Unattended pose monitoring method and system for bucket-wheel stacker-reclaimer

Abstract

The invention provides an unattended pose monitoring method and system for a bucket-wheel stacker-reclaimer, and relates to the technical field of bucket-wheel stacker-reclaimers, and the method comprises the steps: obtaining the first positioning information of at least four non-coplanar ultra-wideband tags on a cantilever based on an ultra-wideband positioning technology; acquiring second positioning information, wherein the second positioning information comprises signal characteristics of at least one ultra-wideband tag acquired by the monitoring device; analyzing a position error between the two reference labels based on the first positioning information to obtain calibration positioning information; fusing the calibration positioning information and the second positioning information based on a deep learning algorithm to obtain a corrected coordinate; and establishing a three-dimensional model of the cantilever based on the corrected coordinates to obtain pose information of the cantilever. The ultra-wideband positioning is calibrated by analyzing the first positioning information, and the second positioning information from different sources is fused at the same time, so that the positioning error is effectively reduced, and the number requirement and the setting difficulty of the near-end monitoring device are reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of bucket wheel stacker reclaimers, and in particular, to a method and system for unattended pose monitoring of a bucket wheel stacker reclaimer. Background Art

[0002] A bucket wheel stacker reclaimer is an efficient loading and unloading machine for continuous conveying that can both stack and reclaim materials in a large dry bulk yard, and it has a boom that can pitch and swing horizontally. Currently, unmanned bucket wheel stacker reclaimers are gradually replacing manually operated ones. Unmanned operation is achieved by presetting programs and parameters in an automated control system to autonomously complete tasks such as material stacking, transportation, and discharge without manual intervention. However, there is a problem of unstable posture during the operation of the boom of the stacker reclaimer. If the pose of the boom cannot be accurately monitored and effectively controlled, it may lead to equipment damage, production interruption, or even safety accidents. Therefore, how to accurately monitor the pose of the boom to avoid operation accidents has become an urgent problem to be solved. Summary of the Invention

[0003] The purpose of the present invention is to provide a method and system for unattended pose monitoring of a bucket wheel stacker reclaimer to improve the above problems. To achieve the above purpose, the technical solutions adopted by the present invention are as follows:

[0004] In a first aspect, the present application provides a method for unattended pose monitoring of a bucket wheel stacker reclaimer, including:

[0005] Obtaining first positioning information of at least four non - coplanar ultra - wideband tags on the boom based on ultra - wideband positioning technology, where at least two ultra - wideband tags are respectively arranged at both ends of the boom, and two ultra - wideband tags on the boom are denoted as reference tags, and the connection line of the two reference tags is perpendicular to the central axis of the boom;

[0006] Obtaining second positioning information, where the second positioning information includes signal characteristics of at least one ultra - wideband tag obtained by a monitoring device, and the ultra - wideband tag corresponding to the second positioning information is denoted as a reference tag; the monitoring device includes at least one of a binocular camera, a photoelectric sensor, or a radar sensor;

[0007] Based on the first positioning information, calibrating the first positioning information by analyzing the position error between the two reference tags to obtain calibrated positioning information;

[0008] Fusing the calibrated positioning information and the second positioning information based on a deep learning algorithm to obtain corrected coordinates;

[0009] Based on the corrected coordinates, establishing a three - dimensional model of the boom to obtain the pose information of the boom.

[0010] In a second aspect, the present application further provides an unattended pose monitoring system for a bucket wheel stacker-reclaimer, including:

[0011] A first acquisition module, configured to obtain first positioning information of at least four non-coplanar ultra-wideband tags on the boom based on ultra-wideband positioning technology, wherein at least two ultra-wideband tags are respectively arranged at both ends of the boom, and the two ultra-wideband tags on the boom are denoted as reference tags, and the connection line of the two reference tags is perpendicular to the central axis of the boom;

[0012] A second acquisition module, configured to obtain second positioning information, where the second positioning information includes signal characteristics of at least one ultra-wideband tag obtained by a monitoring device, and the ultra-wideband tag corresponding to the second positioning information is denoted as a reference tag; the monitoring device includes at least one of a binocular camera, a photoelectric sensor, or a radar sensor;

[0013] A calibration module, configured to calibrate the first positioning information based on the first positioning information by analyzing the position error between the two reference tags to obtain calibrated positioning information;

[0014] A fusion module, configured to fuse the calibrated positioning information and the second positioning information based on a deep learning algorithm to obtain corrected coordinates;

[0015] A building module, configured to build a three-dimensional model of the boom based on the corrected coordinates to obtain the pose information of the boom.

