensity reflected back to the receiver after the laser pulse illuminates the tunnel wall, used to characterize the reflectivity of the wall material, and is affected by both the laser emission power and the receiver sensitivity.
[0021] In this step, firstly, in order to acquire multi-source data simultaneously during the scanning process, the target scanner needs to be controlled to perform a spiral scan of the tunnel cross-section at a fixed angular resolution. The echo signal of each laser pulse is encoded according to the timestamp to generate raw point cloud data containing three-dimensional coordinates and echo intensity values, thus obtaining a complete three-dimensional point cloud dataset of the tunnel. For example, if the fixed angular resolution is set to 0.02 degrees and the laser pulse frequency is 100kHz, approximately 18,000 point cloud data points are generated per scanning cycle. Each point contains X, Y, and Z coordinates and an echo intensity value from 0 to 255.
[0022] Furthermore, in order to capture the real-time state changes of the laser emitter, it is necessary to extract the laser emission power readback value from the laser emitter drive circuit at the moment of each pulse emission, forming a power sequence that corresponds one-to-one with the scanning point. This ensures that each point cloud data has a corresponding emission power value that can be traced. For example, the continuously acquired power sequence shows a trend of 3.5W, 3.5W, 3.4W, 3.4W, 3.3W, etc.
[0023] Furthermore, in order to sense changes in the health status of the rotating motor, the motor drive current needs to be continuously collected from the horizontal rotating motor drive circuit at a sampling rate matching the line scan cycle to form a current sequence associated with each section. For example, with a line scan cycle of 50ms and a sampling rate of 1kHz, approximately 500 current sample values are collected for each section, which can reflect the current fluctuation characteristics of the motor in a complete section scan.
[0024] Finally, in order to monitor the vibration environment of the equipment during the scanning process, an accelerometer needs to be attached to the optomechanical base to continuously collect vibration acceleration values in three orthogonal directions throughout the scanning process. The square root of the sum of the squares of the three direction signals is combined and the maximum value within the sliding window is taken as the vibration amplitude. A time synchronization relationship is established with the scanning section at the corresponding moment. For example, if the sliding window length is 100ms, the accelerations in the three directions are 0.3g, 0.2g, and 0.1g, and the combined amplitude is 0.37g.
[0025] For example, taking a mining 3D laser scanner performing a scheduled inspection scan in an underground roadway as an example, the scanner performs a spiral scan of the roadway cross-section with a fixed angular resolution of 0.02 degrees and a laser pulse frequency of 100kHz. Each scan cycle generates approximately 18,000 point cloud data points, each containing X, Y, and Z coordinates and an echo intensity value from 0 to 255. Simultaneously, power readback values are extracted from the laser emitter drive circuit to form a power sequence, with an average power of 3.4W in the current cycle. A current sequence is acquired from the horizontal rotary motor drive circuit at a sampling rate of 1kHz, with approximately 500 sampling points per cross-section. A three-axis accelerometer is attached to the optomechanical base to acquire vibration accelerations of 0.3g on the X-axis, 0.2g on the Y-axis, and 0.1g on the Z-axis. After summing the square roots of the squares, the maximum value of 0.37g is taken as the vibration amplitude and synchronized with the scanning time of that cross-section. All data is uniformly aligned and stored based on the GPS time stamp built into the scanner.
[0026] In summary, this step generates point cloud data by controlling the scanner to perform a spiral scan at a fixed angular resolution. Simultaneously, it extracts three types of internal operating parameters—power, current, and vibration—from the laser drive circuit, motor drive circuit, and accelerometer, establishing a time synchronization relationship. Compared with existing technologies, this step has the following advantages: First, it establishes a unified temporal and spatial correspondence between the scanning measurement output and the equipment's own operating status parameters, providing a complete and traceable multi-source data foundation for subsequently separating the degradation factors of ranging and angular accuracy. Second, it utilizes the equipment's existing sensors and drive circuits to acquire internal parameters without requiring additional external detection equipment. This achieves holographic data acquisition of the scanning process without increasing hardware costs, solving the problem of traditional offline verification being completely detached from the operating process and unable to detect gradual degradation.
[0027] S200: Based on the original point cloud data, extract the fitting contours of multiple feature sections of the roadway cross section, calculate the contour deviation between the contour of each feature section and the corresponding reference section contour line, and form a set of deviations. In this embodiment, after acquiring the original point cloud data and internal operating parameters such as laser emission power, scanning motor current, and shell vibration, the key to accuracy assessment lies in extracting multiple key cross-sectional contours distributed along the tunnel from the massive point cloud and accurately quantifying and comparing them with the baseline contours under the equipment's healthy state. Traditional methods rely on manual selection of a small number of cross-sections for visual comparison, which is not only inefficient and subjective but also unable to automatically extract and quantify contour deviations across the entire tunnel after each scanning task. This makes it difficult to detect minor degradations in ranging accuracy in a timely manner, hindering subsequent ranging accuracy attenuation analysis. Therefore, a method is needed that can automatically extract multi-section contours and perform robust registration and deviation quantification with the baseline state to solve the problems of low efficiency and inability to comprehensively cover the accuracy analysis of the entire tunnel cross-section using traditional manual comparison.
[0028] Step S200 in the method provided in this application embodiment includes: The scanning timestamp and horizontal rotation angle corresponding to each scanning point are parsed from the raw point cloud data. Combined with the inertial navigation pose data built into the target scanner, multiple mileage positions are determined at fixed spatial intervals along the roadway. Each mileage position corresponds to a feature section. Taking the mileage position of each feature section as the center, all point cloud data within half the distance between the front and rear sections are extracted as the original cross-sectional point cloud slices of the corresponding feature sections. The echo intensity distribution of all point cloud data in the original cross-sectional point cloud slices is statistically analyzed. The preset lower quantile is taken as the dynamic intensity lower limit threshold. Points with echo intensity lower than the dynamic intensity lower limit threshold are removed to obtain the filtered cross-sectional point cloud slices. Points in the filtered cross-sectional point cloud slices are sorted according to their horizontal rotation angle. Outliers are eliminated using median filtering with a moving window. Then, cubic spline interpolation is used to generate a closed tunnel cross-sectional fitting contour. A reference cross-sectional contour at the same mileage location is extracted. This reference contour is the cross-sectional contour collected and saved by the target scanner when scanning the same mileage location in a reference state, confirming that the ranging and angle measurement accuracy is acceptable. The tunnel cross-sectional fitting contour is spatially registered with the reference cross-sectional contour. The shortest distance from each point on the registered tunnel cross-sectional fitting contour to the reference cross-sectional contour is calculated, and the maximum value among all shortest distances is taken as the contour deviation of the feature section. All feature sections are traversed, and the contour deviations of each feature section are arranged in mileage order to form a set of deviations. Detailed explanation follows: In this embodiment, the feature section refers to a key cross-section selected at fixed spatial intervals along the roadway direction for extracting the cross-sectional profile and performing deviation analysis. Each feature section corresponds to a mileage location. The original cross-sectional point cloud slice refers to a local point cloud set composed of point cloud data within half the cross-sectional distance range before and after the mileage location of the feature section, used for subsequent fitting of the complete cross-sectional profile at that location. The dynamic intensity lower limit threshold refers to a filtering threshold determined based on the quartiles of the echo intensity distribution statistics of the current cross-sectional point cloud slice, used to adaptively remove low-reflectivity areas or low-quality data points caused by sensor noise.
[0029] The tunnel cross-section fitting contour line refers to the closed, smooth cross-section contour curve generated by cubic spline interpolation after sorting the effective points in the filtered cross-section point cloud slice by rotation angle, eliminating outliers through moving window mid-range filtering, and sorting them by rotation angle. It represents the reconstructed cross-section shape at the current scan location. The reference cross-section contour line refers to the cross-section contour line collected and saved after the target scanner has been factory calibrated or passed its most recent offline verification, scanning the same mileage location and confirming that the ranging and angle measurement accuracy is qualified. It serves as the reference true value for accuracy evaluation. The contour deviation refers to the maximum value among the shortest distances from all points on the fitted contour line to the reference contour line after spatial registration between the current fitted contour line and the reference contour line. It reflects the degree of cross-sectional deformation distortion caused by the deterioration of scanner accuracy at this feature section.
[0030] In this step, firstly, in order to divide the continuous point cloud data into several spatial reference points that can be used for accuracy assessment, it is necessary to parse the timestamp and horizontal rotation angle of each scanning point from the original point cloud data. Combined with the inertial navigation pose data, multiple mileage positions are determined as feature sections along the roadway at fixed spatial intervals. This can yield evaluation sampling points that are evenly distributed along the roadway. For example, 100 feature sections can be determined from a 200m long roadway at 2m intervals.
[0031] Furthermore, in order to extract the local cross-sectional information at each feature section, it is necessary to take the mileage position of each feature section as the center and extract the point cloud data within the range of half the cross-section spacing before and after to form the original cross-sectional point cloud slice. This can obtain the complete local point cloud around each feature section. For example, if the cross-section spacing is 1m, each slice contains about 12,000 point cloud data within the range of 0.5m before and after.
