JP2012163451A - Shape recognition device, shape recognition method, and program thereof - Google Patents

Abstract

PROBLEM TO BE SOLVED: To properly abstract a recognition object through a three-dimensional model having a variable part of a cable or the like added to a fixed part of a connector or the like.SOLUTION: A shape recognition device 120 includes: a model holding part 210 which holds three-dimensional model information showing a three-dimensional model 122 including the fixed part and the variable part having the internal degree of freedom and having a longitudinal axis; a recognition target information acquisition part 230 which acquires recognition target information showing a three-dimensional shape of the recognition object 112; and a recognition object specification part 234 which specifies the position and attitude of the recognition object by preparing two candidate attitudes symmetrical to the longitudinal axis and pattern-matching the recognition object and the three-dimensional model changing the internal degree of freedom of the variable part on the basis of the recognition target information and the three-dimensional information.

Description

  The present invention relates to a shape recognition apparatus, a shape recognition method, and a program for recognizing the position and orientation of an arbitrary recognition object.

  In assembly processing in factory automation, a connection operation between a plurality of electronic devices is also performed using a cable. When attempting to connect such cables using an industrial robot, the robot autonomously recognizes the connector at the end of the cable and the cable near it, and specifies the position and orientation of the connector. Must be picked.

  However, with the current technology, only a simple connector or a single colored connector other than a cable can recognize the shape. Also, even if the connector is configured with a simple shape or special coloration, the cable is irregularly accommodated, the cable covers part of the connector, or the color colored by ambient light cannot be recognized correctly In this case, the connector could not be accurately identified. Further, various patterns are usually shown on the cable, and the pattern (texture) may be erroneously recognized as the edge of the cable or connector. If the connector cannot be accurately recognized in this manner, the connector may be unconnected or the connector may be erroneously connected to an incorrect connection position.

  As a technique for preventing unconnected or erroneous connection of a connector, for example, an RF-ID tag is built in a connector portion of a cable with a connector, and a reading circuit is arranged on the corresponding jack component, so that the cable with the connector and the jack component are arranged. Has been disclosed (for example, Patent Document 1). In addition, a technique in which a predetermined portion of a connector or a cable emits light by pressing a pressure switch provided at an arbitrary point of the cable among cables with a connector is also known (for example, Patent Document 2). .

JP 2004-349184 A JP 2006-228482 A

  However, the technique of Patent Document 1 is a technique for confirming the connection relationship of connectors afterwards, and the technique of Patent Document 2 is a technique for recognizing an arbitrary cable by a person. In addition, both technologies cannot be applied to existing connectors and cables, assuming that connectors and cables are processed.

  In addition, with the conventional technology, it is necessary to recognize the connector alone, and the cable extending from the connector cannot be physically disconnected, which may be a detrimental effect on shape recognition, and can improve the connector recognition accuracy. There wasn't.

  In view of such problems, the present invention provides a shape recognition device, a shape recognition method, and a program for accurately extracting a recognition object through a three-dimensional model in which a variable portion such as a cable is added to a fixed portion such as a connector. The purpose is to provide.

  In order to solve the above-described problem, a shape recognition device of the present invention includes a model holding unit that holds a three-dimensional model information indicating a three-dimensional model having a longitudinal axis, including a fixed unit and a variable unit having an internal degree of freedom. Based on the recognition target information acquisition unit that acquires the recognition target information indicating the three-dimensional shape of the recognition target, the recognition target information and the three-dimensional model information, two candidate postures that are symmetric with respect to the longitudinal axis are created, and the recognition target A recognition object specifying unit that specifies the position and orientation of the recognition object by pattern matching the object and a three-dimensional model in which the internal degree of freedom of the variable part is changed.

  The shape recognition device further includes an end cutout unit that cuts out a fixed portion side end portion of the variable portion candidate portion when the portion that is a candidate for the variable portion of the recognition object is extended. Also good.

  The end cutout portion may cut out a portion divided into two by a circle determined based on the maximum value of the linear distance between any two points on the cross section perpendicular to the extending direction of the variable portion.

  The recognition object specifying unit specifies a part matching the fixed part of the three-dimensional model in the recognition object, and changes the internal degree of freedom of the recognition object and the variable part using the position and orientation of the specified part. The position and orientation of the recognition target object may be specified by pattern matching with the three-dimensional model.

