JP2015062991A - Coordinate system calibration method, robot system, program, and recording medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a coordinate system calibration method for making it unnecessary for a user to search, by a trial-and-error method, the teaching point at which a robot and such one out of a calibrator and a visual sensor as is supported by the robot do not interfere with an obstacle.SOLUTION: A CAD (Computer Aided Design) data is used to perform calculation for making a robot virtually act on a teaching point, thereby to determine whether or not the robot or a calibrator interferes with an obstacle (S4). If it is determined that the robot or the calibrator interferes with the obstacle when the robot is virtually caused to act on the teaching point (S4: Yes), the position and attitude of the robot, in which the robot and the calibrator do not interfere with the obstacle, is searched (S8). Next, in the case that the position and attitude of the robot in which the robot and the calibrator do not interfere with the obstacle is found (S8: Yes), the teaching point is reset on the basis of the found position and attitude of the robot (S9).

Description

  The present invention relates to a coordinate system calibration method, a robot system, a program, and a recording medium for calibrating a coordinate system based on a robot and a coordinate system based on a visual sensor.

  In recent years, automatic assembly by robots has increased in order to automate and save labor in factory production lines. There, a robot system combining a robot and a visual sensor is used. In particular, recently, there is an increasing need to three-dimensionally measure a workpiece that is a work target using a visual sensor.

  The robot has a robot arm and a robot hand attached to the tip of the robot arm. The visual sensor is fixed to the robot or a frame around the robot.

  When measuring the position and orientation of a workpiece using a visual sensor and correcting the movement of the robot based on it, the position and orientation data of the workpiece in the vision coordinate system, measured using the visual sensor, are stored in the coordinate system where the robot operates. Need to convert to data. Here, the robot coordinate system includes a robot coordinate system based on the robot arm and a tool coordinate system based on the robot hand. Therefore, the relationship between the vision coordinate system and the robot coordinate system or tool coordinate system must be obtained in advance. Calibration between these coordinate systems is generally called hand-eye calibration.

  As a method for performing hand-eye calibration, there is a method using a special calibrator whose accurate design data is clear (see Patent Document 1). The calibrator includes a marker that can be measured using a visual sensor and a calibration jig to which the marker is fixed. The marker here is a work having high measurement accuracy for the visual sensor. There is accurate design data on where the marker is fixed with respect to the calibration jig, and an obvious one is used. Also, there is accurate design data on where the calibration jig is fixed with respect to the robot, and an obvious one is used. By measuring the position and orientation of the calibrator using a visual sensor, the relationship between the coordinate systems becomes clear, and the relative position and orientation between the vision coordinate system and the robot coordinate system (or tool coordinate system) can be obtained.

  The method of Patent Document 1 requires a special calibration jig in which accurate design data is clear, which increases the cost. Therefore, a method not using a special calibration jig in which accurate design data is clear is also proposed. (See Patent Document 2).

  In the method of Patent Document 2, a marker that can be measured using a visual sensor is necessary, but it is not necessarily clear where the marker is fixed on the calibration jig. Instead, three or more relative positions and orientations between the marker and the visual sensor are acquired by operating the robot near the calibration point, and the marker is measured at each point using the visual sensor. Simultaneous equations are established from the relative position and orientation between the robot coordinate system and the tool coordinate system (hand coordinate system) obtained at each point and the relative position and orientation between the vision coordinate system and the marker coordinate system. By solving this by optimization calculation, the relative position and orientation between the vision coordinate system and the robot coordinate system can be obtained.

Japanese Patent No. 2700985 Japanese Patent No. 3644991

  However, Patent Documents 1 and 2 are based on the premise that teaching points are taught in advance, and do not consider that the robot, or a calibrator or visual sensor supported by the robot interferes with an obstacle. . Actually, there are obstacles such as fixing jigs around the robot, and there is a possibility that interference with the obstacle occurs due to the operation of the robot. Therefore, it is necessary for the user to search for teaching points that can be used for calibration without interference with obstacles, and the search operation is complicated, so it is difficult to determine the teaching points. Met.

  Therefore, the present invention eliminates the need for the user to search for a teaching point where the robot and any one of the calibrator and the visual sensor supported by the robot do not interfere with the obstacle by trial and error. For the purpose.

  In the present invention, the robot supports one member of the calibrator and the visual sensor, and the control unit moves the robot to the position and orientation taught at the teaching point, thereby imaging the calibrator on the visual sensor. In the coordinate system calibration method for calibrating the coordinate system based on the robot and the coordinate system based on the visual sensor based on the processing result of the captured image, the control unit sets the teaching point. A setting step, an acquisition step in which the control unit acquires at least the robot, the one member, and an obstacle that may be an obstacle to the operation of the robot, and the control unit includes the solid shape data. Is used to perform an operation to virtually operate the robot at the teaching point, and to determine whether the robot or the one member interferes with the obstacle. When the determination step and the control unit determine that the robot or the one member interferes with the obstacle when the robot is virtually operated to the teaching point in the interference determination step, the robot And a search step for searching for a position and orientation of the robot in which the one member does not interfere with the obstacle, and the control unit includes the robot in which the robot and the one member do not interfere with the obstacle in the search step. And a resetting step of resetting the teaching point based on the found position and orientation of the robot.

  According to the present invention, a user can search by trial and error by automatically searching for a teaching point where the robot and any one of the calibrator and the visual sensor supported by the robot do not interfere with the obstacle. This reduces the burden on the user.

It is explanatory drawing which shows schematic structure of the robot system which concerns on 1st Embodiment. It is a block diagram which shows the structure of the robot system which concerns on 1st Embodiment. It is a schematic diagram which shows the surface of a marker. It is a flowchart of the coordinate system calibration method which concerns on 1st Embodiment. It is the schematic diagram which illustrated the relative position and orientation of the camera and calibrator which are used in 1st Embodiment, and a teaching point. It is the schematic diagram which illustrated the resetting method of the teaching point in case a robot interferes with an obstruction. It is explanatory drawing which shows schematic structure of the robot system which concerns on 2nd Embodiment. It is a flowchart of the coordinate system calibration method which concerns on 2nd Embodiment.

