A belt conveyor belt surface detection method and system based on laser scanning

Abstract

The application provides a kind of laser scanning based belt conveyor belt surface detection method and system, it is related to conveying belt detection technical field.The method includes the following steps: obtaining the current three-dimensional profile data and historical three-dimensional profile data set of each physical region;Obtain the topographic difference of physical region and the time variation characteristics of topographic difference by comparison;According to the topographic difference and the time variation characteristics of topographic difference, distinguish the foreign matter or material residue appearing instantaneously in physical region;Trigger alarm to the physical region identified as foreign matter appearing instantaneously.The method of the application aims to solve the problem that the traditional method is prone to false alarm and missed detection in the process of conveying sticky material without increasing the false alarm rate, thereby effectively distinguishing between dangerous foreign matter appearing instantaneously and harmless material residue protrusions accumulating or changing over time, and significantly reducing the false alarm rate and avoiding missed detection while ensuring detection accuracy.

Description

Technical Field

[0001] This invention relates to the field of conveyor belt inspection technology, and more specifically, to a method and system for inspecting the surface of a belt conveyor based on laser scanning. Background Technology

[0002] In the metallurgical, coal, and chemical industries, belt conveyors are core equipment for long-distance, high-capacity, continuous material transport. To ensure equipment safety and production continuity, a laser-scanning-based belt surface detection system is typically installed on the conveyor. This system projects a laser beam onto the belt surface using a laser emitter, and an image acquisition unit captures images of the laser stripes. By calculating the deformation of the laser stripes in the image, the system can reconstruct the three-dimensional contour data of the belt surface. The data processing unit analyzes the three-dimensional contour data to identify protrusions with abnormal height or volume, classifying them as foreign objects that may damage the belt, such as fallen equipment parts or large rocks, and triggers an alarm.

[0003] Under certain specific operating conditions, such as when conveying wet pulverized coal after washing in a coal preparation plant, or when conveying adhesive materials such as mineral powder and soil in high-humidity environments, the situation changes. Due to the high moisture content and stickiness of the materials, not all materials can completely detach from the conveyor belt surface after being unloaded at the discharge point. Some fine, muddy, or paste-like materials adhere to the conveyor belt surface, forming a layer of material residue of uneven thickness and irregular shape. This residue layer is dynamically affected by ambient temperature, humidity, and the moisture content of subsequent batches of materials transported as the conveyor belt repeatedly runs. For example, in dry weather, the residue layer may partially lose water and solidify, forming hard lumps; while in humid weather or after conveying wetter materials, the residue layer may soften or thicken.

[0004] This dynamically changing material residue layer presents a challenge to 3D contour-based foreign object detection. The residue layer itself is not flat; its surface contains numerous protrusions of varying heights and sizes formed by material accumulation. These protrusions appear as anomalous signals in the 3D contour data, potentially causing confusion with the signal characteristics of genuine hazardous foreign objects (such as a piece of metal) in either the height or volume dimensions. If the detection algorithm's threshold is set too sensitively, the system will frequently misclassify harmless material clumps on the residue layer as hazardous foreign objects, leading to a large number of invalid alarms. This high false alarm rate can disrupt the normal work of on-site maintenance personnel, even causing them to distrust the alarm system and ultimately choose to ignore alarms or increase the alarm threshold.

[0005] However, once maintenance personnel set higher height or volume thresholds for detection to reduce false alarm rates, a new and more insidious risk emerges. A truly destructive foreign object, such as a flat but sharp-edged metal plate, or a rock with extremely high hardness but a size below the alarm threshold, might become embedded in a soft, viscous material residue layer upon impact. In this case, most of the foreign object's volume is encased in the residue layer, and its exposed height may be very small, even less than the height of a harmless, large material agglomerate. At this point, detection systems that primarily rely on height or volume for judgment will fail to identify this dangerous foreign object "disguised" or "concealed" by the residue layer because their detection thresholds have been raised. This foreign object will continue to move with the conveyor belt, causing scratches, tears, and other damage to the belt itself, idlers, rollers, and other critical components, leading to safety accidents and production interruptions. Summary of the Invention

[0006] The purpose of this invention is to provide a laser scanning-based method and system for detecting the surface of a belt conveyor. The aim is to solve the problem of false alarms and missed detections in traditional methods under the condition of conveying viscous and wet materials without increasing the false alarm rate. This effectively distinguishes between dangerous foreign objects that appear instantaneously and harmless material residues that accumulate or change over time, thereby significantly reducing the false alarm rate and avoiding missed detections while ensuring detection accuracy.

[0007] In a first aspect, the present invention provides a method for detecting the surface of a belt conveyor based on laser scanning, comprising the following steps:

[0008] Acquire the current 3D contour data and historical 3D contour data set of each physical region on the surface of the belt conveyor; the historical 3D contour data set contains the 3D contour data of each physical region acquired at different time points;

[0009] By comparing the current 3D contour data with the historical 3D contour data set of the corresponding physical region, the morphological differences of the physical region and the temporal variation characteristics of the morphological differences can be obtained.

[0010] Based on the differences in morphology and the time-varying characteristics of these differences, physical areas can be distinguished as foreign objects or material residues that appear instantaneously.

[0011] An alarm is triggered for the physical area where a foreign object is identified as appearing instantaneously.

[0012] The present invention provides a laser scanning-based belt conveyor surface detection method. By establishing a historical three-dimensional contour data set for each region of the belt and comparing the currently acquired three-dimensional contour data with the historical set in a time sequence, the method distinguishes between instantaneously appearing foreign objects and dynamically changing material residues based on the time characteristics of morphological changes. This method ensures detection accuracy while significantly reducing false alarm rate and avoiding missed detections.

[0013] Secondly, the present invention provides a belt surface detection system for a belt conveyor based on laser scanning, comprising:

[0014] The acquisition module is used to acquire the current three-dimensional contour data and historical three-dimensional contour data set of each physical region on the surface of the belt conveyor; the historical three-dimensional contour data set contains the three-dimensional contour data of each physical region acquired at different time points;

[0015] The comparison module is used to compare the current 3D contour data with the historical 3D contour data set of the corresponding physical region to obtain the morphological differences of the physical region and the temporal variation characteristics of the morphological differences.

[0016] The differentiation module is used to distinguish physical areas as instantaneously appearing foreign objects or material residues based on differences in morphology and the time-varying characteristics of these differences.

[0017] The alarm module is used to trigger an alarm in physical areas where foreign objects are identified as appearing instantaneously.

[0018] As can be seen from the above, the laser scanning-based belt conveyor surface detection method provided by this invention effectively distinguishes between dangerous foreign objects appearing instantaneously on the belt surface and dynamically changing material residue layers by merging historical three-dimensional contour datasets of the belt area and performing time-series comparison. Compared with existing methods, this invention significantly reduces the false alarm rate caused by dynamic changes in the material residue layer, while also identifying low-height but destructive foreign objects embedded or attached to the residue layer, thus avoiding missed detections. This greatly improves the accuracy and reliability of the belt conveyor foreign object detection system, ensuring the safe operation of the conveyor system and related equipment, and preventing belt damage, production interruptions, and economic losses caused by foreign objects.

[0019] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing embodiments of the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description and the accompanying drawings. Attached Figure Description

[0020] Figure 1 This is a flowchart of a laser scanning-based belt conveyor surface detection method provided in an embodiment of the present invention.

[0021] Figure 2 This is a schematic diagram of a laser scanning-based belt conveyor surface detection system provided in an embodiment of the present invention.

[0022] Label Explanation:

[0023] 100. Acquisition module; 200. Comparison module; 300. Differentiation module; 400. Alarm module. Detailed Implementation

[0024] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0025] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this invention, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0026] Reference Appendix Figure 1 This invention provides a laser scanning-based method for detecting the surface of a belt conveyor, comprising the following steps:

[0027] Acquire the current 3D contour data and historical 3D contour data set of each physical region on the surface of the belt conveyor; the historical 3D contour data set contains the 3D contour data of each physical region acquired at different time points;

[0028] By comparing the current 3D contour data with the historical 3D contour data set of the corresponding physical region, the morphological differences of the physical region and the temporal variation characteristics of the morphological differences can be obtained.

