Rapid three-dimensional positioning method for the end face of bundled bars

Abstract

A rapid three-dimensional positioning method for the end face of bundled bars includes the following steps: obtaining the original three-dimensional point cloud of the end face of the bundled bars and completing preprocessing; extracting the end face point cloud of a single bar from the preprocessed point cloud and performing feature dimensionality reduction; translating according to the centroid of the two-dimensional point cloud to move the two-dimensional point cloud of the end face of a single bar to the statistical area; using the normal vector of the end face of a single bar before dimensionality reduction to represent the attitude of the end face center, and combining the position of the two-dimensional point cloud of the end face of a single bar and the depth mean value of the end face point cloud of a single bar after dimensionality reduction to obtain the spatial pose matrix of the bar end face center. This method has high positioning accuracy and fast speed, laying a technical foundation for the production automation of the final bar finishing line.

Description

Technical Field

[0001] The present invention relates to a method for three-dimensional positioning of an end face, and in particular to a method for rapid three-dimensional positioning of the end face of bundled bar materials based on point cloud partition statistics, belonging to the technical field of machine vision. Background Art

[0002] Bar materials are the main products in the current steel industry and are also important raw materials for production and processing in the industrial field. The quality requirements for bar materials in different industries are also different, and the specifications of bar materials produced in the steel industry are also diverse. In order to distinguish bar materials of different types and specifications, it is necessary to paste labels on the bar materials to facilitate the identification of some basic information of the bar materials. Purchasers can learn information such as the diameter, length, production furnace number, composition, and production date of the bar materials through the labels.

[0003] Currently, the production of steel enterprises is gradually moving towards intelligentization, putting forward higher and higher requirements for automation technology. As an important automation equipment, robots need to have strong autonomous intervention capabilities to work under the condition of no human intervention. For industrial robots operating on the end face of bundled bar materials, end face center positioning is the basis for having autonomous intervention capabilities. The end face center positioning of bundled bar materials is based on three-dimensional point cloud data. Compared with traditional two-dimensional vision measurement systems, the measurement accuracy of structured light measurement systems is not easily affected by the environment and is more suitable for complex factory environments. Using a three-dimensional point cloud data acquisition system to scan the end face of bundled bar materials, the acquired field of view point cloud contains a large number of noise points, and there may be partial occlusion on the end face of the bundled bar materials, which all bring difficulties to the end face center positioning of bundled bar materials. Currently, there is no rapid three-dimensional positioning method for the end face of bundled bar materials. Summary of the Invention

[0004] In order to solve the above technical problems, the present invention proposes a method for rapid three-dimensional positioning of the end face of bundled bar materials, providing a technical basis for the automation of bar material production.

[0005] The method for rapid three-dimensional positioning of the end face of bundled bar materials includes the following steps:

[0006] Step 1: On the finishing line of special steel bar materials, use a three-dimensional point cloud data acquisition system to obtain the original three-dimensional point cloud of the end face of bundled bar materials. The original three-dimensional point cloud includes the end face point cloud of bundled bar materials, the support point cloud of bundled bar materials, and some noise points. Perform straight-through filtering and statistical filtering on the obtained original three-dimensional point cloud in sequence to complete preprocessing;

[0007] Step 2: Extract the single bar end face point cloud from the preprocessed point cloud and perform feature dimensionality reduction. The specific steps are as follows: (1) Use clustering segmentation based on Euclidean distance to obtain the non-adhesive single bar end face point cloud and the adhered multiple bar end face point cloud. The non-adhesive single bar end face point cloud is the required single bar end face point cloud; (2) The edge extraction algorithm, improved RANSAC circle fitting algorithm and reverse extraction method are used in turn to obtain the single bar end face point cloud for the adhered multiple bar end face point cloud; (3) According to the principle of rigid body rotation invariance, the feature dimensionality reduction of the obtained single bar end face point cloud is performed to remove Z Direction information is obtained only with X , Y 2D point cloud of the end face of a single bar with directional information;