  Hereinafter, embodiments for carrying out the present invention will be described in detail with reference to the drawings.

[First Embodiment]
FIG. 1 is an explanatory diagram showing a schematic configuration of the robot system according to the first embodiment of the present invention. The robot system 10 includes a robot 300, a control device 400, a camera 500 as a visual sensor, the gantry 20, and a frame 30. The control device 400 includes the robot control device 100 and the image processing device 200.

  The robot 300 has a multi-joint (for example, six joints) robot arm 301 and a robot hand 302 as an end effector attached to the tip of the robot arm 301. The base end of the robot arm 301 is fixed on the gantry 20.

  The robot hand 302 is configured to be able to grip an object such as a workpiece. For example, the robot hand 302 includes a plurality of fingers, and can grip and release an object by opening and closing the plurality of fingers. .

  The frame 30 has a rectangular parallelepiped frame structure, for example, and is fixed on the gantry 20, for example. In FIG. 1, the top of the frame 30 is illustrated.

  The camera 500 is a fixed camera that is fixed to the top of the frame 30. The camera 500 is fixed so that the optical axis is directed vertically downward. The camera 500 is a digital camera having a solid-state image sensor such as a CMOS image sensor or a CCD image sensor. The camera 500 as a visual sensor uses a monocular camera in the first embodiment, but the visual sensor may be a compound eye camera, a laser range finder, or a combination thereof.

  The camera 500 and the image processing apparatus 200 are connected by, for example, a cable, and the image processing apparatus 200 can acquire an image signal from the camera 500 through the cable. Note that the camera 500 and the image processing apparatus 200 may be wirelessly connected so that the image processing apparatus 200 acquires an image signal from the camera 500 wirelessly.

  The robot 300 and the robot control apparatus 100 are connected by a cable or the like, for example, and the operation of the robot 300 is controlled by the robot control apparatus 100. Further, the image processing apparatus 200 and the robot control apparatus 100 are connected by, for example, a cable or the like, and the robot control apparatus 100 is configured to receive a processing result from the image processing apparatus 200.

  The image processing apparatus 200 converts the image signal acquired from the camera 500 into a light / dark signal using a grayscale gray scale, and stores the converted signal in a frame memory. In addition, the image stored in the frame memory is processed to measure the object identification, position and orientation. Further, the processed object identification result and the position and orientation information are transmitted to the robot controller 100.

  Note that illumination (not shown) may be connected to the image processing apparatus 200. At that time, the image processing apparatus 200 also controls illumination.

Here, a coordinate system used for the operation of the robot 300 is defined for the robot 300. More specifically, a robot coordinate system C ₁ based on the robot arm 301 and a tool coordinate system C ₃ based on the robot hand 302 are defined. The camera 500 defines a vision coordinate system C _{2 that} is used when measuring the position and orientation of a workpiece reflected in a captured image captured by the camera 500 and based on the camera 500.

When subjected to machining a workpiece or if the workpiece to be gripped by the robot hand 302 of the robot 300, by imaging the workpiece by the camera 500, the position and orientation of the workpiece measured by the vision coordinate system C _2, based on the measurement results The operation of the robot 300 is corrected. Therefore, prior to starting the production process, it is necessary to obtain the calibration data between the robot coordinate system C ₁ and vision coordinate system C ₂ (data of the relative position and orientation).

  Therefore, in the first embodiment, since the camera 500 is fixed to the frame 30 other than the robot 300, the calibrator 40 used for calibration is supported by the robot 300.

That is, the robot 300 supports the calibrator 40 as one member of the calibrator 40 and the camera 500. Then, the robot controller 100, based on the relative position and orientation of the camera 500 and the calibrator 40 obtains the calibration data between the robot coordinate system C ₁ and vision coordinate system C _2.

The calibrator 40 has a marker 41 formed in a flat plate shape and a calibration jig 42 to which the marker 41 is fixed. The calibration jig 42 is configured to be gripped by the robot hand 302. When the calibration jig 42 is gripped by the robot hand 302, the calibrator 40 is supported by the robot 300. As shown in FIG. 1, a coordinate system based on the calibrator 40 with the marker coordinate system C _4. The calibrator 40 may be fixed to the robot hand 302 or the robot arm 301. Here, where the marker 41 is fixed to the calibration jig 42 may not be known. Moreover, although the case where the calibrator 40 includes the marker 41 and the calibration jig 42 to which the marker 41 is fixed has been described, the marker pattern may be directly formed on the calibration jig.

  FIG. 2 is a block diagram showing the configuration of the robot system 10. Here, although not shown in FIG. 1, the robot system 10 includes a monitor 600 as a display device and a keyboard 700 as an input device.

  The robot control apparatus 100 is configured by a computer and includes a CPU (Central Processing Unit) 101 as a control unit (arithmetic unit). The robot controller 100 also includes a ROM (Read Only Memory) 102, a RAM (Random Access Memory) 103, and an HDD (Hard Disk Drive) 104 as storage units. The robot control device 100 also includes a recording disk drive 105 and various interfaces 107 to 110.

  A ROM 102, a RAM 103, an HDD 104, a recording disk drive 105, and various interfaces 107 to 110 are connected to the CPU 101 via a bus 106. The ROM 102 stores basic programs such as BIOS.

  The RAM 103 temporarily stores a program for controlling robot operation, command transmission to the image processing apparatus 200, reception of image processing results from the image processing apparatus 200, and related setting values. Further, the RAM 103 is also used as a temporary storage memory when the CPU 101 executes operations and as a register area set as necessary.

  The HDD 104 stores calculation processing results of the CPU 101, various data acquired from the outside, and the like, and also records a program for controlling the entire system 10 and a program 120 for executing each step of the coordinate system calibration method. Is. The CPU 101 executes various arithmetic processes (each step of the coordinate system calibration method) based on the program 120 recorded (stored) in the HDD 104.