[0029] Based on the differences in morphology and the time-varying characteristics of these differences, physical areas can be distinguished as foreign objects or material residues that appear instantaneously.

[0030] An alarm is triggered for the physical area where a foreign object is identified as appearing instantaneously.

[0031] Three-dimensional contour data refers to the geometric morphology data of the belt surface acquired through three-dimensional scanning technology. This can be achieved using techniques such as laser triangulation, structured light scanning, or time-of-flight methods. For example, a laser emitter projects a laser beam, and an image acquisition unit captures the laser stripe image. The deformation of the laser stripes in the image is then calculated. The primary purpose is to obtain accurate spatial information about the belt surface. Historical three-dimensional contour data sets refer to the accumulated dataset of multiple three-dimensional scans of the same physical area of ​​the belt surface at different time points. This can be achieved through periodic scanning, continuous monitoring, and time-stamped storage. For example, data can be collected and archived at fixed time intervals or after specific events. The main purpose is to establish a benchmark and trend for the evolution of the belt surface morphology over time. Morphological differences refer to the geometric deviations between the currently acquired three-dimensional contour data of the belt surface and historical data. This can be achieved using methods such as point cloud registration, surface distance calculation, or volume difference. For example, by calculating the Euclidean distance or volume difference between the current contour and the historical average contour, the primary purpose is to quantify the geometric changes between the current and past states of the belt surface. The temporal variation characteristics of morphological differences refer to the dynamic evolution or trend of morphological differences over time. This can be achieved using time series analysis, trend fitting, or anomaly detection algorithms. For example, analyzing the growth rate, duration, or degree of abrupt change of morphological differences is mainly to identify whether the morphological change is an instantaneous event or a cumulative process. Distinguishing between transiently occurring foreign objects or material residues in physical areas refers to the process of judging the nature of abnormal areas on the belt surface based on morphological differences and their temporal variation characteristics. This can be achieved using rule-based expert systems, machine learning classifiers, or threshold judgment logic. For example, setting a threshold to judge the degree of abrupt change of morphological differences is mainly to accurately identify foreign objects that pose a potential threat to the belt and eliminate harmless material residues.

[0032] The working principle of this invention is based on the analysis of the temporal characteristics of the three-dimensional morphological changes on the surface of a belt conveyor. By continuously and periodically collecting three-dimensional contour data of various regions of the belt, the system establishes and maintains a historical morphological database for each region. When the three-dimensional contour data of a belt region at the current moment is obtained, the system compares it with the historical morphological data of that region and calculates the morphological differences. The key lies in analyzing the trend of these differences over time: if the morphological change is sudden and occurs significantly within a short period of time, it is judged as an instantaneously appearing foreign object; if the morphological change is gradual and accumulates or evolves slowly over time, it is judged as a dynamic change in the material residue layer itself. Through this time-series analysis, this invention can effectively distinguish between truly dangerous foreign objects that require alarm and harmless material residue protrusions.

[0033] The core innovation of this application lies in the fact that by introducing the analysis of the time-varying characteristics of the surface morphology differences of the belt and combining it with the morphology differences, it is possible to effectively distinguish between instantaneous foreign objects and cumulative, dynamically changing material residues, thereby reducing the false alarm rate and improving the accuracy of foreign object detection.

[0034] Specifically, this method first continuously or periodically acquires the current three-dimensional contour data of each physical region on the surface of the belt conveyor, and combines this with a pre-accumulated set of historical three-dimensional contour data to lay the data foundation for subsequent dynamic analysis. The historical data set contains morphological information of the same physical region at different points in time, enabling the system to track the evolution trajectory of the belt surface morphology. Subsequently, the newly acquired current three-dimensional contour data is compared with the historical three-dimensional contour data set of the corresponding physical region. This comparison process not only calculates the geometric deviation between the current morphology and the historical morphology, i.e., morphological differences, but more importantly, through dynamic comparison and analysis with historical data from multiple time points, the system can identify the dynamic evolution law of this morphological difference over time, i.e., the temporal variation characteristics of morphological differences. For example, for a foreign object that falls instantaneously, its morphological difference will show a sudden and obvious abrupt change; while for material residue, its morphological difference may show a slow accumulation or fluctuations affected by the environment. After acquiring information on morphological differences and their temporal variations, the system uses this comprehensive information to classify physical areas into those containing transiently appearing foreign objects or those representing normally accumulated and changing material residues. This differentiation mechanism is the core of this solution, enabling the system to effectively avoid misclassifying harmless material clumps as hazardous foreign objects. Ultimately, alarms are triggered only in physical areas identified as transiently appearing foreign objects, ensuring alarm accuracy, preventing invalid alarms from interfering with maintenance work, and guaranteeing the safety of conveyor operation.

[0035] As a preferred embodiment, the solution of this application is implemented as follows: A laser 3D scanning system is installed above the belt conveyor. This system includes a line laser emitter and an industrial camera. The laser emitter projects a line laser onto the belt surface, while the industrial camera simultaneously acquires images formed by the laser line on the belt surface. The image processing unit calculates the deformation of the laser line to generate 3D point cloud data of each physical region on the belt surface in real time, and converts it into 3D contour data. This current 3D contour data is transmitted to a data processing server. Simultaneously, the server continuously stores and maintains a historical 3D contour data set, which contains 3D contour data of the same physical region acquired at different points in time over a period of time (e.g., the last few hours or days). When new current 3D contour data arrives, the data processing server compares it with the historical 3D contour data set of the corresponding physical region. Specifically, the server first performs point cloud registration between the current contour data and the historical data, calculates the geometric deviation between the two, and thus obtains the morphological differences of the physical region. Subsequently, by performing time-series analysis on historical data sets, such as using moving averages or trend fitting algorithms, the evolution trend of morphological differences over time is evaluated, thereby determining the temporal variation characteristics of morphological differences. For example, if morphological differences undergo a large abrupt change in a very short period of time, their temporal variation characteristics are instantaneous; if morphological differences exhibit slow accumulation or periodic fluctuations, their temporal variation characteristics are gradual. Based on the magnitude of morphological differences and their temporal variation characteristics, the data processing server uses preset judgment logic or classification models to distinguish and identify the physical area as either an instantaneously appearing foreign object or material residue. For example, when the abrupt change in morphological differences exceeds a certain threshold, and its temporal variation characteristics indicate non-cumulative, it is determined to be a foreign object. Once a physical area is identified as an instantaneously appearing foreign object, the system immediately sends an alarm signal to the field control system through the communication interface, triggering an audible and visual alarm device, and displays the location information of the foreign object on the operator interface so that maintenance personnel can handle it promptly.

[0036] By employing the above-described solution, this application effectively addresses the problem of false alarms and missed alarms caused by the difficulty of traditional detection methods in accurately distinguishing between transient foreign objects and material residues when a dynamically changing material residue layer exists on the surface of a belt conveyor. This solution, by introducing analysis of the time-varying characteristics of morphological differences, achieves accurate differentiation between foreign objects and material residues, reduces the false alarm rate, and improves the accuracy of foreign object detection, thereby ensuring the safe operation of the belt conveyor and the continuity of production.

[0037] In some embodiments, the step of comparing the current 3D contour data with a set of historical 3D contour data for the corresponding physical region to obtain the morphological differences of the physical region and the temporal variation characteristics of the morphological differences includes:

[0038] By performing time-series analysis on historical 3D contour data sets, the evolution law of the material residue layer morphology in the physical region is obtained; the evolution law characterizes the cumulative or changing trend of the material residue layer morphology over time.

[0039] Based on the evolutionary pattern, the morphological differences of the physical region are obtained by dynamically comparing the current three-dimensional contour data with the historical three-dimensional contour data set.

[0040] Based on the degree of deviation between morphological differences and evolutionary patterns, the temporal variation characteristics of morphological differences are determined; the degree of deviation characterizes the degree of conformity or non-conformity between morphological differences and evolutionary patterns.

