Visual three-dimensional positioning method and system for cylindrical charging barrel

Abstract

The invention provides a visual three-dimensional positioning method and system for a cylindrical charging barrel, and relates to the technical field of machine visual localization, the system comprises a 3D imaging unit and a computing unit, the 3D imaging unit performs 3D imaging of the charging barrel from a top view angle to form a point cloud model of the charging barrel; and the calculation unit is used for calculating the three-dimensional coordinates of each charging barrel according to the three-dimensional model. According to the method, after overlook point clouds of a charging barrel are collected, three-dimensional ROI extraction of the point clouds is firstly carried out, and only the point clouds on a tray are reserved; then, extracting a point cloud at a barrel opening of the charging barrel to extract effective circular ring point cloud data; the circular ring point clouds are mapped to an XOY plane, circular ring feature detection of the point clouds is carried out, the number of circular rings is calculated, and the point clouds contained in each circular ring are segmented; and finally, performing circle fitting on the segmented circular ring point clouds one by one to obtain circle center and radius parameters, and outputting all the parameters in combination with the height of the barrel opening to realize high-precision three-dimensional coordinate identification of the cylindrical charging barrel.

Description

Technical Field

[0001] The present invention relates to the technical field of machine vision positioning, and in particular to a visual three-dimensional positioning method and system for a cylindrical barrel. Background Art

[0002] Cylindrical barrels are widely used in the chemical, energy, and pharmaceutical industries, and are usually used for material storage, transportation, and loading. In the automated production process of these industries, barrels need to be accurately grasped and placed by automated robotic arms. Therefore, barrel positioning and posture recognition are one of the key technologies to achieve automated operations.

[0003] However, in the actual production process, the placement of barrels is often disordered, and the position of barrels on the tray may be irregular, the angle is not uniform, or the shape and size change due to wear and tear due to long-term use. These factors make it difficult for traditional positioning methods based on 2D vision to meet actual needs, especially when there are many barrels piled up or the position is complex, the positioning accuracy is difficult to guarantee.

[0004] Currently, the technology for barrel positioning usually relies on two-dimensional vision systems (such as cameras) or laser ranging devices for position detection. These systems rely on specific visual algorithms to identify the edges or contours of the barrel. However, the performance of two-dimensional systems in complex scenes is usually limited, and it is difficult to accurately identify and locate cylindrical objects, especially when the height, angle and relative position of the barrel are not fixed.

[0005] In order to solve these problems, 3D vision technology has gradually been applied to material positioning. 3D point cloud technology can obtain spatial information of objects from different angles, including multi-dimensional data such as position and posture, and has strong adaptability and accuracy. Through 3D imaging technology, the 3D coordinates of the barrel (XY plane position and Z axis height) can be obtained more accurately, providing accurate data support for subsequent robotic arm grasping and automatic loading.

[0006] At present, although there are some positioning technologies based on 3D imaging, most solutions face the following problems:

[0007] 1. Lack of real-time performance: The existing three-dimensional positioning system has slow positioning and calculation speeds in high-speed production lines and cannot meet real-time operation requirements.

[0008] 2. Low accuracy: Due to the influence of the external environment or hardware limitations, many 3D vision systems cannot provide sufficiently accurate 3D coordinates in complex scenes, especially when there are a large number of barrels and they are stacked in a chaotic manner.

[0009] 3. Complex algorithms and high computational complexity: Existing 3D point cloud processing algorithms are relatively complex, involving a large amount of point cloud data processing and optimization calculations. In practical applications, this may result in a longer system response time and affect overall production efficiency.

[0010] Therefore, there is an urgent need for a new three-dimensional positioning system that can simultaneously solve the problems of high precision and high real-time performance, especially in the application scenario of automatic barrel loading. Summary of the invention

[0011] In view of this, the purpose of the present invention is to propose a visual three-dimensional positioning method and system for a cylindrical barrel. Through three-dimensional imaging and precise positioning algorithms, by collecting three-dimensional point cloud data of the barrel, after point cloud processing and three-dimensional coordinate calculation, the positions of multiple barrels on the pallet can be accurately identified, and provided to a robotic arm for accurate grasping and loading, thereby realizing high-precision three-dimensional coordinate recognition of the cylindrical barrel, thereby providing accurate positioning data support for the automated robotic arm and improving the accuracy and efficiency of the automated loading process.