[0032] Furthermore, in order to eliminate the interference of low-quality data points on subsequent contour fitting, it is necessary to statistically analyze the distribution of echo intensity values within the slice, take a preset lower quartile as the dynamic intensity lower limit threshold, and remove points with echo intensity below this threshold to obtain filtered cross-sectional point cloud slices. This can adaptively match the reflection characteristics of different wall materials. For example, if the lower quartile of echo intensity of 32 is taken as the threshold, after removing points with intensity below 32, approximately 1200 valid points remain.
[0033] Furthermore, in order to reconstruct a continuous cross-sectional profile from the discrete point cloud, the filtered point cloud needs to be sorted by the horizontal rotation angle. After eliminating outliers by moving window median filtering, cubic spline interpolation is used to generate a closed tunnel cross-sectional fitting profile. Median filtering can suppress isolated noise, while cubic spline interpolation ensures the smoothness and continuity of the profile. For example, if the window width is 11 points, the interpolated profile contains 360 uniform angle sampling points.
[0034] Furthermore, in order to obtain a reference benchmark for accuracy assessment, it is necessary to extract the benchmark profile at the same mileage location. This benchmark profile is the profile saved by the target scanner after scanning the same location in the benchmark state and confirming that the accuracy is qualified. It represents the true measurement value of the scanner in a healthy state and can be directly retrieved from the equipment benchmark database by mileage index.
[0035] Finally, in order to quantify the deviation between the current scan and the reference state, the fitted contour line and the reference contour line need to be spatially registered. The shortest distance from each point on the fitted contour line to the reference contour line after registration is calculated, and the maximum value is taken as the contour deviation of the feature section. The deviations of all feature sections are arranged by mileage to form a deviation set.
[0036] Step S200 in the method provided in this application embodiment further includes: Calculate the local curvature value of each point on the fitted contour line, extract the local maxima of curvature as the first feature point set, and simultaneously extract the local maxima of curvature on the reference section contour line as the second feature point set. Establish candidate matching pairs between each point in the first feature point set and the point with the smallest curvature value difference in the second feature point set, and eliminate mismatched pairs, retaining at least three pairs of valid corresponding feature points. Based on the valid corresponding feature points, use the least squares method to solve for the two-dimensional rigid transformation parameters from the fitted contour line coordinate system to the reference section contour line coordinate system. The two-dimensional rigid transformation parameters include a rotation angle and a translation vector. Then, apply the two-dimensional rigid transformation method to all points on the fitted contour line. The coordinate transformation is performed using transformation parameters to obtain the registered fitted contour line. For each point on the registered fitted contour line, the shortest distance from the point to the reference section contour line is calculated. The calculation steps for the shortest distance include: traversing all line segments that constitute the reference section contour line, calculating the shortest distance from the point to each line segment. If the foot of the perpendicular drawn from the point to a line segment falls on a certain line segment, then the shortest distance from the point to the corresponding line segment is the length of the perpendicular; otherwise, it is the smaller value between the distances from the point to the two endpoints of the corresponding line segment. The minimum value among all calculated shortest distances is taken as the shortest distance from the point to the reference section contour line. After traversing all points, the shortest distance corresponding to each point is output. Detailed explanation is as follows: In this embodiment, the first feature point set refers to the set of local maxima of curvature extracted from the fitted contour line. These points represent the most significant geometric inflection points on the cross-sectional contour, such as the arch, side corners, etc. The second feature point set refers to the set of feature points obtained from the reference cross-sectional contour line using the same curvature maxima extraction method, used for matching corresponding points with the first feature point set. Candidate matching pairs refer to the initial correspondence established between each point in the first feature point set and the point in the second feature point set with the smallest difference in curvature value. False matches may occur due to local shape similarity. Valid corresponding feature points refer to the pairs of corresponding points that are actually corresponding in geometric space after the random sampling consensus algorithm removes false matches, used for subsequent rigid transformation parameter solving. Two-dimensional rigid transformation parameters refer to two parameters describing the rotation and translation relationship between the fitted contour line coordinate system and the reference contour line coordinate system, including a rotation angle and a translation vector, maintaining the contour shape unchanged while performing only rotation and translation alignment.
[0037] In this step, firstly, in order to find geometric control points that can be used for registration between the two contour lines, it is necessary to calculate the local curvature value of each point on the fitted contour line and extract the local maxima of curvature as the first feature point set. At the same time, the second feature point set is extracted on the reference contour line in the same way. The curvature maxima correspond to stable geometric features such as the arch apex and turning point of the cross-sectional contour. For example, 15 feature points are extracted on the fitted contour line and 14 feature points are extracted on the reference contour line.
[0038] Furthermore, in order to establish a preliminary correspondence between feature points, candidate matching pairs need to be established between each point in the first feature point set and the point with the smallest curvature value difference in the second feature point set. Since there may be symmetrical or similar structures in the local contour, some erroneous matches are included in the candidate matching pairs, and erroneous matching pairs need to be eliminated by random sampling consistency algorithm.
[0039] The random sample consensus algorithm is a robust parameter estimation method that iteratively estimates mathematical models from datasets containing outliers. In contour registration scenarios, this algorithm randomly samples two pairs of points from candidate matching pairs to calculate an initial rigid transformation. After applying the transformation to the source points in all candidate matching pairs, it counts the number of matching pairs where the Euclidean distance between the target point and the transformed source point is less than a preset inlier distance threshold as the inlier count. The inlier determination distance formula is... ,in As the source point, Let R be the target point, and t be the rotation matrix and translation vector calculated from the current sample. The interior point distance threshold is set as follows: After a preset maximum number of iterations, the transformation with the most interior points is selected as the final result for eliminating false matches, retaining at least 3 pairs of valid identical feature points. The interior point distance threshold τ is related to the nominal ranging accuracy σ of the target scanner, and is preferably set to a value that is more than 100%. =(2~3)σ. This value ensures that points within the normal measurement noise range are considered interior points, while deviations exceeding the nominal accuracy range are considered mismatches or local roadway deformations and are discarded. An excessively large threshold can lead to registration distortion, while an excessively small threshold can result in the loss of valid feature points. In practical applications, the threshold can be adaptively set based on the statistical distribution of historical registration residuals under the baseline condition. For example, with an inlier distance threshold of 0.05m and a maximum of 200 iterations, 5 pairs of valid corresponding feature points can be retained from 15 candidate matching pairs.
[0040] Furthermore, in order to accurately align the fitted contour line to the reference contour line, the two-dimensional rigid transformation parameters, including a rotation angle and a translation vector, need to be solved using the least squares method based on the effective corresponding feature points. This transformation is then applied to all points on the fitted contour line to obtain the registered fitted contour line.
[0041] Finally, in order to quantify the local deviation between the two profiles after registration, it is necessary to traverse each point on the fitted profile after registration, calculate the shortest distance from that point to each line segment on the reference section profile, take the minimum value as the deviation value of that point, and take the maximum value among the shortest distances of all points as the profile deviation of that feature section.
[0042] Step S200 in the method provided in this application embodiment further includes: Points belonging to the fitted contour line among the effective corresponding feature points are denoted as the source point set, and points belonging to the reference section contour line are denoted as the target point set. The centroid coordinates of the source point set and the target point set are calculated respectively. The source and target point sets are then decentered to obtain the source-decentered coordinate set and the target-decentered coordinate set respectively. A covariance matrix is constructed based on the source and target-decentered coordinate sets, and singular value decomposition (SVD) is performed on the covariance matrix to obtain three SVD matrices. A rotation matrix is calculated based on the three SVD matrices, and the rotation angle is extracted from the rotation matrix. The translation vector is obtained by subtracting the rotation matrix multiplied by the source point set's centroid coordinates from the target point set's centroid coordinates. The rotation angle and translation vector are combined as the output parameters for a two-dimensional rigid transformation. A detailed explanation follows: In this embodiment, the source point set and the target point set refer to the coordinate sets composed of valid corresponding feature points belonging to the fitted contour line and the reference section contour line, respectively, and are the direct inputs for solving the rigid transformation. The covariance matrix refers to the matrix constructed from the source decentralized coordinate set and the target decentralized coordinate set, whose elements reflect the correlation distribution of the two coordinate sets in different directions. The singular value decomposition matrix refers to the three matrices obtained after performing singular value decomposition on the covariance matrix, including a left singular vector orthogonal matrix, a singular value diagonal matrix, and the transpose of a right singular vector orthogonal matrix, used to mathematically decouple the rotation component and the scaling component.
[0043] In this step, firstly, in order to solve the optimal rigid transformation, the x and y coordinates of each point in the source point set and the target point set are summed and divided by the number of points to obtain the centroid coordinates. Then, the two sets of points are decentered so that the two sets of points are translated into a local coordinate system with their respective centroids as the origin.
[0044] Furthermore, in order to construct a matrix that reflects the spatial relationship between the two sets of points, the source decentralized coordinate set and the target decentralized coordinate set need to be arranged in columns, the covariance matrix needs to be calculated, and the singular value decomposition of the matrix is performed to obtain three matrices.
[0045] Finally, in order to extract the rotation component from the decomposition result, the left singular vector matrix and the right singular vector matrix are multiplied to obtain the rotation matrix. The rotation angle is extracted from the rotation matrix, and the translation vector is obtained by subtracting the rotation matrix from the centroid of the target point set and multiplying it by the centroid of the source point set. These are then combined to form the two-dimensional rigid transformation parameters.