[0041] Time-series analysis of historical 3D contour datasets refers to the processing and pattern recognition of continuously acquired historical 3D contour data over a period of time. This can employ statistical methods, such as moving averages, exponential smoothing, and regression analysis, or machine learning methods, such as recurrent neural networks and long short-term memory networks, to identify trends, periodicity, or randomness in the data over time. The aim is to extract the inherent patterns of material residue morphology changes over time from historical data. The evolutionary pattern of material residue morphology characterizes the cumulative or changing trend of material residue morphology over time. It can be a mathematical model, a statistical model, or a rule-based knowledge base, used to describe the expected change patterns of characteristics such as residue height, volume, and surface roughness at different time scales, providing a benchmark for judging the nature of current morphological differences. Dynamic comparison refers to adjusting or predicting historical data by combining the evolution patterns of material residue morphology when comparing current 3D contour data with historical 3D contour data sets. This results in a benchmark that better reflects the current situation. It can be achieved by comparing current 3D contour data with the residue morphology predicted based on evolution patterns, or by comparing current data with historical data corrected for evolution patterns. The aim is to filter out morphological differences caused by normal evolution of the material residue morphology and more accurately identify abnormal morphologies. The degree of deviation characterizes the degree to which the morphological difference conforms to or does not conform to the evolution patterns of the material residue morphology. This can be achieved by calculating the absolute difference, relative difference, or mean square error between the morphological difference and the predicted value, or by defining a threshold range to determine whether it falls within the normal fluctuation range. This quantifies the abnormality of the morphological difference and thus determines its temporal variation characteristics.

[0042] The overall operational logic of this solution lies in optimizing the identification and judgment process of surface morphology differences on conveyor belts by introducing in-depth analysis of the evolution law of material residue layer morphology. First, the system performs time-series analysis on a long-term accumulated set of historical 3D contour data to learn and establish the accumulation or change trend of material residue layer morphology in the physical area over time, i.e., to obtain its evolution law. This law is key to understanding the dynamic behavior of the residue layer, characterizing the expected change pattern of the residue layer under normal operating conditions. Subsequently, when acquiring current 3D contour data, the system no longer simply compares it with a fixed historical benchmark, but dynamically adjusts or predicts the historical data based on the acquired evolution law, thus performing dynamic comparison. This dynamic comparison can effectively predict the morphology that the residue layer should have at the current moment, allowing the calculated morphology differences to more accurately reflect the actual changes caused by abnormal factors (such as foreign objects), rather than the normal accumulation or dissipation of the residue layer itself. Finally, to determine the temporal variation characteristics of the morphology differences, the system further evaluates the degree of deviation between the currently identified morphology differences and the expected evolution law. If the morphological differences significantly deviate from the evolutionary pattern, indicating a large deviation, it can be identified as a momentary anomaly, such as a foreign object. Conversely, if the morphological differences largely conform to the evolutionary pattern, indicating a small deviation, it can be identified as the normal evolution of the material residue layer. Through this step-by-step and interconnected processing, this solution can effectively distinguish between momentary foreign objects and the normal dynamic changes of the material residue layer. Therefore, when calculating morphological differences and their temporal variation characteristics, it can effectively adapt to the dynamic evolution of the material residue layer, avoiding misjudging normal changes in the residue layer as foreign objects, or obscuring actual foreign objects due to excessive changes in the residue layer. This refined comparison and judgment mechanism enables the system to achieve higher accuracy and robustness in distinguishing between momentary foreign objects and material residue in physical areas, thereby improving the reliability of triggering alarms for momentary foreign objects.

[0043] In a specific embodiment, this method can be implemented as follows: First, for each physical region on the surface of the belt conveyor, the system can continuously acquire and store its historical three-dimensional contour data, for example, recording the three-dimensional contour data of the region once every fixed time interval (e.g., every minute). To obtain the evolution law of the material residue layer morphology, the system can perform time-series analysis on these historical three-dimensional contour data sets. For example, for each physical region, using its historical three-dimensional contour data over a past period (e.g., the past 24 hours), the system can calculate the moving average of its features such as average height, volume, or surface roughness, or train a time series prediction model (e.g., an autoregressive moving average model) to establish the cumulative or changing trend of the material residue layer morphology of the region over time. When the current three-dimensional contour data is acquired, the system can predict the expected residue layer morphology of the physical region at the current moment based on the previously established evolution law. Subsequently, the system can perform dynamic comparison, calculating the geometric difference between the current three-dimensional contour data and the predicted expected residue layer morphology point by point or region by region, for example, calculating the height difference of each corresponding grid point, thereby obtaining the morphological difference of the physical region. Finally, to determine the temporal variation characteristics of the morphological differences, the system can assess the degree of deviation between this morphological difference and the evolutionary pattern. For example, a threshold can be set: if the current value of the morphological difference or its rate of change exceeds the normal fluctuation range defined by the evolutionary pattern, it is considered a large deviation, indicating that it may be a transient foreign object; conversely, if the morphological difference is within the normal fluctuation range, it is considered a small deviation, indicating that it may be a normal accumulation or change of material residue. In this way, the system can more accurately identify and judge abnormalities on the belt surface.

[0044] This solution incorporates analysis of the evolution patterns of material residue morphology and, based on this, performs dynamic comparison and deviation assessment, effectively adapting to the dynamic evolution of material residue on the surface of belt conveyors. This allows the system to accurately distinguish between normal accumulation or change in the material residue layer and instantaneously appearing foreign matter when calculating morphological differences and their temporal variation characteristics. This avoids misjudging normal changes in the residue layer as foreign matter and prevents excessive changes in the residue layer from masking actual foreign matter. Therefore, this solution significantly reduces false alarms while effectively preventing missed alarms, improving the robustness and accuracy of detection.

[0045] In some embodiments described above, this application proposes to obtain the evolution pattern of the material residue layer morphology in a physical area by performing time-series analysis on historical three-dimensional contour data sets, which can then be used as a benchmark for comparing with current three-dimensional contour data. Specifically, obtaining this evolution pattern can be achieved by analyzing the cumulative or changing trends of historical three-dimensional contour data over time. For example, a time series model or regression model can be established to predict the expected morphology of the residue layer at a future point in time, thus providing a dynamic benchmark for distinguishing between instantaneously appearing foreign objects or material residue. However, in its implementation, in the specific scenario of conveying viscous and wet materials on a belt conveyor, the morphological evolution pattern of the material residue layer is not fixed; it is affected by dynamic environmental factors such as ambient temperature, humidity, and the moisture content of the conveyed material. For example, in dry weather, the residue layer may rapidly lose water and solidify, leading to a faster rate of morphological change; while in humid weather or after conveying wetter materials, the residue layer may soften or thicken, and its evolution pattern changes accordingly. Therefore, how to obtain adaptive evolutionary patterns that can adapt to or reflect the influence of these dynamic environmental factors, and avoid misjudgment or missed detection due to the system using fixed patterns that are not adapted to the current environment when environmental conditions change, has become a problem that needs to be solved.

[0046] In some embodiments, the step of obtaining the evolution pattern of the material residue layer morphology in a physical region by performing time-series analysis on a historical three-dimensional contour data set includes:

[0047] Obtain the set of environmental parameters for the time point corresponding to the historical 3D contour data set; the set of environmental parameters includes at least one of the temperature, humidity, and moisture content of the conveyed material at the time point.

[0048] Based on historical 3D contour data sets and environmental parameter sets, by analyzing the correlation between the morphology of the material residue layer in the physical region and the environmental parameters, parameter adjustment rules for the evolution law of the material residue layer morphology are obtained; the parameter adjustment rules characterize the trend of the evolution law of the material residue layer morphology with the change of environmental parameters.

[0049] Get the environmental parameters at the current time point;

[0050] Based on the environmental parameters and parameter adjustment rules at the current time point, the parameters for the evolution law of the material residue layer morphology are adjusted to obtain the evolution law of the material residue layer morphology adapted to the current environment.