  In order to solve the above-mentioned problem, a shape recognition method of the present invention preliminarily holds three-dimensional model information indicating a three-dimensional model having a longitudinal axis including a fixed portion and a variable portion having an internal degree of freedom. Recognition target information indicating the three-dimensional shape of the object, and based on the recognition target information and the three-dimensional model information, create two candidate postures symmetric with respect to the longitudinal axis. The position and orientation of the recognition target object are specified by pattern matching with a three-dimensional model in which the angle is changed.

  In order to solve the above problem, the computer includes a recognition target information acquisition unit that acquires recognition target information indicating the three-dimensional shape of the recognition target, a recognition target information, a fixed unit, and a variable unit having internal degrees of freedom. And two candidate postures symmetric with respect to the longitudinal axis based on the 3D model information indicating the 3D model having the longitudinal axis and the internal degrees of freedom of the recognition object and the variable part are changed. A program for functioning as a recognition target specifying unit that specifies the position and orientation of a recognition target by pattern matching with a dimensional model is provided.

  According to the present invention, a recognition object can be accurately extracted through a three-dimensional model in which a variable part such as a cable is added to a fixed part such as a connector.

It is explanatory drawing which showed the rough connection relation of the picking system. It is explanatory drawing for demonstrating the outline | summary of the pattern matching of a recognition target object. It is a functional block diagram showing a schematic configuration of a shape recognition device It is explanatory drawing which illustrated the three-dimensional model which integrally formed the fixing | fixed part and the variable part. It is explanatory drawing for demonstrating operation | movement of an edge part cutting part. It is explanatory drawing for demonstrating an example of the concrete cut-out operation | movement of an edge part cut-out part. It is the flowchart which showed the flow of the process of the shape recognition method.

  Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. The dimensions, materials, and other specific numerical values shown in the embodiments are merely examples for facilitating the understanding of the invention, and do not limit the present invention unless otherwise specified. In the present specification and drawings, elements having substantially the same function and configuration are denoted by the same reference numerals, and redundant description is omitted, and elements not directly related to the present invention are not illustrated. To do.

(Picking system 100)
FIG. 1 is an explanatory diagram showing a schematic connection relationship of the picking system 100. The picking system 100 includes a three-dimensional shape measuring device 110, a shape recognition device 120, a robot control device 130, and a robot 140. The recognition object 112 targeted by the present embodiment includes not only a fixed part that does not change the model itself but also a unit in which a fixed part and a variable part having an internal degree of freedom are integrally formed. Such an internal degree of freedom means not a spin or the like of the appearance of the fixed part used in normal pattern matching but a relative change inside the fixed part. Here, in order to facilitate the understanding, description will be made using a recognition object 112 in which a connector (fixed part) and a cable (variable part) extended from the connector are integrally formed. Needless to say, it is not limited to cables. In such shape recognition in the picking system 100, the recognition target object 112 generally has a longitudinal axis (longitudinal direction) due to a shape deviation except for a sphere-like object that is completely symmetric with respect to all directions. The longitudinal axis refers to an axis along a direction in which the shape spread is large.

  The three-dimensional shape measuring apparatus 110 is constituted by, for example, a laser scanner type three-dimensional distance sensor, and the cable as the variable portion is randomly accommodated in a state where the cable is irregularly bent (wrapped with a tuna) 1 or The three-dimensional coordinates of the plurality of detection points on the measurement surface (surface) of the plurality of recognition objects (work: here connectors and cables) 112 are detected.

  Specifically, the three-dimensional distance sensor as the three-dimensional shape measurement apparatus 110 includes a Z-axis coordinate indicating the distance from the three-dimensional shape measurement apparatus 110 of a plurality of detection points on the measurement surface and a surface perpendicular to the detection direction. The X and Y axis coordinates on the (XY plane) are combined into three-dimensional coordinates, and recognition target information that is information indicating the three-dimensional coordinates of all of the plurality of detection points is generated. The recognition target information generated in this way is transmitted to the shape recognition device 120. Here, three-dimensional coordinates at a plurality of detection points on the recognition target object are transmitted as the recognition target information. However, the present invention is not limited to this, and various information representing the three-dimensional shape of the recognition target object is transmitted. Can do. Moreover, although the three-dimensional distance sensor is mentioned as the three-dimensional shape measuring apparatus 110, it is possible to employ | adopt various existing sensors which can grasp | ascertain a three-dimensional shape.

  The shape recognition device 120 extracts the recognition target object 112 based on the recognition target information received from the three-dimensional shape measurement device 110, and the recognition target 112 is indicated by the three-dimensional model 122 by pattern matching or the like. The object 112 is recognized and the position and orientation of the recognition object 112 are specified.

