Mineral Screening System and Method Based on Robotic Arm

Abstract

This invention relates to the field of robotic arm control technology, and discloses a mineral screening system and method based on a robotic arm. The method includes: acquiring three-dimensional morphological information of each mineral on a conveyor belt using a scanning module, determining the elemental fingerprint of each mineral based on short-wave infrared spectroscopy, and establishing the morphological composition characteristics of the minerals based on mineral coordinates; establishing mineral models based on the three-dimensional morphology of each mineral, and determining the classification and specifications of each mineral on the conveyor belt based on the mineral models and the morphological composition characteristics; screening each mineral based on preset mineral types and specifications, and determining the minerals to be screened; acquiring the grasping posture score of each mineral to be screened, determining the grasping order of each mineral to be screened based on the grasping posture score, generating grasping instructions, and controlling the robotic arm to grasp each mineral to be screened. This invention achieves efficient and accurate mineral classification and automated screening through three-dimensional morphology, spectral composition analysis, and intelligent screening and grasping strategies.

Description

Technical Field

[0001] This invention relates to the field of robotic arm control technology, and more specifically, to a mineral screening system and method based on a robotic arm. Background Technology

[0002] With the continuous development of automation technology, robotic arms are being used more and more widely in the industrial field, especially in the mineral screening process, where their application has improved screening efficiency and accuracy.

[0003] However, current mineral screening tasks typically face challenges such as a wide variety of mineral types and morphologies, and complex screening environments. Therefore, how to accurately and efficiently classify, screen, and grasp minerals has become a crucial issue in current technological research. Furthermore, the mineral screening process requires classification based on the three-dimensional morphological information and elemental composition characteristics of the minerals. Traditional mineral screening methods rely on manual labor or simple robotic arms, lacking precise mineral data support, resulting in low screening accuracy and inefficiency. Simultaneously, when using robotic arms for grasping, the grasping strategy is the most critical technology; however, traditional grasping strategies mostly rely on human experience or simple algorithms, failing to fully utilize precise mineral data, leading to low mineral grasping efficiency and certain operational risks.

[0004] Therefore, there is an urgent need to invent a mineral screening technology using robotic arms to solve the problems of low screening accuracy, poor efficiency, complex screening environment, reliance on human experience in grasping strategies, and insufficient grasping stability in existing technologies for mineral screening based on robotic arms. Summary of the Invention

[0005] In this invention, the "elemental fingerprint" refers to the spectral characteristics of a mineral obtained through short-wave infrared spectroscopy analysis. These characteristics are extracted based on the mineral's spectral absorption and reflection properties and mapped to specific element types through a mineral spectral model, uniquely identifying the mineral's chemical composition. Specifically, determining the elemental fingerprint includes: projecting short-wave infrared light onto the mineral, acquiring spectral data based on the reflected light, extracting spectral feature information, and matching the mineral's element type using a pre-established mineral spectral model and iterative clustering based on elemental correlations.

[0006] In view of this, the present invention proposes a mineral screening system and method based on a robotic arm, aiming to solve the problems of low screening accuracy, poor efficiency, complex screening environment, reliance on human experience in grasping strategy and insufficient grasping stability in the prior art when screening minerals based on robotic arms.

[0007] This invention proposes a mineral screening method based on a robotic arm, comprising:

[0008] The scanning module acquires the three-dimensional morphological information of each mineral on the transport belt, and determines the elemental fingerprint of each mineral based on the short-wave infrared spectrum. The morphological composition characteristics of the mineral are established based on the mineral coordinates.

[0009] Based on the three-dimensional morphology of each mineral, a mineral model is established, and based on the mineral model and the morphological composition characteristics of each mineral, the classification and specifications of each mineral on the transport belt are determined.

[0010] Based on preset mineral types and specifications, each mineral is screened, and the minerals to be screened are determined.

[0011] Obtain the grasping posture score of each candidate mineral, and determine the grasping order of each candidate mineral based on the grasping posture score;

[0012] Based on the grasping order, grasping instructions are generated, and the robotic arm is controlled to grasp each mineral to be screened according to the grasping instructions.

[0013] Furthermore, when acquiring the three-dimensional morphological information of each mineral on the transport belt based on the scanning module, it includes:

[0014] Acquire deformed images of the structured light stripes projected onto each mineral by the scanning module;

[0015] Phase analysis of deformed images based on the phase-shifting method is used to determine the phase distribution information of the mineral surface;

[0016] Based on the baseline distance, projection angle, and imaging angle between the projection unit and the image acquisition unit in the triangulation and scanning module, as well as the phase distribution information of the mineral surface, the three-dimensional morphological information of the mineral is generated.

[0017] Furthermore, when determining the elemental fingerprints of each mineral based on shortwave infrared spectroscopy, the following are included:

[0018] Short-wave infrared light is projected onto each mineral on the transport belt, and the spectral data of each mineral are determined based on the reflected light from the surface of each mineral.

[0019] Based on the spectral absorption and reflection properties of minerals, spectral feature information of minerals is extracted;

[0020] A mineral spectral model is pre-established, and spectral feature information is substituted into the mineral spectral model to determine the elemental fingerprint of the mineral.

[0021] Furthermore, when establishing a mineral spectral model in advance, it includes:

[0022] Acquire the spectral data of each mineral and establish the elemental correlation formula for each mineral based on the spectral data. The spectral data includes spectral characteristics and elemental types of the minerals.