[0051] To address the aforementioned issues, this application, while performing time-series analysis on historical 3D contour data sets to obtain the evolution patterns of material residue layer morphology in physical regions, further incorporates environmental parameters. Specifically, this step includes acquiring a set of environmental parameters corresponding to the historical 3D contour data set at specific time points. This set of environmental parameters may include at least one of the following: temperature, humidity, and moisture content of the conveyed material at that time point. The set of environmental parameters refers to data related to the environmental conditions at that time, recorded or associated synchronously when acquiring the historical 3D contour data. These parameters can be collected in real time by environmental sensors (e.g., temperature sensors, humidity sensors) installed near the belt conveyor, or obtained through material analysis equipment (e.g., online moisture content analyzers). This data can be matched with the timestamps of the 3D contour data to form the environmental parameter set at the corresponding time point. The purpose of introducing environmental parameters is to capture the external influences on the evolution of the material residue layer morphology, thereby enabling the establishment of a more comprehensive evolution model that considers environmental factors, providing a data foundation for subsequent adaptive adjustments.

[0052] Based on this, and using historical 3D contour data sets and environmental parameter sets, by analyzing the correlation between the morphology of the material residue layer in the physical region and environmental parameters, parameter adjustment rules for the evolution of the material residue layer morphology can be obtained. These parameter adjustment rules characterize the trend of the evolution of the material residue layer morphology with changes in environmental parameters. The correlation refers to the statistical or physical relationship between specific characteristics of the material residue layer morphology (e.g., thickness, roughness, degree of solidification) and environmental parameters (temperature, humidity, moisture content). The parameter adjustment rule is a model or function that quantifies this correlation; it describes how which parameters (e.g., evolution rate, final stable thickness, solidification time, etc.) in the evolution of the material residue layer morphology will adjust when environmental parameters change. The analysis of the correlation can employ various data analysis and modeling techniques, such as multiple regression analysis, machine learning algorithms, time series analysis, and cross-analysis of environmental factors. Through these methods, a mathematical model can be established that can predict or adjust the specific parameters of the material residue layer morphology evolution based on the input environmental parameters. This step aims to learn the mechanism by which environmental factors influence the evolution of residual layer morphology from historical data, thereby generating a basis for dynamic adjustment. This transforms the evolutionary pattern from a static model to a dynamically adjustable model, which is the core of achieving adaptability.

[0053] Subsequently, the environmental parameters at the current time point are acquired. These parameters refer to data such as ambient temperature, humidity, and moisture content of the conveyed material, collected in real-time or near real-time during the current 3D contour data comparison. This can be achieved using the same sensor system as historical data acquisition; for example, environmental sensors installed in the conveyor's detection area can read temperature and humidity data in real time, and an online material analyzer can be used to obtain the current moisture content of the conveyed material. This data is transmitted to the data processing unit in real time. Acquiring the current environmental parameters provides the latest environmental information as input for application parameter adjustment rules, enabling real-time adjustments to the evolution of the material residue layer morphology and ensuring the accuracy and timeliness of the comparison benchmark.

[0054] Finally, based on the environmental parameters and parameter adjustment rules at the current time point, the parameters for the evolution law of the material residue layer morphology are adjusted to obtain the evolution law of the material residue layer morphology adapted to the current environment. Adjusting the parameters for the evolution law of the material residue layer morphology means applying the previously learned parameter adjustment rules to the currently acquired environmental parameters, thereby calculating the specific parameter values ​​of the evolution law of the material residue layer morphology under the current environment. The resulting "evolution law of the material residue layer morphology adapted to the current environment" is a dynamically updated model that can accurately reflect the expected behavior of the residue layer under the current actual working conditions. The data processing unit receives the current environmental parameters and inputs them into the pre-established parameter adjustment rule model. This model outputs the adjusted evolution law parameters based on the current environmental parameters. This step is crucial for achieving the adaptability of the evolution law. It ensures that the benchmark model used for foreign object detection can dynamically adapt to constantly changing environmental conditions.

[0055] The detection method of this application, after acquiring the current three-dimensional contour data and historical three-dimensional contour data sets of each physical region on the surface of the belt conveyor, needs to compare the current three-dimensional contour data with the historical three-dimensional contour data sets of the corresponding physical regions in order to accurately distinguish instantaneously appearing foreign objects or material residues. This comparison yields the morphological differences of the physical regions and their temporal variation characteristics. During this comparison process, time-series analysis of the historical three-dimensional contour data sets is performed to obtain the evolution law of the material residue layer morphology in the physical regions. This evolution law characterizes the accumulation or change trend of the material residue layer morphology over time. Considering that the morphological evolution law of the material residue layer is not fixed when the belt conveyor is conveying viscous and wet materials, but is affected by dynamic environmental factors such as ambient temperature, humidity, and the moisture content of the conveyed material, this application further optimizes the steps for obtaining the morphological evolution law of the material residue layer to ensure that the acquired evolution law can adapt to or reflect the influence of these dynamic environmental factors. This avoids misjudgment or missed detection due to the system using fixed laws that are not adapted to the current environment when environmental conditions change.

[0056] Specifically, firstly, the system acquires a set of environmental parameters corresponding to the historical 3D contour data set at a given time point. This set of environmental parameters may include at least one of the following: temperature, humidity, and moisture content of the transported material at that time point. By correlating the historical 3D contour data with the environmental parameters at that time, a foundation is laid for subsequent analysis of the interaction between the material residue layer morphology and environmental parameters. Secondly, based on the acquired historical 3D contour data set and the corresponding environmental parameter set, the system conducts an in-depth analysis of the correlation between the material residue layer morphology and environmental parameters in the physical area. This analysis aims to reveal how environmental factors affect the morphological evolution process of the residue layer, such as accumulation, solidification, softening, or thickening. Through this analysis, parameter adjustment rules for the evolution law of the material residue layer morphology can be obtained. These parameter adjustment rules quantitatively characterize the trend of the evolution law of the material residue layer morphology with changes in environmental parameters. For example, when the temperature increases, the solidification rate of the residue layer may accelerate; when the humidity increases, the thickening rate of the residue layer may increase. The establishment of this rule transforms the evolution law from a static model to a dynamically adjustable model, which is the core of achieving adaptability. Next, before comparing the current 3D contour data, the system acquires the environmental parameters at the current time point in real time. These parameters reflect the actual environmental conditions under the current operating conditions. Finally, the system adjusts the parameters of the evolution law of the material residue layer morphology based on the environmental parameters at the current time point and the previously obtained parameter adjustment rules. By substituting the current environmental parameters into the parameter adjustment rules, the most suitable evolution law parameter values ​​for the current environment can be calculated, thus obtaining an evolution law that accurately reflects the expected morphology of the material residue layer under the current actual operating conditions. Through the above steps, this application makes the evolution law of the material residue layer morphology adaptive. This means that no matter how the ambient temperature, humidity, or material moisture content changes, the system can dynamically adjust its expectation of the material residue layer morphology, thereby ensuring that the benchmark used to compare the current three-dimensional contour data is always accurate and in line with reality. This dynamic adjustment capability improves the accuracy and reliability of the detection system in distinguishing between instantaneously appearing foreign objects and material residues, avoiding the problems of misjudgment or missed detection caused by environmental changes in traditional methods.

[0057] In one specific embodiment, this method can be implemented as follows: Multiple environmental sensors, such as a digital temperature sensor and a humidity sensor, can be deployed above and to the side of the belt conveyor's detection area to collect ambient temperature and humidity data in real time. Simultaneously, a microwave or infrared-based online moisture content detector can be installed near the material unloading point to obtain the moisture content of the conveyed material in real time. The data acquisition frequency of these sensors and detectors can be set to once per minute and synchronized with the timestamp of the three-dimensional contour data acquired by the laser scanning system. Historical data is stored in a database, where each historical three-dimensional contour data record is associated with a set of environmental parameters for its corresponding time point. To obtain parameter adjustment rules for the evolution of the material residue layer morphology, a machine learning model can be used for training. For example, a multi-input multi-output neural network model can be constructed. The input layer of this model receives key morphological features from the historical three-dimensional contour data set (e.g., average residue layer thickness, surface roughness, volume change rate, etc.) and the corresponding set of environmental parameters (temperature, humidity, moisture content). The model's output layer corresponds to parameters representing the evolution of the material residue layer morphology, such as a parameter indicating the residue layer thickness growth rate and a parameter indicating the degree of residue layer solidification. Through training on a large amount of historical data, this neural network model can learn the non-linear relationship between morphological features and environmental parameters, thus forming parameter adjustment rules. For example, if training data shows that the residue layer thickness growth rate slows down and the degree of solidification increases under high temperature and low humidity conditions, the model will learn and encode this trend. In actual operation, when it is necessary to inspect the current conveyor belt surface, the system first uses the aforementioned environmental sensors and an online moisture content detector to obtain the ambient temperature, humidity, and moisture content of the conveyed material in real time. These real-time environmental parameters are input into the pre-trained neural network model. Based on these inputs, the neural network model calculates and outputs adjusted parameters for the evolution of the material residue layer morphology in real time. For example, if the current ambient temperature is high and the humidity is low, the model may output a higher solidification rate parameter and a lower thickness growth rate parameter, allowing the evolution law to more accurately predict the expected morphology of the residue layer under the current dry environment. Thus, the system obtains an evolution law for the material residue layer morphology adapted to the current environment. This dynamically adjusted evolution pattern was then compared with the current 3D contour data to more accurately identify real foreign objects and effectively distinguish material residues that change due to environmental influences.