[0012] To achieve the above object, the present invention provides the following technical solutions:

[0013] Based on the above purpose, in a first aspect, the present invention provides a visual three-dimensional positioning method for a cylindrical barrel, comprising the following steps:

[0014] Step 1: Point cloud data acquisition: Obtain the top view point cloud data of the barrel through 3D imaging equipment;

[0015] Step 2: ROI extraction and point cloud cutting: According to the set region of interest (ROI), extract the point cloud data of the barrel area, cut the point cloud data, and retain the point cloud data of the tray area;

[0016] Step 3, barrel mouth point cloud extraction and ring segmentation: According to the point cloud data of the barrel area, the ring point cloud data of the barrel mouth is extracted, the ring point cloud data is mapped to the XOY plane, the ring feature detection of the point cloud is performed, the number of rings is calculated and the ring point cloud set contained in each ring is segmented;

[0017] Step 4: Circle fitting and circle center coordinate calculation: perform circle fitting on each segmented circular point cloud, obtain the circle center coordinate and radius parameters, and calculate the Z-axis coordinate in combination with the height information to obtain the three-dimensional coordinates of each barrel;

[0018] Step 5: Correction of circle center offset: In view of the fitting error caused by surface deformation, an iterative optimization algorithm is used to correct the center position offset and optimize the center coordinates;

[0019] Step 6. Summarize the final results: Calculate the precise three-dimensional position information of each barrel through the above steps, including the center coordinates (X, Y) and height coordinates (Z), and transmit the three-dimensional position information results to the automation equipment to provide a positioning reference for the automation equipment.

[0020] As a further solution of the present invention, extracting the circular point cloud data where the barrel mouth is located includes the following steps:

[0021] According to the collected overhead point cloud data, calculate the point cloud height of the barrel body, and set the current cutting point at a height of 1 / 2 of the distance between the previous cutting point and the highest point;

[0022] Count the number of point clouds above the current cutting point. If the number is greater than the preset threshold, set the next cutting point at a height 1 / 2 of the distance between the current cutting point and the highest point, and count the number of point clouds above the next cutting point.

[0023] This cycle continues until the number of point clouds above the nth cutting point is less than the preset threshold, the bucket mouth is located at the nth cutting point, and the bucket height is obtained.

[0024] As a further solution of the present invention, the circular point cloud segmentation step in step 3 includes projecting the point cloud data onto the XOY plane to form a point set of circular features, and calculating the center position of each circular ring through matrix mapping.

[0025] As a further solution of the present invention, the annular point cloud segmentation comprises the following steps:

[0026] Project the point cloud into the XOY plane to form a plane point set with 1-4 circular ring features;

[0027] Each point in the projected point cloud contains a set of coordinate data, which represent the integer position of each point in the XOY plane;

[0028] Map all points to an m*n matrix M, where m and n are the spans of the point cloud on the X and Y axes;

[0029] Calculate the maximum and minimum values ​​of the point cloud on the X and Y axes, traverse each point in the point set, and calculate the position of each point in the matrix M based on the coordinates of each point;

[0030] Create a matrix M of the same size as the mapping matrix h , calculate the center distance of the points in the matrix M and generate the matrix M h Visualization results of

[0031] For the matrix M h Perform histogram analysis, extract the area representing the ring, and calculate the center position of the circle.

[0032] As a further solution of the present invention, a matrix M having the same size as the mapping matrix is ​​created. h Initialize all positions to 0, traverse all points with a value of 1 in the matrix M, and use this point as the center of the circle to convert the matrix M h All the distances from the center of the circle are r 0 Add 1 to the value in the matrix coordinates of .

[0033] As a further solution of the present invention, when calculating the center position of the circle, the radius r=r 0 +r Δ Draw circles, classify the points in each circle into one category, segment the point set, and transform the matrix M h The points belonging to each circle divided in the figure are mapped back to the XOY plane to obtain the coordinates of the point set represented by each circle.

[0034] As a further solution of the present invention, a circle fitting is performed on each segmented circular point cloud, comprising the following steps:

[0035] Calculate the mean distance from each point in the circle to the center of the circle based on the statistical center of the circle;

[0036] Draw a ring containing inner and outer radii, and count the ratio of the number of inner points to the number of outer points;

[0037] Calculate the distance between each outer point and the center of the circle. When the distance is greater than the average radius, generate a unit vector from the center of the circle to the outer point; when the distance is less than the average radius, generate a unit vector from the outer point to the center of the circle.

[0038] Traverse all external points to obtain several unit vectors, and adjust the center position through iteration until the ratio of the number of internal and external points reaches the maximum value or the maximum number of iterations is reached.

[0039] As a further solution of the present invention, when an iterative optimization algorithm is used to correct the offset of the center position, the center offset is calculated, the distance between the outer point and the center is used to generate a unit vector, the center is moved based on the vector direction, and the center position is optimized.