  The robot control device 130 receives information indicating the position and orientation of the one recognition object 112 recognized as described above from the shape recognition device 120, and issues a control command for causing the robot 140 to pick the recognition object 112. . In response to the control command, the robot 140 passes the specified one recognition object 112, for example, a connector or a cable in the vicinity thereof, from the plurality of recognition objects 112 stored irregularly through the suction band 142. Pick.

  In conventional picking systems, it is necessary to recognize the connector alone, and the cable extending from the connector cannot be physically disconnected, which may be a detrimental to shape recognition, and the connector recognition accuracy cannot be increased. It was. In this embodiment, the connection position of the connector and the shape of the cable are also used as elements that characterize the connector, so that the position and orientation of the connector are specified with high accuracy.

  However, since a variable part such as a cable changes flexibly, if an attempt is made to perform pattern matching on all the extended cables, the amount of calculation becomes enormous, which may result in a problem that matching is impossible. Here, it is also intended to incorporate the change range of the variable portion into the pattern matching parameters by limiting the change range (degree of freedom) of the variable portion.

  FIG. 2 is an explanatory diagram for explaining an outline of pattern matching of the recognition target object 112. In particular, FIG. 2A shows a case where the recognition object 112 is a single connector, FIG. 2B shows a case where a part of the connector and the cable is used as the recognition object 112, and FIG. FIG. 2B shows a case where pattern matching is performed by further changing the cable within the change range.

  For example, it is assumed that there is a recognition target object 112 and a three-dimensional model 122 that image only a cylindrical connector having symmetry and a longitudinal axis. However, in FIG. 2, since the recognition object 112 and the three-dimensional model 122 are represented in a side view, the shape is rectangular. Here, in order to facilitate understanding, translational conversion has already been performed, and the longitudinal axis of the connector which is the recognition target object 112 is 270 in the clockwise direction in FIG. 2A with respect to the longitudinal axis of the three-dimensional model 122. It is in a rotated state (a state where the rotational angle is 0 ° in FIG. 2A), and the spin angle around the longitudinal axis is not considered.

  In pattern matching, first, the direction in which the degree of coincidence between the recognition object 112 and the three-dimensional model 122 increases with reference to the three-dimensional model 122 (indicated by an arrow along the locus of coincidence in FIG. 2A). Next, the recognition object 112 is rotated. When the degree of coincidence settles to an extreme value, that point is used as a convergence point for pattern matching.

  In FIG. 2A, a convergence point (extreme value) having a high degree of coincidence can be obtained between the two postures when the recognition object 112 has a 90 ° posture and the 270 ° posture. . Of these, the 270 ° posture is the correct posture, and the 90 ° posture is the wrong posture. Here, a correct posture can be extracted by evaluating the degree of coincidence.

  FIG. 2B shows pattern matching between the three-dimensional model 122 obtained by adding a part of the cable to the connector and the recognition target object 112 corresponding thereto. Here, as the elements of the pattern matching, the connection position with the cable in the connector and the shape of the cable are included. Therefore, pattern matching is performed more strictly, and at least an incorrect posture (of two extreme positions) 90 °) is low. Further, in the correct posture (270 °), the degree of coincidence increases depending on the cable posture of the recognition target object 112, and the difference from the wrong posture (90 °) increases, so compared with FIG. It becomes easier to specify the position and orientation of the recognition object 112. In the recognition target object having the longitudinal axis, the degree of coincidence increases when the longitudinal direction is aligned in the forward direction and the reverse direction. When either the forward direction or the reverse direction is found, a candidate posture in the opposite direction is created and the two final matching results are compared, so that it is possible to avoid obtaining an incorrect posture.

  Further, in FIG. 2C, not only pattern matching is performed on the three-dimensional model 122 obtained by adding a part of the cable to the connector and the recognition target object 112 corresponding thereto, but the cable 114 portion of the three-dimensional model 122 is known. The range is changed within the bending range (change range) and is adjusted to the recognition object 112. At the time of pattern matching, if the cable is bent except for the correct posture (270 °), the degree of matching is not so high. However, when the cable approaches the correct posture (270 °), the degree of matching is remarkably improved due to the change in cable bending. In this way, the difference between the correct posture (270 °) and the wrong posture (90 °) is further increased, so that it becomes easier to specify the position and posture of the recognition object 112.

  The shape recognition device 120 of the present embodiment can accurately identify the recognition target object 112 by integrally forming a fixed portion and a variable portion as the three-dimensional model 122 and performing pattern matching. Below, the specific structure of the shape recognition apparatus 120 is shown, and the flow of a specific process (shape recognition method) is explained in full detail after that.