  The HDD 104 is arranged around the robot such as the robot 300, the camera 500, the calibrator 40, and the fixing jig, and the three-dimensional shape data of the entire system including the obstacle 60 (FIG. 1) that can interfere with the operation of the robot 300. (CAD data) is stored.

  The recording disk drive 105 can read various data and programs recorded on the recording disk 800.

  The image processing apparatus 200 is connected to the interface 107, and the CPU 101 can acquire position and orientation information received from the image processing apparatus 200. Further, the CPU 101 outputs a trigger signal indicating the imaging timing to the camera 500 via the interface 107 and the image processing device 200.

  A robot 300 (specifically, a servo circuit that controls each axis of the robot 300) is connected to the interface 108, and the CPU 101 can control the operation of the robot 300 based on the taught teaching points. . A monitor 600 is connected to the interface 109, and the CPU 101 can display an image on the monitor 600. A keyboard 700 which is an input device is connected to the interface 110, and the CPU 101 performs various settings by receiving input from the keyboard 700. The robot control device 100 may be connected to an external storage device such as a rewritable nonvolatile memory or an external HDD, an offline programming device, a teaching device for teaching, and the like (not shown).

  FIG. 3 is a schematic diagram showing the surface of the marker 41. A plurality of dots 43 are attached to the surface of the marker 41 as a pattern, and positions between the dots 43 are stored in advance in the image processing apparatus 200 as set values. Thereby, the image processing apparatus 200 can measure the position and orientation of the calibrator 40.

  Further, the posture of the calibrator 40 can be uniquely obtained by the triangular mark 44 attached to the marker 41. The marker 41 of the calibrator 40 only needs to be able to measure the position and orientation of the calibrator 40 with the image processing apparatus 200 when it is reflected in the captured image captured by the camera 500, and is not a flat plate but a three-dimensional shape. It may be a thing.

  The operation of the robot 300 during calibration is controlled by the program 120 stored in the memory of the robot control apparatus 100. By operating the robot 300, the calibrator 40 can be positioned at an arbitrary position and orientation within the operation range of the robot 300.

Hereinafter, a method for performing hand eye calibration in the robot system 10 will be described. In the first embodiment, the calibrator is supported by the robot 300, the camera 500 is fixed, it is necessary to calibrate the robot coordinate system C ₁ and vision coordinate system C _2.

As a precondition, a case where it is not known where the marker 41 is fixed with respect to the calibration jig 42 is considered. In that case, in general, the relative position and orientation of the camera 500 and the calibrator 40 are at least three different relative positions and orientations. And the relative position and orientation of the robot coordinate system C ₁ and the tool coordinate system C ₃ by each point, it is necessary to obtain a relative position and orientation of the vision coordinate system C ₂ and the marker coordinate system C _4. Thereby, it is possible to obtain the robot coordinate system C ₁ and vision coordinate system C ₂ and the relative position and orientation to be calibrated.

FIG. 4 is a flowchart of the coordinate system calibration method according to the first embodiment. The CPU 101 reads and executes the program 120 stored in the HDD 104, and automatically sets the teaching points used for calibration according to the algorithm shown in FIG. The teaching point includes six parameters, that is, three parameters indicating a three-dimensional teaching position and three parameters indicating a three-dimensional teaching posture. Then, the CPU 101 causes the camera 500 to image the calibrator 40 by operating the robot 300 (for example, the tip of the robot 300, that is, the robot hand 302) to the position and orientation taught by the plurality of teaching points. Next, the CPU 101 calibrates the robot coordinate system C ₁ and the vision coordinate system C ₂ based on the processing result of the captured image in the image processing apparatus 200.

  Hereinafter, it will be described in detail along the flowchart of FIG. First, the CPU 101 generates a plurality of teaching points for calibration distributed in the maximum range of the region to be calibrated (S1). That is, in step S1, the CPU 101 sets at least three teaching points as a plurality of teaching points (setting process, setting process). In this step S1, the CPU 101 may acquire a teaching point from an external teaching device (not shown) and set it (that is, store it in a memory or the like so that it can be used in later control), or a predetermined program May be automatically generated and set. Here, three teaching points are used for calibration. A method for setting the three teaching points will be described.

FIG. 5 is a schematic view illustrating the relative positions and orientations and teaching points between the camera 500 and the calibrator 40 used in the first embodiment. 5A shows the center of the calibrator 40 (marker 41) and the calibration range when the robot 300 is operated at each teaching point, and FIG. 5B shows the camera 500, the calibrator 40, and the calibration range. The relative position and orientation of are shown. Here, the plane A shown in FIG. 5, the optical axis of the camera 500 is set inside the range of the angle of view of the camera 500 a plane having a normal line, this plane robot coordinate system C ₁ and vision coordinate system It shows a calibration region with C _2. The range of the calibration area A is a range in which a marker 41 on the calibrator 40 to be described later can be measured, and shows the range that the center of the calibrator 40 can take by the operation of the robot 300.

As shown in FIG. 5B, the first teaching point (first teaching point) P ₁ is on the plane M ₁ having the normal to the design optical axis (camera optical axis) L ₀ of the camera 500. The teaching point is to operate the robot 300 so that the surface of the calibrator 40 is located on the first plane. Here, the optical axis L _{0 in} the design of the camera 500 is set before calibration, and thus the CPU 101 does not know the actual optical axis of the camera 500.

  The surface of the calibrator 40 is a surface on which a pattern such as a dot 43 is formed in the marker 41 in the first embodiment. When a pattern such as a dot is directly formed on the surface of the calibration jig 42, the surface of the calibrator 40 is the surface of the calibration jig 42. That is, the surface of the calibrator 40 is a surface on which feature points such as dots are formed.

Secondly taught points (second teaching point) P _2, to the teaching points P _1, an axis parallel L around ₁ and calibrator 40 is the optical axis of the camera L _0, the relative angle α designated by the user in advance taken, and a teaching point for operating the robot 300 so as to translate in the plane M ₁ in the plane direction (xy direction). That is, the teaching point P ₁ and the teaching point P ₂ are teaching points for operating the robot 300 so that the surface of the calibrator 40 is positioned on the plane M ₁ and the position and orientation of the calibrator 40 are different from each other. Note that the above-described axis L ₁ may be parallel to the camera optical axis L ₀ or in the range of the axis tilt of ± 2 deg. The axis L ₁ is set so that the center O ₂ of the calibrator 40 falls within the range of the angle of view of the camera 500 when the relative angle α is taken.