[0008] Step 3: Calculate the centroid of the 2D point cloud of the end face of a single bar ( x , y ), according to the center of gravity of the two-dimensional point cloud, the two-dimensional point cloud of the end face of a single bar is translated to the statistical area. According to the degree of missing of the two-dimensional point cloud of the end face of a single bar, it is divided into a complete two-dimensional point cloud of the end face of a single bar and an incomplete two-dimensional point cloud of the end face of a single bar, and the positioning is performed respectively. The specific steps for positioning the complete two-dimensional point cloud of the end face of a single bar are: (1) iterative translation is performed according to the regional statistical results to complete precise positioning; (2) the step length of each translation is accumulated to calculate the center position of the two-dimensional point cloud of the end face of a single bar; the specific steps for positioning the incomplete two-dimensional point cloud of the end face of a single bar are: (1) according to the statistical values ​​of each statistical area, the incomplete two-dimensional point cloud of the end face of the bar is translated to one side of the complete arc to complete rough positioning; (2) iterative translation is performed according to the statistical results of the corresponding area on one side of the arc to complete precise positioning; (3) the step length of each translation is accumulated to calculate the center position of the two-dimensional point cloud of the end face of a single bar;

[0009] Step 4: Use the normal vector of the end face of a single bar before dimensionality reduction to represent the center posture of the end face, combined with the position of the two-dimensional point cloud of the end face of a single bar ( x , y ) and the depth mean of the point cloud of the end face of a single bar after dimensionality reduction z , and obtain the spatial pose matrix of the center of the bar end face H , the expression is:

[0010]

[0011] In the formula , R is the rotation matrix during feature dimensionality reduction.

[0012] Beneficial effects of the method of the present invention:

[0013] According to the environment of the bar finishing line production line and the characteristics of the end face of bundled bars, the three-dimensional point cloud of the end face of bundled bars is collected through a three-dimensional point cloud data acquisition system. The point cloud of the end face of a single bar is obtained through the proposed segmentation algorithm. Based on the principle of rigid body rotation invariance, the feature dimension of the obtained point cloud of the end face of a single bar is reduced. A fast positioning method based on partition statistics is used to locate the two-dimensional center of the bar end face. Combining with the center attitude, the positioning of the center of the end face of bundled bars is completed. This method has high positioning accuracy and speed, laying a technical foundation for the final production automation of the bar finishing line. Brief Description of the Drawings

[0014] Figure 1 is a schematic structural diagram of the three-dimensional point cloud data acquisition system adopted by the present invention;

[0015] Figure 2 is a flowchart of the fast three-dimensional positioning method for the end face of bundled bars of the present invention;

[0016] Figure 3 is a statistical area grid division diagram of the present invention. Detailed Embodiment

[0017] Combined with the attached Figures 1-3 Explain the system structure and operation steps of the present invention.

[0018] The fast three-dimensional positioning method for the end face of bundled bars includes the following steps:

[0019] Step 1: On the special steel bar finishing line, use a three-dimensional point cloud data acquisition system to obtain the original three-dimensional point cloud of the end face of bundled bars. The original three-dimensional point cloud includes the point cloud of the end face of bundled bars, the point cloud of the support of bundled bars, and some noise points. The obtained original three-dimensional point cloud is sequentially subjected to direct filtering and statistical filtering to complete preprocessing;

[0020] Step 2: Extract the point cloud of the end face of a single bar from the preprocessed point cloud and reduce the feature dimension. The specific steps are as follows: (1) Use clustering segmentation based on Euclidean distance to obtain the non-adhesive point cloud of the end face of a single bar and the adhesive point cloud of multiple bars. The non-adhesive point cloud of the end face of a single bar is the required point cloud of the end face of a single bar; (2) Sequentially use the edge extraction algorithm, the improved RANSAC circle fitting algorithm, and the reverse extraction method on the adhesive point cloud of multiple bars to obtain the point cloud of the end face of a single bar; (3) According to the principle of rigid body rotation invariance, reduce the feature dimension of the obtained point cloud of the end face of a single bar, remove Z direction information to obtain a single bar end face two-dimensional point cloud with only X , Y direction information;