(Shape recognition device 120)
FIG. 3 is a functional block diagram illustrating a schematic configuration of the shape recognition device 120. The shape recognition apparatus 120 includes a model holding unit 210, a data buffer 212, and a central control unit 214.

  The model holding unit 210 includes a ROM, a nonvolatile RAM, a flash memory, an HDD (Hard Disk Drive), and the like, and holds 3D model information indicating the 3D model 122 corresponding to the recognition target object 112. In the present embodiment, the model holding unit 210 holds 3D model information indicating a 3D model 122 in which a fixed part and a variable part are integrally formed, in addition to 3D model information indicating a 3D model having only a fixed part. To do.

  FIG. 4 is an explanatory view illustrating a three-dimensional model 122 in which a fixed portion and a variable portion are integrally formed. For example, in FIG. 4A, a connector 250 that does not have a flexible element is used as a fixed portion, and a cable 252 that can be bent in an arc shape within a change range of an arrow is used as a variable portion. Further, as shown in FIG. 4B, the connector 250 may be provided with a locking mechanism 254 to a receptacle which is a connection destination when the connector 250 is a plug. After the connector 250 is connected to the receptacle, The lock mechanism 254 is rotated in the direction of the arrow to prevent it from falling off. Such a lock mechanism 254 can also be handled as a variable portion having a predetermined rotation range (change range) with respect to the fixed portion.

  Other examples of the variable portion with respect to the fixed portion include other latch mechanisms (variable portions) and slide bars (variable portions) of the connector (fixed portion), the presence / absence of a lid (variable portion) of the connector (fixed portion), and the lid. Various recognition objects 112 such as a restraining member (variable portion) such as a loss prevention chain, a volume in an electronic device (fixed portion), a switch, and a receptacle locking mechanism (variable portion) can be assumed.

  The data buffer 212 is configured by SRAM, DRAM, or the like, and temporarily holds recognition target information received from the three-dimensional shape measuring apparatus 110.

  The central control unit 214 includes a central processing unit (CPU), a signal processing unit (DSP: Digital Signal Processor), a ROM and a memory storing programs, a semiconductor integrated circuit including a RAM as a work area, and the like. Manage and control the entire 120. In the present embodiment, the central control unit 214 also functions as a recognition target information acquisition unit 230, an end cutout unit 232, and a recognition target object specifying unit 234.

  The recognition target information acquisition unit 230 acquires recognition target information indicating the three-dimensional shape of an arbitrary recognition target object 112 from the three-dimensional shape measurement apparatus 110 and stores it in the data buffer 212.

  When the part which becomes the candidate for the variable part of the recognition object 112 is extended, the end part cutout part 232 uses the part (label) at the fixed part side end part (label) of the part which becomes the candidate for the variable part as the variable part. cut. Here, a label shows the curved surface connected continuously and smoothly. Originally, when a series of labels is specified, where all the labels should be variable parts, here, the end cutout part 232 dares to divide the series of labels and define the fixed part side end part. Only the label shown is a variable part.

  FIG. 5 is an explanatory diagram for explaining the operation of the end cutout 232. Here, the connector 250 is taken as an example of the fixed portion, and the cable 252 is taken as an example of the variable portion. When the connector 250 is connected to the end of the cable 252 and the cables 252 are little overlapped or submerged, a portion (label) regarded as a series of cables 252 is judged to be long as shown in FIG. Sometimes. However, if the label of the cable 252 is large, the range of change as the variable part is also large, and the calculation becomes enormous. Therefore, in this embodiment, in order to execute pattern matching with the recognition object 112 and the three-dimensional model 122 in which the fixed portion and the variable portion are integrally formed as shown in FIG. Are cut out only the necessary variable portion, in this case, as shown in FIG. 5C, at the connector 250 (fixed portion) side end portion of the cable 252 and the others are excluded.

  By doing this, it is possible to exclude the label of the cable having a large change range from the candidates before the pattern matching. In addition, by preparing the variable unit 252 having the same size as the three-dimensional model 122, it is possible to perform pattern matching more accurately. With this configuration, it is possible to increase the speed and robustness of calculation processing.

  As an example of cutting out a candidate portion for the variable portion, the end cutout portion 232 is a circle determined based on the maximum value of the linear distance between any two points in the cross section perpendicular to the extending direction of the variable portion. A part that is divided into two parts (a part that is not divided into three parts) is regarded as an end and cut out.