  The relative angle α is most preferably 90 deg, but is preferably selectable by the user in the range of 0 deg <α <180 deg. Here, the designation of the relative angle α may be performed by the user using the keyboard 700, for example, or may be set in the program 120 in advance.

The third teaching point (third teaching point) P ₃ takes two angles other than the camera optical axis L ₀ , that is, an angle designated in advance by the user around the xy axis in the vision coordinate system, and the calibrator 40 The teaching points P ₁ and P ₂ are teaching points set to take different positions and orientations. More specifically, the teaching point P ₃ is a teaching point for operating the robot 300 so that the surface of the calibrator 40 is positioned on the plane M ₂ (on the second plane) intersecting the plane M ₁ . . That is, the surface of the calibrator 40 taking the teaching point P ₃ as not on the plane M _1.

  Here, the angle specified by the user may be about 5 to 10 degrees. Further, this angle may be set in the program 120 in advance.

The three teaching points P ₁ , P ₂ , and P ₃ are located at different positions on the plane M ₁ where the centers O ₁ , O ₂ , and O ₃ of the calibrators 40 are normal to the camera optical axis L _0. It is desirable to arrange so as to be in a posture and to maximize the range to be calibrated. In the first embodiment, as shown in FIG. 5A, two points are arranged at the corners of the calibration area A, and one point is arranged at the midpoint on the diagonal line of the calibration area A.

  Next, the CPU 101 reads out CAD data (three-dimensional shape data) stored in the HDD 104 from the HDD 104 (S2). That is, the CPU 101 acquires CAD data from the HDD 104 (acquisition process, acquisition process). The CAD data includes shape data of the obstacle 60 such as the robot 300, the camera 500, the calibrator 40, and the fixing jig. This CAD data is stored in the HDD 104 in advance, but may be acquired from an external device.

  Then, the CPU 101 determines whether or not the teaching point set in step S1 is within the movable range of the robot 300 (specifically, each joint of the robot arm 301) (S3). When making the determination, CAD data read by the CPU 101 is used.

  When the CPU 101 determines that each set teaching point is within the movable range of the robot 300 as a result of the determination in step S3 (S3: Yes), the CPU 101 executes an interference determination process (interference determination process) in step S4.

In step S < _b > 4, the CPU 101 uses the acquired CAD data to perform a calculation that causes the robot 300 to virtually operate the teaching points P ₁ , P ₂ , and P ₃ , and the robot 300 or the calibrator 40 moves the obstacle 60. Determine whether to interfere.

As a result of the determination in step S4, the CPU 101 determines that the robot 300 and the calibrator 40 do not interfere with the obstacle 60 by the operation of the robot 300 based on the set teaching points P _{1 to} P ₃ (S4: No). The process of step S5 is executed.

CPU101, in step S5, determines the teaching point _P 1 to P ₃ used in the actual calibration. The determined teaching points P _{1 to} P ₃ are stored in a storage unit (for example, HDD 104) of the robot control apparatus 100 as a robot operation program.

Next, the CPU 101 executes hand eye calibration using the teaching points P _{1 to} P ₃ determined in step S5 (S6). Specifically explaining the processing of step S6, first, CPU 101 operates the robot 300 on the basis of teaching points _P 1 to P ₃ determined in step S5, is operated on the teaching points _P 1 to P ₃ position In the posture, the calibrator 40 is imaged by the camera 500. The image processing apparatus 200 captures the captured image, measures the position and orientation of the calibrator 40 from the image of the calibrator 40 captured in the captured image, and sends the measurement result (processed result of the captured image) to the robot control apparatus 100. Send.

Here, the relative position and orientation of the robot coordinate system C ₁ and the tool coordinate system C ₃ is so synonymous with the position and orientation of the teaching point P ₁ to P _3, obtained from the robot controller 100. The vision coordinate system C ₂ , the marker coordinate system C _4, and the relative position and orientation are obtained from the image processing apparatus 200 because they are obtained by measuring the calibrator 40.

When the relative position and orientation are expressed by a translation component and Euler angle, (X, Y, Z, R _x , R _y , R _z ) is obtained. It is defined as the following equation in the form of a homogeneous transformation matrix H.

Here, R _otx , R _oty , and R _otz are 3 × 3 rotation matrices representing rotations about the x, y, and z axes, respectively.

Robot coordinate system C ₁ and vision coordinate system C ₂ and the relative position and orientation of the homogeneous transformation matrix H _rv, the relative position and orientation of the tool coordinate system C ₃ and the marker coordinate system C ₄ and homogeneous transformation matrix H _tp. Furthermore, homogeneous transformation matrix relative position and orientation of the robot coordinate system _{C 1} and the tool coordinate system _{C 3} in each three taught points _P 1 to P _{3 H rt1,} and H _{rt2, H rt3.} Also, the vision coordinate system _{C 2} and the marker coordinate system _{C 4} homogeneous transformation matrix relative position and orientation of the _{H vp1,} when the H _{vp2, H vp3,} simultaneous equations shown by the following equation is obtained.

The relative position and orientation H _rv between the robot coordinate system C ₁ to be hand-eye calibrated and the vision coordinate system C ₂ can be obtained by solving the optimization calculation for H _rv and H _tp of Equation (2). It is known that the optimization calculation can be solved by, for example, the least square method, although not limited thereto.

The CPU 101 outputs the obtained relative position and orientation H _rv between the robot coordinate system C ₁ and the vision coordinate system C ₂ to the HDD 104 or the like as a storage device of the robot control apparatus 100 and stores (updates) it as a set value (S7). ). Thereby, the calibration process is completed.