[0023] The distance metric between the correlation expressions of each element is obtained based on Euclidean distance, and a distance matrix between the correlation expressions of each element is constructed based on the distance metric.

[0024] Iterative clustering of the correlations between elements is performed based on the distance matrix, and a mineral spectral model is established based on the correlations between elements after clustering.

[0025] Furthermore, when establishing the morphological composition characteristics of minerals based on mineral coordinates, this includes:

[0026] The three-dimensional coordinate information of each mineral on the transport belt is obtained, and the three-dimensional data of each mineral is determined based on the three-dimensional morphology information of the mineral. The three-dimensional data includes the aspect ratio, sphericity, surface roughness and volume of the mineral.

[0027] Matching minerals based on their elemental fingerprints and spatial coordinates;

[0028] Based on the matching results, the morphological composition characteristics of each mineral on the transport belt are established.

[0029] Furthermore, when establishing mineral models based on the three-dimensional morphology of each mineral, the following steps are included:

[0030] Obtain the three-dimensional morphological features of each mineral, and establish three-dimensional point cloud data of each mineral based on the three-dimensional morphological features;

[0031] Redundant, noisy, and isolated points in the 3D point cloud data of each mineral are removed based on noise filtering.

[0032] The iterative nearest point algorithm is used to align point cloud data of minerals from different perspectives or at different times to generate three-dimensional point cloud models of each mineral.

[0033] Furthermore, based on the mineral models and morphological composition characteristics of each mineral, the determination of the mineral classification and specifications on the transport belt includes:

[0034] Based on the three-dimensional model of minerals and clustering algorithms, the morphological classification of minerals is determined;

[0035] The chemical composition classification of minerals is determined based on the elemental fingerprint and morphological classification in the morphological composition characteristics of minerals.

[0036] Based on the morphological and chemical composition classification of minerals, and using a decision tree algorithm, the classification of each mineral on the transport belt is determined, and based on the classification results, the size classification and mineral composition classification of each mineral on the transport belt are determined.

[0037] Obtain the maximum percentage of each specification category and the maximum percentage of each mineral composition category among the various specification categories of the conveyor belt, and identify the minerals with the maximum percentage of each specification category and mineral composition category as suspected minerals.

[0038] Furthermore, based on preset mineral types and specifications, each mineral is screened, and when determining the minerals to be screened, the process includes:

[0039] Obtain the mineral composition type and mineral specifications of suspected minerals, and determine the minerals to be screened based on the relationship between the mineral composition type and mineral specifications of suspected minerals and the preset mineral composition type and mineral specifications;

[0040] If the mineral composition type of the suspected mineral is consistent with the preset mineral composition type, and the mineral size of the suspected mineral is less than or equal to the preset mineral size, then the suspected mineral is determined not to be screened.

[0041] When the mineral size of a suspected mineral is larger than the preset mineral size, and / or the mineral composition type of a suspected mineral is inconsistent with the preset mineral composition type, the suspected mineral is determined to be a mineral to be screened.

[0042] Furthermore, when obtaining the grasping posture score of each candidate mineral and determining the grasping order of each candidate mineral based on the grasping posture score, the process includes:

[0043] Based on the three-dimensional model of the mineral to be screened, the initial gripping score of the mineral to be screened is determined by calculating the surface normal vector, gripping surface shape and surface features of the mineral to be screened, and using the geometric feature scoring method.

[0044] Based on finite element analysis, the minerals to be screened are simulated and grasped. Based on the stress conditions during the simulated grasping process, the grasping force distribution is determined and a stability score is generated.

[0045] Based on the relationship between the stability score and the configured first and second preset stability scores, an adjustment coefficient is determined, and the initial grasping score of the mineral to be screened is adjusted according to the adjustment coefficient, wherein:

[0046] When the stability score is lower than or equal to the first preset stability score, the adjustment coefficient is determined to be L3;

[0047] When the stability score is higher than the first preset stability score and lower than or equal to the second preset stability score, the adjustment coefficient is determined to be L2.

[0048] When the stability score is higher than the second preset stability score, the adjustment coefficient is determined to be L1;

[0049] Among them, the first preset stability score is less than the second preset stability score, and 0.8 < L1 < L2 < L3 < 1.2;

[0050] The adjusted initial grasp score is determined as the grasp score of the minerals to be screened, and the grasp order of each mineral to be screened is determined based on the reverse order.