  FIG. 6 is an explanatory diagram for explaining an example of a specific cutout operation of the end cutout unit 232. The end cutout unit 232 first sets a circle 256 as an index for determining the cutout position shown in FIG. The radius of the circle 256 is not less than the maximum value of the linear distance between any two points on the cross section perpendicular to the extending direction of the variable portion. Here, the cross section perpendicular to the extending direction of the variable part is, for example, a cross section (circle) perpendicular to the extending direction when the variable part is shown in a cylindrical shape, and the maximum linear distance between any two points. The value is the diameter of the surface (circle) when the variable part is shown in a cylindrical shape. That is, the end cutout 232 has a circle 256 with a radius equal to or greater than the width of the variable portion in the short direction.

  Then, as shown in FIG. 6A, the end cutout portion 232 forms a circle 256 around an arbitrary point 258 on the label indicating the cable 252 as the variable portion. In the example of FIG. 6A, the label of the cable 252 is divided into three (1), (2), and (3) by a circle 256. The end cutout 232 does not determine the point 258 at the center of the circle 256 that divides the label of the cable 252 into three as the end of the cable.

  The end cutout 232 shifts an arbitrary point 258 to form, for example, a circle 256 shown in FIG. In the example of FIG. 6B, the label of the cable 252 is divided into only two of (1) and (2) by the circle 256. The end cutout 232 determines that the center point 258 of the circle 256 where the label of the cable 252 is not divided into three in this way is the end of the cable. Here, for convenience of explanation, it has been described that it is determined whether or not it is an end at every point 258 of the cable 252, but in practice, in order to reduce the calculation load, a circle from a portion close to the connector 250, which is a fixed portion. 256 are sequentially formed, and the cut-out range is determined based on the position where the divided areas are changed from two to three. However, the method of cutting out the end portion is not limited to such a case, and various existing methods can be employed.

  Based on the recognition target information and the three-dimensional model information, the recognition target specifying unit 234 performs pattern matching between the recognition target object 112 and the three-dimensional model 122 in which the internal degree of freedom of the variable part is changed to recognize the recognition target object 112. Identify the position and posture of

  Here, the recognition target specifying unit 234 changes the position and orientation of the other in the pattern matching (axis specification) between the recognition target 112 and the three-dimensional model 122 with reference to either one as a reference. Which is the standard is not limited. In the present embodiment, since the three-dimensional model 122 has more information than the recognition target object 112, the position and orientation of the recognition target object 112 are changed based on the three-dimensional model 122 in order to reduce the calculation load. To do.

  For such axis identification, various methods such as ICP (Iterative Closest Points) and PCA (Principal Component Analysis) can be used. Here, ICP is used as an example. The ICP determines the position and orientation of the three-dimensional model 122 having six degrees of freedom until the total distance (norm) between corresponding points for identifying the recognition object 112 and the three-dimensional model 122 is equal to or less than a threshold value. This is a method of updating iteratively using a conjugate gradient method or the like. PCA is a method (principal component analysis) for examining how input data has a longitudinal axis in space.

  ICP, for example, is different from PCA and the like, and aims at matching points between the recognition object 112 and the three-dimensional model 122. Therefore, the recognition object 112 is partially missing due to the overlap of the recognition objects 112, etc. Therefore, the position and orientation of the recognition target object 112 can be specified with relatively high accuracy.

  Further, as described above, the recognition object specifying unit 234 performs pattern matching with the three-dimensional model 122 while changing the internal degree of freedom of the variable part of the recognition object 112 within an assumed change range. Such internal degrees of freedom include bending, rotation of only the variable portion, and the like as described with reference to FIG. Furthermore, it is possible to consider a rigidity model of a variable part such as a cable. Such a change range of the variable part is held in the model holding unit 210 together with the three-dimensional model 122 as a change parameter. The recognition target specifying unit 234 converges the change parameter by changing the position of each corresponding point in the later-described three-dimensional point group data of the three-dimensional model 122.

  In this way, by performing pattern matching with the recognition target object 112 while changing the internal degree of freedom of the variable part of the three-dimensional model 122 within the change range, the change of the variable part can be taken into account. The degree can be increased dramatically. Further, by using only a part that can limit the change range, such as having regularity, as a variable part, it is possible to suppress an increase in calculation load, and under an appropriate calculation load, the exact position of the recognition object 112 and It becomes possible to specify the posture.