[0021] Step 3: Calculate the centroid of the two-dimensional point cloud of the end face of a single bar ( x , y), perform translation according to the centroid of the two-dimensional point cloud, move the two-dimensional point cloud of the end face of a single bar to the statistical area, and distinguish it into a complete two-dimensional point cloud of the end face of a single bar and an incomplete two-dimensional point cloud of the end face of a single bar according to the degree of missing of the two-dimensional point cloud of the end face of a single bar, and perform positioning respectively. The specific steps for positioning the complete two-dimensional point cloud of the end face of a single bar are as follows: (1) Perform iterative translation according to the regional statistical results to complete precise positioning; (2) Accumulate the step size of each translation and calculate the central position of the two-dimensional point cloud of the end face of a single bar. The specific steps for positioning the incomplete two-dimensional point cloud of the end face of a single bar: (1) According to the statistical values of each statistical area, move the incomplete two-dimensional point cloud of the end face of the bar to the side of the complete arc to complete rough positioning; (2) Perform iterative translation according to the statistical results of the corresponding area on one side of the arc to complete precise positioning; (3) Accumulate the step size of each translation and calculate the central position of the two-dimensional point cloud of the end face of a single bar.

[0022] Step 4: Characterize the end face center attitude with the normal vector of the end face of a single bar before dimensionality reduction, and combine the position of the two-dimensional point cloud of the end face of a single bar ( x , y ) and the depth mean value of the point cloud of the end face of a single bar after dimensionality reduction z to obtain the spatial pose matrix of the end face center of the bar H , and the expression is:

[0023]

[0024] In the formula , R is the rotation matrix during feature dimensionality reduction.

[0025] The three-dimensional point cloud data acquisition system adopted by the present invention includes an industrial stereo camera 1, a camera bracket 2 and a three-dimensional data acquisition control system 3. The industrial stereo camera 1 is arranged on the camera bracket 2. The camera bracket 2 is placed directly in front of the end face of the bundled bars to be measured. The optical axis of the industrial stereo camera 1 is parallel to the axis of the bundled bars. The field of view of the industrial stereo camera 1 covers the end face of the bundled bars. The power supply and data line of the industrial stereo camera 1 are connected to the three-dimensional data acquisition control system 3.

[0026] In order to verify the feasibility of the positioning method, the present invention simulates the steel mill production environment in the laboratory, obtains the three-dimensional point cloud of the end face of the bundled bars through the three-dimensional point cloud data acquisition system, and in order to position the end face center pose, preprocess, segment, reduce the dimension, position, etc. the three-dimensional point cloud of the end face of the bundled bars collected, and finally complete the positioning of the end face center of the bundled bars.

[0027] 1. Preprocessing of the original point cloud

[0028] Due to the complex industrial site environment, the original point cloud collected by the three-dimensional point cloud data acquisition system contains the end face point cloud of bundled bars, the support point cloud of bundled bars, and some noise points. The present invention preprocesses the original point cloud by successive application of pass-through filtering and statistical filtering to obtain a point cloud with clear boundaries and easy segmentation. The specific steps are as follows:

[0029] (1)Pass-through filtering

[0030] Since there are a large number of fixed point clouds in the point cloud background that are irrelevant to the main structure and will affect target extraction, pass-through filtering is usually the simplest method to remove fixed point clouds. Set upper and lower limit thresholds for the points in the point cloud ([[]] x m , y m , z m ), x for the [[[]] X 1 , X 2 ) axis, y for the [[[]] Y 1 , Y 2 ) axis, z for the [[[]] Z 1 , Z 2 ) axis, and retain the points within the range. Then the point coordinates after pass-through filtering are:

[0031]

[0032] (2)Statistical filtering

[0033] There are still a large number of noise points in the point cloud after pass-through filtering, which seriously affects the segmentation process and will cause errors in locating the center of the end face. Statistical filtering is used to remove the noise points, and the discrete point removal operation of the point cloud is completed without destroying the geometric structure and shape characteristics of the target.

[0034] First, perform a statistical analysis on the neighborhood of each point and calculate the average distance between each point and its nearest [[[]] k neighboring points; second, calculate the mean [[[]] μ and standard deviation [[[]] σ of all the average distances, then the distance threshold [[[]] d max can be expressed as:

[0035]

[0036] αis a constant and a proportionality coefficient; finally, a statistical analysis is performed on the neighborhood of each point again to eliminate points whose average distance from k neighboring points is greater than d max .