[0051] Compared with existing technologies, the advantages of this invention are as follows: By utilizing a scanning module to acquire the three-dimensional morphological information of minerals on the transport belt and combining it with short-wave infrared spectroscopy to extract the elemental fingerprints of the minerals, morphological composition features are established based on the minerals' coordinate information. This method, combining morphological information and compositional analysis, significantly improves the accuracy of mineral identification compared to traditional single image recognition or manual sorting methods, reducing misjudgments caused by factors such as changes in lighting and morphological complexity, thus making mineral classification more refined. Secondly, by establishing a three-dimensional model of the minerals and combining it with morphological composition features to classify and determine the size of the minerals, the accuracy of mineral screening is improved. Traditional mineral screening methods usually rely on human experience or simple rule judgments, while this method, with the help of computer vision and spectral analysis technology, can accurately determine the type and size of minerals, improve the automation of the screening process, reduce human intervention, and reduce screening errors. At the same time, the use of three-dimensional modeling methods can more comprehensively describe the spatial characteristics of minerals, making screening decisions more scientific and reasonable. Furthermore, after mineral screening, by calculating the grasping posture score of the candidate minerals and determining the grasping order based on the score results, the robotic arm can ensure that it executes the optimal grasping strategy, improving the stability and success rate of grasping. Compared with traditional robotic arm grasping methods, this method not only considers the morphological characteristics of minerals but also incorporates factors such as grasping stability analysis and force calculation, enabling the robotic arm to maintain efficient grasping even in complex environments and effectively avoiding drops or damage caused by improper grasping angles or unstable gripping. Finally, intelligent control of the robotic arm is achieved through grasping sequence optimization and automatic generation of grasping instructions. By automatically planning the optimal grasping path based on mineral classification information and grasping scores, the ineffective movement of the robotic arm is reduced, improving overall work efficiency. Compared with existing technologies, this method improves mineral screening accuracy while reducing energy consumption and time costs, enhances the intelligence level of the screening system, and enables it to adapt to different mineral types and production needs, demonstrating high application value.

[0052] On the other hand, this application also provides a mineral screening system based on a robotic arm, comprising:

[0053] The scanning module is configured to acquire the three-dimensional morphological information of each mineral on the transport belt, determine the elemental fingerprint of each mineral based on the short-wave infrared spectrum, and establish the morphological composition characteristics of the mineral based on the mineral coordinates.

[0054] The central control module is electrically connected to the scanning module. The central control module is configured to build a model of each mineral based on its three-dimensional morphology, and to determine the classification and specifications of each mineral on the conveyor belt based on the mineral model and the morphological composition characteristics of each mineral. The central control module is also configured to screen each mineral based on preset mineral types and mineral specifications, and to determine the minerals to be screened.

[0055] The output module is electrically connected to the central control module. The output module is configured to acquire the grasping posture score of each candidate mineral and determine the grasping order of each candidate mineral based on the grasping posture score. The output module is also configured to generate grasping instructions based on the grasping order and control the robotic arm to grasp each candidate mineral based on the grasping instructions.

[0056] It is understood that the mineral screening system and method based on robotic arms in the above embodiments of the present invention have the same beneficial effects, and will not be described again. Attached Figure Description

[0057] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings:

[0058] Figure 1 A flowchart illustrating a mineral screening method based on a robotic arm provided in an embodiment of the present invention;

[0059] Figure 2 This is a functional block diagram of a mineral screening system based on a robotic arm, provided in an embodiment of the present invention. Detailed Implementation

[0060] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the disclosure to those skilled in the art. It should be noted that, unless otherwise specified, embodiments and features in the embodiments of the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0061] like Figure 1 As shown in some embodiments of this application, this embodiment provides a mineral screening method based on a robotic arm, including:

[0062] Step S100: Obtain the three-dimensional morphological information of each mineral on the transport belt based on the scanning module, determine the elemental fingerprint of each mineral based on the short-wave infrared spectrum, and establish the morphological composition characteristics of the mineral based on the mineral coordinates.

[0063] Specifically, when acquiring the three-dimensional morphological information of each mineral on the transport belt based on the scanning module, the process includes: acquiring deformed images of structured light stripes projected by the scanning module onto each mineral; performing phase analysis on the deformed images based on the phase shift method to determine the phase distribution information of the mineral surface; and generating the three-dimensional morphological information of the mineral based on triangulation, the baseline distance between the projection unit and the image acquisition unit in the scanning module, the projection angle and the imaging angle, and the phase distribution information of the mineral surface.

[0064] Understandably, when structured light stripes are projected onto the mineral via a scanning module, the unevenness of the mineral surface causes deformation of the light stripes, which is then captured by the image acquisition unit. The key to this process is ensuring uniform projection of the light stripes and accurate image capture of the deformed image for subsequent analysis of the mineral surface's height variations. Secondly, to further analyze the morphological information of the mineral surface, phase-shifting analysis is performed on the deformed image. Phase-shifting is a common optical measurement technique that calculates the phase distribution information of each pixel by projecting structured light stripes with different phases multiple times. Specifically, by adjusting the phase of the projected light stripes and combining the brightness changes of multiple acquired images, the phase value of each point on the mineral surface is calculated, thus obtaining the phase distribution information of the mineral surface. This phase distribution information accurately reflects the height variations of different areas on the object's surface and is a crucial foundation for subsequent 3D reconstruction. After obtaining the phase distribution information, based on the principle of triangulation and combining parameters such as the baseline distance between the projection unit and the image acquisition unit, the projection angle, and the imaging angle, the 3D coordinates of each point on the mineral surface are calculated. The basic idea of ​​triangulation is to calculate the depth (Z-direction coordinates) of a target point using the known baseline distance and two angles between two points. By traversing all measurement points on the mineral surface, complete three-dimensional morphological information can be reconstructed, thus providing accurate morphological data support for subsequent mineral classification, screening, and grasping.

[0065] Specifically, determining the elemental fingerprint of each mineral based on shortwave infrared spectroscopy includes: projecting shortwave infrared light onto each mineral on the transport belt and determining the spectral data of each mineral based on the reflected light from the surface of each mineral; extracting the spectral feature information of the mineral based on its spectral absorption and reflection characteristics; and pre-establishing a mineral spectral model and substituting the spectral feature information into the mineral spectral model to determine the elemental fingerprint of the mineral.