[0058] This application introduces environmental parameters and establishes parameter adjustment rules, enabling the evolution of the material residue layer morphology to be dynamically adjusted according to actual environmental conditions. This solves the problem of traditional methods where the evolution pattern is fixed and cannot adapt to dynamic changes in environmental temperature, humidity, and the moisture content of the transported material. By acquiring environmental parameters corresponding to historical data and analyzing their correlation with morphology evolution, an adjustment rule characterizing the trend of evolution pattern changes with environmental parameters can be obtained. During actual detection, the evolution pattern is adjusted in real time according to the current environmental parameters, ensuring that the benchmark model used for comparison can always accurately reflect the actual state of the material residue layer under the current operating conditions. This avoids the system misclassifying normal material residue as foreign matter or vice versa when environmental conditions change due to the use of fixed rules that are not adapted to the current environment, thereby improving the accuracy and reliability of the detection system in distinguishing between foreign matter and material residue.

[0059] In some embodiments, the step of adjusting the parameters of the evolution law of the material residue layer morphology according to the environmental parameters at the current time point and the parameter adjustment rules to obtain the evolution law of the material residue layer morphology adapted to the current environment includes:

[0060] Acquire local environmental feature information for each physical region; the local environmental feature information characterizes at least one of the following: local location, local ventilation conditions, and material accumulation thickness of the physical region.

[0061] Based on local environmental characteristics, the parameter adjustment rules are modified in a regionalized manner to obtain the regionalized parameter adjustment rules for each physical region.

[0062] For each physical region, based on the environmental parameters at the current time point and the corresponding regional parameter adjustment rules for the physical region, the parameters for the evolution law of the material residue layer morphology in the physical region are adjusted to obtain the evolution law of the material residue layer morphology adapted to the current environment of the physical region.

[0063] Local environmental characteristic information refers to data characterizing the micro-environmental properties of a physical region. It can reflect at least one of the following: local location, local ventilation conditions, and material accumulation thickness. This information can be acquired by deploying multiple local sensor arrays along the belt conveyor. For example, local airflow can be measured using a small anemometer, local material accumulation height can be measured using an ultrasonic sensor or laser rangefinder, or the precise coordinates of the physical region on the belt can be determined using encoders and position sensors. Alternatively, material accumulation thickness or local solidification degree can be indirectly inferred from historical 3D contour data through local texture or morphological analysis, thus serving as part of the local environmental characteristic information. Parameter adjustment rules refer to the logic or model used to guide how the parameters governing the evolution of the material residue layer morphology change according to environmental parameters. This can be a pre-defined lookup table, a mathematical function model (e.g., a linear or nonlinear regression model), or a predictive model trained using machine learning (e.g., a neural network model or a support vector machine model). The purpose of these rules is to ensure that the evolution of the material residue layer morphology dynamically adapts to changes in the overall environment. Regionalization correction refers to the process of adjusting or optimizing existing parameter adjustment rules based on local environmental characteristics of a specific physical region, making them more suitable for that region. This correction can employ various methods, such as locally adjusting the weight coefficients in the parameter adjustment rules, introducing a correction factor related to local environmental characteristics, or dynamically selecting different sets of sub-rules based on local environmental characteristics. The goal of correction is to enable the adjustment rules to reflect the impact of microenvironmental differences in different regions on the evolution of material residue layers. Regionalized parameter adjustment rules are parameter adjustment rules specifically generated for a particular physical region after regionalization correction. This can manifest as generating an independent set of adjustment parameters for each physical region, or adding a specific correction term to each region based on the original global rules. This rule can more accurately guide the parameter adjustment of the material residue layer morphology evolution law in the corresponding physical region, thereby improving the precision of predicting the behavior of material residue layers in that region.

[0064] This solution aims to regionalize and refine the evolution pattern of material residue layer morphology by introducing local environmental characteristic information, thereby making the obtained evolution pattern more accurately adaptable to the actual environment of each physical region and improving the accuracy of morphology difference judgment. First, the system acquires local environmental characteristic information for each physical region. This information reflects the micro-environmental differences in different areas of the conveyor belt, such as local location, local ventilation conditions, or material accumulation thickness. Traditional environmental parameters (such as temperature and humidity) are usually global, but different locations on the conveyor belt may form unique local environments due to differences in airflow, structural obstruction, or material accumulation. This local characteristic information can reflect the differences in the drying, solidification, or thickening rate of the material residue layer in different regions, providing basic data for subsequent refined adjustments. Second, based on this local environmental characteristic information, the system regionally modifies the previously established parameter adjustment rules, thereby generating a unique set of regional parameter adjustment rules for each physical region. This process no longer uses a single global rule to adjust the evolution pattern of all regions, but rather modifies the original parameter adjustment rules according to the unique local environmental characteristics of each physical region, generating a set of rules applicable to that region. This correction considers the influence of local factors on the evolution of the material residue layer morphology, making the adjustment rules more targeted and able to more accurately reflect the behavior of the material residue layer in a specific area. Finally, for each physical area on the conveyor belt, the system adjusts the parameters of the evolution of the material residue layer morphology in that area based on the overall environmental parameters at the current time point and the corresponding regional parameter adjustment rules. In this way, the resulting evolution pattern not only considers the influence of the overall environment, but more importantly, it precisely adapts to the local environmental conditions of that physical area. This regional adjustment mechanism is closely integrated with the overall detection process. By providing a more refined and accurate evolution pattern of the material residue layer morphology, this solution can significantly improve the accuracy of dynamic comparison between current 3D contour data and historical data in subsequent steps. When the degree of deviation between the morphological difference and this refined evolution pattern is determined, the system can more effectively identify real transient foreign objects and distinguish them from normally accumulated or changing material residue layers. This consideration of local environmental differences makes the prediction of the expected behavior of the material residue layer morphology in each physical area more accurate. As a result, in subsequent morphology difference comparisons, it is possible to more effectively distinguish instantaneous foreign objects or material residues, reduce false alarm rates, and improve the reliability of detection.

[0065] In one specific embodiment, to achieve regionalized adjustment of the evolution law of the material residue layer morphology, a micro-sensor array can be deployed in different physical areas of the belt conveyor surface. For example, a sensor node integrating temperature, humidity, a small anemometer, and a laser ranging module can be installed at regular intervals above the belt. These sensor nodes can periodically collect local environmental characteristic information of their respective physical areas, such as local temperature, local humidity, local airflow velocity, and local thickness of material accumulation. This data can be transmitted to a central processing unit. After receiving this local environmental characteristic information, the central processing unit can regionally modify the parameter adjustment rules according to a pre-established correction model. For example, the parameter adjustment rule can be a polynomial function whose coefficients are related to the overall environmental parameters. During regionalized modification, a correction factor matrix can be introduced, where each element of the matrix is ​​associated with the local environmental characteristic information of the corresponding physical area. For example, if the local ventilation conditions of a certain physical area are poor, a decreasing correction factor can be applied to the material drying rate parameter of that area; if the material accumulation thickness is large, an increasing correction factor can be applied to the material solidification rate parameter. This correction factor matrix can be trained and optimized using regression analysis or machine learning algorithms based on historical data and local environmental characteristics. Once the regionalized parameter adjustment rules for each physical region are obtained, the central processing unit can adjust the parameters of the evolution law of the material residue layer morphology for each physical region based on the overall environmental parameters at the current time point (e.g., global temperature and humidity obtained from the main environmental sensor) and the specific regionalized parameter adjustment rules for that physical region. For example, for a physical region at the front end of the conveyor belt, where local ventilation conditions may be better and the material residue layer dries faster, the regionalized parameter adjustment rules will adjust the drying rate parameter in the evolution law to be higher; while for a physical region in the middle of the conveyor belt with thicker material accumulation, the regionalized parameter adjustment rules will adjust the solidification rate parameter in the evolution law to be higher. In this way, the evolution law of the material residue layer morphology in each physical region can accurately reflect its unique local environmental conditions, thus providing a more accurate benchmark for subsequent morphology difference judgment.