On the other hand, if the CPU 101 determines in step S3 that the set teaching points P _{1 to} P ₃ are not within the movable range of the robot 300 (each joint of the robot arm 301) (S3: No), the search processing in step S8 (Search process) is executed. Further, CPU 101, in step S4, when the robot 300 or the calibrator 40 determines that interferes with the obstacle 60 when the virtually operating the robot 300 to the teaching points _{P 1} and / or teaching point _{P 2,} the step The process of S8 is executed. Furthermore, CPU 101 is executed in step S4, even when the robot 300 or the calibrator 40 determines that interferes with the obstacle 60 when the virtually operating the robot 300 to the teaching point _{P 3,} the processing in step S8 To do. That is, if the CPU 101 determines Yes in step S4, it executes the process of step S8.

CPU101, in step S8, not within a movable range of the robot 300, or interferes teaching point and the obstacle 60, whether rotatable about a first axis parallel to the optical axis _{L 0} of the camera 500 of the design Determine.

In the determination, a rotation range around the first axis parallel to the optical axis L ₀ of the designed camera 500 preset in the robot controller 100 is used. The rotation range is about -180 deg to 180 deg and is searched with a predetermined step size (angle). The predetermined step size is, for example, about 5 deg or 10 deg. Further, the rotation angle around the axis parallel to the optical axis of the camera 500 already taken by the teaching point is excluded from this rotation range.

Here, the teaching point is determined in steps S3, S4 is, in the case of teaching points P ₁ and / or teaching point P _2, to search as follows. That is, the CPU 101 performs an operation for virtually rotating the calibrator 40 positioned on the robot 300 operating based on the teaching points P ₁ and P ₂ around the first axis based on the CAD data. Do. At that time, CPU 101 includes a calibrator 40 at the time of operating the robot 300 to the teaching point P _1, the relative position and orientation between the calibrator 40 at the time of operating the robot 300 to the teaching point P ₂ While maintaining, the calculation for rotating the calibrator 40 is performed. That is, in the CAD data, the CPU 101 rotates the calibrator 40 relative to the camera 500 by a predetermined step size (angle) so that the robot 300 and the calibrator 40 do not interfere with the obstacle 60. Explore.

Taught point is determined in steps S3, S4 is the case of the teaching point P _3, to search as follows. That CPU101, based on the CAD data, the calibrator 40 that is positioned in the robot 300 operates based on the teaching point _{P 3,} virtually rotated relative to the first axis with respect to the camera 500. At that time, the CPU 101 rotates the calibrator 40 relative to the camera 500 by a predetermined step size (angle) to search for a first rotation angle at which the robot 300 and the calibrator 40 do not interfere with the obstacle 60. To do.

  In the first embodiment, since the camera 500 is fixed, the calibrator 40 is rotated. In this way, since the search is performed by relatively rotating the calibrator 40 and the camera 500, it is possible to search in a narrower range than in the case of relatively moving in parallel.

CPU101 as a result of the search in step S8, if the teaching point is rotatable about a first axis parallel to the camera optical axis _{L 0} (S8: Yes), i.e. when Locate the proper first rotation angle, in step S9 A resetting process (resetting process) is executed.

CPU101, in step S9, by rotating the taught point determined in step S8 to the first axis parallel to the camera optical axis _{L 0,} resets. That is, when the CPU 101 finds the first rotation angle, the CPU 101 teaches the teaching points P ₁ , P ₂ or the teaching so that the calibrator 40 rotates relative to the camera 500 around the first axis at the first rotation angle. to re-set the point _{P 3.}

FIG. 6 is a schematic view illustrating a method for resetting teaching points when the robot 300 interferes with the obstacle 60. The amount of rotation follows the first rotation angle within the rotation range searched in step S8. However, when the teaching points to be rotated are the teaching points P ₁ and P ₂ , as shown in FIG. 6, the calibrator 40 is rotated so as to maintain the relative angle α of (teaching points P ₁ and P ₂ ).

In the case of the teaching point P ₃ with an inclination with respect to the same plane of the camera optical axis and the normal line, the two relative angles of teaching points P _1, P ₂ with the camera optical axis on the same plane with the normal Is not taken into consideration, and it rotates around the first axis parallel to the optical axis of the camera and resets. Similarly, the amount of rotation follows the rotation range used in step S8. When the resetting of the teaching point is completed, the process returns to step S3.

  On the other hand, if it is determined in step S8 that the set teaching point cannot be rotated about an axis parallel to the camera optical axis (S8: No), the CPU 101 executes the search process (search process) in step S10.

In step S10, the CPU 101 determines whether or not the teaching point determined to be unrotatable about the first axis parallel to the camera optical axis in step S8 is not on the same plane having the camera optical axis as a normal line. Judging. That is, the CPU 101 virtually rotates the calibrator 40 relative to the camera 500 around the second axis intersecting the first axis when there is no first rotation angle in the search in step S8. At that time, two axes other than the camera optical axis set in advance in the robot controller 100 are used as the second axis, and a rotation range around the two axes is used. The rotation range is searched with a predetermined step size (angle) determined in a range of about −45 deg to 45 deg. The predetermined step width (angle) is, for example, about 5 deg or 10 deg. Further, the rotation angle around two axes other than the camera optical axis already taken by the teaching point is excluded from this rotation range. With the above operation, the robot 300 and the calibrator 40 search for the second rotation angle that does not interfere with the obstacle 60. Also in step S10, as in step S8, in the case of the teaching points P ₁ and P ₂ , the search is performed while maintaining the relative angle α.

  If the result of the determination in step S10 is that the teaching point is movable so that it is not on the same plane with the camera optical axis as the normal line (S10: Yes), the resetting process in step S11 (resetting process) Execute.

In step S11, the CPU 101 moves and resets the teaching point determined in step S10 so that it is not on the same plane with the camera optical axis as a normal. That is, when the CPU 101 finds the second rotation angle, the CPU 101 sets the teaching points P ₁ and P ₂ so that the calibrator 40 rotates relative to the camera 500 around the second axis at the second rotation angle. Reset it. The amount of rotation follows the rotation range used in step S10. When the resetting of the teaching point is completed, the process proceeds to step S3.