[0046] For example, taking a mining 3D laser scanner performing a scheduled inspection scan in an underground roadway as an example, a feature section is taken every 2 meters along a 200m roadway, for a total of 100 feature sections. The original cross-sectional point cloud slice of the 25th feature section contains approximately 12,000 points, with a lower quartile of 32 for echo intensity. After removing low-intensity points using a dynamic lower threshold, approximately 1,200 valid points are retained. The valid points are sorted by rotation angle, and outliers are eliminated by moving window mid-range filtering with a window width of 11. A closed fitted contour line containing 360 angle sampling points is generated by cubic spline interpolation. A baseline contour line at the same location is extracted from the baseline database. 15 curvature maxima are extracted from the fitted contour line and 14 curvature maxima are extracted from the baseline contour line. False matches are removed using the minimum curvature difference matching and random sampling consistency algorithms. The in-point distance threshold is 0.05m. After 200 iterations, 5 pairs of valid corresponding feature points are retained. The rotation angle was calculated to be 0.015 rad using SVD, and the translation vectors were 0.10 m and 0.05 m. The fitted contour line was then transformed and registered with the reference contour line. 360 sampling points on the registered contour line were traversed, and the shortest distance from each point to each segment of the reference contour line was calculated. The maximum value of 22 mm was taken as the contour deviation of that feature section. The deviations of all 100 feature sections were arranged by mileage to form a deviation set.
[0047] In summary, this step automatically extracts multiple feature cross-sectional contours evenly distributed along the roadway, employs curvature feature points combined with a random sampling consensus algorithm for robust registration, and quantifies the maximum deviation as the contour deviation. Compared with existing technologies, this step has the following advantages: First, it transforms the accuracy degradation problem of large-scale point clouds into a set of comparable deviation sequences arranged by mileage, enabling continuous quantitative evaluation of the scanning accuracy of the entire roadway along the spatial direction; second, by eliminating mismatches through the random sampling consensus algorithm, stable registration results can still be obtained even under conditions of local deformation or occlusion of the cross-sectional contours, providing direct and objective quantitative input for the subsequent calculation of the ranging health factor, and solving the problems of low efficiency, strong subjectivity, and inability to cover the full cross-sectional accuracy analysis of traditional manual visual comparison of a small number of cross-sections.
[0048] S300: Using the laser emission power readback value and the vibration amplitude of the equipment housing as correction factors, the contour deviation in the deviation set is compensated to obtain the first health factor characterizing the ranging accuracy. In this embodiment, after obtaining the set of profile deviations for each characteristic section, these deviations include not only profile deformation caused by the actual degradation of the device's ranging accuracy, but also the effects of non-accuracy factors such as the decrease in echo signal-to-noise ratio due to laser emission power attenuation and point cloud jitter caused by device housing vibration. Directly using the raw deviations for health assessment will overestimate or underestimate the actual degree of device degradation. For example, power attenuation weakens the echo signal, increases measurement noise, and artificially inflates the profile fitting deviation; while vibration causes point cloud position jitter, similarly increasing the deviation. If these non-accuracy factors are not removed, the health assessment of ranging accuracy will lose its accuracy, making it impossible to distinguish between the device's own accuracy degradation and measurement fluctuations caused by environmental interference. Therefore, a method is needed that can remove power attenuation and vibration interference from the deviations and extract the core indicators that simply reflect changes in the device's ranging capability.
[0049] Step S300 in the method provided in this application embodiment includes: From the power sequence, based on the time synchronization relationship between the scanning section and the feature section, the laser emission power readback value corresponding to the scanning point of each feature section is extracted, and the arithmetic mean of the power corresponding to each feature section is calculated; from the vibration amplitude of the equipment housing associated with each feature section, the maximum vibration amplitude within the corresponding scanning period of each feature section is extracted; from the original accuracy parameters calibrated by the target scanner in the reference state, the reference laser emission power and reference vibration amplitude are read; based on the arithmetic mean of the power of each feature section and the reference laser emission power, the power attenuation rate is determined, and based on the maximum vibration amplitude of each feature section and the reference vibration amplitude, the vibration interference rate is determined; the contour deviation in the deviation set is compensated using the power attenuation rate and the vibration interference rate to obtain the compensated contour deviation of each feature section; the arithmetic mean of the compensated contour deviation of all feature sections is calculated, and the arithmetic mean is used as the first health factor. Detailed explanation follows: In this embodiment, the power arithmetic mean refers to the arithmetic mean of the laser emission power readback values corresponding to all scanning points constituting a certain characteristic cross section, used to characterize the overall output energy level of the laser when scanning that cross section. The power attenuation rate is the ratio of the power arithmetic mean to the reference laser emission power, reflecting the degree of attenuation of the current laser emission capability relative to the factory or calibration state; the smaller this value, the more severe the laser aging. The vibration interference rate is the ratio of the maximum vibration amplitude to the reference vibration amplitude, reflecting the amplification factor of the current scanning environment vibration level relative to the reference state.
[0050] The reference laser emission power refers to the nominal laser emission power measured in a standard environment under factory calibration or the most recent offline verification condition of the target scanner. The reference vibration amplitude refers to the upper limit of the shell vibration amplitude collected by the target scanner under reference conditions in a standard test environment, representing the normal vibration environment level. The compensated profile deviation is the result obtained by multiplying the original profile deviation by the power attenuation rate and then dividing by the vibration interference rate. Mathematically, this is equivalent to separating the contributions of power attenuation and vibration interference to the deviation, restoring the deviation component caused by the actual change in the equipment's ranging accuracy. The first health factor refers to the arithmetic mean of the compensated profile deviations of all characteristic sections, comprehensively reflecting the ranging accuracy health level of the scanner throughout the entire tunnel; the smaller the value, the closer the ranging accuracy is to the reference state.
[0051] In this step, firstly, in order to spatially correspond the power data with the deviation, it is necessary to extract the laser emission power readback value corresponding to the scanning point that makes up each feature section based on the time synchronization relationship between the scanning section and the feature section, and calculate the arithmetic mean of all power values in the section to obtain the overall power level of each feature section. For example, the 25th feature section contains 1200 scanning points, and the corresponding power value arithmetic mean is 3.3W.
[0052] Furthermore, in order to obtain the vibration environment of each characteristic section, it is necessary to extract the maximum vibration amplitude during the scanning period of that section from the vibration records with established time synchronization. Vibration interference mainly affects the instantaneous fluctuation of the distance measurement value. Taking the maximum value within the time period can reflect the worst working conditions. For example, the maximum vibration amplitude during the scanning period of the 25th section is 0.37g.
[0053] Furthermore, in order to determine the benchmark reference for compensation, the benchmark laser emission power and benchmark vibration amplitude need to be read from the original accuracy parameters calibrated under the benchmark state. The benchmark power represents the optimal output level of the laser when the equipment leaves the factory, and the benchmark vibration amplitude represents the normal vibration level under the standard test environment, such as a benchmark power of 5.0W and a benchmark vibration amplitude of 0.1g.
[0054] Furthermore, to quantify the relative degrees of power attenuation and vibration deviation, it is necessary to calculate that the power attenuation rate equals the arithmetic mean of the power output divided by the reference laser emission power, and the vibration interference rate equals the maximum vibration amplitude divided by the reference vibration amplitude. For example, the power attenuation rate = 3.3. 5.0 = 0.66, vibration disturbance rate = 0.37 0.1 = 3.7.
[0055] Furthermore, to eliminate the contribution of non-precision factors to the deviation, the original contour deviation needs to be compensated using the power attenuation rate and vibration interference rate. The compensated contour deviation is equal to the original deviation multiplied by the power attenuation rate and then divided by the vibration interference rate. When the laser power attenuates, the echo signal weakens and the measurement noise increases. The original deviation includes a false high component caused by power attenuation. Multiplying by a power attenuation rate less than 1 can reduce and restore this false high component. When vibration intensifies, the point cloud position jitter causes the fitted contour to deviate further from the reference. The original deviation also includes a false high component caused by vibration. Dividing by a vibration interference rate greater than 1 can similarly reduce and restore this false high component. The design intent of the compensation formula is to eliminate the amplification effect of environmental interference on the deviation and extract the true contour deviation caused by the degradation of the equipment's own precision.
[0056] It should be noted that if the current power is not lower than the reference power, i.e., the power attenuation rate is greater than 1, it indicates that the laser emission capability has not attenuated. In this case, the power attenuation rate is truncated to 1, and no power attenuation compensation is performed. Similarly, if the current vibration is lower than the reference vibration, i.e., the vibration interference rate is less than 1, it indicates that the vibration environment is better than the calibration state. In this case, the vibration interference rate is truncated to 1, and no vibration interference compensation is performed. This truncation process avoids erroneous amplification compensation caused by reference value deviation or short-term environmental fluctuations.
[0057] Finally, to obtain the overall ranging accuracy index for the entire tunnel, the arithmetic mean of the compensated profile deviations of all characteristic sections needs to be calculated as the first health factor. For example, the original deviation of the 25th characteristic section is 22mm, and after compensation, it is 22 × 0.66. 3.7≈3.92mm, traversing 100 characteristic sections, the power attenuation rate of each section is between 0.55 and 0.72, the vibration interference rate is between 2.1 and 4.5, the compensation deviation is between 3.2mm and 4.8mm, and the arithmetic mean of 4.0mm is taken as the first health factor.