  The bending angle of the cable 252 and the like obtained by such matching can also be used for estimating the extension direction of the cable 252 and the like. For example, a flexible elastic tube (for example, a rubber hose) is registered as the three-dimensional model 122 of the variable portion, and the node of the model and the angle of the tangent at the node are set as parameters. Then, since the degree of bending of the three-dimensional model 122 can be changed by adjusting the angle, not only the rough direction of the elastic tube but also the angle at important points are matched, and the undulation posture of the elastic tube, etc. Can also be recognized.

  In addition, when the recognition target object 112 has a longitudinal axis, the recognition target specifying unit 234 first specifies the longitudinal axis, and two posture candidates arranged symmetrically (opposite to each other) along the longitudinal axis. And the pattern matching with the three-dimensional model 122 is performed for both of the two candidate postures, and the position and posture of the recognition target object 112 may be specified.

  As described with reference to FIG. 2, the recognition target object 112 having a longitudinal axis and the connector and the cable integrally formed with the connector often has two or more extreme values in pattern matching. In this case, if the longitudinal axes of the recognition object 112 and the three-dimensional model 122 are matched, the matching degree of pattern matching is increased even if the recognition object 112 and the three-dimensional model 122 are in an incorrect posture.

  Therefore, the recognition object specifying unit 234 divides pattern matching into two stages, and in the first stage, identifies and specifies only the axes on which the pattern matching results between the recognition object 112 and the three-dimensional model 122 can converge. Two posture candidates based on the axis are generated, and pattern matching is strictly performed using the posture candidates in the second stage. Specifically, the recognition object specifying unit 234 rotates the recognition object 112 clockwise from 0 ° around the Z axis, and the predetermined degree before the coincidence reaches the first extreme value (90 °). When the threshold value is reached, the pattern matching is once stopped. Then, the recognition target object specifying unit 234 sets the recognition target object 112 when stopped as the second-stage posture candidate, and also recognizes the recognition target object 112 rotated 180 ° as the posture candidate. Subsequently, strict pattern matching is performed in parallel based on the two orientation candidates of the recognition target object 112.

  In the example of FIG. 2, such two posture candidates show extreme values at 90 ° and 270 °, respectively, and the posture of 270 ° having a high degree of coincidence is determined to be the true posture of the recognition object 112. Is done. Here, since there are two posture candidates for one longitudinal axis, pattern matching is avoided by converging pattern matching only in a wrong posture by performing pattern matching for both of them.

  Moreover, since it is grasped that the relationship between the two postures, for example, a symmetric (reverse direction to each other) relationship along the axis, with respect to specifying the axis, if one of the posture axes is used, Often, strict pattern matching is not required at that time. That is, at the first stage, only one of the axes needs to be extracted with a gentler threshold than that of strict pattern matching.

  Here, since the optimum convergence point (extreme value) is derived for at least one posture candidate, the pattern matching is completed without deriving the other convergence point. Therefore, the position and orientation of the recognition object 112 are accurately derived. Further, since the pattern matching until the axis is extracted is not executed for each of the two extreme values, but only one of them is performed, the processing load can be reduced.

  Furthermore, the recognition target object specifying unit 234 specifies a part that matches only the fixed part of the three-dimensional model 122 in the recognition target object 112, and uses the position and orientation of the specified part to recognize the recognition target object 112 and the variable part. It is also possible to specify the position and orientation of the recognition object 112 by strictly pattern matching with the three-dimensional model 122 in which the internal degrees of freedom are changed.

  For example, when the cable 252 as the variable portion is too flexible, it may be difficult to obtain a convergence point only by pattern matching of the three-dimensional model 122 in which the connector 250 and the cable 252 described above are integrally formed. In that case, a three-dimensional model of only the connector 250 that is a fixed part is used independently, pattern matching of only the connector 250 is performed, only the longitudinal axis of the connector 250 is estimated, and the posture along the longitudinal axis is determined. By performing pattern matching with the three-dimensional model 122 in which the connector 250 and the cable 252 are integrally formed as a starting posture, the position and posture of the recognition target object 112 are specified.

  Here, it is possible to estimate the general posture of the three-dimensional model 122 in which the connector 250 and the cable 252 are integrally formed by performing pattern matching of the fixed portion that is difficult to be distorted in advance, and the change range of the variable portion can be determined. Since it is possible to limit, it is possible to reduce the processing load and quickly specify the position and orientation of the recognition target object 112.