[0040] In a second aspect, the present invention provides a visual three-dimensional positioning system for a cylindrical barrel, comprising:

[0041] A 3D imaging unit, used to obtain three-dimensional point cloud data of the barrel by scanning;

[0042] The computing unit is used to extract the circular point cloud of the barrel from the point cloud data and perform segmentation and circle fitting, calculate the center coordinates and barrel mouth height of each barrel, and transmit the results to the automation equipment.

[0043] As a further solution of the present invention, the 3D imaging unit is a 3D camera, a multi-line laser radar scanning 3D imaging device or a device for acquiring precise three-dimensional data.

[0044] As a further solution of the present invention, the computing unit includes an image processing module, a circle detection module, a circle fitting optimization module and a result output module, wherein the circle detection module is used to perform circle feature detection and segmentation on the point cloud data mapped to the XOY plane, and the circle fitting optimization module is used to optimize the center coordinates (X, Y) and height coordinates (Z) of each barrel; the result output module is used to transmit the calculated three-dimensional position information to the automation equipment.

[0045] Compared with the prior art, the visual three-dimensional positioning method and system of a cylindrical barrel proposed by the present invention has the following beneficial effects:

[0046] 1. The present invention obtains the top view point cloud data of the barrel through the 3D imaging unit, and combines the annular point cloud segmentation and circle fitting technology to accurately calculate the three-dimensional coordinates of each barrel. By calculating the xy axis center coordinates and z axis height center of the cylindrical barrel, it can provide accurate position information for the robot arm to ensure accurate grasping and dumping of the barrel during the automated loading and unloading process.

[0047] 2. In traditional visual positioning systems, circular objects are often not accurately identified due to reflection, occlusion, etc., especially on barrels with metal reflective characteristics. The present invention can effectively reduce the impact of interference points on positioning accuracy and improve the robustness of the system by performing fine processing on the point cloud, such as extraction of barrel mouth point cloud, segmentation of circular ring point cloud, and circle fitting. In particular, the three-dimensional coordinates of the barrel can still be accurately extracted in the face of irregular placement or material reflection interference.

[0048] 3. The system of the present invention can not only detect the number of barrels through ROI extraction, ring feature detection and circle fitting, but also accurately calculate the center and radius parameters according to the circle fitting results, and further optimize the accuracy of the center. This is of great significance for the automatic positioning, grasping and transportation of barrels in automated production lines, and can greatly improve production efficiency and automation level.

[0049] 4. The point cloud processing technology used in the system of the present invention is highly adaptable and can handle barrels placed at different angles and positions, and can cope with complex scenarios where the barrel shape and reflective characteristics change. Whether it is a single barrel or multiple barrels mixed, the system can automatically identify and accurately locate it. By accurately calculating the three-dimensional coordinates of each barrel, it can provide accurate operating data for the robot arm, thereby ensuring that the robot arm can accurately clamp the barrel and reduce errors and deviations. Especially in automatic feeding tasks, accurate positioning is crucial for the smooth dumping of materials to the feeding port.

[0050] 5. The present invention can efficiently and automatically provide accurate three-dimensional positioning data for the robot arm, greatly reducing manual intervention and improving the automation level of the production line. Through precise positioning, the system can prevent the robot arm from colliding or failing to grasp due to inaccurate barrel positions when multiple barrels are placed at the same time. The present invention is applicable to multiple industry fields, especially in the material transportation and storage links in the chemical, energy, and pharmaceutical industries. Whether in the loading and unloading, handling of barrels, or automatic loading and unloading, it can greatly improve production efficiency, save labor costs, and reduce human errors.

[0051] In summary, the present invention realizes accurate three-dimensional positioning of cylindrical barrels through high-precision three-dimensional visual algorithms combined with point cloud segmentation and circle fitting technology. This can not only improve the efficiency and accuracy of automated production lines, but also effectively cope with various complex operating environments, and has broad application prospects and high commercial value.

[0052] These and other aspects of the present application will be more concise and understandable in the following description of the embodiments. It should be understood that the above general description and the following detailed description are only exemplary and explanatory and cannot limit the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or related technologies, the following briefly introduces the drawings required for use in the exemplary embodiments or related technical descriptions. The drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation to the present invention. In the drawings:

[0054] Figure 1 The present invention is a flowchart of a method for visual three-dimensional positioning of a cylindrical barrel according to an embodiment of the present invention.