[0037] 2. End face point cloud segmentation and feature dimensionality reduction of bundled bars

[0038] Due to the depth difference on the end face of the bundled bars and the certain angle between the camera lens and the end face of the bundled bars, the end face point cloud of the bundled bars will have varying degrees of missing. Direct positioning will cause some missing point clouds to be unable to be accurately positioned. Therefore, it is necessary to segment the end face point cloud of the bundled bars to obtain the end face point cloud of a single bar. The obtained end face point cloud of a single bar has a certain angle with the XOY coordinate plane, which is not conducive to the positioning of the end face center. Therefore, it is necessary to perform dimensionality reduction processing on it. The specific steps are as follows:

[0039] (1)Cluster segmentation based on Euclidean distance to distinguish adhesive point clouds

[0040] Cluster segmentation based on Euclidean distance is Euclidean cluster segmentation. Euclidean distance is the Euclidean metric, which refers to the true distance between two points in m -dimensional space. Suppose the point cloud S 0 ={ p 1 , p 2 ,…, p n}, where the p i and p j of the k neighborhood Euclidean distance d ij is:

[0041]

[0042] Points with a distance less than the distance threshold d 0 are clustered together, and this process is repeated until the number of points in the cluster no longer increases, and the entire clustering process ends.

[0043] Through cluster segmentation, the non-adhesive end face point cloud of a single bar and the adhesive end face point cloud of multiple bars are distinguished. Among them, the non-adhesive single bar is the required end face point cloud of a single bar, and dimensionality reduction processing is directly performed.

[0044] (2)Realize the segmentation of the adhesive point cloud of multiple bars based on multiple fusion algorithms

[0045] First, for the end-face point clouds of multiple adhered bars, an edge extraction algorithm based on normal vectors is used to extract the edge point clouds of the end-face point clouds of multiple bars;

[0046] Secondly, for the extracted edge point clouds, the RANSAC circle fitting algorithm with a given radius threshold is used to perform several rounds of circle fitting on the edge point clouds in sequence, and then the required center data is selected according to the given radius threshold;

[0047] Finally, since the center data obtained above has a large difference from the theoretical center, a circle slightly larger than the radius is taken with the obtained center data as the center to perform reverse extraction on the end-face point clouds of multiple adhered bars. Reverse extraction: obtaining a certain subset within a set by removing its complement.

[0048] (3)Feature Dimensionality Reduction Based on Rigid Body Rotation Invariance

[0049] The end-face point cloud of a single bar obtained by segmentation is a spatial three-dimensional point cloud. To facilitate the positioning of the end-face center, it is necessary to perform feature dimensionality reduction, that is, without changing its spatial geometric features, rotate it to a position parallel to the XOY coordinate plane, and then only extract the X , Y direction information to obtain the two-dimensional point cloud of the end-face of a single bar. The specific steps of feature dimensionality reduction are as follows:

[0050] 1) Calculate the normal vector of the end-face point cloud of a single bar m ;

[0051] 2) Determine the normal vector of the XOY coordinate plane of the end-face of a single bar n ;

[0052] 3) Calculate the rotation axis through the cross product of the normal vector m and the normal vector n ;

[0053] 4) Calculate the angle m between the normal vector n and the normal vector φ ;

[0054] 5) Calculate the rotation matrix R using the Rodriguez formula;

[0055] 6) Multiply each point in the end-face point cloud of a single bar on the left by the rotation matrix R .

[0056] 3. Calculation of the Two-Dimensional Point Cloud Center Position of the End-Face of a Single Bar Based on Partition Statistics

[0057] The statistical area grid division of the present invention is as shown in Figure 3As shown in the figure, it is divided into radius division and angle division. In the radial direction, three concentric circles are used as the division basis. The radius of the middle circle is the same as the radius of the bar, which is r , and the radii of the other two circles are respectively r-t and r+ t, t are the radius division parameters of the statistical area. t The smaller it is, the narrower the neighborhood and the higher the positioning accuracy. According to the equal division principle, the statistical area is divided into angle divisions, and the area is divided into several area grids. In the radial direction, it is stipulated that r < d ≤ r + t is the area A , r-t < d≤r is the area B , d ≤ r - t is the area C , d represents the distance from any point in the area to the origin. In the angular direction, the counterclockwise direction is specified as positive. Starting from the X positive semi-axis, the areas A , area B , area C are numbered in sequence.