[0016] The beneficial effects of the present invention are as follows:

[0017] The present invention makes a special layout of the ultra-wideband tags, calibrates the ultra-wideband positioning by analyzing the first positioning information of the ultra-wideband tags, and at the same time fuses the second positioning information from different sources, effectively reducing the positioning error. The ultra-wideband positioning technology can simultaneously obtain the coordinates of ultra-wideband tags at different positions on the surface of the boom and directly build a three-dimensional model of the boom. The proximal monitoring device is used to improve the positioning accuracy, reduce the quantity requirement and installation difficulty of the proximal monitoring device, and is applicable to most stacking sites.

[0018] Other features and advantages of the present invention will be described in the subsequent description, and some of them will become obvious from the description, or can be understood by implementing the embodiments of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the embodiments. It should be understood that the following drawings only show some embodiments of the present invention, and therefore should not be regarded as limiting the scope. For those of ordinary skill in the art, other related drawings can be obtained based on these drawings without creative efforts.

[0020] Figure 1 This is the flow chart of the unattended pose monitoring method for the bucket wheel stacker-reclaimer in the embodiment of the present application;

[0021] Figure 2 This is the structural schematic diagram of the unattended pose monitoring system for the bucket wheel stacker-reclaimer in the embodiment of the present application;

[0022] Markings in the figure: 100 - First acquisition module; 200 - Second acquisition module; 300 - Calibration module; 310 - Drawing unit; 320 - Comparison unit; 330 - Calibration unit; 400 - Fusion module; 410 - First construction unit; 420 - Second construction unit; 430 - Conversion unit; 440 - Analysis unit; 450 - Stitching unit; 460 - Correction unit; 500 - Establishment module; 600 - First calculation module; 700 - Second calculation module; 800 - Third calculation module; 900 - Update module. Detailed implementation manners

[0023] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Components of the embodiments of the present invention generally described and illustrated in the figures herein can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0024] It should be noted that: like reference numerals and letters denote like items in the following figures, and thus, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0025] The pose monitoring of existing bucket wheel stacker reclaimers generally adopts visual monitoring, ultrasonic sensors, photoelectric sensors, etc. These monitoring methods generally require installing multiple monitoring devices near the bucket wheel stacker, and multiple reference points need to be set on the fixed structure around the boom as the basis for measurement. Due to the limitations of the material yard operation environment, it is not easy to select and design the installation positions of these monitoring devices. UWB (Ultra Wideband) technology is a new type of wireless communication technology. A UWB positioning system usually consists of multiple positioning base stations and positioning tags. The positions of the positioning base stations are known, and the positioning tags are carried by personnel or installed on mobile devices. The tag periodically emits UWB pulse signals and communicates with the positioning base stations. By measuring the time of flight (TOF) of the signals, the distance between the tag and each base station can be calculated, and then the position of the tag can be calculated using the positioning algorithm. If several tags are set on the boom, the overall position and attitude of the boom can be obtained through UWB positioning. However, since the base stations are far from the bucket wheel stacker, there will be uncertain obstacles on the signal transmission path, which will interfere with the positioning accuracy and may result in a large error. Therefore, the present invention effectively improves the positioning accuracy by autonomously calibrating the positioning coordinates of UWB and performing collaborative calculation with the positioning information of a detection device near the end. Installing a detection device near the bucket wheel stacker can implement this method, greatly reducing the layout difficulty and complexity of on-site equipment compared with traditional methods.

[0026] Embodiment 1

[0027] See Figure 1 , this application provides a monitoring method for unmanned operation of a bucket wheel stacker, including steps S100, S200, S300, S400, and S500.