[0058] For example, taking a mining 3D laser scanner performing a scheduled inspection scan in an underground roadway as an example, the 25th feature section contains 1200 scan points. The arithmetic mean of the extracted corresponding power readback values is 3.3W. The maximum vibration amplitude during the scanning period of this section is 0.37g. The reference laser emission power is 5.0W and the reference vibration amplitude is 0.1g, read from the original accuracy parameters. The power attenuation rate is 3.3. 5.0 = 0.66, vibration disturbance rate is 0.37 0.1 = 3.7, vibration disturbance rate greater than 1 does not require truncation. The original profile deviation of this section is 22mm, and the compensated profile deviation = 22 × 0.66 3.7≈3.92mm. Traversing 100 characteristic sections, the power attenuation rate of each section is between 0.55 and 0.72, the vibration interference rate is between 2.1 and 4.5, and the deviation after compensation is between 3.2mm and 4.8mm. The arithmetic mean of 4.0mm is taken as the first health factor.
[0059] In summary, this step extracts the arithmetic mean of power and the maximum vibration amplitude cross-section by cross-section, and calculates the power attenuation rate and vibration interference rate by combining the reference power and reference vibration value. After compensating for the contour deviation, the average value is taken to obtain the first health factor. Compared with the prior art, this step has the following advantages: First, by performing bidirectional correction on the original deviation through the power attenuation rate and vibration interference rate, the interference of two non-precision factors, laser power attenuation and shell vibration, on the contour measurement is effectively eliminated, making the first health factor more purely reflect the true attenuation level of the equipment's ranging accuracy. Second, the cases where the power attenuation rate is greater than 1 and the vibration interference rate is less than 1 are respectively truncated to prevent inverse compensation errors caused by reference value deviation or environmental fluctuations, thereby enhancing the robustness and engineering practicality of the correction algorithm.
[0060] S400: Based on the fluctuation variance of the scanning motor drive current and the attenuation rate of the laser emission power readback value, calculate the second health factor characterizing the accuracy of angle measurement; In this embodiment, after obtaining the first health factor characterizing the ranging accuracy, it is also necessary to independently assess the health status of the scanner's angle measurement accuracy. Fluctuations in the drive current of the scanning motor reflect the smoothness of the rotating mechanism's motion during scanning. Mechanical faults such as motor bearing wear, increased gear clearance, or lubrication failure can directly cause irregular current fluctuations, leading to deviations between the actual and nominal step distances of the horizontal rotation angle. Simultaneously, the attenuation of laser emission power reduces the signal-to-noise ratio of the echo signal, decreasing the scanner's ability to distinguish angle boundaries. Traditional offline calibration methods typically use external angle reference devices for measurement, making it impossible to perform online assessments of angle accuracy using the device's own parameters during routine scanning tasks. This often leads to the gradual degradation of angle accuracy being overlooked until significant measurement distortion occurs. Therefore, a method is needed that can indirectly quantify changes in angle measurement accuracy using existing operating parameters to address the lack of online assessment methods for angle measurement accuracy and the difficulty in separating the coupling effects of mechanical degradation and signal attenuation.
[0061] Step S400 in the method provided in this application embodiment includes: The scanning motor drive current value corresponding to each characteristic section is extracted from the current sequence, and the fluctuation variance of the current value of each characteristic section during the scanning period is calculated. The laser emission power readback value corresponding to each characteristic section is obtained from the power sequence, and the power attenuation rate of each characteristic section is determined based on the relationship between the laser emission power readback value and the reference power value. Based on the fluctuation variance and power attenuation rate of each characteristic section, the angle measurement deviation characteristic value of each characteristic section is determined. The arithmetic mean of the angle measurement deviation characteristic values of all characteristic sections is calculated as the second health factor. Detailed explanation follows: In this embodiment, the current fluctuation variance refers to the variance of all scanning motor drive current sampling values within the scanning period of a certain characteristic section. It is used to quantify the motion stability of the rotating motor when scanning the section. The larger the variance, the more uneven the motor speed, and the worse the consistency of the inferred angle measurement step. The angle measurement deviation characteristic value is a composite index obtained by multiplying the current fluctuation variance by a power degradation factor, where the power degradation factor is the ratio of the reference power value to the current laser emission power readback value. Its physical meaning is that when the laser power decays, the signal-to-noise ratio of the echo signal decreases, the ambiguity of the angle boundary increases, and the same current fluctuation will lead to a larger angle measurement deviation under weak signal conditions. Therefore, multiplying by the power degradation factor amplifies and quantifies the impact of signal attenuation. The second health factor refers to the arithmetic mean of the angle measurement deviation characteristic values of all characteristic sections, which comprehensively reflects the health level of the scanner's angle measurement accuracy throughout the entire tunnel. The smaller the value, the closer the angle measurement accuracy is to the reference state.
[0062] In this step, firstly, in order to extract the motor running stability index of each characteristic section from the current sequence, it is necessary to extract all current sampling values within the scanning period of the characteristic section based on the time synchronization relationship between the characteristic section and the scanning section, and calculate the current fluctuation variance. This can quantify the degree of fluctuation of the motor speed during the scanning process of the section. For example, the scanning period corresponding to the 25th characteristic section contains 500 current sampling values with a mean of 0.85A and a fluctuation variance of 0.0025.
[0063] Furthermore, to obtain the laser power attenuation at each characteristic section, it is necessary to extract the power readback values corresponding to all scanning points within that section from the power sequence, and calculate the ratio of the arithmetic mean of the power to the reference power value as the power attenuation rate. For example, the arithmetic mean of the power at the 25th characteristic section is 3.3W, the reference power value is 5.0W, and the power attenuation rate is 3.3. 5.0 = 0.66.
[0064] Furthermore, to comprehensively quantify the coupled impact of current fluctuations and power attenuation on angle measurement accuracy, the current fluctuation variance needs to be multiplied by a power degradation factor to obtain the characteristic value of the angle measurement deviation. When power attenuation is severe, the current power value decreases, the power degradation factor increases, and the same current fluctuation variance multiplied by a larger degradation factor will yield a larger characteristic value, amplifying the impact of current fluctuations on angle measurement accuracy under weak signal conditions. For example, the arithmetic mean power of the 25th characteristic section is 3.3W, the reference power value is 5.0W, and the power degradation factor is 5.0. 3.3≈1.52, the characteristic value of the angle measurement deviation is equal to 0.0025 multiplied by 1.52, which is approximately equal to 0.0038.
[0065] Finally, in order to obtain the comprehensive angle measurement accuracy index of the entire tunnel, the arithmetic mean of the angle measurement deviation characteristic values of all characteristic sections needs to be calculated as the second health factor. For example, if the angle measurement deviation characteristic values of 100 characteristic sections are between 0.0028 and 0.0065, the arithmetic mean of 0.0042 is taken as the second health factor.
[0066] For example, taking the same mining 3D laser scanner performing a scheduled inspection scan in an underground roadway, the scanning time corresponding to the 25th feature section contains 500 current sampling values with a mean of 0.85A and a variance of 0.0025. The arithmetic mean power of this section is 3.3W, the baseline power value is 5.0W, and the power degradation factor is 5.0. 3.3 ≈ 1.52. Angle measurement deviation characteristic value = 0.0025 × 1.52 = 0.0038. Traversing 100 characteristic sections, the current fluctuation variance of each section is between 0.0018 and 0.0042, the power degradation factor is between 1.39 and 1.82, and the angle measurement deviation characteristic value is between 0.0028 and 0.0065. Taking the arithmetic mean of 0.0042 gives the second health factor.
[0067] In summary, this step extracts the fluctuation variance of the scanning motor drive current and the laser emission power attenuation rate cross-section by cross-section, divides the two to obtain the characteristic value of the angle measurement deviation, and takes the average of all cross-sections as the second health factor. Compared with the prior art, this step has the following advantages: First, by utilizing the current signal and laser power signal of the device's own drive circuit, online quantitative evaluation of angle measurement accuracy is achieved without adding external sensors; second, by dividing by the power attenuation rate, the influence of current fluctuations on angle measurement under weak signal conditions is amplified, making the angle health assessment adaptively sensitive to laser aging, and solving the problem that traditional offline verification cannot detect the degradation of operating conditions.
[0068] S500: Integrates the first health factor and the second health factor, and compares them with the original accuracy parameters calibrated by the target scanner under the baseline state to generate the full life cycle health index of the current scanning task. When the full life cycle health index is lower than the dynamic failure threshold, it is determined that the target scanner has experienced a sudden accuracy failure.