[0055] Figure 2 The present invention is a flowchart of calibrating and matching drilling data and remote sensing data in a visual three-dimensional positioning method for a cylindrical barrel in an embodiment of the present invention.

[0056] Figure 3 The present invention is a flowchart of generating geological profile data of geological features based on drilling data in a visual three-dimensional positioning method of a cylindrical barrel in an embodiment of the present invention.

[0057] Figure 4 The present invention is a flowchart of generating landform contour map data corresponding to terrain features according to remote sensing data in a visual three-dimensional positioning method for a cylindrical barrel in an embodiment of the present invention.

[0058] Figure 5The present invention is a flowchart of a gridding process of a three-dimensional space in a method for visual three-dimensional positioning of a cylindrical barrel according to an embodiment of the present invention.

[0059] Figure 6 The present invention provides a flowchart for constructing a preliminary three-dimensional geological grid model in a method for visual three-dimensional positioning of a cylindrical barrel according to an embodiment of the present invention. Figure 7 This is a schematic diagram of a circular ring drawn in a visual three-dimensional positioning method for a cylindrical barrel according to an embodiment of the present invention. Figure 8 The present invention is a schematic diagram of calculating the distance between each external point and the center of a circle in a visual three-dimensional positioning method for a cylindrical barrel according to an embodiment of the present invention. Fig. 9 It is a structural schematic diagram of a visual three-dimensional positioning system for a cylindrical barrel according to an embodiment of the present invention. DETAILED DESCRIPTION

[0060] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0061] In conjunction with the accompanying drawings, some embodiments of the present application are described in detail below. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0062] In the automated industrial process, since the position of the barrel is unknown, the visual system needs to provide the robot arm with the precise three-dimensional position of the barrel, including the coordinates of the center of the cylinder on the xy axis and the height center on the z axis. The present invention proposes a visual three-dimensional positioning method and system for cylindrical barrels. Through three-dimensional imaging and precise positioning algorithms, by collecting three-dimensional point cloud data of the barrel, after point cloud processing and three-dimensional coordinate calculation, the positions of multiple barrels on the pallet can be accurately identified, and provided to the robot arm for accurate grasping and loading, realizing high-precision three-dimensional coordinate recognition of cylindrical barrels, thereby providing accurate positioning data support for the automated robot arm and improving the accuracy and efficiency of the automated loading process.

[0063] See also Figure 1 As shown, an embodiment of the present invention provides a visual three-dimensional positioning method for a cylindrical barrel, the method comprising the following steps:

[0064] Step S10, point cloud data acquisition: obtaining the top view point cloud data of the barrel through a 3D imaging device.

[0065] Step S20, ROI extraction and point cloud cutting: according to the set region of interest (ROI), the point cloud data of the area where the barrel is located is extracted, and the point cloud data is cut to retain the point cloud data of the tray area.

[0066] Step S30, barrel mouth point cloud extraction and ring segmentation: according to the point cloud data of the barrel area, the ring point cloud data of the barrel mouth is extracted, the ring point cloud data is mapped to the XOY plane, the ring feature detection of the point cloud is performed, the number of rings is calculated and the ring point cloud set contained in each ring is segmented.

[0067] Step S40, circle fitting and circle center coordinate calculation: perform circle fitting on each segmented circular point cloud, obtain the circle center coordinate and radius parameters, and calculate the Z-axis coordinate in combination with the height information to obtain the three-dimensional coordinates of each barrel.

[0068] Step S50, center offset correction: for the fitting error caused by surface deformation, an iterative optimization algorithm is used to perform offset correction on the center position to optimize the center coordinates.

[0069] Step S60, summarize the final results: calculate the precise three-dimensional position information of each barrel through the above steps, including the center coordinates (X, Y) and height coordinates (Z), and transmit the three-dimensional position information results to the automation equipment to provide a positioning reference for the automation equipment.

[0070] The visual three-dimensional positioning method for a cylindrical barrel of the present embodiment, after collecting the top-view point cloud of the barrel, first performs a three-dimensional ROI (region of interest) extraction of the point cloud, retaining only the point cloud on the tray to reduce the influence of useless point clouds; then, extracts the point cloud at the barrel mouth to extract valid circular point cloud data; then maps these circular point clouds to the XOY plane, performs circular feature detection of the point cloud, and is used to calculate the number of circular rings and segment the point cloud contained in each circular ring; finally, performs circle fitting on the segmented circular point clouds one by one to obtain the center and radius parameters, and completes the output of all parameters in combination with the height of the barrel mouth.