[0028] S100. Obtain the first positioning information of at least four non-coplanar UWB tags on the boom based on UWB positioning technology, where at least two UWB tags are respectively arranged at both ends of the boom. Denote the two UWB tags on the boom as reference tags, and the connection line of the two reference tags is perpendicular to the central axis of the boom;

[0029] By setting four non-coplanar ultra-wideband tags on the cantilever and obtaining the coordinates of these ultra-wideband tags, a three-dimensional model of the cantilever can be constructed. Among them, two ultra-wideband tags need to be set at both ends of the cantilever respectively, so as to understand the torsional deformation of the cantilever. Two reference tags are preferably symmetrically set on the left and right sides of the cantilever. During the working process, the width of the cantilever will not change, so the actual distance between the two reference tags will not change. If the measured distance between the two reference tags is different from the actual distance, it indicates that there is an error in the ultra-wideband positioning. Therefore, the positioning data of the ultra-wideband can be calibrated based on the error situation. The fourth ultra-wideband tag is preferably set at a position that is easy to monitor (as a reference tag), such as the upper end of one side of the cantilever, to ensure that there is no obstacle between the fourth ultra-wideband tag and the monitoring device at the proximal end, so that the monitoring device can continuously obtain the position information of the fourth ultra-wideband tag;

[0030] The above arrangement is only an example. In fact, an ultra-wideband tag can serve as both a reference tag and a reference tag at the same time, and the position of the reference tag is selected to be easily monitored.

[0031] By setting at least three ultra-wideband base stations in the stockyard, all bucket wheel stacker reclaimers in all stockyards can be globally monitored at the same time. The first positioning information includes the distance from the ultra-wideband tag to each ultra-wideband base station, the coordinate data of the ultra-wideband tag, the signal strength of the UWB, and the timestamp, etc.

[0032] According to the positions of the ultra-wideband base stations, an ultra-wideband coordinate system is established.

[0033] S200. Obtain the second positioning information, where the second positioning information includes the signal characteristics of at least one ultra-wideband tag obtained by the monitoring device, and the ultra-wideband tag corresponding to the second positioning information is denoted as the reference tag; the monitoring device includes at least one of a binocular camera, a photoelectric sensor, or a radar sensor;

[0034] The monitoring device is set at the proximal end of the bucket wheel stacker reclaimer and is used to monitor the position of at least one ultra-wideband tag on the cantilever. If, due to the layout limitations of the stockyard, it is impossible to set a large number of monitoring devices, then arranging one monitoring device is sufficient, and this monitoring device can detect the position of one ultra-wideband tag; this method has low requirements for the number of monitoring devices, small layout difficulty, and can be applied to most stockyards;

[0035] Specifically, if a binocular camera is used, color features can be designed on the reference tag to facilitate the binocular camera to quickly identify and track the reference tag;

[0036] If an optoelectronic sensor (laser sensor) is adopted, a reflector can be set on the reference label or a coating for enhancing reflection can be applied to quickly identify the reflection signal of the reference label and improve the tracking ability of the reference label.

[0037] If a radar sensor is adopted, a corner reflector can be set on the reference label to enhance the reflection signal of the reference label and improve the tracking ability of the reference label.

[0038] The above-mentioned monitoring devices can be used in combination, each having its own advantages. The radar sensor is suitable for bad weather with low visibility. The optoelectronic sensor has high measurement accuracy. The binocular camera can obtain more global information and can assist in the monitoring of stockpiling. And as an alternative implementation, the monitoring device can be installed on the traveling mechanism of the wheel stacker-reclaimer, and its relative position with the cantilever rotation center remains unchanged and it moves along the track together with the traveling mechanism.

[0039] As an alternative implementation, the monitoring device and the reference label can be set together, and the monitoring device measures the distance from it to a fixed reference point to determine the position of the reference label.

[0040] According to the position of the monitoring device, a detection device coordinate system is established. In addition, a reference coordinate system is established with the rotation center of the cantilever as the origin O. The X-axis of the reference coordinate system is along the traveling track of the wheel stacker-reclaimer, the Y-axis is perpendicular to the X-axis, and the XOY plane is parallel to the ground, and the Z-axis points upward.

[0041] S300. Based on the first positioning information, by analyzing the position error between the two reference tags, the first positioning information is calibrated to obtain calibrated positioning information.

[0042] The first positioning information includes the coordinate data of the ultra-wideband tag. The reference tags include a first reference tag and a second reference tag.