[0069] In this embodiment, after obtaining the first health factor characterizing ranging accuracy and the second health factor characterizing angle measurement accuracy, the key to determining whether the equipment has experienced accuracy failure lies in how to scientifically integrate the two into a full lifecycle index characterizing the overall health status of the device, and how to set a failure judgment threshold that can adapt to the dynamic changes in equipment usage time and power aging. Traditional methods typically use fixed tolerance thresholds for acceptance judgment, such as specifying that the contour deviation must not exceed a certain fixed value. Such static thresholds cannot distinguish between normal aging and abnormal degradation of the equipment. Setting the threshold too strictly will lead to frequent false alarms for normal aging, while setting it too loosely may miss real accuracy failures, and it does not consider the differentiated performance of ranging accuracy and angle accuracy under different operating conditions of the equipment. Therefore, a method is needed that can dynamically integrate the two-dimensional accuracy indicators and adaptively adjust the judgment threshold to solve the problems of poor applicability of fixed thresholds throughout the entire equipment lifecycle and the inability to simultaneously consider normal aging tolerance and abnormal degradation detection.
[0070] Step S500 in the method provided in this application embodiment includes: The process involves extracting reference ranging accuracy and reference angle accuracy values from the original accuracy parameters; determining the normalized ranging health component based on the reference ranging accuracy and the first health factor; determining the normalized angle health component based on the reference angle accuracy and the second health factor; calculating the arithmetic mean of the equipment housing vibration amplitude throughout the scanning process; determining the vibration influence weighting coefficient based on the ratio of the arithmetic mean to the reference vibration amplitude, where a larger ratio results in a smaller vibration influence weighting coefficient; determining the fusion weights of the normalized ranging health component and the normalized angle health component based on the vibration influence weighting coefficients; and weighting and fusing the normalized ranging health component and the normalized angle health component according to their respective fusion weights to obtain the full lifecycle health index. A detailed explanation follows: In this embodiment, the reference ranging accuracy value refers to the ranging accuracy index calibrated by the target scanner under factory calibration or offline verification conditions, for example, a nominal ranging accuracy of 3mm. The reference angle accuracy value refers to the variance of the scanning motor drive current fluctuation measured in a standard environment under factory calibration or offline verification conditions, used to characterize the reference fluctuation level of the angle measurement system in a healthy state, for example, a reference angle accuracy value of 0.002, in amperes (A). 2The normalized ranging health component refers to the ratio obtained by dividing the reference ranging accuracy value by the first health factor. When the first health factor is greater than the reference ranging accuracy value, this component is less than 1, indicating that the ranging capability has deteriorated to the reference level; conversely, it is greater than 1, indicating that the ranging capability is better than the reference. The normalized angle health component refers to the ratio obtained by dividing the reference angle accuracy value by the second health factor. Its meaning is similar to that of the ranging health component; the smaller the value, the more severe the attenuation of angle accuracy relative to the reference.
[0071] The vibration impact weighting coefficient is a dynamic coefficient determined by the ratio of the arithmetic mean of the vibration amplitude over the entire range to the reference vibration amplitude. Its physical basis lies in the fact that the interference of shell vibration on laser ranging is far greater than its interference on angle measurement. Therefore, when vibration is severe, appropriately reducing the fusion weight of the ranging component can prevent the health index from being excessively affected by vibration noise. As a specific implementation method, the vibration impact weighting coefficient can be taken as the ratio of the reference vibration amplitude to the arithmetic mean of the entire vibration range. When this ratio is greater than 1, it is truncated to 1 to ensure that the weighting coefficient is always between 0 and 1. The full life cycle health index refers to the comprehensive index obtained by weighting and fusing the normalized ranging health component and the normalized angle health component according to the vibration impact weighting coefficient. The higher the value, the healthier the overall accuracy of the machine. Under the reference state, this value is approximately 1.
[0072] In this step, firstly, in order to normalize the two health factors into components comparable to the baseline state, the baseline ranging accuracy value and the baseline angle accuracy value need to be extracted from the original accuracy parameters. The baseline value is divided by the corresponding health factor to obtain the normalized ranging health component and the normalized angle health component. This ratio ensures that when the health factor is equal to the baseline accuracy, the health component is 1, and when the health factor is larger, i.e., the accuracy is worse, the health component is smaller. For example, the baseline ranging accuracy value of 3mm divided by the first health factor of 4mm yields a normalized ranging health component of 0.75, and the baseline angle accuracy value of 0.002 degrees divided by the second health factor of 0.0042 is approximately equal to 0.476.
[0073] Furthermore, to reflect the differentiated impact of vibration on the two types of accuracy, it is necessary to calculate the arithmetic mean of the equipment housing vibration amplitude throughout the scanning process. The vibration influence weighting coefficient is determined based on the ratio of this average value to the reference vibration amplitude. A larger ratio indicates more severe vibration in the scanning environment, and a smaller vibration influence weighting coefficient. Since vibration interferes with ranging accuracy far more than it interferes with angle measurement, reducing the fusion weight of the ranging component during periods of severe vibration can prevent the health index from being dominated by vibration noise from the ranging channel. As a specific implementation method, the vibration influence weighting coefficient can be taken as the ratio of the reference vibration amplitude to the arithmetic mean of the entire vibration range. When this ratio is greater than 1, it is truncated to 1 to ensure that the weighting coefficient is always between 0 and 1. For example, if the average vibration range is 0.42g, the reference vibration amplitude is 0.1g, the ratio is 4.2, and the vibration influence weighting coefficient is 0.1. 0.42≈0.238.
[0074] Furthermore, in order to determine the fusion weight of the two health components, the vibration influence weight coefficient should be used as the fusion weight of the normalized ranging health component, and the value of 1 minus the vibration influence weight coefficient should be used as the fusion weight of the normalized angle health component. The sum of the two is always 1, ensuring that the weighted fusion health index remains stable in magnitude. For example, the ranging weight is 0.238 and the angle weight is 0.762.
[0075] Finally, in order to generate a comprehensive health index for the whole machine, the normalized ranging health component and the normalized angle health component need to be weighted and fused according to the fusion weight. The full life cycle health index is equal to the normalized ranging health component multiplied by the ranging fusion weight plus the normalized angle health component multiplied by the angle fusion weight, for example, 0.75×0.238+0.476×0.762≈0.54.
[0076] Step S500 in the method provided in this application embodiment further includes: From historical scan task records, the historical full-lifecycle health index generated by each scan task of the target scanner in a healthy state is extracted, and the scan duration corresponding to each scan task is extracted to form a historical health sequence and a historical duration sequence. The mean and standard deviation of the historical health sequence are calculated to determine the baseline failure threshold. The arithmetic mean of the historical duration sequence is calculated as the reference scan duration. The cumulative scan duration of the target scanner since the last offline verification is obtained, as well as the overall power attenuation degree determined based on the power sequence and the baseline laser emission power. The duration degradation factor is determined based on the relationship between the cumulative scan duration and the reference scan duration, and the power degradation factor is determined based on the overall power attenuation degree. The baseline failure threshold is corrected using the duration degradation factor and the power degradation factor to obtain the dynamic failure threshold. Detailed explanation is as follows: In this embodiment, the baseline failure threshold refers to a lower limit critical value calculated based on the statistical distribution of historical health sequences. It is expressed as the mean minus k times the standard deviation. For example, when k is 2, it corresponds to a lower bound with approximately 95% confidence, meaning that the health index has only about a 2.5% probability of falling below this threshold under normal conditions. The cumulative scan time refers to the cumulative number of working hours of the device since its most recent offline calibration, used to quantify the accumulated mechanical wear of the device. The overall power attenuation level refers to the ratio of the overall arithmetic mean of all laser emission power readback values in the power sequence to the baseline laser emission power, comprehensively reflecting the overall aging level of the laser emitter.
[0077] The duration degradation factor is the ratio of the cumulative scan duration to the reference scan duration. When the cumulative duration exceeds the reference duration, this factor is greater than 1, reflecting that the equipment usage intensity has exceeded the typical cycle, and the failure judgment criteria need to be relaxed. When the cumulative duration is less than the reference duration, this factor is less than 1, indicating that the equipment is still within the typical usage cycle, and the judgment criteria can be relatively strict. The power degradation factor is the reciprocal of the overall power attenuation. The more severe the overall power attenuation, the smaller this value, and the larger its reciprocal, the larger the power degradation factor, reflecting that the signal quality degradation caused by laser aging requires a more lenient failure threshold. The dynamic failure threshold is the result obtained by multiplying the baseline failure threshold by the duration degradation factor and then by the power degradation factor. When the equipment usage time is within the normal range and the power attenuation is small, the dynamic threshold is lower than the baseline threshold, and the judgment criteria are strict. As the equipment usage time accumulates and the power attenuation intensifies, the degradation factor gradually increases, the dynamic threshold increases accordingly, and the judgment criteria are moderately relaxed. Failure judgment is only triggered when the decline in the health index exceeds the normal aging expectation.
[0078] The procedure for this step is as follows (see attached document). Figure 2 As shown; In this step, firstly, in order to establish a normal fluctuation baseline for the device in a healthy state, it is necessary to extract the full life cycle health index and scan duration of each scan task in a healthy state from the historical records to form a historical health sequence and a historical duration sequence. Based on the mean and standard deviation of the health sequence, the baseline failure threshold is calculated. For example, if the mean of the health sequence is 0.95 and the standard deviation is 0.15, and k is 2, the baseline failure threshold is equal to 0.95 minus 2 multiplied by 0.15, which equals 0.65.