[0071] Among them, the top view point cloud model of the barrel is as follows: Figure 2 The shape shown is an uneven cylinder, and the material in the barrel reaches about 90% of the barrel height at most, so there will be a stable ring at the barrel mouth. At the same time, due to the metal reflective characteristics, there will be a lot of stray points in the point cloud at the barrel mouth, which will affect the calculation of the barrel mouth height. Based on this feature, the point cloud data of the ring where the barrel mouth is located is extracted, including the following steps:

[0072] Step S301, after each point cloud is collected, the point cloud height h of the barrel body is calculated according to the collected top view point cloud data, and the current cutting point is set at a height of 1 / 2 of the distance between the previous cutting point and the highest point, and the slice thickness is 2 cm;

[0073] Step S302: Count the number of point clouds above the current cutting point. If the number is greater than the preset threshold m, set the next cutting point at a height 1 / 2 of the height between the current cutting point and the highest point, and count the next cutting point s. 2 The number of point clouds above;

[0074] Step S303, repeat this cycle until the cutting point s n The number of point clouds above is less than the preset threshold m, and the barrel mouth is located at the cutting point s n Position, the barrel height H is:

[0075]

[0076] Where n is the count value of the cutting point. At the same time, at the cutting point s n The image is cut into 1-4 circular point clouds with a thickness of 2 cm.

[0077] In this embodiment, the circular point cloud segmentation step in step 3 includes projecting the point cloud data onto the XOY plane to form a point set of circular features, and calculating the center position of each circular ring through matrix mapping.

[0078] In this embodiment, since the distances between the circular rings in the plane point set are uncertain, the number is uncertain, and the circular rings are occasionally stuck together, it is necessary to first segment the circular rings that meet the circular features. The radius of the barrel is known to be r. 0 , the ring point cloud segmentation includes the following steps:

[0079] Project the point cloud into the XOY plane, such as Figure 3 As shown, a plane point set with 1-4 circular ring features is formed;

[0080] Projected point cloud P xy Each point in contains a set of coordinate data (x, y), which represent the integer position of each point in the XOY plane, in units of mm;

[0081] Map all points to an m*n matrix M, where m and n are point clouds P. xy The span on the X and Y axes, in mm;

[0082] Calculate the point cloud P xy The maximum value x on the X and Y axes max ,y max and the minimum value x mix ,y min , then:

[0083] m=x max -x min ;

[0084] n=y max -y min ;

[0085] Then, traverse each point in the point set and calculate the position (row row, column col) of the point in the matrix M according to the coordinates (x, y) of each point. The specific calculation method is:

[0086] row=xx min ;

[0087] col=yy min ;

[0088] The values ​​of the rowth row and colth column of the matrix are set to 1, and all positions in the matrix where no mapping points exist are set to 0.

[0089] Then, create a matrix M of the same size as the mapping matrix M h , and initialize all position values ​​to 0; calculate the center distance of the points in the matrix M to generate the matrix M h Visualization result; then traverse all points with a value of 1 in the matrix M, and use this point as the center of the circle to transform the matrix M h All the distances from the center of the circle are r 0 Add 1 to the value in the matrix coordinates, and finally get the matrix M h The visualization results are as follows Figure 4 As shown, a value of 0 in the matrix is ​​white, and a larger value indicates a darker color.

[0090] Finally, for the matrix M h Perform a histogram analysis on the values ​​of all coordinates in the image, extract the area with the top 20% values, and calculate its average value to get the following result: Figure 5 The statistical center position of each circle shown. According to the statistically obtained coordinates of the center of each circle, with radius r = r 0 +r Δ Draw a circle, and the points in each circle are classified into one category, that is, the point set can be segmented. The segmentation results are as follows Figure 6 Finally, the points belonging to each ring divided in the matrix are mapped back to the XOY plane to obtain the coordinates of the point set represented by each ring. The coordinate reverse mapping method is:

[0091] x=row+x min

[0092] y=col+y min .

[0093] Since the barrels in industrial sites are not standard circles after long-term use, the statistical center of each circle obtained by segmenting the ring point cloud is not an accurate value, but an approximate value. It is necessary to perform circle fitting on the ring to improve the accuracy of the circle center calculation. The specific implementation method is as follows:

[0094] First, according to the statistical center (x c ,y c ), calculate the distance between each point in the ring and the center (x c ,y c ) Where k is the number of points in the circular point set, and i is the index value. Then, Figure 7 As shown, (x c ,y c ) is the center of the circle, is the inner radius, Draw a ring with the outer radius as the inner point and the outer point as the outer point. Count the ratio λ of the number of inner points to the number of outer points.