As described above, in steps S 8 and S 10, the CPU 101 maintains the relative position and orientation between the calibrator 40 at the teaching point P _{1 and} the calibrator 40 at the teaching point P _2. The position and orientation of the robot 300 that does not interfere with the obstacle 60 are searched. If the robot 300 and the calibrator 40 find the position and orientation of the robot 300 that does not interfere with the obstacle 60 in steps S8 and S10, steps S9 and S11 are executed. That is, the CPU 101 resets the teaching points P ₁ and P ₂ based on the position and orientation of the robot 300 found in steps S8 and S10. In these steps S8 and S10, the search is performed under the condition that the position and orientation of the robot 300 (that is, each joint of the robot arm 301) is within the movable range by the determination in step S3.

  On the other hand, if it is determined in step S10 that the camera 101 is not movable so as not to be on the same plane with the camera optical axis as a normal line (S10: No), the teaching point search range that satisfies the predetermined calibration accuracy. The fact that there is no error is displayed on the monitor 600 (S12). This notifies the user that there is no teaching point search range. Thereafter, the process ends.

The processes in steps S3 to S5 and S8 to S11 described above are operations (simulations) performed by the CPU 101 using CAD data, and are performed without operating the actual robot 300. On the other hand, in step S6, the robot 300 is actually operated based on the teaching points P _{1 to} P ₃ to perform calibration. That is, step S3 to S5, the process of S8~S11 is an arithmetic processing for obtaining a teaching point _P 1 to P ₃ to be used for calibration.

  As described above, in hand-eye calibration using the camera 500 and the calibrator 40, a special jig whose accurate design data is clear is not required, and high-precision calibration at a narrower position is possible. Furthermore, a point that does not interfere with the obstacle 60 can be automatically searched for the calibration teaching point. Then, the robot 300 and the calibrator 40 automatically search for teaching points that do not interfere with the obstacle 60, thereby eliminating the need for the user to search by trial and error and reducing the burden on the user.

In particular, when the position of the marker 41 of the calibrator 40 (that is, the feature point such as the dot 43) is unknown in the CPU 101, the three teaching points P _{1 to} P ₃ are necessary, and the teaching point P ₁ and the teaching point it is necessary to search while maintaining the relative position and orientation of the P _2. In the first embodiment, since this search operation is automatically performed, there is no need for the user to perform a complicated search by trial and error, and the burden on the user is effectively reduced.

[Second Embodiment]
Next, a robot system according to a second embodiment of the invention will be described. FIG. 7 is an explanatory diagram showing a schematic configuration of a robot system according to the second embodiment of the present invention. Note that in the second embodiment, identical symbols are assigned to configurations similar to those in the first embodiment and descriptions thereof are omitted.

  In the first embodiment, the case where the calibrator is supported by the robot has been described. In the second embodiment, the case where the visual sensor is supported by the robot will be described. That is, in the first embodiment, the visual sensor is fixed and the calibrator is rotated with respect to the visual sensor. However, in the second embodiment, the calibrator is fixed and the visual sensor is rotated with respect to the calibrator. ing. Therefore, the first and second embodiments are the same in that they are relatively rotated between the visual sensor and the calibrator.

  In the first embodiment, calibration is performed between the robot coordinate system of the robot arm and the vision coordinate system of the visual sensor as the robot coordinate system. In the second embodiment, calibration is performed between the tool coordinate system of the robot hand and the vision coordinate system of the visual sensor as the robot coordinate system.

The robot system 10A includes a camera 500A as a visual sensor instead of the camera 500 of the first embodiment. The camera 500 </ b> A is supported by the robot 300 (specifically, the tip of the robot arm 301), and can be positioned at any position and orientation within the operation range of the robot 300 by operating the robot 300. The camera 500A is connected to the image processing apparatus 200. The calibrator 40 is installed so that the gantry 20 rather than the robot 300, that is, the robot coordinate system C ₁ in a fixed position.

  The image processing apparatus 200 converts an image signal captured via the camera 500A into a light / dark signal based on grayscale grayscale, and stores it in a frame memory. In addition, the image stored in the frame memory is processed, the identification of the target object, the position and orientation are measured, and the identification result and the position and orientation of the processed target object are transmitted to the robot controller 100. This function is the same as that of the first embodiment.

  The camera 500 </ b> A captures an object such as the calibrator 40 and transmits an image signal to the image processing apparatus 200. The camera 500A uses a monocular camera, but may be any visual sensor that can measure a three-dimensional position and orientation, and may be a compound eye camera, a laser range finder, or a combination thereof.

  Hereinafter, a method for performing hand eye calibration in the robot system 10A will be described.

FIG. 8 is a flowchart of the coordinate system calibration method according to the second embodiment. The CPU 101 (FIG. 2) as the control unit reads and executes the program 120 stored in the HDD 104, and automatically sets the teaching points used for calibration according to the algorithm shown in FIG. The CPU 101 operates the robot 300 (for example, the tip of the robot 300, that is, the robot hand 302) to the position and orientation taught at the plurality of teaching points, and causes the camera 500A to image the calibrator 40. Next, the CPU 101 calibrates the tool coordinate system C ₃ and the vision coordinate system C ₂ based on the processing result of the captured image in the image processing apparatus 200. Hereinafter, differences from the first embodiment will be described.

First, the CPU 101 generates and sets three teaching points used for calibration (S1 ′). In the first embodiment, the position and orientation of the calibrator 40 is changed by the operation of the robot 300, whereas in the second embodiment, the position and orientation of the camera 500A is changed. At that time, the teaching points P _{1 to} P ₃ are set so that the relative positions and orientations of the camera 500A and the calibrator 40 have a positional relationship as shown in FIG. The setting method is the same as in the first embodiment.

In step S8, CPU 101, based on the CAD data, thereby virtually relative rotation calibrator 40 to the camera 500A to the first axis parallel to the optical axis _{L 0} of the design. In the second embodiment, since the calibrator 40 is fixed, the camera 500A is virtually rotated. Then, the CPU 101 rotates the camera 500A by a predetermined step width (angle), and searches for a first rotation angle at which the robot 300 and the camera 500A do not interfere with the obstacle 60.