[0066] Specifically, when establishing a mineral spectral model in advance, the process includes: acquiring spectral data of each mineral and establishing elemental correlations for each mineral based on the spectral data, wherein the spectral data includes spectral features and elemental types of the minerals; acquiring distance metrics between elemental correlations based on Euclidean distance and constructing a distance matrix between elemental correlations based on the distance metrics; performing iterative clustering of elemental correlations based on the distance matrix and establishing a mineral spectral model based on the clustered elemental correlations.

[0067] Understandably, by projecting short-wave infrared light onto minerals on a transport belt, the different absorption and reflection characteristics of each mineral surface result in unique spectral data of the reflected light. Collecting the reflected light data of minerals through a spectral detection module allows for the acquisition of spectral characteristic information for each mineral, laying the foundation for subsequent mineral identification. Secondly, after acquiring the spectral data, spectral characteristic information can be extracted based on the spectral absorption and reflection characteristics. The spectral characteristics of minerals typically manifest as absorption and reflection peaks in specific wavelength bands, and these peaks are closely related to the elemental composition of the mineral. By analyzing the shape of the spectral curve, the position of the absorption peaks, reflectivity, and other information, the compositional characteristics of the mineral can be preliminarily determined. However, since different minerals may have similar spectral characteristics, further spectral modeling methods are needed to improve the accuracy of mineral identification. Furthermore, to achieve high-precision mineral elemental fingerprint identification, a mineral spectral model needs to be established beforehand. This involves acquiring spectral data of different minerals and establishing elemental correlations based on this data, i.e., analyzing the relationship between spectral data and mineral element types. Spectral data includes the spectral characteristics of minerals (such as absorption peaks and reflectance) and their corresponding elemental types. This data can be used to train a mineral spectral model, enabling it to accurately match the elemental fingerprints of minerals. Finally, when constructing the spectral model, a method based on Euclidean distance is used to measure the similarity between different mineral elemental correlations. Euclidean distance is a common metric that calculates the numerical differences between different spectral data, thereby quantifying the relative distance between elemental correlations. By calculating the distance matrix between elemental correlations and performing iterative clustering based on this matrix, mineral categories with similar spectral characteristics can be discovered, and a mineral spectral model can be constructed accordingly. Ultimately, this model can be used to match newly acquired mineral spectral data, thereby accurately identifying the elemental fingerprints of minerals and providing a reliable basis for subsequent mineral screening and classification.

[0068] Specifically, when establishing the morphological composition characteristics of minerals based on mineral coordinates, the process includes: acquiring the three-dimensional coordinate information of each mineral on the transport belt, and determining the three-dimensional data of each mineral based on the three-dimensional morphological information of the minerals, wherein the three-dimensional data includes the aspect ratio, sphericity, surface roughness and volume of the minerals; matching the elemental fingerprint of the minerals with the spatial coordinates of the minerals; and establishing the morphological composition characteristics of each mineral on the transport belt based on the matching results.

[0069] Understandably, the three-dimensional coordinates of minerals on the transport belt are acquired using technologies such as structured light scanning or laser point cloud scanning, and three-dimensional point cloud data of the minerals is constructed based on these coordinates. Then, key geometric features of the minerals, such as aspect ratio, sphericity, surface roughness, and volume, are extracted using point cloud processing techniques. These features accurately reflect the physical morphology of the minerals, providing crucial data support for subsequent analysis. After acquiring the three-dimensional morphological information of the minerals, their compositional characteristics can be further analyzed by combining it with the elemental fingerprint data. The elemental fingerprint of a mineral is obtained through short-wave infrared spectroscopy analysis, reflecting its chemical composition. To establish the correspondence between morphology and composition, the three-dimensional coordinates of the minerals need to be matched with the elemental fingerprints obtained from spectral analysis. The matching process typically employs spatial location mapping technology, that is, a one-to-one correspondence is made between the elemental fingerprint data in the spectral detection area and the three-dimensional coordinate information acquired by the scanning module, thereby ensuring that the morphological features of each mineral are accurately associated with its compositional information. Finally, after completing the matching, statistical analysis or machine learning methods can be used to integrate the three-dimensional morphological features and elemental fingerprint information to establish a mineral morphological composition feature library. This feature database contains the geometric morphological parameters and chemical composition information of each mineral, providing data support for mineral classification, screening, and identification. This approach not only improves the accuracy of mineral identification but also enables the selection of suitable minerals based on morphological and compositional characteristics during the actual screening process, thereby enhancing the precision and efficiency of the screening process.