[0079] Furthermore, in order to determine the reference benchmark for normal aging of the equipment, the arithmetic mean of the historical duration series needs to be calculated as the reference scan duration. This value represents a typical working cycle from healthy to requiring maintenance under normal usage intensity for this model of equipment, for example, a reference scan duration of 800h.
[0080] Furthermore, to quantify the current cumulative aging of the equipment, it is necessary to obtain the cumulative scan time since the last offline verification, and calculate the ratio of the overall arithmetic mean of all power readback values in the power sequence to the reference laser emission power as the overall power attenuation level. For example, if the cumulative scan time is 320 hours, the overall arithmetic mean power is 3.2W, and the overall power attenuation level is 3.2W. 5.0 = 0.64.
[0081] Furthermore, to take into account the impact of usage time and power attenuation on failure determination, the usage time degradation factor needs to be calculated as the cumulative scan time divided by the reference scan time, and the power degradation factor needs to be calculated as the reciprocal of the overall power attenuation level. For example, the usage time degradation factor = 320. 800 = 0.4, power degradation factor = 1 0.64 ≈ 1.56. A duration degradation factor of 0.4, which is less than 1, indicates that the cumulative usage time of the equipment is still within the typical cycle, and the judgment standard can be relatively strict. When the overall power attenuation is greater than 1, that is, the current power is not lower than the reference power, the power degradation factor is truncated to 1 to avoid the threshold error being amplified due to the deviation of the reference value.
[0082] Finally, to obtain a dynamic judgment threshold that adapts to the current aging state of the equipment, the baseline failure threshold needs to be multiplied by the duration degradation factor and then by the power degradation factor to obtain the dynamic failure threshold. For example, the baseline failure threshold is 0.65 × 0.4 × 1.56 = 0.41. After obtaining the dynamic failure threshold, the life cycle health index is compared with the dynamic failure threshold. When the health index is lower than the dynamic failure threshold, it indicates that even under reasonable expectations of the current usage time and power attenuation, the accuracy degradation still exceeds the normal aging range, and it is judged as a sudden accuracy failure.
[0083] For example, taking a mining 3D laser scanner performing a scheduled inspection scanning task in an underground roadway as an example, the reference distance accuracy is 3mm, and the reference angle accuracy is 0.002A. 2 The first health factor is 4.00 mm, and the second health factor is 0.0042. Normalized distance measurement health component = 3. 4.00 = 0.75, the normalized health component equals 0.002. 0.0042 ≈ 0.476. The arithmetic mean of the vibration amplitude throughout the scan is 0.42g, the reference vibration amplitude is 0.1g, and the vibration influence weighting coefficient is 0.1. 0.42 ≈ 0.238, the vibration influence weighting coefficient is taken as the baseline vibration amplitude divided by the arithmetic mean of the entire vibration range, i.e., 0.1. 0.42 ≈ 0.238, not exceeding 1, no truncation required. Ranging fusion weight 0.238, angle fusion weight 0.762. Lifecycle health index = 0.75 × 0.238 + 0.476 × 0.762 ≈ 0.54. Historical health sequence mean 0.95, standard deviation 0.15, k = 2, baseline failure threshold 0.65. Historical duration sequence mean 800h is the reference scan duration, cumulative scan duration 320h, duration degradation factor 0.4. Overall power arithmetic mean 3.2W, overall power attenuation 0.64, power degradation factor 1.56. Dynamic failure threshold = 0.65 × 0.4 × 1.56 equals 0.41. The current health index 0.54 is greater than the dynamic failure threshold 0.41, failure judgment has not been triggered, the equipment can still be used normally.
[0084] In summary, this step normalizes the first and second health factors, dynamically allocates fusion weights using vibration impact weighting coefficients to generate a full lifecycle health index, and constructs a dynamic failure threshold using historical health statistics, cumulative usage time, and power attenuation. Compared with existing technologies, this step has the following advantages: First, by dynamically adjusting the fusion ratio of ranging and angle components through vibration impact weighting coefficients, the health index can maintain the robustness of the assessment under vibration interference, solving the problem of potential distortion in traditional fixed-weight fusion under harsh operating conditions; second, by dynamically correcting the baseline failure threshold through duration degradation factors and power degradation factors, the judgment criteria are appropriately relaxed as the equipment ages naturally, triggering alarms only for abnormal degradation exceeding normal aging expectations, solving the dilemma of traditional fixed thresholds either premature false alarms or late false alarms throughout the equipment's lifecycle.
[0085] The embodiments of this application, through the specific implementation methods described above, achieve the following technical effects: This application proposes a method for full lifecycle health management of mining equipment. First, it simultaneously collects scanner point cloud data along with multi-source operating parameters such as laser emission power, scanning motor current, and casing vibration amplitude. Then, it extracts multiple feature profile contours from the point cloud, registers them with a reference contour, and calculates the set of contour deviations. Next, it corrects the deviations using laser power attenuation and vibration interference to obtain a first health factor characterizing ranging accuracy; simultaneously, it calculates a second health factor characterizing angular accuracy using motor current fluctuations and power attenuation. Finally, it fuses the two health factors to generate a full lifecycle health index, which is compared with a dynamic failure threshold. When the index falls below the threshold, a sudden accuracy failure is determined. This method, through multi-source parameter fusion and dynamic threshold determination, achieves online monitoring and failure early warning of the full lifecycle health status of mining scanners.
[0086] Example 2, as shown in the appendix Figure 3 As shown, based on the inventive concept of a mining equipment lifecycle health management method provided in Embodiment 1, this application also provides a mining equipment lifecycle health management system, specifically including: The data acquisition module 11 is used to acquire the raw point cloud data output in real time when the target scanner performs scanning tasks in the mine roadway, and simultaneously acquire the laser emission power readback value, scanning motor drive current and equipment housing vibration amplitude at the corresponding time. The contour deviation module 12 is used to extract the fitted contour of multiple feature sections of the roadway cross section based on the original point cloud data, calculate the contour deviation between the contour of each feature section and the corresponding reference section contour line, and form a set of deviations. The ranging evaluation module 13 is used to compensate for the contour deviation in the deviation set by using the laser emission power readback value and the vibration amplitude of the equipment housing as correction factors, so as to obtain the first health factor characterizing the ranging accuracy. Angle evaluation module 14 is used to calculate a second health factor characterizing the accuracy of angle measurement based on the fluctuation variance of the scanning motor drive current and the attenuation rate of the laser emission power readback value. The health assessment module 15 is used to fuse the first health factor and the second health factor, and compare them with the original accuracy parameters calibrated by the target scanner in the baseline state to generate the full life cycle health index of the current scanning task. When the full life cycle health index is lower than the dynamic failure threshold, it is determined that the target scanner has experienced a sudden accuracy failure.
[0087] In one embodiment, the data acquisition module 11 is further configured to: control the target scanner to perform a spiral scan of the tunnel cross-section at a fixed angular resolution, and generate raw point cloud data by encoding the echo signal of each scanning cycle according to the timestamp, wherein each point cloud data includes three-dimensional coordinates and the corresponding echo intensity value; extract the laser emission power readback value from the laser emitter drive circuit of the target scanner at the instant of each laser pulse emission to form a power sequence corresponding to each scanning point; continuously acquire the scanning motor drive current from the horizontal rotation motor drive circuit of the target scanner at a sampling rate matching the scanning cycle of the scanner line to form a current sequence associated with each scanning cross-section; attach the accelerometer to the optomechanical base of the target scanner, continuously acquire the vibration acceleration values in three orthogonal directions throughout the scanning process, take the square root of the sum of the squares of the vibration acceleration values in the three directions and take the maximum value within the sliding window as the vibration amplitude of the equipment housing, and establish a time synchronization relationship with the scanning cross-section at the corresponding moment.
[0088] In one embodiment, the contour deviation module 12 is further configured to: parse the scanning timestamp and horizontal rotation angle corresponding to each scanning point from the original point cloud data; combine the inertial navigation pose data built into the target scanner; determine multiple mileage positions along the lane direction at fixed spatial intervals, wherein each mileage position corresponds to a feature section; take the mileage position of each feature section as the center; extract all point cloud data within the half-section spacing range before and after each feature section as the original cross-sectional point cloud slice; statistically analyze the echo intensity value distribution of all point cloud data in the original cross-sectional point cloud slice; take a preset lower quantile as the dynamic intensity lower limit threshold; remove points with echo intensity lower than the dynamic intensity lower limit threshold to obtain the filtered cross-sectional point cloud slice; and process the filtered cross-sectional point cloud slice... The points are sorted according to the horizontal rotation angle. After eliminating outliers using median filtering with a moving window, a closed tunnel cross-section fitting contour is generated using cubic spline interpolation. The reference cross-section contour at the same mileage location is extracted. The reference cross-section contour is the cross-section contour collected and saved by the target scanner when scanning the same mileage location in the reference state and confirming that the ranging and angle measurement accuracy is qualified. The tunnel cross-section fitting contour is spatially registered with the reference cross-section contour. The shortest distance from each point on the registered tunnel cross-section fitting contour to the reference cross-section contour is calculated, and the maximum value among all shortest distances is taken as the contour deviation of the feature section. All feature sections are traversed, and the contour deviations of each feature section are arranged in mileage order to form a set of deviations.