[0066] This scheme acquires local environmental characteristic information of each physical region and uses this information to regionally modify the parameter adjustment rules, thereby generating regionalized parameter adjustment rules for each physical region that adapt to its unique local environment. This method allows the parameter adjustment of the evolution law of the material residue layer morphology to no longer be limited to uniform global environmental parameters, but can fully consider the micro-environmental differences of different physical regions on the conveyor belt. Therefore, the obtained evolution law of the material residue layer morphology can more accurately and regionally reflect the actual situation of each physical region, thus significantly improving the accuracy of judging morphological differences in specific physical regions. This helps to more effectively identify transient foreign objects and distinguish them from normally accumulated or changing material residue layers, thereby reducing false alarm rates and improving the reliability of the entire detection system.

[0067] In some embodiments, the step of regionalizing the parameter adjustment rules based on local environmental feature information to obtain regionalized parameter adjustment rules for each physical region includes:

[0068] Obtain local environmental feature information of the physical region to be corrected;

[0069] Acquire local environmental feature information of physical regions that have spatial proximity to the physical region to be corrected;

[0070] Based on the relative positional relationship between the physical area to be corrected and the spatially adjacent physical areas, as well as the conveying direction of the material on the belt conveyor, determine the degree of local environmental impact of the spatially adjacent physical areas on the physical area to be corrected.

[0071] By combining the local environmental characteristics of the physical region to be corrected, the local environmental characteristics of physical regions with spatial proximity, and the degree of local environmental influence, the parameter adjustment rules are corrected to obtain the regionalized parameter adjustment rules for the physical region to be corrected.

[0072] The physical region to be corrected refers to the specific physical region on the belt conveyor surface where the parameter adjustment rules for the evolution law of the material residue layer morphology need to be corrected. A spatially adjacent physical region refers to a physical region that is spatially adjacent to or close to the physical region to be corrected. This can include directly adjacent regions, regions located upstream of the physical region to be corrected, or regions indirectly affected by media such as airflow or water flow. The relative positional relationship refers to the geometrical positional relationship between the physical region to be corrected and the spatially adjacent physical regions on the belt conveyor surface. This can include upstream, downstream, left, right, or diagonal directions. The degree of local environmental impact refers to the quantitative assessment of the impact of the local environmental characteristics of the spatially adjacent physical regions on the local environmental characteristics of the physical region to be corrected. This can be expressed using weighting coefficients, influence factors, or numerical values ​​calculated based on physical models.

[0073] Based on the above-described features, the overall working principle of this scheme is as follows: To address the aforementioned issues, this scheme incorporates consideration of cross-regional correlations or influences when regionalizing parameter adjustment rules. Specifically, it first acquires the local environmental characteristics of the physical region to be corrected, providing fundamental environmental data for subsequent corrections. Simultaneously, to overcome the limitations of a single region, the system further acquires the local environmental characteristics of physical regions spatially adjacent to the physical region to be corrected. This allows the correction process to consider potential mutual influences between adjacent regions, such as material accumulation, moisture evaporation, or diffusion. Based on this, the key to this scheme lies in quantifying the impact of this proximity relationship. The system determines the degree of local environmental influence of the spatially adjacent physical regions on the physical region to be corrected based on the relative positional relationship between the physical region to be corrected and the spatially adjacent physical regions, as well as the material conveying direction on the belt conveyor. For example, material spillage or moisture evaporation in the upstream region may have a greater impact on the downstream region than on the lateral region. By considering relative position and conveying direction, the actual contribution or interference of adjacent regions to the target region's environment can be more accurately assessed, thereby avoiding unnecessary or inaccurate corrections. Finally, the system combines the local environmental characteristics of the physical region to be corrected, the local environmental characteristics of spatially adjacent physical regions, and the quantified degree of local environmental influence to comprehensively correct the parameter adjustment rules, resulting in regionalized parameter adjustment rules for the physical region to be corrected. This comprehensive correction method fully utilizes the environmental data of the target region itself and creatively incorporates the quantified influence of neighboring regions, enabling the final regionalized parameter adjustment rules to more accurately reflect the evolution of the actual material residue morphology in the physical region under complex dynamic environments. In this way, this solution can more accurately distinguish between transient foreign objects and normal material residues on the conveyor belt surface, thereby effectively reducing the false alarm rate and improving the reliability and practicality of the detection system.

[0074] To further clarify the specific implementation of this solution, an embodiment is provided below. In a specific embodiment, to correct the parameter adjustment rules of a physical area (e.g., area A01) on the belt surface of a belt conveyor, the system first acquires the current local environmental characteristics of area A01 using a sensor array deployed in area A01. These characteristics include real-time temperature, humidity, airflow velocity, and material accumulation thickness data obtained through laser scanning. Simultaneously, the system identifies physical areas spatially adjacent to area A01, such as the upstream area A00, the downstream area A02, and the areas B01 and C01 on either side. The system acquires the local environmental characteristics of these adjacent areas, which can also come from sensors deployed in their respective areas. Then, based on the relative positional relationship between area A01 and these adjacent areas, and the material conveying direction on the belt conveyor, the system determines the degree of local environmental influence of each adjacent area on the physical area A01 to be corrected. For example, if area A00 is directly upstream of area A01, and the material transport direction is from A00 to A01, then the impact of material spillage or moisture evaporation in area A00 on area A01 can be given a higher weight. The impact on areas B01 and C01 (sideways) may be relatively low, or limited to factors such as airflow diffusion. This impact level can be calculated based on a preset physical model, empirical rules, or a machine learning model trained on historical data; for example, it could be an impact factor between 0 and 1. Finally, a processing module combines the local environmental characteristics of area A01 itself, the local environmental characteristics of all spatially adjacent physical areas, and the calculated local environmental impact level to correct the original parameter adjustment rules. For example, a weighted average method can be used, weighting the environmental characteristics of neighboring areas according to their impact level and fusing them with the environmental characteristics of area A01 itself, thereby obtaining a more comprehensive environmental parameter reflecting the actual environment of area A01. Based on this comprehensive environmental parameter, the system can adjust the parameter adjustment rules of region A01 to more accurately adapt to the current complex environmental conditions of region A01, which are affected by multiple factors, and finally obtain the regionalized parameter adjustment rules of region A01.

[0075] In summary, this solution, through the aforementioned technical means, achieves the following technical effects: By incorporating considerations of cross-regional correlations or influences, this solution can more comprehensively assess the actual environmental conditions of the physical region. By acquiring local environmental characteristic information of the physical region to be corrected and its spatially adjacent regions, and quantifying the influence of adjacent regions based on relative position and conveying direction, the parameter adjustment rules can be more accurately corrected. This allows the obtained regionalized parameter adjustment rules to more accurately reflect the evolution of the actual material residue morphology in the physical region under complex dynamic environments. Therefore, the system can more accurately distinguish between transient foreign objects and normal material residues on the conveyor belt surface, thereby effectively reducing the false alarm rate and improving the reliability and practicality of the detection system.

[0076] In some embodiments, the step of determining the time-varying characteristics of morphological differences based on the degree of deviation between morphological differences and evolutionary patterns includes:

[0077] Acquire the local topographic features of the physical region; the local topographic features include at least one of the following: surface normal direction variation, local curvature, and edge sharpness.