[0070] As can be seen, this method acquires the three-dimensional morphological information of minerals through structured light scanning and phase-shifting methods, and generates a high-precision three-dimensional mineral model by combining it with the principle of triangulation. Compared with traditional two-dimensional image recognition methods, this scheme can accurately acquire the geometric features of minerals, such as aspect ratio, sphericity, and surface roughness, thereby significantly improving the accuracy of mineral morphology recognition and providing more reliable morphological data support for subsequent screening. Furthermore, by using short-wave infrared spectroscopy to analyze the elemental fingerprints of minerals, it is possible to effectively distinguish minerals with similar chemical compositions but different morphologies, improving the accuracy of mineral screening. Secondly, by establishing a mineral spectral model and using Euclidean distance metric and iterative clustering algorithm to analyze elemental correlations, this scheme can accurately establish a spectral feature classification system for different minerals. Traditional mineral screening methods usually rely on fixed elemental thresholds for judgment, while this scheme, through a data-driven approach, makes mineral composition analysis more intelligent and adaptive, dynamically adjusting classification criteria according to changes in mineral spectra, improving the recognition ability in complex mineral mixture environments. In addition, the morphological composition characteristics of minerals are jointly established by three-dimensional coordinates, morphological data, and elemental fingerprint information, making mineral classification not only dependent on a single information source but also based on comprehensive features, thereby improving screening accuracy. Finally, in practical applications, this solution enables full automation of the mineral screening process, effectively reducing manual intervention and improving production efficiency. Traditional mineral screening methods often rely on manual screening or simple mechanical sorting, while this solution acquires the three-dimensional coordinates of minerals and matches them with elemental fingerprints to form complete morphological-compositional characteristic data, making the screening process more accurate and stable. Simultaneously, this method can adapt to different types of mineral screening needs, improving the level of intelligence in mineral sorting, helping to increase overall screening efficiency, and reducing losses caused by misjudgment or operational errors.

[0071] Step S200: Establish mineral models based on the three-dimensional morphology of each mineral, and determine the mineral classification and specifications on the transport belt based on the mineral models and the morphological composition characteristics of each mineral.

[0072] Specifically, when establishing a model for each mineral based on its three-dimensional morphology, the process includes: acquiring the three-dimensional morphological features of each mineral and establishing three-dimensional point cloud data for each mineral based on these features; removing redundant, noisy, and isolated points from the three-dimensional point cloud data of each mineral based on noise filtering; and aligning the point cloud data of each mineral from different perspectives or at different times based on the iterative nearest point algorithm to generate a three-dimensional point cloud model for each mineral.

[0073] Specifically, when determining the classification and specifications of minerals on the transport belt based on each mineral model and its morphological composition characteristics, the process includes: determining the morphological classification of minerals based on their 3D models and clustering algorithms; determining the chemical composition classification of minerals based on their elemental fingerprints and morphological classifications; determining the classification among minerals on the transport belt based on their morphological and chemical composition classifications, using a decision tree algorithm, and determining the specification classification and mineral composition classification of each mineral on the transport belt based on the classification results; obtaining the maximum percentage of specification classification and the maximum percentage of mineral composition classification among the various specification classifications on the transport belt, and identifying the minerals with the maximum percentage of specification classification and the maximum percentage of mineral composition classification as suspected minerals.

[0074] Understandably, 3D point cloud data processing technology is used to model the 3D morphology of minerals. The acquired 3D morphological features are represented by point cloud data. Since point cloud data typically contains noise points, redundant points, and isolated points, noise filtering algorithms are needed to remove these interfering data, thus ensuring the accuracy of the mineral morphology data. Furthermore, because mineral scanning data from different viewpoints or at different times may have positional deviations, the Iterative Closest Point (ICP) algorithm is used to register and align the point cloud data, ensuring that the final generated 3D mineral point cloud model can completely and accurately represent the morphological features of the mineral. Based on the constructed 3D mineral model, clustering algorithms are used for morphological classification. Since mineral morphological features are similar, clustering algorithms (such as K-means or DBSCAN) can be used to classify the 3D morphological features of minerals and determine their morphological categories. In addition, since minerals not only have geometric morphological features but also contain different elemental composition features, the spectral data of the minerals is further combined to extract elemental fingerprint information, and chemical composition classification is performed using spectral clustering methods. This allows for more accurate mineral classification based on morphological features and elemental fingerprints. Based on mineral morphological and chemical composition classifications, a decision tree algorithm is employed to further optimize the mineral classification process. The decision tree algorithm establishes classification rules based on mineral morphological and chemical composition categories, thereby automatically determining the final category of each mineral. Furthermore, to determine mineral size classifications, this technical solution calculates the proportion of each size category, identifies the mineral with the highest proportion, and classifies it as a potential mineral, ensuring efficient classification of major mineral types during the mineral screening process. This mineral classification method based on multi-dimensional data fusion improves the accuracy and reliability of the screening process.

[0075] It can be seen that by modeling the three-dimensional morphological features of minerals, accurate acquisition and processing of mineral morphology data are achieved. By acquiring the three-dimensional morphological features of minerals and generating three-dimensional point cloud data, and combining noise filtering and the Iterative Closest Point (ICP) algorithm to align point cloud data collected from different perspectives and times, redundant points, noise points, and isolated points can be effectively removed, generating accurate three-dimensional point cloud models and ensuring high precision and stability of mineral morphology data. This provides a reliable data foundation for subsequent mineral classification and screening. Secondly, by combining clustering algorithms to classify the three-dimensional models of minerals, this scheme can perform preliminary screening based on the geometric morphological features of minerals, and further classify the chemical composition of minerals by combining elemental fingerprint analysis. This classification method that combines the morphological features and chemical composition features of minerals greatly improves the accuracy and effectiveness of mineral classification, avoiding the limitations of traditional methods that rely on only a single feature for classification. Finally, by using the decision tree algorithm to comprehensively classify minerals, the size classification and composition classification of minerals on the transport belt can be automatically determined based on the morphological classification and chemical composition classification results. By calculating the maximum percentage of each size classification and mineral composition classification, it is determined as a suspected mineral.