[0058] The offset parameter characterizes the step offset of the point cloud data in the statistical area, which is divided into X direction offset and Y direction offset. The included angle θ is introduced to calculate the offset of each direction. The included angle θ is the included angle between the line r connecting the center point M of any partition with a radius of O to the origin MO and the X positive semi-axis in the clockwise direction. Then the offset calculation formula is:

[0059]

[0060] In the formula, k is the offset coefficient. When A i is not zero, k =1; when A i is zero, 1< k <5. When θ is in the first and third quadrants, it is a negative sign, and in the second and fourth quadrants, it is a negative sign.

[0061] The two-dimensional point cloud of the end face of a single bar after dimensionality reduction is randomly distributed in the XOY coordinate system. According to the centroid characteristics, the centroid of the two-dimensional point cloud of the end face of a single bar is calculated ( x c , yc ), translate according to the centroid of the two-dimensional point cloud to move the two-dimensional point cloud of the end face of a single bar to the statistical area. The two-dimensional point cloud of the end face of a single bar obtained by segmentation is divided into a complete two-dimensional point cloud of the end face of a single bar and an incomplete two-dimensional point cloud of the end face of a single bar according to the degree of missingness.

[0062] The specific steps for locating the complete two-dimensional point cloud of the end face of a single bar are as follows:

[0063] Step 1: Statistical grid area A 1 - A n , B 1 -B n Count the number of interior points. If the number of points in the outermost grid area A 1 -A n is less than the given threshold K it means that the precise positioning is completed, and the two-dimensional bar center coordinates are ( x c -∑ s x ,y c -∑ s y ), otherwise go to Step 2;

[0064] Step 2: Determine the offset parameters A 1 - A n , B 1 - B n according to the number of interior points in the statistical grid area s x , s y . Add the offset value to each point in the point cloud to obtain the translated coordinates ( x i - x c + s x , y i - y c + s y ), and repeat Step 1;

[0065] Step 3: According to the centroid ( x c , y c), and the offset parameter calculation is to calculate the two-dimensional center coordinates of the end face of a single bar in the statistical area. The center coordinate calculation formula is:

[0066]

[0067] In the formula n represents the number of repetitions of step 1.

[0068] The specific steps for the two-dimensional point cloud positioning of the end face of a single defective bar are as follows:

[0069] Step 1: Count A 1 - A n , B 1 - B n , C 1 - C n the number of inliers, determine the orientation of the arc side of the end face point cloud of the single defective bar in the statistical area, record the corresponding grid area of the arc side A i , B i , C i , calculate the offset parameters corresponding to the grid areas A i , B i , C i , translate the end face point cloud of the single defective bar. If the value of the grid area A i is zero, repeat step 1 until A i the value is not zero, which means the preliminary positioning is completed;

[0070] Step 2: Count A i , B i and the number of inliers in its multi-neighborhoods. If the number of inliers in the grid area A i and its multi-neighborhood points is less than the given threshold K and B i and the number of inliers in its multi-neighborhoods is approximately equal to the number of points when the corresponding area is completely filled with points, it means the precise positioning is completed. Otherwise, proceed to step 3;

[0071] Step 3: According to the statistical grid area A i , Bi Determining the offset parameter based on the number of multi-neighborhood points s x 、 s y Each point in the point cloud is added with an offset value to obtain the translated coordinates ( x i - x c + s x , y i -y c + s y ), and repeat step 2;

[0072] Step 4: Calculate the two-dimensional center coordinates of the end face of a single bar in the statistical region plane according to the centroid ( x c , y c ), and the calculation formula for the center coordinates is:

[0073]

[0074] In the formula m represents the number of repetitions of step 2, n represents the number of repetitions of step 3.

[0075] 4. Calculation of the pose of the center of the bar end face

[0076] In order to fully describe the pose of any coordinate system in the world coordinate system in space, it can be divided into two parts. One part is to describe the relative position relationship of the coordinate system with respect to the world coordinate system, and the other part is to describe the attitude of the coordinate system relative to the world coordinate system, so that the pose of the coordinate system in the world coordinate system can be determined.

[0077] To determine the relative position relationship of the center of the end face, it is necessary to determine the information of the center of the end face in X 、 Y 、 Z directions.

[0078] The position of the two-dimensional point cloud of the end face of a single bar obtained by the positioning in step 3 ( x , y ), combined with the depth mean value z of the point cloud of the end face of the single bar after dimensionality reduction, the relative position of the center of the end face is ( x , y, z ).