[0078] Based on the evolution pattern, the degree of deviation between the local morphological features and the expected local features of the material residue layer is evaluated to obtain local deviation information;

[0079] Based on local deviation information and the degree of deviation between morphological differences and evolutionary patterns, the temporal variation characteristics of morphological differences are determined.

[0080] Local morphological features of a physical region refer to the three-dimensional geometric attributes of a specific small area or detail within that region. These can be obtained through local geometric analysis, feature point extraction, or local surface fitting of 3D point cloud data. Variation in surface normal direction refers to the degree of difference in the direction of the normal vectors of adjacent points or small regions on a 3D surface. This can be characterized by calculating the angle between the normals of adjacent triangular facets, the directional divergence after estimating the point cloud normals, or the rate of gradient change. Local curvature refers to the degree of bending of a 3D surface at a point or small region. This can be quantified using principal curvature, Gaussian curvature, average curvature, or by calculating curvature parameters through fitting a local quadratic surface. Edge sharpness refers to the sharpness of boundaries or contours on a 3D surface. This can be evaluated by calculating the gradient magnitude of edge points, edge strength, edge point density, or by applying edge detection algorithms to depth maps. The expected local features of the material residue layer morphology refer to the local geometric properties that the material residue layer should exhibit at a specific physical region and time point, predicted based on the evolution law of the material residue layer morphology. These can be determined using predictive models trained on historical data, statistical averages, or dynamic threshold ranges. Local deviation information refers to the degree of inconsistency or difference between the currently acquired local morphology features and the expected local features of the material residue layer morphology. This can be obtained by calculating feature vector distance, similarity measures, statistical significance tests, or rule-based logical judgments.

[0081] This scheme, in determining the temporal variation characteristics of morphological differences, introduces the analysis of local morphological features of the physical region and combines this with the evolution law of the material residue layer morphology to more precisely determine the temporal variation characteristics of morphological differences. Specifically, it first acquires local morphological features of the physical region, including at least one of the following: surface normal direction change, local curvature, and edge sharpness. These local features can extract more refined and discriminative local geometric information from macroscopic morphological differences, providing basic data for subsequent accurate assessment. Next, based on the evolution law of the material residue layer morphology, it assesses the degree of deviation between the currently acquired local morphological features and the expected local features of the material residue layer morphology, thereby obtaining local deviation information. This assessment process is no longer a simple overall comparison of the current morphology with historical data, but rather a comparison of the currently acquired local features with the expected local features predicted by the evolution law of the material residue layer morphology, thereby identifying abnormal local features that do not conform to the normal evolution law of the residue. For example, if the evolution law indicates that the edge of the residue should be blurred, but a sharp edge has appeared in the current local area, then this degree of deviation will be high, indicating an anomaly. Finally, the local deviation information obtained above is comprehensively considered in conjunction with the degree of deviation between the overall morphological differences and evolutionary patterns to determine the temporal variation characteristics of the morphological differences. This comprehensive judgment mechanism means that determining whether a morphological change is an instantaneous foreign object or material residue no longer relies solely on the overall difference in "height and size," but also incorporates the differences in "shape features" of local details. In this way, even an object with a small overall height can be identified as a foreign object if its local morphological features deviate from the expected residue characteristics. Conversely, a material accumulation with a large overall height will not be misjudged as a foreign object if its local features conform to the evolutionary patterns of residues. This method, combined with the previous approach of determining the temporal variation characteristics of morphological differences based on the degree of deviation between morphological differences and evolutionary patterns, allows the system to consider not only macroscopic overall changes but also in-depth analysis of microscopic local details when making judgments, thereby improving the accuracy and robustness of detection and enabling the identification of instantaneous foreign objects partially obscured or deformed by material residue layers.

[0082] In one specific embodiment, the method can be implemented as follows: First, after acquiring the current three-dimensional contour data of the physical region, a data processing module can process the three-dimensional contour data to obtain local topographic features of the physical region. For example, the module can perform normal estimation on the three-dimensional point cloud data and calculate the change in surface normal direction of adjacent points or small regions to identify abrupt surface changes or sharp edges. Simultaneously, local curvature can be calculated, for example, by fitting a local quadratic surface to quantify the degree of surface unevenness. Furthermore, edge sharpness can be evaluated by analyzing the gradient information of the depth image or the density changes of the point cloud. The extraction of these local topographic features aims to capture the geometric details that foreign objects may possess.

[0083] Next, an analysis unit can assess the degree of deviation between the acquired local morphological features and the expected local features of the material residue layer morphology, based on pre-established evolution patterns of the material residue layer morphology, thereby obtaining local deviation information. For example, the evolution pattern might indicate that under specific environmental conditions, the edges of the material residue layer are typically blurred and have gentle curvature changes. If the currently detected local area shows high edge sharpness or curvature changes, a higher local deviation value can be calculated. This assessment process can utilize statistical or machine learning models to compare the current local features with the statistical distribution of historical residue features to quantify the degree of non-compliance.

[0084] Finally, a decision-making unit can comprehensively determine the temporal variation characteristics of morphological differences based on the obtained local deviation information and the previously calculated overall deviation from the morphological differences and evolution patterns. For example, the decision-making unit can use a weighted fusion algorithm to sum the local deviation information and the overall deviation, or use a classifier to classify these comprehensive features. In this way, even if a foreign object is partially obscured by the material residue layer, resulting in insignificant overall morphological differences, the system can still identify it as a momentarily appearing foreign object as long as its local morphological features deviate from the expected residue characteristics. Conversely, if the overall morphological difference of a material accumulation is large, but its local features perfectly match the evolution pattern of the material residue layer, it can be correctly identified as material residue, avoiding false alarms.

[0085] This solution introduces the analysis of local morphological features of the physical region and combines it with the evolution law of the material residue layer morphology to more precisely determine the temporal variation characteristics of morphological differences. This allows the system to identify transient foreign objects or material residues that are partially obscured or deformed by the material residue layer, rather than relying solely on macroscopic overall morphological differences. By acquiring local features such as changes in surface normal direction, local curvature, and edge sharpness, and comparing them with the expected local features of the material residue layer morphology, geometric details that may still exist even when foreign objects are obscured can be captured. Finally, combining local deviation information with the degree of deviation of overall morphological differences improves the accuracy and robustness of detection, thereby avoiding the problem of missing real foreign objects with low exposure height or those that are partially obscured.

[0086] In some embodiments, the step of evaluating the degree of deviation between local morphological features and expected local features of the material residue layer morphology, based on evolutionary patterns, to obtain local deviation information includes:

[0087] The local morphological features are smoothed to obtain the smoothed local morphological features;

[0088] Based on the evolutionary pattern, determine the normal fluctuation range of the expected local characteristics of the material residue layer morphology;

[0089] The smoothed local morphological features are compared with the normal fluctuation range. By determining whether the smoothed local morphological features are outside the normal fluctuation range, local deviation information is obtained.

[0090] Local morphological features refer to data characterizing the geometric shape or texture of a specific area on the surface of a conveyor belt. These features can include variations in the surface normal direction, local curvature, and edge sharpness, and can be extracted from 3D contour data. Smoothing refers to filtering or averaging the raw data to reduce random noise and irregular fluctuations. This can be achieved using algorithms such as moving average filtering, Gaussian filtering, or median filtering. Evolutionary patterns refer to rules or models obtained through time-series analysis of historical 3D contour data sets, characterizing the cumulative or changing trend of the material residue layer morphology over time. These can be expressed as mathematical functions, statistical models, or machine learning models. The expected normal fluctuation range of local features of the material residue layer morphology refers to the reasonable range of possible variations in local morphological features under normal operating conditions. This can be determined based on evolutionary patterns obtained from historical data analysis using statistical methods (e.g., calculating the mean and standard deviation, or determining percentile intervals). Local deviation information refers to the judgment result indicating whether local morphological features exceed the normal fluctuation range. This can be expressed as a Boolean value (yes / no), a numerical value indicating the degree of deviation, or anomaly level.