[0076] Step S300: Screen each mineral based on preset mineral types and mineral specifications, and determine the minerals to be screened.

[0077] Specifically, when screening minerals based on preset mineral types and specifications and determining the minerals to be screened, the process includes: obtaining the mineral composition type and specifications of suspected minerals, and determining the minerals to be screened based on the relationship between the mineral composition type and specifications of suspected minerals and preset mineral composition types and specifications; when the mineral composition type of suspected minerals is consistent with the preset mineral composition type, and the mineral specifications of suspected minerals are less than or equal to the preset mineral specifications, then the suspected minerals are determined not to be screened; when the mineral specifications of suspected minerals are greater than the preset mineral specifications, and / or the mineral composition type of suspected minerals is inconsistent with the preset mineral composition type, then the suspected minerals are determined to be screened.

[0078] Understandably, the decision to include suspected minerals in the screening process is based on a comparison of their composition type and specifications. First, the composition type and specifications of suspected minerals are obtained and compared with preset mineral composition types and specifications to establish a relationship. This initial screening effectively reduces the number of minerals that do not meet the screening criteria from entering subsequent processing stages. Second, a rule for judging the consistency of mineral specifications and composition type is used: when the mineral composition type of a suspected mineral matches the preset mineral composition type, and the mineral specification is less than or equal to the preset specification, the suspected mineral does not need to be screened. This helps reduce the probability of misjudgment during the screening process and improves screening efficiency. Conversely, when the mineral specification of a suspected mineral is greater than the preset specification, or the mineral composition type is inconsistent, it needs to be identified as a mineral to be screened. Finally, by establishing a comparison mechanism between mineral composition type and specifications, precise mineral screening is achieved. The matching rule settings ensure that only minerals that meet specific conditions are selected, preventing unqualified minerals from entering the screening process. This principle not only optimizes screening efficiency but also reduces the risk of misselection or omission, improving the overall accuracy and reliability of mineral screening.

[0079] Step S400: Obtain the grasping posture score of each candidate mineral, and determine the grasping order of each candidate mineral based on the grasping posture score.

[0080] Specifically, when obtaining the grasping posture score of each candidate mineral and determining the grasping order of each mineral based on the grasping posture score, the process includes: determining the initial grasping score of each mineral based on a 3D model of the mineral by calculating the surface normal vector, grasping surface shape, and surface features of the mineral using a geometric feature scoring method; simulating the grasping of the mineral based on finite element analysis, determining the grasping force distribution and generating a stability score based on the force conditions during the simulated grasping process; determining an adjustment coefficient based on the relationship between the stability score and the configured first and second preset stability scores, and then determining the grasping order of each mineral based on the adjustment coefficient. The initial grasping score for screening minerals is adjusted as follows: when the stability score is lower than or equal to the first preset stability score, the adjustment coefficient is determined to be L3; when the stability score is higher than the first preset stability score and lower than or equal to the second preset stability score, the adjustment coefficient is determined to be L2; when the stability score is higher than the second preset stability score, the adjustment coefficient is determined to be L1; wherein the first preset stability score is less than the second preset stability score, and 0.8 < L1 < L2 < L3 < 1.2; the adjusted initial grasping score is determined as the grasping score of the minerals to be screened, and the grasping order of each mineral to be screened is determined based on the reverse order.

[0081] Understandably, by calculating the surface normal vector, gripping surface shape, and surface features of the mineral's 3D model, a geometric feature scoring method is used to initially score the gripping of the mineral to be screened. This score assesses the feasibility of gripping based on the geometric characteristics of the mineral's surface, providing basic data for determining the subsequent gripping order. Secondly, finite element analysis is used to simulate the gripping process, evaluating the forces acting on the mineral during gripping. By simulating the mechanical properties of the mineral surface during gripping, the gripping force distribution is obtained, and a stability score is generated. This step ensures the stability of the mineral during gripping through mechanical simulation, thus avoiding problems such as slippage or deformation in actual operation. The stability score directly affects the feasibility and accuracy of gripping; therefore, it is a key factor in determining the mineral gripping order. Finally, an adjustment coefficient is determined based on the relationship between the stability score and the preset first and second stability scores. This coefficient dynamically adjusts the initial gripping score, improving the stability and accuracy of the gripping operation. When the stability score is low, the adjustment coefficient is large, resulting in a lower gripping score; conversely, when the stability score is high, the adjustment coefficient is small, resulting in a higher gripping score. Finally, based on the adjusted grasping score, the minerals to be screened are arranged in reverse order to ensure that minerals with higher stability are grasped first. This method effectively optimizes the grasping strategy for mineral screening and improves the safety and efficiency of the grasping operation.

[0082] Step S500: Generate grasping instructions according to the grasping order, and control the robotic arm to grasp each mineral to be screened according to the grasping instructions.