[0095] Then, if Figure 8 As shown, calculate the distance between each external point and the center of the circle (x c ,y c ) is the distance d, if Then generate a unit vector from the center of the circle to the outer point, if Then generate a unit vector from the outer point to the center of the circle. Traverse all the outer points to obtain several unit vectors. Add these unit vectors together to get the direction in which the center of the circle needs to be offset. The meaning is that the center of the circle moves in this direction with the greatest probability of making λ larger.

[0096] Then, along The calculation method of moving the center of the circle in the direction of the vector and the distance l moved is:

[0097]

[0098] Among them, t is the number of external points, i is the index value, l ε =5mm, if Then l=l ε ,on the contrary

[0099] Finally, the loop calculation follows The center of the circle after the vector moves a distance l (x' c ,y' c ), the ratio of the number of inner points to the number of outer points λ', if λ'>λ, the next cycle will be performed until λ no longer increases or the number of moves reaches the maximum threshold, at which time the center of the circle (x' c ,y'c ) is the final parameter of a ring.

[0100] The final result contains the parameters of 1-4 circles {(x c ,y c ,h c ),…}, where x c ,y c The center data from the circle fitting, h c =H / 2 comes from the bucket height calculated by the ring point cloud segmentation.

[0101] In this step, the present invention can effectively reduce the influence of interference points on positioning accuracy and improve the robustness of the system by performing fine processing on the point cloud, such as extraction of barrel mouth point cloud, segmentation of annular point cloud and circle fitting. In particular, the three-dimensional coordinates of the barrel can still be accurately extracted in the face of irregular placement or material reflection interference.

[0102] The visual three-dimensional positioning method of cylindrical barrels of the present invention can not only detect the number of barrels through ROI extraction, ring feature detection and circle fitting, but also accurately calculate the center and radius parameters according to the circle fitting results, and further optimize the accuracy of the center. This is of great significance for the automatic positioning, grasping and transportation of barrels in automated production lines, and can greatly improve production efficiency and automation level.

[0103] The point cloud processing technology used in the system of the present invention is highly adaptable and can handle barrels placed at different angles and positions, and can cope with complex scenarios where the barrel shape and reflective characteristics change. Whether it is a single barrel or multiple barrels mixed, the system can automatically identify and accurately locate it. By accurately calculating the three-dimensional coordinates of each barrel, it can provide accurate operation data for the robot arm, thereby ensuring that the robot arm can accurately clamp the barrel and reduce errors and deviations. Especially in automatic feeding tasks, accurate positioning is crucial for the smooth dumping of materials to the feeding port.

[0104] It should be noted that the above figures are only schematic illustrations of the processes included in the method according to an exemplary embodiment of the present invention, and are not intended to be limiting. It is easy to understand that the processes shown in the above figures do not indicate or limit the time sequence of these processes. In addition, it is also easy to understand that these processes can be performed synchronously or asynchronously, for example, in multiple modules.

[0105] It should be understood that, although described in a certain order, these steps are not necessarily performed in sequence in the above order. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be performed in other orders. Moreover, a part of the steps of the present embodiment may include a plurality of steps or a plurality of stages, and these steps or stages are not necessarily performed at the same time, but can be performed at different times, and the execution order of these steps or stages is not necessarily performed in sequence, but can be performed in turn or alternately with at least a part of the steps or stages in other steps or other steps.

[0106] In the second aspect of the embodiment of the present invention, see Fig. 9 As shown, the present invention also provides a visual three-dimensional positioning system for a cylindrical barrel, comprising:

[0107] A 3D imaging unit is used to obtain three-dimensional point cloud data of the barrel by scanning; the 3D imaging unit is a 3D camera, a multi-line laser radar scanning 3D imaging device or a device for obtaining accurate three-dimensional data, wherein the 3D imaging unit is a 3D imaging device 1, which is installed on a tray 4 through a bracket 2, and a cylindrical barrel 3 is placed on the tray 4, and the field of view of the 3D imaging device 1 covers the cylindrical barrel 3.

[0108] The computing unit is used to extract the circular point cloud of the barrel from the point cloud data and perform segmentation and circle fitting, calculate the center coordinates and barrel mouth height of each barrel, and transmit the results to the automation equipment.

[0109] Among them, the computing unit includes an image processing module, a circle detection module, a circle fitting optimization module and a result output module, wherein the circle detection module is used to perform circle feature detection and segmentation on the point cloud data mapped to the XOY plane, and the circle fitting optimization module is used to optimize the center coordinates (X, Y) and height coordinates (Z) of each barrel; the result output module is used to transmit the calculated three-dimensional position information to the automation equipment.