  As described above, in the shape recognition device 120 described above, even if the posture of the fixed part alone cannot be specified due to symmetry, the connection position of the fixed part with the variable part and the shape of the variable part are used as elements that characterize the fixed part. Therefore, the position and orientation of the recognition target object 112 can be specified with high accuracy. Furthermore, not only pattern matching of the three-dimensional model 122 in which the variable part is added to the fixed part and the recognition target object 112 corresponding thereto, but also by changing the flexible variable part within a known change range, the position and orientation Can be specified with higher accuracy.

  Furthermore, in this embodiment, since ICP is used as the pattern matching method, the position and orientation can be specified for the recognition target object 112 having various shapes. In the ICP, even when a part of the recognition target object 112 is hidden, the robustness is high, so that an effective result can be obtained. Therefore, even if a part of the cable as the variable portion is covered with another object, the traveling direction of the cable can be grasped, and the tip of each cable can be estimated even if the cables are overlapped.

  In addition, a program that functions as the shape recognition device 120 by a computer and a storage medium that stores the program are also provided. Further, the program may be read from a storage medium and taken into a computer, or may be transmitted via a communication network and taken into a computer.

(Shape recognition method)
A shape recognition method using the shape recognition device 120 is also provided. Hereinafter, such a shape recognition method will be described in detail. FIG. 7 is a flowchart showing the flow of processing of the shape recognition method. In particular, FIG. 7A shows the overall processing flow of the shape recognition method, and FIG. 7B shows a subroutine related to ICP. Here, the model holding unit 210 holds 3D model information in advance.

  In the pattern matching using the three-dimensional model 122 in which the fixed portion and the variable portion are integrally formed in this embodiment, the fixed portion has symmetry and the direction cannot be uniquely determined. This is effective when the connected variable part changes flexibly and it is desired to pick the flexible change part. On the other hand, when the shape of the fixed part is characteristic and it is not necessary to specify the variable part, the pattern matching can be executed by the three-dimensional model 122 using only the fixed part.

  The recognition target information acquisition unit 230 acquires recognition target information indicating the three-dimensional shape of an arbitrary recognition target object 112 from the three-dimensional shape measurement apparatus 110 (S300). Then, the end cutout unit 232 determines whether or not the size of the label of the variable part of the recognition object 112 is greater than a predetermined threshold, that is, whether or not the part that is a candidate for the variable part is extended. If the size of the label of the variable portion is larger than the predetermined threshold (YES in S302), the portion (label) at the fixed portion side end portion that is a candidate for the variable portion is cut out (S304). At this time, if the size of the label of the variable part is equal to or smaller than the predetermined threshold (NO in S302), the process proceeds to the identifying step S306 of the recognition target object 112 as it is.

  The recognition target specifying unit 234 performs pattern matching by ICP based on the recognition target information and the three-dimensional model information (S306).

  The flow of ICP processing by the recognition target specifying unit 234 will be briefly described with reference to FIG. In the recognition target information acquired from the three-dimensional shape measuring apparatus 110, three-dimensional coordinates (three-dimensional point group data) relating to a plurality of detection points arranged on the measurement surface of the recognition target object 112 are shown. The number of points in the three-dimensional point group data is determined according to the size of the recognition object 112 and the required accuracy of pattern matching. The ICP engine excludes any clear background measurement points such as floors, columns, containers, etc., from the three-dimensional point cloud data in the recognition target information before performing the ICP repetition process (S350). ).

  The ICP engine prepares an initial posture and performs posture change processing such as rotation and parallel movement conversion on the initial posture (S352). The posture changing process can be performed on the recognition target object 112 or can be performed on the three-dimensional model 122 side. For example, the calculation processing may be reduced by moving the one with the smaller number of points. Subsequently, corresponding points between the recognition target object 112 after the posture change process and the three-dimensional model 122 are obtained (S354: pairing), and a total distance (norm) between the corresponding points is calculated (S356). It can be said that the smaller the norm, the higher the degree of coincidence of shapes. A posture with a high degree of coincidence is extracted (S358). While the degree of coincidence is less than or equal to a threshold value (norm is greater than or equal to a threshold value) (NO in S360), it is replaced with an initial posture and the above process is repeated. As described above, pattern matching is performed by repetitively updating using the conjugate gradient method or the like until the degree of coincidence exceeds the threshold (norm becomes equal to or less than the threshold) (YES in S360).