[0083] In the above embodiments, the three-dimensional morphological information of minerals on the transport belt is acquired using a scanning module, and the elemental fingerprints of the minerals are extracted using short-wave infrared spectroscopy. Then, morphological composition features are established based on the minerals' coordinate information. This method, combining morphological information and compositional analysis, significantly improves the accuracy of mineral identification compared to traditional single-image recognition or manual sorting methods. It reduces misjudgments caused by factors such as changes in lighting and morphological complexity, making mineral classification more refined. Secondly, by establishing a three-dimensional model of the minerals and combining it with morphological composition features to classify and determine their specifications, the accuracy of mineral screening is improved. Traditional mineral screening methods typically rely on human experience or simple rule-based judgments, while this method, using computer vision and spectral analysis technology, can accurately determine the type and specifications of minerals, improving the automation of the screening process, reducing human intervention, and lowering screening errors. Simultaneously, the three-dimensional modeling method can more comprehensively describe the spatial characteristics of minerals, making screening decisions more scientific and reasonable. Furthermore, after mineral screening is completed, by calculating the grasping posture score of the candidate minerals and determining the grasping order based on the score results, it is possible to ensure that the robotic arm executes the optimal grasping strategy, improving the stability and success rate of grasping. Compared to traditional robotic arm grasping methods, this method not only considers the morphological characteristics of minerals but also incorporates factors such as grasping stability analysis and force calculations. This allows the robotic arm to maintain efficient grasping even in complex environments, effectively preventing drops or damage caused by improper grasping angles or unstable gripping. Finally, intelligent control of the robotic arm is achieved through optimized grasping sequence and automatic generation of grasping instructions. By automatically planning the optimal grasping path based on mineral classification information and grasping scores, the ineffective movement of the robotic arm is reduced, improving overall work efficiency. Compared to existing technologies, this method improves mineral screening accuracy while reducing energy consumption and time costs, enhancing the intelligence level of the screening system and enabling it to adapt to different mineral types and production needs, thus possessing high application value.

[0084] In another preferred embodiment based on the above embodiments, such as Figure 2 As shown, this embodiment provides a mineral screening system based on a robotic arm, including a scanning module, a central control module, and an output module.

[0085] Specifically, the scanning module is configured to acquire the three-dimensional morphological information of each mineral on the conveyor belt, determine the elemental fingerprint of each mineral based on short-wave infrared spectroscopy, and establish the morphological composition characteristics of the minerals based on mineral coordinates; the central control module is electrically connected to the scanning module, and is configured to establish a mineral model based on the three-dimensional morphology of each mineral, and determine the classification and specifications of each mineral on the conveyor belt based on the mineral model and the morphological composition characteristics of each mineral; the central control module is also configured to screen each mineral based on preset mineral types and mineral specifications, and determine the minerals to be screened; the output module is electrically connected to the central control module, and is configured to acquire the grasping posture score of each mineral to be selected, and determine the grasping order of each mineral to be screened based on the grasping posture score; the output module is also configured to generate grasping instructions based on the grasping order, and control the robotic arm to grasp each mineral to be screened based on the grasping instructions.

[0086] It is understood that the mineral screening system and method based on robotic arms in the above embodiments of the present invention have the same beneficial effects, and will not be described again.

[0087] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program goods. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program goods embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0088] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program goods according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0089] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1The function specified in one or more boxes.

[0090] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0091] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method of mineral sorting based on a robotic arm, characterized in that, The method comprises the following steps: acquiring three-dimensional shape information of each mineral on the conveying belt based on a scanning module, and determining an element fingerprint of each mineral based on a short-wave infrared spectrum, and establishing a shape composition feature of the mineral based on a mineral coordinate, wherein the element fingerprint is information reflecting a chemical composition of the mineral obtained through short-wave infrared spectrum analysis; establishing a mineral model based on the three-dimensional shape of each mineral, and determining a classification of each mineral and a specification of each mineral on the conveying belt based on the mineral model and the shape composition feature of each mineral; screening each mineral based on a preset mineral type and mineral specification, and determining a mineral to be screened; acquiring a grabbing posture score of each candidate mineral, and determining a grabbing order of each mineral to be screened based on the grabbing posture score; generating a grabbing instruction according to the grabbing order, and controlling a mechanical arm to grab each mineral to be screened according to the grabbing instruction.

2. The robotic arm-based mineral sorting method of claim 1, wherein, When the three-dimensional shape information of each mineral on the conveying belt is acquired based on the scanning module, the method comprises the following steps: acquiring a deformation image of a structured light fringe projected by the scanning module to each mineral; performing phase analysis on the deformation image based on a phase shift method to determine phase distribution information of the mineral surface; generating three-dimensional shape information of the mineral based on triangulation, a baseline distance between a projection unit and an image acquisition unit in the scanning module, a projection angle and an imaging angle, and the phase distribution information of the mineral surface.

3. The robotic arm-based mineral sorting method of claim 2, wherein, When the element fingerprint of each mineral is determined based on the short-wave infrared spectrum, the method comprises the following steps: projecting short-wave infrared light to each mineral on the conveying belt, and determining spectral data of each mineral based on reflected light of the mineral surface; extracting spectral feature information of the mineral based on spectral absorption and reflection characteristics of the mineral; pre-establishing a mineral spectral model, substituting the spectral feature information into the mineral spectral model, and determining the element fingerprint of the mineral.

4. The robotic arm-based mineral sorting method of claim 3, wherein, When the mineral spectral model is pre-established, the method comprises the following steps: acquiring spectral data of each mineral, and establishing an element correlation formula of each mineral based on the spectral data, wherein the spectral data comprises spectral features and an element type of the mineral; acquiring distance measurement between each element correlation formula based on Euclidean distance, and constructing a distance matrix between each element correlation formula based on the distance measurement; iteratively clustering each element correlation formula based on the distance matrix, and establishing the mineral spectral model based on each element correlation formula after clustering.