[0110] For example, see Figure 4 As shown in the figure, cylindrical barrels are placed on the pallet in irregular positions. The pallet is transported to the inspection area by the conveyor line. The 3D imaging device scans the 3D model (point cloud) of the area where the pallet is located from a top-down perspective, and calculates the number of barrels on the pallet and the 3D coordinates of each barrel based on the 3D model. The 3D coordinates of these barrels will be sent to the robot mechanism, which grabs the barrel and dumps the material to complete the automatic loading task.

[0111] Through the above detailed steps, the visual three-dimensional positioning system of the cylindrical barrel of the present invention is used to execute the steps of the visual three-dimensional positioning method of the cylindrical barrel in the above embodiment, which will not be repeated here. The visual three-dimensional positioning system of the cylindrical barrel of the present invention obtains the top-down point cloud data of the barrel through the 3D imaging unit, and combines the circular point cloud segmentation and circle fitting technology to accurately calculate the three-dimensional coordinates of each barrel. By calculating the xy axis center coordinates and z axis height center of the cylindrical barrel, accurate position information can be provided to the robotic arm to ensure accurate grasping and dumping of the barrel during automated loading and unloading.

[0112] The present invention can efficiently and automatically provide accurate three-dimensional positioning data for the robot arm, greatly reducing manual intervention and improving the automation level of the production line. Through precise positioning, the system can prevent the robot arm from colliding or failing to grasp due to inaccurate barrel positions when multiple barrels are placed at the same time. The present invention is applicable to multiple industry fields, especially in the material transportation and storage links in the chemical, energy, and pharmaceutical industries. Whether in the loading and unloading, handling of barrels, or automatic loading and unloading, it can greatly improve production efficiency, save labor costs, and reduce human errors.

[0113] In summary, this system achieves accurate 3D positioning of cylindrical barrels through high-precision 3D vision algorithms combined with point cloud segmentation and circle fitting technology. This can not only improve the efficiency and accuracy of automated production lines, but also effectively cope with various complex operating environments, and has broad application prospects and high commercial value.

[0114] The above are exemplary embodiments disclosed in the present invention, but it should be noted that various changes and modifications may be made without departing from the scope disclosed in the embodiments of the present invention as defined in the claims. The functions, steps and / or actions of the method claims according to the disclosed embodiments described herein do not need to be performed in any particular order. In addition, although the elements disclosed in the embodiments of the present invention may be described or required in individual form, they may also be understood as multiple unless explicitly limited to the singular.

[0115] It should be understood that, as used herein, the singular form "a" or "an" is intended to include the plural form as well, unless the context clearly supports an exception. It should also be understood that, as used herein, "and / or" refers to any and all possible combinations of one or more of the items listed in association. The serial numbers of the embodiments disclosed in the above embodiments of the present invention are for description only and do not represent the advantages and disadvantages of the embodiments.

[0116] A person skilled in the art should understand that the discussion of any of the above embodiments is only exemplary and is not intended to imply that the scope of the disclosure of the embodiments of the present invention (including the claims) is limited to these examples; under the concept of the embodiments of the present invention, the technical features in the above embodiments or different embodiments can also be combined, and there are many other changes in different aspects of the embodiments of the present invention as above, which are not provided in detail for the sake of simplicity. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the embodiments of the present invention should be included in the protection scope of the embodiments of the present invention.

Claims

1. A visual three-dimensional positioning method for a cylindrical barrel, characterized in that: The method comprises the following steps: Obtain the top view point cloud data of the barrel through 3D imaging equipment; According to the set area of ​​interest, extract the point cloud data of the barrel area, crop the point cloud data, and retain the point cloud data of the pallet area; According to the point cloud data of the area where the barrel is located, the circular point cloud data of the barrel mouth is extracted, the circular point cloud data is mapped to the XOY plane, the circular feature detection of the point cloud is performed, the number of circular rings is calculated and the circular point cloud set contained in each circular ring is divided; Perform circle fitting on each segmented circular point cloud to obtain the coordinates of the circle center and radius parameters, and calculate the Z-axis coordinates in combination with the height information to obtain the three-dimensional coordinates of each barrel; In view of the fitting error caused by surface deformation, an iterative optimization algorithm is used to correct the center position and optimize the center coordinates. The three-dimensional position information of each barrel is calculated, including the center coordinates and height coordinates, and the three-dimensional position information results are transmitted to the automation equipment.