[0043] S310. In the reference coordinate system, a reference circle is drawn with the distance from the vertical axis (Z-axis) to the first reference tag as the radius. According to the actual installation positions of the first reference tag and the second reference tag, the standard coordinates of the second reference tag are determined on the reference circle. The first reference tag and the reference label are on the same side of the cantilever.

[0044] Since the UWB base station measures each ultra-wideband tag at the same time and in the same environment, the measurement error between different ultra-wideband tags mainly comes from different obstacles on the signal path. By making the first reference tag and the reference label on the same side of the cantilever, it can ensure that the obstacle situations on the paths of each UWB base station signal to the first reference tag and the reference label are basically the same, reducing the measurement error between the two.

[0045] S320. Compare the standard coordinates with the coordinate data of the second reference tag to obtain the measurement offset information of the second reference tag. The measurement offset information includes the offset distance and the offset direction, from which the measurement errors on both sides of the cantilever caused by different signal paths can be known.

[0046] Keep the coordinate data of the first reference tag unchanged. Based on the measurement offset information, calibrate the coordinate data of other ultra-wideband tags to obtain calibrated positioning information. That is, modify the coordinates of the ultra-wideband tags on the same side of the second reference tag according to the offset distance and the offset direction to achieve calibration.

[0047] Based on the same idea, in this step, the perpendicular bisector connecting the first reference tag and the standard coordinates can also be drawn in the reference coordinate system. This perpendicular bisector is the longitudinal axis of the cantilever. Based on this, calculate the longitudinal distance of the ultra-wideband tags at both ends of the cantilever along the perpendicular bisector and compare it with the standard distance. Since the rigid cantilever cannot be extended during operation, if the longitudinal distance is greater than the standard distance, it indicates that there is an error in the ultra-wideband positioning. Based on this error, the coordinates of other ultra-wideband tags can be calibrated.

[0048] S400. Based on the deep learning algorithm, fuse the calibrated positioning information and the second positioning information to obtain the corrected coordinates, including:

[0049] Establish a reference coordinate system, an ultra-wideband coordinate system, and a monitoring device coordinate system. The reference coordinate system takes the rotation center of the cantilever as the origin. The reference coordinate system, the ultra-wideband coordinate system, and the monitoring device coordinate system have been established in the previous steps, so they will not be elaborated in this step.

[0050] Construct a coordinate transformation matrix based on the spatial position relationship among the rotation center of the cantilever, the ultra-wideband base station, and the monitoring device.

[0051] Based on the coordinate transformation matrix, convert the positions of the ultra-wideband tags obtained by the ultra-wideband base station and the monitoring device into the reference coordinate system to obtain the first transformed coordinates and the second transformed coordinates.

[0052] The first transformed coordinates are the coordinates obtained by the ultra-wideband base station, and the second transformed coordinates are the coordinates obtained by the monitoring device.

[0053] Based on the first transformed coordinates and the second transformed coordinates, analyze the position differences of the reference tags measured by the ultra-wideband base station and the monitoring device to obtain the difference features. The difference features include distance differences and direction differences. If there are multiple monitoring devices, it is also necessary to analyze the tag position differences among the monitoring devices.

[0054] Extract data features from the calibration positioning information and the second positioning information by using a convolutional neural network, and perform feature splicing on the data features and the difference features. Among them, an attention mechanism is used to dynamically weight the data features of the calibration positioning information and the second positioning information;

[0055] For a radar sensor or a lidar sensor, its second positioning information includes second conversion coordinates, distance, speed, azimuth angle, reflection intensity, etc.; for a binocular camera, its second positioning information includes second conversion coordinates, feature points, disparity map, depth map, etc.; extract the data features from these information, and perform denoising, normalization and encoding processing, and perform feature splicing together with the difference features;

[0056] Input the spliced features into a preset coordinate correction model to obtain corrected coordinates. The preset coordinate correction model is trained by using historical calibration positioning information and second positioning information, and the real coordinate data (model output) used for training is obtained by manual measurement using a total station.

[0057] S500. Based on the corrected coordinates, establish a three-dimensional model of the cantilever to obtain the pose information of the cantilever.