When the first rotation angle is found, in step S9, the CPU 101 teaches the teaching points P ₁ , P so that the calibrator 40 rotates relative to the camera 500A around the first axis at the first rotation angle. to reset the ₂ or teaching point _{P 3.} That is, the CPU 101 resets the teaching points P ₁ and P ₂ or the teaching point P ₃ so that the camera 500A rotates around the first axis at the first rotation angle. The same applies to steps S10 and S11.

Next, in step S6 ′, the CPU 101 operates the robot 300 on the _three teaching points P _{1 to} P ₃ determined in step S5. Then, CPU 101 obtains each of the relative position and orientation of the robot coordinate system _{C 1} and the tool coordinate system _{C 3} at the teaching point _P 1 to P _3, the relative position and orientation of the vision coordinate system _{C 2} and the marker coordinate system _{C 4.}

Here, homogeneous transformation matrix relative position and orientation between the tool coordinate system C ₃ and vision coordinate system C ₂ H _tv, homogeneous transformation matrix relative position and orientation of the robot coordinate system C ₁ and the marker coordinate system C ₄ H _rp And Further, in the respective three teaching points, the robot coordinate system _{C 1} and the tool coordinate system _{C 3} and the relative position and orientation of the homogeneous transformation matrix _{H rt1,} and H _{rt2, H rt3.} Also, the vision coordinate system _{C 2} and the marker coordinate system _{C 4} homogeneous transformation matrix relative position and orientation of the _{H vp1,} when the H _{vp2, H vp3,} simultaneous equations shown by the following equation is obtained.

The relative position and orientation H _tv between the tool coordinate system C ₃ to be hand-eye calibrated and the vision coordinate system C ₂ can be obtained by solving the optimization calculation for H _tv and H _{rp in} Equation (3). Then, CPU 101 is a tool coordinate system _{C 3} and of the relative position and orientation _{H tv} vision coordinate system _{C 2} obtained updating as a set value to the robot controller 100 (S7 '). Thereby, the process ends. Note that steps S2 to S5 and S8 to S12 execute substantially the same processing as in the first embodiment.

  As described above, in hand-eye calibration using the camera 500 and the calibrator 40, a special jig whose accurate design data is clear is not required, and high-precision calibration at a narrower position is possible. Furthermore, a point that does not interfere with the obstacle 60 can be automatically searched for the calibration teaching point. Then, the robot 300 and the calibrator 40 automatically search for teaching points that do not interfere with the obstacle 60, thereby eliminating the need for the user to search by trial and error and reducing the burden on the user.

In particular, when the position of the marker 41 of the calibrator 40 (that is, the feature point such as the dot 43) is unknown in the CPU 101, the three teaching points P _{1 to} P ₃ are necessary, and the teaching point P ₁ and the teaching point it is necessary to search while maintaining the relative position and orientation of the P _2. In the second embodiment, since this search operation is automatically performed, there is no need for the user to perform a complicated search by trial and error, and the burden on the user is effectively reduced.

  The present invention is not limited to the embodiments described above, and many modifications can be made by those having ordinary knowledge in the art within the technical idea of the present invention.

  In the first and second embodiments, the case where there are only three teaching points used for calibration has been described, but other teaching points may be included in addition to these three teaching points. That is, as long as three points satisfying the above conditions are included, the number is not limited to three, and three or more points may be set. At that time, a robot coordinate system or a tool coordinate system to be hand-eye calibrated is established by establishing the relationship of the formula (2) or the formula (3) at those points and solving the optimization calculation such as the least square method. And the relative position and orientation relative to the vision coordinate system.

  In step S8, if the calibration area A is a rectangle, if another vertex of the calibration area A can be selected as a teaching point that does not interfere, the teaching point can be reset in step S9.

  Furthermore, in step S12, the teaching point may be reset to the inside of the calibration area A by sacrificing the calibration accuracy at the user's judgment.

In steps S8 and S10, the calibrator 40 (or camera 500A) is rotated based on the CAD data to search for the teaching point. However, the teaching point is not limited to rotation, and translation may be performed. At that time, when the teaching points P ₁ and P ₂ are reset, the teaching points P ₁ and P ₂ may be translated while maintaining the relative angle α of the calibrator 40 viewed from the cameras 500 and 500A.

  In the first and second embodiments, the case where the position of the mark of the calibrator is unknown has been described. However, the present invention is also applied to the case where the position of the mark of the calibrator is known in the CPU as the control unit. Is possible. In that case, the teaching point used for calibration may be one.

  Each processing operation of the above embodiment is specifically executed by the CPU 101 as a control unit. Therefore, it may be achieved by supplying a recording medium recording a program for realizing the above-described function to the control device, and reading and executing the program stored in the recording medium by a computer (CPU or MPU) of the control device. Good. In this case, the program itself read from the recording medium realizes the functions of the above-described embodiments, and the program itself and the recording medium recording the program constitute the present invention.

  In the above embodiment, the computer-readable recording medium is the HDD 104, and the program 120 is stored in the HDD 104. However, the present invention is not limited to this. The program may be recorded on any recording medium as long as it is a computer-readable recording medium. For example, as a recording medium for supplying the program, the ROM 102, the recording disk 800, an external storage device (not shown) and the like shown in FIG. 2 may be used. As a specific example, a flexible disk, a hard disk, an optical disk, a magneto-optical disk, a CD-ROM, a CD-R, a magnetic tape, a rewritable nonvolatile memory (for example, a USB memory), a ROM, etc. Can be used.

  Further, the program in the above embodiment may be downloaded via a network and executed by a computer.

  Further, the present invention is not limited to the implementation of the functions of the above-described embodiment by executing the program code read by the computer. This includes a case where an OS (operating system) or the like running on the computer performs part or all of the actual processing based on the instruction of the program code, and the functions of the above-described embodiments are realized by the processing. .

  Furthermore, the program code read from the recording medium may be written in a memory provided in a function expansion board inserted into the computer or a function expansion unit connected to the computer. This includes a case where the CPU of the function expansion board or function expansion unit performs part or all of the actual processing based on the instruction of the program code, and the functions of the above embodiments are realized by the processing.