[0089] Furthermore, the contour deviation module 12 is also used for: calculating the local curvature value of each point on the fitted contour line, extracting the local curvature maxima as the first feature point set, and simultaneously extracting the local curvature maxima on the reference section contour line as the second feature point set; establishing candidate matching pairs between each point in the first feature point set and the point with the smallest curvature value difference in the second feature point set, and eliminating mismatched pairs, retaining at least three pairs of valid corresponding feature points; based on the valid corresponding feature points, using the least squares method to solve for the two-dimensional rigid transformation parameters from the fitted contour line coordinate system to the reference section contour line coordinate system, wherein the two-dimensional rigid transformation parameters include a rotation angle and a translation vector; and transferring the local curvature maxima of each point on the fitted contour line to the reference section contour line coordinate system. All points undergo coordinate transformation according to the two-dimensional rigid transformation parameters to obtain the registered fitted contour line. For each point on the registered fitted contour line, the shortest distance from the point to the reference section contour line is calculated. The calculation steps for the shortest distance include: traversing all line segments that constitute the reference section contour line and calculating the shortest distance from the point to each line segment. If the foot of the perpendicular drawn from the point to the line segment falls on a certain line segment, the shortest distance from the point to the corresponding line segment is the length of the perpendicular; otherwise, it is the smaller value between the distances from the point to the two ends of the corresponding line segment. The minimum value among all calculated shortest distances is taken as the shortest distance from the point to the reference section contour line. After traversing all points, the shortest distance corresponding to each point is output.
[0090] Furthermore, the contour deviation module 12 is also used to: denote the points belonging to the fitted contour line among the effective corresponding feature points as the source point set, and the points belonging to the reference section contour line as the target point set; calculate the centroid coordinates of the source point set and the target point set respectively; perform decentering processing on the source point set and the target point set respectively to obtain the source decentering coordinate set and the target decentering coordinate set; construct a covariance matrix based on the source decentering coordinate set and the target decentering coordinate set, and perform singular value decomposition on the covariance matrix to obtain three singular value decomposition matrices; calculate the rotation matrix based on the three singular value decomposition matrices, and extract the rotation angle from the rotation matrix; subtract the result of multiplying the rotation matrix by the centroid coordinates of the source point set from the centroid coordinates of the target point set to obtain the translation vector; and combine the rotation angle and the translation vector as the output of the two-dimensional rigid transformation parameters.
[0091] In one embodiment, the ranging evaluation module 13 is further configured to: extract the laser emission power readback value corresponding to the scanning point of each feature section from the power sequence according to the time synchronization relationship between the scanning section and the feature section, and calculate the arithmetic mean of the power corresponding to each feature section; extract the maximum vibration amplitude within the scanning period corresponding to each feature section from the vibration amplitude of the device housing associated with each feature section; read the reference laser emission power and reference vibration amplitude from the original accuracy parameters calibrated by the target scanner in the reference state; determine the power attenuation rate according to the arithmetic mean of the power of each feature section and the reference laser emission power, and determine the vibration interference rate according to the maximum vibration amplitude of each feature section and the reference vibration amplitude; compensate the contour deviation in the deviation set using the power attenuation rate and the vibration interference rate to obtain the compensated contour deviation of each feature section; calculate the arithmetic mean of the compensated contour deviation of all feature sections, and use the arithmetic mean as the first health factor.
[0092] In one embodiment, the angle evaluation module 14 is further configured to: extract the scanning motor drive current value corresponding to each characteristic section from the current sequence, calculate the fluctuation variance of the current value of each characteristic section during the scanning period; obtain the laser emission power readback value corresponding to each characteristic section from the power sequence, determine the power attenuation rate of each characteristic section according to the relationship between the laser emission power readback value and the reference power value; determine the angle measurement deviation characteristic value of each characteristic section based on the fluctuation variance and power attenuation rate of each characteristic section; and calculate the arithmetic mean of the angle measurement deviation characteristic values of all characteristic sections as a second health factor.
[0093] In one embodiment, the health assessment module 15 is further configured to: extract the reference ranging accuracy value and the reference angle accuracy value from the original accuracy parameters; determine the normalized ranging health component based on the reference ranging accuracy value and the first health factor; determine the normalized angle health component based on the reference angle accuracy value and the second health factor; calculate the arithmetic mean of the vibration amplitude of the device housing throughout the scanning process, and determine the vibration influence weight coefficient according to the ratio of the arithmetic mean to the reference vibration amplitude, wherein the larger the ratio, the smaller the vibration influence weight coefficient; determine the fusion weight of the normalized ranging health component and the fusion weight of the normalized angle health component based on the vibration influence weight coefficient; and perform weighted fusion of the normalized ranging health component and the normalized angle health component according to the corresponding fusion weight to obtain the full life cycle health index.
[0094] Furthermore, the health assessment module 15 is also used to: extract the historical full-lifecycle health index generated by each scanning task of the target scanner in a healthy state from the historical scanning task records, and extract the scanning duration corresponding to each scanning task to form a historical health sequence and a historical duration sequence; calculate the mean and standard deviation of the historical health sequence and determine the baseline failure threshold; calculate the arithmetic mean of the historical duration sequence as a reference scanning duration; obtain the cumulative scanning duration of the target scanner since the last offline verification, and the overall power attenuation degree determined according to the power sequence and the baseline laser emission power; determine the duration degradation factor according to the relationship between the cumulative scanning duration and the reference scanning duration, and determine the power degradation factor according to the overall power attenuation degree; and use the duration degradation factor and the power degradation factor to correct the baseline failure threshold to obtain the dynamic failure threshold.
[0095] This application provides a full lifecycle health management system for mining equipment. This system enables intelligent full lifecycle health management in scenarios such as 3D laser scanning measurement in mines, goaf deformation monitoring, and roadway engineering acceptance. It covers everything from synchronous acquisition of point cloud data and operating parameters to extraction of cross-sectional contour deviations and independent evaluation of ranging and angle measurement accuracy, culminating in dual-factor fusion and dynamic failure determination. The system can be integrated into the embedded control unit of a mining 3D laser scanner or a mining IoT monitoring platform. It effectively improves the online detection capability of sudden attenuation of the scanner's ranging and angle measurement accuracy, reducing the risk of measurement data distortion and misjudgment in mining engineering due to accuracy failure. Simultaneously, it provides structured data support for equipment maintenance planning and calibration cycle optimization, including a full lifecycle health index, a first health factor, a second health factor, and a dynamic failure threshold. For specific detection procedures and identification details of this system, please refer to Example 1.
[0096] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.
Claims
1. A mine equipment full life cycle health management method, characterized in that, The method includes: The system acquires raw point cloud data output in real time when the target scanner performs scanning tasks in the mine roadway, and simultaneously collects the laser emission power readback value, scanning motor drive current and equipment housing vibration amplitude at the corresponding time. Based on the original point cloud data, the fitted contours of multiple feature sections of the roadway cross section are extracted, and the contour deviation between the contour of each feature section and the corresponding reference section contour line is calculated to form a set of deviations. Using the laser emission power readback value and the vibration amplitude of the device housing as correction factors, the contour deviation in the deviation set is compensated to obtain the first health factor characterizing the ranging accuracy; Based on the fluctuation variance of the scanning motor drive current and the attenuation rate of the laser emission power readback value, a second health factor characterizing the accuracy of angle measurement is calculated. The first health factor and the second health factor are fused together and compared with the original accuracy parameters calibrated by the target scanner under the baseline state to generate the full life cycle health index of the current scanning task. When the full life cycle health index is lower than the dynamic failure threshold, it is determined that the target scanner has experienced a sudden accuracy failure.
2. The mine equipment whole life cycle health management method according to claim 1, characterized in that, Acquire raw point cloud data output in real time by the target scanner during scanning operations in a mine roadway, and simultaneously collect the laser emission power readback value, scanning motor drive current, and equipment housing vibration amplitude at corresponding moments, including: The target scanner is controlled to perform a spiral scan of the tunnel cross-section at a fixed angular resolution. The echo signal of each scanning cycle is encoded according to the timestamp to generate the original point cloud data. Each point cloud data includes three-dimensional coordinates and the corresponding echo intensity value. From the laser emitter drive circuit of the target scanner, the laser emission power readback value is extracted at the moment of each laser pulse emission to form a power sequence corresponding to each scanning point. The drive current of the scanning motor is continuously acquired from the horizontal rotation motor drive circuit of the target scanner at a sampling rate matching the scanning cycle of the scanner, forming a current sequence associated with each scanning section; An accelerometer is attached to the optomechanical base of the target scanner. Vibration acceleration values in three orthogonal directions are continuously collected throughout the scanning process. The square root of the sum of the squares of the vibration acceleration values in the three directions is taken as the maximum value within the sliding window and used as the vibration amplitude of the device housing. A time synchronization relationship is established with the scanning section at the corresponding moment.