[0058] The corrected coordinates are non-coplanar points distributed on the surface of the cantilever. Based on the installation position parameters of the ultra-wideband tags on the cantilever, a three-dimensional model of the cantilever can be established with the ultra-wideband tags as the base points. The three-dimensional model is displayed in real time on the remote monitoring device to realize the real-time monitoring of the cantilever pose;

[0059] At the same time, a belt scale can be installed under the cantilever conveyor belt to obtain the load information of the cantilever. Based on the load information and pose information of the cantilever, the center of gravity position of the cantilever can be calculated, so as to adjust the counterweight mechanism of the cantilever and predict the equipment overturning risk.

[0060] For a relatively long bucket wheel cantilever, under the influence of a large load and wind force, the cantilever may be significantly distorted. In this case, there is a greater risk to continue operating. Therefore, the method further includes step S600:

[0061] Based on the height difference between two ultra-wideband tags at both ends of the cantilever, calculate the angle between the cantilever and the horizontal plane, and calculate the cantilever inclination;

[0062] Based on the distance between the ultra-wideband tags at both ends of the cantilever, calculate the vertical offset amplitude and the up and down offset amplitude of the cantilever respectively to obtain the cantilever deflection; the deflection is the linear displacement of the rod axis in the direction perpendicular to the axis, which is used to reflect the distortion degree of the cantilever;

[0063] Obtain wind load information and cantilever load information, and calculate the distortion risk coefficient based on the cantilever inclination and cantilever deflection through a preset relationship function. The preset relationship function is obtained by fitting historical data, and the historical distortion risk coefficient is given by expert evaluation;

[0064] Map the distortion risk coefficient to a color, and modify the pixels of the cantilever image in the 3D model with the mapped color to obtain an updated 3D model of the cantilever. Thus, the distortion risk of the cantilever can be visually seen through the 3D model, and the monitoring personnel can take corresponding measures in a timely manner.

[0065] When the monitoring device includes a binocular camera, due to the continuous change of the cantilever movement and the influence of occlusion by personnel and other operating equipment, the binocular camera may temporarily lose the monitoring target. To help the binocular camera quickly lock the target, the method further includes S700:

[0066] Convert the coordinate data of the reference label in the first positioning information into the coordinate system of the binocular camera to obtain the third conversion coordinate;

[0067] In the image obtained by the binocular camera, find the target pixel corresponding to the third conversion coordinate; segment the image within a preset range centered on the target pixel to obtain the region of interest;

[0068] Perform feature recognition on the region of interest, extract the feature points of the reference label, and track the reference label based on the feature points.

[0069] Based on this method, the binocular camera can quickly re-track the reference label after losing the target, greatly reducing the influence of the surrounding environment on the target recognition ability of the camera.

[0070] When the monitoring device includes a binocular camera, a photoelectric sensor, and a radar sensor at the same time; the calibration positioning information and the second positioning information are fused based on the deep learning algorithm to obtain the corrected coordinates, including:

[0071] Perform multi-modal late fusion based on the first positioning information and the second positioning information obtained by each monitoring device; late fusion is to perform fusion and decision-making after the separate processing of each modal feature (respectively extract the features of each modality through the model, that is, convert each modality into a vector by using a model good at processing each modality). Simply put, each modality is independently processed, and finally their results are fused, and the feature processing and learning processes do not interfere. Late fusion can maintain the independence of each modality, avoid information loss, and the most suitable model architecture can be selected for different modal inputs.

[0072] Obtain environmental features, determine the first positioning information and the fusion weights of each monitoring device based on the environmental features, and obtain the corrected coordinates. The measurement performance of some devices is greatly affected by environmental factors. For example, in poor light conditions, the target recognition ability of a binocular camera is very poor, and factors such as weather, temperature, and humidity will have different effects on sensors. Therefore, it is necessary to dynamically adjust the fusion weights of each modality according to environmental features to obtain the most accurate fusion information.