5. The robotic arm-based mineral sorting method of claim 4, wherein, When the shape composition feature of the mineral is established based on the mineral coordinate, the method comprises the following steps: acquiring three-dimensional coordinate information of each mineral on the conveying belt, and determining three-dimensional data of each mineral based on the three-dimensional shape information of the mineral, wherein the three-dimensional data comprises an aspect ratio, a sphericity, a surface roughness of the mineral, and a volume of the mineral; matching the element fingerprint of the mineral with the spatial coordinate of the mineral; establishing the shape composition feature of each mineral on the conveying belt according to the matching result.

6. The robotic arm-based mineral sorting method of claim 1, wherein, When the mineral model is established based on the three-dimensional shape of each mineral, the method comprises the following steps: acquiring three-dimensional shape features of each mineral, and establishing three-dimensional point cloud data of each mineral based on the three-dimensional shape features; removing redundant points, noise points and isolated points in the three-dimensional point cloud data of each mineral based on noise filtering; aligning each mineral point cloud data from different angles or different times based on an iterative closest point algorithm to generate a three-dimensional point cloud model of each mineral.

7. The robotic arm-based mineral sorting method of claim 6, wherein, When determining the classification of each mineral on the conveying belt and the specification of each mineral based on the mineral model and the morphological component characteristics of each mineral, the method comprises: determining the morphological classification of the mineral based on the three-dimensional model of the mineral and the clustering algorithm; determining the chemical component classification of the mineral according to the element fingerprint and the morphological classification in the morphological component characteristics of the mineral; determining the classification between each mineral on the conveying belt based on the morphological classification and the chemical component classification of the mineral based on the decision tree algorithm, and determining the specification classification and the mineral component classification of each mineral on the conveying belt according to the classification result; obtaining the maximum specification classification proportion and the maximum mineral component classification proportion between each specification classification on the conveying belt, and determining the mineral species of the maximum specification classification proportion and the mineral species of the maximum mineral component classification proportion as the suspected mineral.

8. The robotic arm-based mineral sorting method of claim 7, wherein, When screening each mineral based on the preset mineral type and mineral specification and determining the mineral to be screened, the method comprises: obtaining the mineral component type and the mineral specification of the suspected mineral, and determining the mineral to be screened according to the relationship between the mineral component type and the mineral specification of the suspected mineral and the preset mineral component type and the preset mineral specification; when the mineral component type of the suspected mineral is consistent with the preset mineral component type, and the mineral specification of the suspected mineral is less than or equal to the preset mineral specification, it is determined that the suspected mineral is not the mineral to be screened; when the mineral specification of the suspected mineral is greater than the preset mineral specification, and / or the mineral component type of the suspected mineral is inconsistent with the preset mineral component type, it is determined that the suspected mineral is the mineral to be screened.

9. The robotic arm-based mineral sorting method of claim 1, wherein, When obtaining the grabbing posture score of each candidate mineral and determining the grabbing order of each mineral to be screened according to the grabbing posture score, the method comprises: based on the three-dimensional model of the mineral to be screened, the initial grabbing score of the mineral to be screened is determined by calculating the surface normal vector of the mineral to be screened, the grabbing surface shape and the surface characteristics of the mineral to be screened using the geometric feature scoring method; based on the finite element analysis, the mineral to be screened is simulated to be grabbed, and the grabbing force distribution is determined based on the stress condition in the simulation grabbing process to generate a stability score; based on the relationship between the stability score and the first preset stability score and the second preset stability score, an adjustment coefficient is determined, and the initial grabbing score of the mineral to be screened is adjusted according to the adjustment coefficient, wherein: when the stability score is less than or equal to the first preset stability score, the adjustment coefficient is L3; when the stability score is higher than the first preset stability score, and the stability score is less than or equal to the second preset stability score, the adjustment coefficient is L2; when the stability score is higher than the second preset stability score, the adjustment coefficient is L1; wherein the first preset stability score is less than the second preset stability score, and 0.8 the adjusted initial grabbing score is determined as the grabbing score of the mineral to be screened, and the grabbing order of each mineral to be screened is determined based on the reverse order.

10. A robotic mineral sorting system adapted to perform a robotic mineral sorting method according to any one of claims 1 to 9, wherein, It comprises: a scanning module configured to obtain three-dimensional morphological information of each mineral on the conveying belt, and determine the element fingerprint of each mineral based on short-wave infrared spectrum, and establish the morphological component characteristics of the mineral based on the mineral coordinates; The central control module is electrically connected with the scanning module, and is configured to establish a mineral model according to the three-dimensional shape of each mineral, and determine the classification and specification of each mineral on the conveying belt based on the mineral model and the shape and component characteristics of each mineral; The central control module is further configured to screen each mineral based on the preset mineral type and mineral specification, and determine a mineral to be screened; The output module is electrically connected with the central control module, and is configured to obtain a grabbing posture score of each candidate mineral, and determine a grabbing order of each mineral to be screened according to the grabbing posture score; the output module is further configured to generate a grabbing instruction according to the grabbing order, and control the mechanical arm to grab each mineral to be screened according to the grabbing instruction.

Images