  For example, there are L1 norm and L2 norm as the norm taking method (metric), and the taking method suitable for the application is adopted. At this time, the distance between the corresponding points is derived while changing the change parameter corresponding to the variable part of the three-dimensional model 122 within a predetermined change range. This is equivalent to increasing the dimension of the convergence parameter in ICP. In this way, by performing pattern matching with the recognition target object 112 while changing the internal degree of freedom of the variable part of the three-dimensional model 122, it is possible to take into account the change of the variable part, so that the degree of matching is significantly increased. be able to. Further, by using only a part that can limit the change range, such as having regularity, as a variable part, it is possible to suppress an increase in calculation load, and under an appropriate calculation load, the exact position of the recognition object 112 and It becomes possible to specify the posture.

  The position and orientation thus derived are finally converted from the coordinate system of the shape recognition device 120 to the coordinate system of the robot 140.

  Even in such a shape recognition method, even if it is not possible to specify the forward direction or the reverse direction with respect to the longitudinal axis due to symmetry of the fixed unit alone, the connection position of the fixed unit with the variable unit and the variable unit Since the shape is used as an element that characterizes the fixed portion, the position and orientation of the recognition target object 112 can be specified with high accuracy. Furthermore, not only pattern matching of the three-dimensional model 122 in which the variable part is added to the fixed part and the recognition target object 112 corresponding thereto, but also by changing the flexible variable part within a known change range, the position and orientation Can be specified with higher accuracy.

  As mentioned above, although preferred embodiment of this invention was described referring an accompanying drawing, it cannot be overemphasized that this invention is not limited to this embodiment. It will be apparent to those skilled in the art that various changes and modifications can be made within the scope of the claims, and these are naturally within the technical scope of the present invention. Is done.

  Note that each step in the shape recognition method of the present specification does not necessarily have to be processed in time series in the order described in the flowchart, and may include processing in parallel or by a subroutine.

  The present invention can be used for a shape recognition device, a shape recognition method, and a program for recognizing the position and posture of an arbitrary recognition target.

DESCRIPTION OF SYMBOLS 100 ... Picking system 112 ... Recognition object 120 ... Shape recognition apparatus 122 ... Three-dimensional model 210 ... Model holding part 230 ... Recognition object information acquisition part 232 ... End part extraction part 234 ... Recognition object specific part 250 ... Connector 252 ... cable

Claims (6)

A model holding unit for holding three-dimensional model information indicating a three-dimensional model having a longitudinal axis, including a fixed part and a variable part having an internal degree of freedom;
A recognition target information acquisition unit for acquiring recognition target information indicating a three-dimensional shape of the recognition target;
Based on the recognition target information and the three-dimensional model information, two candidate postures that are symmetric with respect to the longitudinal axis are created, and the recognition target and a three-dimensional model in which the internal degrees of freedom of the variable part are changed, A recognition target specifying unit that specifies the position and orientation of the recognition target by pattern matching,
A shape recognition device comprising:   An end cutout section for cutting out the fixed portion side end portion of the variable portion candidate portion when the variable portion candidate portion of the recognition object is extended; The shape recognition apparatus according to claim 1, characterized in that:   The end cutout portion cuts out a portion divided into two by a circle determined based on the maximum value of the linear distance between any two points in the cross section perpendicular to the extending direction of the variable portion. The shape recognition apparatus according to claim 2, wherein   The recognition object specifying unit specifies a part that matches the fixed part of the three-dimensional model in the recognition object, and uses the position and orientation of the specified part to identify the recognition object and the variable part. The shape recognition apparatus according to any one of claims 1 to 3, wherein the position and orientation of the recognition target object are specified by pattern matching with a three-dimensional model in which internal degrees of freedom are changed. 3D model information indicating a 3D model having a longitudinal axis including a fixed part and a variable part having an internal degree of freedom is held in advance,
Obtaining recognition object information indicating the three-dimensional shape of the recognition object;
Based on the recognition target information and the three-dimensional model information, two candidate postures that are symmetric with respect to the longitudinal axis are created, and the recognition target and a three-dimensional model in which the internal degrees of freedom of the variable part are changed, A pattern recognition method is used to specify the position and orientation of the recognition object. Computer
A recognition target information acquisition unit for acquiring recognition target information indicating a three-dimensional shape of the recognition target;
Based on the recognition target information, and a three-dimensional model information indicating a three-dimensional model having a longitudinal axis including a fixed portion and a variable portion having an internal degree of freedom, two candidate postures symmetric with respect to the longitudinal axis are obtained. A recognition target specifying unit that pattern-matches the recognition target and a three-dimensional model in which the internal degree of freedom of the variable unit is changed, and specifies the position and orientation of the recognition target;
Program to make it function.

Images