[0073] Embodiment 2

[0074] Refer to Figure 2 , this application also provides a pose monitoring system for unattended bucket wheel stacker-reclaimers, including:

[0075] A first acquisition module, configured to obtain the first positioning information of at least four non-coplanar ultra-wideband tags on the boom based on ultra-wideband positioning technology, where at least two ultra-wideband tags are respectively arranged at both ends of the boom, and the two ultra-wideband tags on the boom are denoted as reference tags, and the connection line of the two reference tags is perpendicular to the central axis of the boom;

[0076] A second acquisition module, configured to obtain second positioning information, where the second positioning information includes the signal characteristics of at least one ultra-wideband tag obtained by a monitoring device, and the ultra-wideband tag corresponding to the second positioning information is denoted as a reference tag; the monitoring device includes at least one of a binocular camera, a photoelectric sensor, or a radar sensor;

[0077] A calibration module, configured to calibrate the first positioning information based on the first positioning information by analyzing the position error between the two reference tags to obtain calibrated positioning information;

[0078] A fusion module, configured to fuse the calibrated positioning information and the second positioning information based on a deep learning algorithm to obtain corrected coordinates;

[0079] A building module, configured to build a three-dimensional model of the boom based on the corrected coordinates to obtain the pose information of the boom.

[0080] As an optional implementation manner, the fusion module includes:

[0081] A first construction unit, configured to establish a reference coordinate system, an ultra-wideband coordinate system, and a monitoring device coordinate system, where the reference coordinate system takes the rotation center of the boom as the origin;

[0082] A second construction unit, configured to construct a coordinate transformation matrix based on the spatial position relationship between the rotation center of the boom, the ultra-wideband base station, and the monitoring device;

[0083] A conversion unit, configured to convert the positions of UWB tags obtained by the UWB base station and the monitoring device into the reference coordinate system based on the coordinate conversion matrix, to obtain a first converted coordinate and a second converted coordinate;

[0084] An analysis unit, configured to analyze the position differences of the reference tags measured by the UWB base station and the monitoring device based on the first converted coordinate and the second converted coordinate, to obtain difference features;

[0085] A splicing unit, configured to extract data features from the calibration positioning information and the second positioning information by using a convolutional neural network, and splice the data features and the difference features, wherein the data features of the calibration positioning information and the second positioning information are dynamically weighted by using an attention mechanism;

[0086] A correction unit, configured to input the spliced features into a preset coordinate correction model to obtain corrected coordinates.

[0087] As an optional implementation manner, the calibration module includes:

[0088] A drawing unit, configured to draw a reference circle in the reference coordinate system with the distance from the vertical axis to the first reference tag as the radius, and determine the standard coordinates of the second reference tag on the reference circle according to the actual installation positions of the first reference tag and the second reference tag; the first reference tag and the reference tag are on the same side of the cantilever;

[0089] A comparison unit, configured to compare the standard coordinates with the coordinate data of the second reference tag to obtain the measurement offset information of the second reference tag;

[0090] A calibration unit, configured to keep the coordinate data of the first reference tag unchanged, and calibrate the coordinate data of other UWB tags based on the measurement offset information to obtain calibration positioning information.

[0091] As an optional implementation manner, the system further includes:

[0092] A first calculation module, configured to calculate the cantilever inclination based on the height difference between two UWB tags at both ends of the cantilever;

[0093] A second calculation module, configured to calculate the vertical offset amplitude and the up-and-down offset amplitude of the cantilever respectively based on the distance between the UWB tags at both ends of the cantilever to obtain the cantilever deflection;

[0094] A third calculation module, configured to obtain the wind load information and the cantilever load information, and calculate the distortion risk coefficient based on the cantilever inclination and the cantilever deflection through a preset relationship function;

[0095] An update module maps the twist risk coefficient to a color, and modifies the pixels of the cantilever image in the three-dimensional model with the mapped color to obtain an updated three-dimensional model of the cantilever.

[0096] It should be noted that in this article, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or further includes elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "comprising an..." does not exclude the presence of additional identical elements in the process, method, article or device comprising the element.