3. The mine equipment full life cycle health management method of claim 1, wherein, Based on the original point cloud data, the fitted contours of multiple feature sections of the roadway cross-section are extracted, and the contour deviation between the contour of each feature section and the corresponding reference section contour line is calculated to form a set of deviations, including: The scanning timestamp and horizontal rotation angle corresponding to each scanning point are parsed from the original point cloud data. Combined with the inertial navigation pose data built into the target scanner, multiple mileage positions are determined at fixed spatial intervals along the alleyway direction. Each mileage position corresponds to a feature section. Using the mileage position of each feature section as the center, all point cloud data within the distance range of half the section before and after are extracted as the original section point cloud slices of the corresponding feature section. The distribution of echo intensity values of all point cloud data in the original cross-sectional point cloud slice is statistically analyzed. A preset lower quantile is taken as the dynamic intensity lower limit threshold. Points with echo intensity lower than the dynamic intensity lower limit threshold are removed to obtain the filtered cross-sectional point cloud slice. The points in the filtered cross-sectional point cloud slice are sorted according to the horizontal rotation angle. After eliminating outliers by median filtering with a moving window, cubic spline interpolation is used to generate a closed tunnel cross-sectional fitting contour line. Extract the reference cross-sectional profile at the same mileage location, wherein the reference cross-sectional profile is the cross-sectional profile collected and saved by the target scanner when scanning the same mileage location in the reference state and confirming that the distance measurement and angle measurement accuracy are qualified. Spatially register the fitted profile of the roadway cross section with the reference profile, calculate the shortest distance from each point on the fitted profile of the roadway cross section to the reference profile, and take the maximum value among all the shortest distances as the profile deviation of the feature section. Traverse all feature sections and arrange the contour deviations of each feature section in mileage order to form the deviation set.
4. The mine equipment full life cycle health management method of claim 3, wherein, Extract the baseline profile of the same mileage location, spatially register the fitted profile of the roadway profile with the baseline profile, and calculate the shortest distance from each point on the fitted profile of the roadway profile to the baseline profile, including: Calculate the local curvature value of each point on the fitted contour line, extract the local curvature maxima points as the first feature point set, and simultaneously extract the local curvature maxima points on the reference cross-section contour line as the second feature point set. Candidate matching pairs are established between each point in the first feature point set and the point with the smallest curvature value difference in the second feature point set, and mismatched pairs are eliminated, retaining at least three pairs of valid feature points with the same name. Based on the effective corresponding feature points, the least squares method is used to solve the two-dimensional rigid transformation parameters from the fitted contour line coordinate system to the reference section contour line coordinate system, wherein the two-dimensional rigid transformation parameters include a rotation angle and a translation vector. All points on the fitted contour line are transformed according to the two-dimensional rigid transformation parameters to obtain the registered fitted contour line. For each point on the fitted contour line after registration, the shortest distance from the point to the reference section contour line is calculated, wherein the calculation steps for the shortest distance include: Traverse all line segments that constitute the baseline profile and calculate the shortest distance from a point to each line segment. If the foot of the perpendicular line drawn from a point to a line segment falls on a certain line segment, the shortest distance from the point to the corresponding line segment is the length of the perpendicular line; otherwise, it is the smaller of the distances from the point to the two endpoints of the corresponding line segment. Take the minimum value among all the calculated shortest distances as the shortest distance from the point to the reference section profile line; After traversing all points, output the shortest distance for each point.
5. The mine equipment full life cycle health management method of claim 4, wherein, Based on the effective corresponding feature points, the two-dimensional rigid transformation parameters from the fitted contour line coordinate system to the reference section contour line coordinate system are solved using the least squares method, including: Points belonging to the fitted contour line among the effective corresponding feature points are denoted as the source point set, and points belonging to the reference section contour line are denoted as the target point set. The centroid coordinates of the source point set and the centroid coordinates of the target point set are calculated respectively. The source point set and the target point set are respectively decentralized to obtain the source decentralized coordinate set and the target decentralized coordinate set; Based on the source decentralized coordinate set and the target decentralized coordinate set, a covariance matrix is constructed, and singular value decomposition is performed on the covariance matrix to obtain three singular value decomposition matrices. Calculate the rotation matrix based on the three singular value decomposition matrices, and extract the rotation angle from the rotation matrix; The translation vector is obtained by subtracting the rotation matrix multiplied by the centroid coordinates of the source point set from the centroid coordinates of the target point set. The rotation angle and the translation vector are combined and output as the two-dimensional rigid transformation parameters.
6. The mine equipment full life cycle health management method of claim 1, wherein, Using the laser emission power readback value and the vibration amplitude of the device housing as correction factors, the contour deviation in the deviation set is compensated to obtain a first health factor characterizing the ranging accuracy, including: From the power sequence, based on the time synchronization relationship between the scanning section and the feature section, the laser emission power readback value corresponding to the scanning point that makes up each feature section is extracted, and the arithmetic mean of the power corresponding to each feature section is calculated. Extract the maximum vibration amplitude within the scanning period corresponding to each characteristic section from the vibration amplitudes of the device housing associated with each characteristic section; The reference laser emission power and reference vibration amplitude are read from the original accuracy parameters calibrated by the target scanner in the reference state; The power attenuation rate is determined based on the arithmetic mean of the power of each characteristic section and the reference laser emission power, and the vibration interference rate is determined based on the maximum vibration amplitude of each characteristic section and the reference vibration amplitude. The profile deviation in the deviation set is compensated using the power attenuation rate and the vibration disturbance rate to obtain the compensated profile deviation of each feature section. Calculate the arithmetic mean of the compensated contour deviations of all feature sections, and use the arithmetic mean as the first health factor.
7. The mine equipment full life cycle health management method of claim 1, wherein, Based on the fluctuation variance of the scanning motor drive current and the attenuation rate of the laser emission power readback value, a second health factor characterizing the angle measurement accuracy is calculated, including: Extract the scanning motor drive current value corresponding to each characteristic section from the current sequence, and calculate the fluctuation variance of the current value of each characteristic section during the scanning period; The laser emission power readback value corresponding to each characteristic section is obtained from the power sequence, and the power attenuation rate of each characteristic section is determined according to the relationship between the laser emission power readback value and the reference power value. Based on the fluctuation variance of each characteristic section and the power attenuation rate, the characteristic value of the angle measurement deviation of each characteristic section is determined; Calculate the arithmetic mean of the angular measurement deviation characteristics of all characteristic sections, and use it as the second health factor.
8. The mine equipment full life cycle health management method of claim 1, wherein, By fusing the first health factor and the second health factor, and comparing them with the original accuracy parameters calibrated by the target scanner under baseline conditions, a life-cycle health index for the current scanning task is generated, including: Extract the reference distance accuracy value and the reference angle accuracy value from the original accuracy parameters; Based on the benchmark ranging accuracy value and the first health factor, the normalized ranging health component is determined; Based on the reference angle accuracy value and the second health factor, the normalized angle health component is determined; Calculate the arithmetic mean of the vibration amplitude of the device housing throughout the entire scanning process, and determine the vibration influence weighting coefficient based on the ratio of the arithmetic mean to the reference vibration amplitude. The larger the ratio, the smaller the vibration influence weighting coefficient. The fusion weights of the normalized ranging health component and the normalized angle health component are determined based on the vibration influence weighting coefficient. The normalized ranging health component and the normalized angle health component are weighted and fused according to the corresponding fusion weights to obtain the full life cycle health index.
9. The mine equipment full life cycle health management method of claim 1, wherein, The calculation steps for the dynamic failure threshold include: From the historical scan task records, extract the historical full life cycle health index generated by the target scanner in a healthy state for each scan task, and extract the scan duration corresponding to each scan task to form a historical health sequence and a historical duration sequence. Calculate the mean and standard deviation of the historical healthy sequences, and determine the baseline failure threshold; Calculate the arithmetic mean of the historical duration sequence as a reference scan duration; The cumulative scanning time of the target scanner since the last offline verification is obtained, as well as the overall power attenuation determined based on the power sequence and the reference laser emission power; A duration degradation factor is determined based on the relationship between the cumulative scan duration and the reference scan duration, and a power degradation factor is determined based on the overall power attenuation. The dynamic failure threshold is obtained by correcting the baseline failure threshold using the duration degradation factor and the power degradation factor.
10. A mine equipment full life cycle health management system, characterized in that, The system is used to execute the full life-cycle health management method for mining equipment according to any one of claims 1-9, the system comprising: The data acquisition module is used to acquire the raw point cloud data output in real time when the target scanner performs scanning tasks in the mine roadway, and simultaneously acquire the laser emission power readback value, scanning motor drive current and equipment housing vibration amplitude at the corresponding time. The contour deviation module is used to extract the fitted contours of multiple feature sections of the roadway cross section based on the original point cloud data, calculate the contour deviation between the contour of each feature section and the corresponding reference section contour line, and form a set of deviations. The ranging evaluation module is used to compensate for the contour deviation in the deviation set by using the laser emission power readback value and the vibration amplitude of the device housing as correction factors, so as to obtain a first health factor characterizing the ranging accuracy. An angle evaluation module is used to calculate a second health factor characterizing the accuracy of angle measurement based on the fluctuation variance of the scanning motor drive current and the attenuation rate of the laser emission power readback value. The health assessment module is used to integrate the first health factor and the second health factor, and compare them with the original accuracy parameters calibrated by the target scanner under the baseline state to generate the full life cycle health index of the current scanning task. When the full life cycle health index is lower than the dynamic failure threshold, it is determined that the target scanner has experienced a sudden accuracy failure.