[0091] This solution aims to suppress minor fluctuations caused by noise and non-foreign objects by preprocessing local morphological features and introducing the concept of a normal fluctuation range. This allows for a more accurate and robust assessment of the deviation between local morphological features and the expected local features of the material residue layer morphology, thereby obtaining reliable local deviation information and reducing the false alarm rate. Specifically, when assessing the deviation between local morphological features and the expected local features of the material residue layer morphology, the original local morphological features are first smoothed. This processing step effectively filters out noise in the 3D contour data acquired by the laser scanning system, as well as minor fluctuations or textures on the surface of the material residue layer caused by non-foreign objects, resulting in smoothed local morphological features that are more stable and better reflect the true morphological trend. Simultaneously, based on the evolution of the material residue layer morphology obtained through time-series analysis of historical 3D contour data sets, the system can accurately determine the normal fluctuation range of the expected local features of the material residue layer morphology under normal operating conditions. This fluctuation range considers the dynamic influence of factors such as ambient temperature, humidity, and material moisture content on the residue layer morphology, providing a dynamic and realistic benchmark. Subsequently, the smoothed local morphological features are compared with a pre-determined normal fluctuation range. This comparison is no longer a simple point-to-point comparison with a fixed or ideal value, but rather a determination of whether the smoothed feature exceeds its allowable fluctuation range under normal conditions. Only when the smoothed local morphological features are outside the normal fluctuation range is a true "local deviation" considered to exist, thus obtaining more accurate and reliable local deviation information. It is precisely because of the preprocessing of local morphological features and the introduction of a dynamic normal fluctuation range as a judgment benchmark that this scheme can effectively distinguish signals caused by noise or normal fluctuations from signals caused by actual foreign matter or significant residue changes. This processing method makes the obtained local deviation information more robust and accurate, significantly reducing the false alarm rate. By providing this optimized local deviation information, this scheme further enhances the accuracy of determining the temporal variation characteristics of morphological differences. When local deviation information is more reliable, combining the degree of deviation between morphological differences and evolution patterns to determine time-varying characteristics can more accurately identify instantaneous foreign objects or material residues, avoiding misjudging normal fluctuations or noise as abnormalities, thereby improving the reliability and practicality of the entire belt surface detection method.

[0092] Reference Appendix Figure 2 This invention provides a laser scanning-based belt conveyor surface inspection system, comprising:

[0093] The acquisition module 100 is used to acquire the current three-dimensional contour data and historical three-dimensional contour data set of each physical region on the surface of the belt conveyor; the historical three-dimensional contour data set contains the three-dimensional contour data of each physical region acquired at different time points;

[0094] The comparison module 200 is used to compare the current three-dimensional contour data with the historical three-dimensional contour data set of the corresponding physical region to obtain the morphological differences of the physical region and the temporal variation characteristics of the morphological differences.

[0095] The differentiation module 300 is used to differentiate physical areas as instantaneously appearing foreign objects or material residues based on morphological differences and the time-varying characteristics of morphological differences.

[0096] The alarm module 400 is used to trigger an alarm in a physical area that is identified as a foreign object that appears momentarily.

[0097] In this document, relational terms such as first and second are used only to distinguish one entity or operation from another entity or operation, without necessarily requiring or implying any such actual relationship or order between these entities or operations.

[0098] The above description is merely an embodiment of the present invention and is not intended to limit the scope of protection of the present invention. For those skilled in the art, the present invention can have various modifications and variations. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for detecting the surface of a belt conveyor based on laser scanning, characterized in that, Includes the following steps: Acquire the current 3D contour data and historical 3D contour data set of each physical region on the surface of the belt conveyor; the historical 3D contour data set contains the 3D contour data of each physical region acquired at different time points; By comparing the current 3D contour data with the historical 3D contour data set of the corresponding physical region, the morphological differences of the physical region and the temporal variation characteristics of the morphological differences can be obtained. Based on the differences in morphology and the time-varying characteristics of these differences, physical areas can be distinguished as foreign objects or material residues that appear instantaneously. An alarm is triggered in the physical area where a foreign object is identified as appearing instantaneously; The steps to compare the current 3D contour data with the historical 3D contour data set of the corresponding physical region to obtain the morphological differences of the physical region and the temporal variation characteristics of the morphological differences include: By performing time-series analysis on historical 3D contour data sets, the evolution pattern of material residue layer morphology in physical regions can be obtained. Based on the evolutionary pattern, the morphological differences of the physical region are obtained by dynamically comparing the current three-dimensional contour data with the historical three-dimensional contour data set. Based on the degree of deviation between morphological differences and evolutionary patterns, the temporal variation characteristics of morphological differences are determined; The steps for obtaining the evolution pattern of material residue morphology in a physical region by performing time-series analysis on historical 3D contour data sets include: Obtain the set of environmental parameters for the time points corresponding to the historical 3D contour data set; Based on historical 3D contour data sets and environmental parameter sets, by analyzing the correlation between the morphology of the material residue layer in the physical region and environmental parameters, parameter adjustment rules for the evolution law of the material residue layer morphology are obtained. Get the environmental parameters at the current time point; Based on the environmental parameters and parameter adjustment rules at the current time point, the parameters of the evolution law of the material residue layer morphology are adjusted to obtain the evolution law of the material residue layer morphology adapted to the current environment. The steps for adjusting the parameters of the evolution law of the material residue layer morphology according to the environmental parameters and parameter adjustment rules at the current time point, to obtain the evolution law of the material residue layer morphology adapted to the current environment, include: Obtain local environmental feature information for each physical region; Based on local environmental characteristics, the parameter adjustment rules are modified in a regionalized manner to obtain the regionalized parameter adjustment rules for each physical region. For each physical region, based on the environmental parameters at the current time point and the corresponding regional parameter adjustment rules for the physical region, the parameters for the evolution law of the material residue layer morphology in the physical region are adjusted to obtain the evolution law of the material residue layer morphology adapted to the current environment of the physical region.

2. The method for detecting the belt surface of a belt conveyor based on laser scanning according to claim 1, characterized in that, The set of environmental parameters includes at least one of the following at a given time point: temperature, humidity, and moisture content of the conveyed material.

3. The method for detecting the belt surface of a belt conveyor based on laser scanning according to claim 1, characterized in that, Local environmental characteristics information represents at least one of the following: local location of the physical area, local ventilation conditions, and material accumulation thickness.

4. The method for detecting the belt surface of a belt conveyor based on laser scanning according to claim 1, characterized in that, The steps for determining the temporal variation characteristics of morphological differences based on the degree of deviation from the evolutionary pattern include: Obtain the local topographic features of the physical region; Based on the evolution pattern, the degree of deviation between the local morphological features and the expected local features of the material residue layer morphology is evaluated to obtain local deviation information; Based on local deviation information and the degree of deviation between morphological differences and evolutionary patterns, the temporal variation characteristics of morphological differences are determined.

5. The method for detecting the belt surface of a belt conveyor based on laser scanning according to claim 4, characterized in that, Local morphological features include at least one of the following: surface normal direction variation, local curvature, and edge sharpness.

6. The method for detecting the belt surface of a belt conveyor based on laser scanning according to claim 4, characterized in that, Based on the evolutionary pattern, the steps to assess the degree of deviation between local morphological features and the expected local features of the material residue layer morphology, and to obtain local deviation information, include: The local morphological features are smoothed to obtain the smoothed local morphological features; Based on the evolutionary pattern, determine the normal fluctuation range of the expected local characteristics of the material residue layer morphology; The smoothed local morphological features are compared with the normal fluctuation range. By determining whether the smoothed local morphological features are outside the normal fluctuation range, local deviation information is obtained.

7. A belt conveyor belt surface detection system based on laser scanning, employing the laser scanning-based belt conveyor belt surface detection method as described in any one of claims 1-6, characterized in that, include: The acquisition module is used to acquire the current three-dimensional contour data and historical three-dimensional contour data set of each physical region on the surface of the belt conveyor; the historical three-dimensional contour data set contains the three-dimensional contour data of each physical region acquired at different time points; The comparison module is used to compare the current 3D contour data with the historical 3D contour data set of the corresponding physical region to obtain the morphological differences of the physical region and the temporal variation characteristics of the morphological differences. The differentiation module is used to distinguish physical areas as instantaneously appearing foreign objects or material residues based on differences in morphology and the time-varying characteristics of these differences. The alarm module is used to trigger an alarm in physical areas where foreign objects are identified as appearing instantaneously.

Images