[0097] The above is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims

1. A method for unattended pose monitoring of a bucket wheel stacker-reclaimer, characterized in that, Including: Obtaining first positioning information of at least four non-coplanar ultra-wideband tags on the cantilever based on ultra-wideband positioning technology, where at least two ultra-wideband tags are respectively arranged at both ends of the cantilever. Denote the two ultra-wideband tags on the cantilever as reference tags, and the connection line of the two reference tags is perpendicular to the central axis of the cantilever. Obtaining second positioning information, where the second positioning information includes signal characteristics of at least one ultra-wideband tag obtained by a monitoring device. Denote the ultra-wideband tag corresponding to the second positioning information as a reference tag; the monitoring device includes at least one of a binocular camera, a photoelectric sensor, or a radar sensor. Based on the first positioning information, calibrate the first positioning information by analyzing the position error between the two reference tags to obtain calibrated positioning information. Fuse the calibrated positioning information and the second positioning information based on a deep learning algorithm to obtain corrected coordinates. Based on the corrected coordinates, establish a three-dimensional model of the cantilever to obtain the pose information of the cantilever.

2. The unattended pose monitoring method for a bucket wheel stacker-reclaimer according to claim 1, wherein The fusing the calibrated positioning information and the second positioning information based on a deep learning algorithm to obtain corrected coordinates includes: Establish a reference coordinate system, an ultra-wideband coordinate system, and a monitoring device coordinate system, where the reference coordinate system takes the rotation center of the cantilever as the origin. Construct a coordinate transformation matrix based on the spatial position relationship of the rotation center of the cantilever, the ultra-wideband base station, and the monitoring device. Based on the coordinate transformation matrix, convert the positions of the ultra-wideband tags obtained by the ultra-wideband base station and the monitoring device into the reference coordinate system to obtain a first transformed coordinate and a second transformed coordinate. Based on the first transformed coordinate and the second transformed coordinate, analyze the position difference of the reference tag measured by the ultra-wideband base station and the monitoring device to obtain a difference feature. Use a convolutional neural network to extract data features from the calibrated positioning information and the second positioning information, and splice the data features and the difference features, where an attention mechanism is used to dynamically weight the data features of the calibrated positioning information and the second positioning information. Input the spliced features into a preset coordinate correction model to obtain corrected coordinates.

3. A method for unattended pose monitoring of a bucket wheel stacker-reclaimer according to claim 2, characterized in that, The first positioning information includes coordinate data of the ultra-wideband tags; the reference tags include a first reference tag and a second reference tag; the calibrating the first positioning information by analyzing the position error between the two reference tags based on the first positioning information to obtain calibrated positioning information includes: In the reference coordinate system, draw a reference circle with the distance from the vertical axis to the first reference tag as the radius, and determine the standard coordinate of the second reference tag on the reference circle according to the actual installation positions of the first reference tag and the second reference tag; the first reference tag and the reference tag are on the same side of the cantilever. Compare the standard coordinate with the coordinate data of the second reference tag to obtain the measurement offset information of the second reference tag. Keep the coordinate data of the first reference tag unchanged, and based on the measurement offset information, calibrate the coordinate data of other ultra-wideband tags to obtain calibrated positioning information.

4. A posture monitoring method for unattended bucket wheel stacker-reclaimer according to claim 2, characterized in that, The method further includes: Calculating the cantilever inclination based on the height difference between two ultra-wideband tags at both ends of the cantilever. Based on the distance between the ultra-wideband tags at both ends of the cantilever, calculate the vertical offset amplitude and the up-and-down offset amplitude of the cantilever respectively to obtain the cantilever deflection; Obtain the wind load information and the cantilever load information, and calculate the distortion risk coefficient based on the cantilever inclination and the cantilever deflection through a preset relationship function; Map the distortion risk coefficient to a color, and modify the cantilever image pixels in the 3D model with the mapped color to obtain an updated 3D model of the cantilever.

5. The unattended pose monitoring method for a bucket wheel stacker-reclaimer according to claim 1, characterized in that, The monitoring device includes a binocular camera, and the method further includes: Convert the coordinate data of the reference tag in the first positioning information into the coordinate system of the binocular camera to obtain the third conversion coordinate; In the image acquired by the binocular camera, find the target pixel corresponding to the third conversion coordinate; segment the image within a preset range centered on the target pixel to obtain the region of interest; Perform feature recognition on the region of interest, extract the feature points of the reference tag, and track the reference tag based on the feature points.

6. The unattended pose monitoring method for a bucket wheel stacker-reclaimer according to claim 1, characterized in that The monitoring device includes a binocular camera, a photoelectric sensor, and a radar sensor; the fusion of the calibration positioning information and the second positioning information based on the deep learning algorithm to obtain the corrected coordinates includes: Perform multi-modal late fusion based on the first positioning information and the second positioning information acquired by each monitoring device; Obtain the environmental features, determine the fusion weights of the first positioning information and each monitoring device based on the environmental features, and obtain the corrected coordinates.