2. The visual three-dimensional positioning method of a cylindrical barrel as claimed in claim 1, characterized in that: Extracting the circular point cloud data of the barrel mouth includes the following steps: According to the collected overhead point cloud data, calculate the point cloud height of the barrel body, and set the current cutting point at a height of 1 / 2 of the distance between the previous cutting point and the highest point; Count the number of point clouds above the current cutting point. If the number is greater than the preset threshold, set the next cutting point at a height 1 / 2 of the distance between the current cutting point and the highest point, and count the number of point clouds above the next cutting point. This cycle continues until the number of point clouds above the nth cutting point is less than the preset threshold, the bucket mouth is located at the nth cutting point, and the bucket height is obtained.

3. The visual three-dimensional positioning method of a cylindrical barrel as claimed in claim 1, characterized in that: When calculating the number of circular rings and segmenting the circular ring point cloud set contained in each circular ring, the circular ring point cloud segmentation step includes projecting the point cloud data onto the XOY plane to form a point set of circular ring features, and calculating the center position of each circular ring through matrix mapping.

4. The visual three-dimensional positioning method of a cylindrical barrel as claimed in claim 3, characterized in that: The ring point cloud segmentation comprises the following steps: Project the point cloud into the X0Y plane to form a plane point set with 1-4 circular ring features; Each point in the projected point cloud contains a set of coordinate data, which represent the integer position of each point in the XOY plane; Map all points to an m*n matrix M, where m and n are the spans of the point cloud on the X and Y axes; Calculate the maximum and minimum values ​​of the point cloud on the X and Y axes, traverse each point in the point set, and calculate the position of each point in the matrix M based on the coordinates of each point; Create a matrix M of the same size as the mapping matrix h , calculate the center distance of the points in the matrix M and generate the matrix M h Visualization results of For the matrix M h Perform histogram analysis, extract the area representing the ring, and calculate the center position of the circle.

5. The visual three-dimensional positioning method of a cylindrical barrel as claimed in claim 4, characterized in that: Create a matrix M of the same size as the mapping matrix h Initialize all positions to 0, traverse all points with a value of 1 in the matrix M, and use this point as the center of the circle to convert the matrix M h Add 1 to the values ​​in the matrix coordinates of all those whose distance from the center of the circle is r0.

6. The visual three-dimensional positioning method of a cylindrical barrel as claimed in claim 5, characterized in that: When calculating the center position of a circle, draw a circle with a radius of r = r0 + rΔ, classify the points in each circle into one category, divide the point set, and convert the matrix M h The points belonging to each circle divided in the figure are mapped back to the XOY plane to obtain the coordinates of the point set represented by each circle.

7. The visual three-dimensional positioning method of a cylindrical barrel as claimed in claim 6, characterized in that: Performing circle fitting on each segmented ring point cloud includes the following steps: Calculate the mean distance from each point in the circle to the center of the circle based on the statistical center of the circle; Draw a ring containing inner and outer radii, and count the ratio of the number of inner points to the number of outer points; Calculate the distance between each outer point and the center of the circle. When the distance is greater than the average radius, generate a unit vector from the center of the circle to the outer point; when the distance is less than the average radius, generate a unit vector from the outer point to the center of the circle. Traverse all external points to obtain several unit vectors, and adjust the center position through iteration until the ratio of the number of internal and external points reaches the maximum value or the maximum number of iterations is reached.

8. The visual three-dimensional positioning method of a cylindrical barrel as claimed in claim 7, characterized in that: When the iterative optimization algorithm is used to correct the offset of the center position of the circle, the center offset is calculated, the distance between the outer point and the center of the circle is used to generate a unit vector, and the center of the circle is moved based on the direction of the vector to optimize the center position.

9. A visual three-dimensional positioning system for a cylindrical barrel, characterized in that: A method for performing a visual three-dimensional positioning of a cylindrical barrel as claimed in any one of claims 1 to 8, the system comprising: A 3D imaging unit, used to obtain three-dimensional point cloud data of the barrel by scanning; The computing unit is used to extract the circular point cloud of the barrel from the point cloud data and perform segmentation and circle fitting, calculate the center coordinates and barrel mouth height of each barrel, and transmit the results to the automation equipment.

10. The visual three-dimensional positioning system for a cylindrical barrel as claimed in claim 9, characterized in that: The computing unit includes an image processing module, a circle detection module, a circle fitting optimization module and a result output module, wherein the circle detection module is used to perform circle feature detection and segmentation on the point cloud data mapped to the XOY plane, and the circle fitting optimization module is used to optimize the center coordinates and height coordinates of each barrel; the result output module is used to transmit the calculated three-dimensional position information to the automation equipment.

Images