  Moreover, although the said embodiment demonstrated the case where a computer performs a process by running the program recorded on recording media, such as HDD, it is not limited to this. A part or all of the functions of the arithmetic unit that operates based on the program may be configured by a dedicated LSI such as an ASIC or FPGA. Note that ASIC is an acronym for Application Specific Integrated Circuit, and FPGA is an acronym for Field-Programmable Gate Array.

DESCRIPTION OF SYMBOLS 10 ... Robot system, 40 ... Calibrator, 60 ... Obstacle, 101 ... CPU (control part), 120 ... Program, 300 ... Robot, 500 ... Camera (visual sensor)

Claims (9)

The robot supports one member of the calibrator and the visual sensor, and the control unit moves the robot to the position and orientation taught at the teaching point, causing the visual sensor to image the calibrator, and the captured image. In the coordinate system calibration method for calibrating the coordinate system based on the robot and the coordinate system based on the visual sensor based on the processing result of
A setting step in which the control unit sets the teaching point;
An acquisition step in which the control unit acquires at least the robot, the one member, and three-dimensional shape data relating to an obstacle that may be an obstacle to the operation of the robot;
The control unit performs an operation for virtually operating the robot at the teaching point using the solid shape data, and determines whether the robot or the one member interferes with the obstacle. A decision process;
When the control unit determines that the robot or the one member interferes with the obstacle when the robot is virtually operated at the teaching point in the interference determination step, the robot and the one of the robot A search step for searching for a position and orientation of the robot in which a member does not interfere with the obstacle;
When the control unit finds the position and orientation of the robot in which the robot and the one member do not interfere with the obstacle in the searching step, the control unit determines the teaching point based on the found position and orientation of the robot. A coordinate system calibration method comprising: a resetting step for resetting. In the setting step, the control unit has the surface of the calibrator positioned as a teaching point at least on a first plane whose normal is the optical axis of the visual sensor design, and Operate the robot so that the surface of the calibrator is positioned on a first plane that intersects the first plane and a first teaching point that causes the robot to move so that the position and orientation are different from each other. Set the third teaching point to be
In the interference determination step, the control unit uses the solid shape data to perform an operation for virtually operating the robot at the first teaching point and the second teaching point, and the robot or the one member Determine whether or not it interferes with the obstacle,
In the searching step, when the control unit virtually operates the robot to at least one of the first teaching point and the second teaching point in the interference determination step, the robot or the robot When it is determined that one member interferes with the obstacle, the calibrator when the robot is operated to the first teaching point and the calibrator when the robot is operated to the second teaching point And searching for the position and orientation of the robot in which the robot and the one member do not interfere with the obstacle, while maintaining the relative position and orientation between
In the resetting step, when the control unit finds the position and orientation of the robot in which the robot and the one member do not interfere with the obstacle in the searching step, the control unit sets the position and orientation of the found robot. The coordinate system calibration method according to claim 1, wherein the first teaching point and the second teaching point are reset based on the first teaching point. In the searching step, the controller rotates the calibrator virtually relative to the visual sensor around a first axis parallel to the designed optical axis based on the solid shape data. And searching for a first rotation angle at which the robot and the one member do not interfere with the obstacle,
In the resetting step, when the control unit finds the first rotation angle, the calibrator rotates relative to the visual sensor around the first axis at the first rotation angle. Resetting the first teaching point and the second teaching point to
The coordinate system calibration method according to claim 2, wherein: In the searching step, the control unit virtually rotates the calibrator around the second axis intersecting the first axis relative to the visual sensor when there is no first rotation angle. Searching for a second rotation angle at which the robot and the one member do not interfere with the obstacle when
In the resetting step, when the control unit finds the second rotation angle, the calibrator rotates relative to the visual sensor around the second axis at the second rotation angle. Resetting the first teaching point and the second teaching point to
The coordinate system calibration method according to claim 3. In the interference determination step, the control unit performs an operation for virtually operating the robot at the third teaching point using the solid shape data, and the robot or the one member interferes with the obstacle. Decide whether or not to
In the search step, when the control unit determines in the interference determination step that the robot or the one member interferes with the obstacle when the robot is operated at the third teaching point, the robot And searching for the position and orientation of the robot in which the one member does not interfere with the obstacle,
In the resetting step, when the control unit finds the position and orientation of the robot in which the robot and the one member do not interfere with the obstacle in the searching step, the control unit sets the position and orientation of the found robot. The coordinate system calibration method according to any one of claims 2 to 4, wherein the third teaching point is reset based on the third teaching point.   In the searching step, the control unit searches for a position and orientation of the robot in which the robot and the one member do not interfere with the obstacle under a condition that the robot is within a movable range. The coordinate system calibration method according to claim 1. A visual sensor that images the calibrator;
A robot that supports any one member of the calibrator and the visual sensor;
The robot is operated to the position and orientation taught at the teaching point, the calibrator is imaged by the visual sensor, and the coordinate system based on the robot and the visual sensor based on the processing result of the captured image are used as a reference. A control unit that calibrates the coordinate system
The controller is
A setting process for setting the teaching point;
An acquisition process for acquiring at least three-dimensional shape data related to an obstacle that may be an obstacle to the operation of the robot, the one member, and the robot;
Using the three-dimensional shape data, performing an operation for virtually operating the robot at the teaching point, and determining whether the robot or the one member interferes with the obstacle;
In the interference determination process, when it is determined that the robot or the one member interferes with the obstacle when the robot is virtually operated at the teaching point, the robot and the one member are the obstacle. A search process for searching for the position and orientation of the robot that does not interfere with
In the search process, when the robot and the one member find the position and orientation of the robot that does not interfere with the obstacle, the resetting is performed to reset the teaching point based on the position and orientation of the found robot. A robot system characterized by executing processing.   The program for making a computer perform each process of the coordinate system calibration method of any one of Claims 1 thru | or 6.   A computer-readable recording medium on which the program according to claim 8 is recorded.

Images