7. An unattended pose monitoring system for a bucket wheel stacker-reclaimer, characterized in that, Includes: The first acquisition module is used to obtain the first positioning information of at least four non-coplanar ultra-wideband tags on the cantilever based on the ultra-wideband positioning technology, where at least two ultra-wideband tags are respectively arranged at both ends of the cantilever, and the two ultra-wideband tags on the cantilever are denoted as reference tags, and the line connecting the two reference tags is perpendicular to the central axis of the cantilever; The second acquisition module is used to obtain the second positioning information, and the second positioning information includes the signal features of at least one ultra-wideband tag acquired by the monitoring device, and the ultra-wideband tag corresponding to the second positioning information is denoted as the reference tag; the monitoring device includes at least one of a binocular camera, a photoelectric sensor, or a radar sensor; The calibration module is used to calibrate the first positioning information based on the first positioning information by analyzing the position error between the two reference tags to obtain the calibrated positioning information; The fusion module is used to fuse the calibrated positioning information and the second positioning information based on the deep learning algorithm to obtain the corrected coordinates; The establishment module is used to establish a 3D model of the cantilever based on the corrected coordinates to obtain the pose information of the cantilever.

8. The unattended pose monitoring system for the bucket wheel stacker-reclaimer according to claim 7, wherein The fusion module includes: The first construction unit is used to establish a reference coordinate system, an ultra-wideband coordinate system, and a monitoring device coordinate system, and the reference coordinate system takes the rotation center of the cantilever as the origin; The second construction unit is used to construct a coordinate transformation matrix based on the spatial position relationship of the rotation center of the cantilever, the ultra-wideband base station, and the monitoring device; The conversion unit is used to convert the positions of the ultra-wideband tags acquired by the ultra-wideband base station and the monitoring device into the reference coordinate system based on the coordinate transformation matrix to obtain the first conversion coordinate and the second conversion coordinate; An analysis unit, configured to analyze the position difference of the reference tag measured by the ultra-wideband base station and the monitoring device based on the first conversion coordinate and the second conversion coordinate, so as to obtain a difference feature; A splicing unit, configured to extract data features from the calibration positioning information and the second positioning information by using a convolutional neural network, and splice the data features and the difference features, wherein an attention mechanism is used to dynamically weight the data features of the calibration positioning information and the second positioning information; A correction unit, configured to input the spliced features into a preset coordinate correction model to obtain corrected coordinates.

9. The unattended pose monitoring system for the bucket wheel stacker-reclaimer according to claim 7, wherein The calibration module includes: A drawing unit, configured to draw a reference circle in the reference coordinate system with the distance from the vertical axis to the first reference tag as the radius, and determine the standard coordinates of the second reference tag on the reference circle according to the actual installation positions of the first reference tag and the second reference tag; the first reference tag and the reference tag are on the same side of the cantilever; A comparison unit, configured to compare the standard coordinates with the coordinate data of the second reference tag to obtain the measurement offset information of the second reference tag; A calibration unit, configured to keep the coordinate data of the first reference tag unchanged, and calibrate the coordinate data of other ultra-wideband tags based on the measurement offset information to obtain calibration positioning information.

10. The unattended pose monitoring system for a bucket wheel stacker-reclaimer according to claim 7, characterized in that, It further includes: A first calculation module, configured to calculate the cantilever inclination based on the height difference between two ultra-wideband tags at both ends of the cantilever; A second calculation module, configured to calculate the vertical offset amplitude and the up-and-down offset amplitude of the cantilever respectively based on the distance between the ultra-wideband tags at both ends of the cantilever to obtain the cantilever deflection; A third calculation module, configured to obtain the wind load information and the cantilever load information, and calculate the distortion risk coefficient through a preset relationship function based on the cantilever inclination and the cantilever deflection; An update module, configured to map the distortion risk coefficient to a color, and modify the cantilever image pixels in the three-dimensional model with the mapped color to obtain an updated three-dimensional model of the cantilever.

Images