[0079] Since there may be no actual center for a defective single bar, the normal vector of the end face of the single bar mTo represent the attitude of the end face center, the rotation matrix is the same as the rotation matrix during dimensionality reduction, that is, the center attitude matrix of the end face is:

[0080]

[0081] Combining the obtained center position and attitude of the end face, the spatial pose matrix of the center of the bar end face is obtained H , and the expression is:

[0082]

Claims

1. A method for rapid three-dimensional positioning of the end face of bundled bars, comprising the following steps: Step 1: On the finishing line of special steel bars, a three-dimensional point cloud data acquisition system is used to obtain the original three-dimensional point cloud of the end face of the bundled bars. The original three-dimensional point cloud includes the end face point cloud of the bundled bars, the support point cloud of the bundled bars, and some noise points. The obtained original three-dimensional point cloud is sequentially subjected to direct filtering and statistical filtering to complete preprocessing; Step 2: Extract the end face point cloud of a single bar from the preprocessed point cloud and reduce the features. The specific steps are as follows: (1) Use clustering segmentation based on Euclidean distance to obtain the end face point cloud of non-adhesive single bars and the end face point cloud of adhesive multiple bars. The end face point cloud of non-adhesive single bars is the required end face point cloud of a single bar; (2) For the end face point cloud of adhesive multiple bars, use an edge extraction algorithm, an improved RANSAC circle fitting algorithm, and a reverse extraction method in sequence to obtain the end face point cloud of a single bar; (3) Perform feature dimensionality reduction on the end-face point cloud of the obtained single bar, removing Z direction information to obtain a two-dimensional end-face point cloud of a single bar that only has X , Y direction information. The specific steps of feature dimensionality reduction are as follows: 1) Calculate the normal vector of the end-face point cloud of a single bar m ; 2) Determine XOY the normal vector of the coordinate plane n ; 3) By the normal vector m Cross-multiply the normal vectors n Calculate the axis of rotation; 4) Calculate the normal vector m and the normal vector n angle φ ; 5) Calculate the rotation matrix using Rodriguez's formula R ; 6) Multiply each point in the end-face point cloud of a single bar by the rotation matrix R ; Step 3: Calculate the centroid of the 2D point cloud of the end face of a single bar ( x , y ), according to the center of gravity of the 2D point cloud, the 2D point cloud of the end face of a single bar is translated to the statistical area. According to the degree of loss of the 2D point cloud of the end face of a single bar, it is distinguished into a complete 2D point cloud of the end face of a single bar and an incomplete 2D point cloud of the end face of a single bar, and the positioning is performed separately. The specific steps for positioning the complete 2D point cloud of the end face of a single bar are as follows: (1) Perform iterative translation according to the regional statistical results to complete precise positioning; (2) Accumulate the step sizes of each translation and calculate the center position of the two-dimensional point cloud of the end face of a single bar; Specific steps for positioning the incomplete two-dimensional point cloud of the end face of a single bar: (1) According to the statistical values of each statistical region, translate the incomplete two-dimensional point cloud of the bar end face towards the side of the complete arc to complete rough positioning; (2) Perform iterative translation according to the statistical results of the corresponding region on one side of the arc to complete precise positioning; (3) Accumulate the step sizes of each translation and calculate the center position of the two-dimensional point cloud of the end face of a single bar; Step 4: Characterize the attitude of the end face center with the normal vector of the end face of a single bar before dimensionality reduction, and combine the position of the two-dimensional point cloud of the end face of a single bar ( x , y ) and the depth mean value of the point cloud of the end face of a single bar after dimensionality reduction z to obtain the spatial pose matrix of the end face center of the bar H , the expression is: ， In the formula , R is the rotation matrix during feature dimensionality reduction.

2. The method for rapid three-dimensional positioning of the end face of bundled bars according to claim 1, characterized in that: The three-dimensional point cloud data acquisition system used in Step 1 of the method includes an industrial stereo camera (1), a camera support (2), and a three-dimensional data acquisition control system (3). The industrial stereo camera (1) is arranged on the camera support (2). The camera support (2) is placed directly in front of the end face of the measured object, the bundled bars. The optical axis of the industrial stereo camera (1) is parallel to the axis of the bundled bars. The field of view of the industrial stereo camera (1) covers the end face of the bundled bars. The power supply and data line of the industrial stereo camera (1) are connected to the three-dimensional data acquisition control system (3).

Images