WO2015037222A1 - 三次元物体認識装置、三次元物体認識方法、及び移動体 - Google Patents
三次元物体認識装置、三次元物体認識方法、及び移動体 Download PDFInfo
- Publication number
- WO2015037222A1 WO2015037222A1 PCT/JP2014/004618 JP2014004618W WO2015037222A1 WO 2015037222 A1 WO2015037222 A1 WO 2015037222A1 JP 2014004618 W JP2014004618 W JP 2014004618W WO 2015037222 A1 WO2015037222 A1 WO 2015037222A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- dimensional
- dimensional shape
- obstacle
- shape
- measurement
- Prior art date
Links
- 238000000034 method Methods 0.000 title claims description 23
- 238000005259 measurement Methods 0.000 claims abstract description 122
- 230000002093 peripheral effect Effects 0.000 claims description 8
- 238000010586 diagram Methods 0.000 description 9
- 238000009825 accumulation Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 230000005484 gravity Effects 0.000 description 2
- 241000282412 Homo Species 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000004321 preservation Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/70—Determining position or orientation of objects or cameras
- G06T7/73—Determining position or orientation of objects or cameras using feature-based methods
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/02—Control of position or course in two dimensions
- G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
- G05D1/0231—Control of position or course in two dimensions specially adapted to land vehicles using optical position detecting means
- G05D1/0246—Control of position or course in two dimensions specially adapted to land vehicles using optical position detecting means using a video camera in combination with image processing means
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
- G06V20/58—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/60—Type of objects
- G06V20/64—Three-dimensional objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2200/00—Indexing scheme for image data processing or generation, in general
- G06T2200/04—Indexing scheme for image data processing or generation, in general involving 3D image data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30248—Vehicle exterior or interior
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30248—Vehicle exterior or interior
- G06T2207/30252—Vehicle exterior; Vicinity of vehicle
- G06T2207/30261—Obstacle
Definitions
- the present invention relates to a three-dimensional object recognition device, a three-dimensional object recognition method, and a moving object.
- Patent Document 1 discloses a vehicle equipped with an image recognition device for the purpose of accurately and quickly recognizing a pedestrian within a certain range ahead of the vehicle in a captured image of a camera.
- This image recognition apparatus assigns identification information ID to a target detected beyond a predetermined pedestrian detectable area ahead of the host vehicle by laser radar exploration. Accordingly, the identification information ID is given to a target that is not a pedestrian and that absorbs less laser light, such as a vehicle or an electric pole of a roadside object.
- this image recognition apparatus recognizes a pedestrian by performing the image recognition process of pattern matching only about the target which does not have identification information ID of the pedestrian detectable area of a camera. Accordingly, it is stated that the load of the image recognition processing is reduced, and a pedestrian in the pedestrian detectable region can be recognized accurately and quickly.
- the three-dimensional shape of the obstacle is taken into consideration while recognizing the three-dimensional shape of the obstacle so that the moving body does not contact the obstacle. It is necessary to plan the movement route and move.
- a method of mounting a sensor for measuring the three-dimensional shape of the surrounding obstacle on the moving body and measuring the three-dimensional shape of the obstacle using the sensor can be considered.
- many sensors that measure a three-dimensional shape have a limited measurement range. Therefore, the obstacle that has entered the blind spot of the sensor cannot be measured.
- the present invention has been made based on the above-described knowledge, and is capable of grasping a wide range of three-dimensional shapes of surrounding objects while suppressing cost, and a three-dimensional object recognition method,
- the object is to provide a moving body.
- An apparatus for recognizing a three-dimensional object is an omnidirectional sensor that measures peripheral objects in all directions and generates position information that can specify the position of the object as a measurement result;
- a three-dimensional measurement device that measures an object within a certain measurement range and generates three-dimensional shape information that can specify the three-dimensional shape of the object as a measurement result, and the object is the third order
- a control unit that updates the shape recognized as the three-dimensional shape of the object based on the three-dimensional shape information generated by the three-dimensional measurement device when within the measurement range of the original measurement device, and the control unit Is to track the object based on the position information generated by the omnidirectional sensor even when the object goes out of the measurement range of the three-dimensional measuring apparatus.
- a three-dimensional object is measured by measuring a three-dimensional shape of an object within a certain measurement range of the three-dimensional measurement apparatus among peripheral objects. Based on the result, the step of updating the shape recognized as the three-dimensional shape of the object within the measurement range of the three-dimensional measuring device, and the position of the surrounding object in all directions by the omnidirectional sensor, And tracking an object that has come out of the measurement range of the three-dimensional measurement apparatus.
- the moving body includes an omnidirectional sensor that measures peripheral objects in all directions and generates position information capable of specifying the position of the object as a measurement result, and the peripheral objects
- a three-dimensional measuring device that measures an object within a certain measurement range and generates three-dimensional shape information capable of specifying the three-dimensional shape of the object as a measurement result, and the object is the three-dimensional measuring device If it is within the measurement range, the shape recognized as the three-dimensional shape of the object is updated based on the three-dimensional shape information generated by the three-dimensional measuring device, and the three-dimensional shape of the recognized object is updated.
- control unit configured to plan a movement path so as not to contact the object and to move the planned movement path, and the control unit includes a measurement range of the three-dimensional measurement apparatus. Even when they go out, it is to track the object based on the position information the omnidirectional sensor is generated.
- FIG. 1 is an external configuration diagram of an autonomous mobile body 1 according to the present embodiment.
- FIG. 2 is a diagram illustrating an example of an object measurement state by the autonomous mobile body 1 according to the embodiment.
- the autonomous mobile body 1 has a three-dimensional distance sensor 2, an omnidirectional sensor 3, a main body 4, and wheels 5a and 5b.
- the autonomous mobile body 1 recognizes obstacles (objects) around it and moves autonomously while avoiding the obstacles.
- the three-dimensional distance sensor 2 measures obstacles within a certain measurement range among obstacles around the autonomous mobile body 1, and uses the measurement surface (third order) of those obstacles as a measurement result.
- Three-dimensional shape information that can specify the three-dimensional shape on the surface facing the original distance sensor 2 is generated.
- the three-dimensional distance sensor 2 is disposed at the front part of the main body 4 so that an obstacle ahead of the autonomous mobile body 1 can be measured.
- the three-dimensional distance sensor 2 is, for example, a sensor that measures the distance to an obstacle for each direction within the measurement range. That is, the three-dimensional distance sensor 2 is a sensor that generates distance information indicating the distance to each point on the measurement surface of the obstacle within the measurement range as the three-dimensional shape information. Note that the three-dimensional distance sensor may use an arbitrary one of various methods such as an optical method, an ultrasonic method, and a laser beam method.
- the three-dimensional distance sensor 2 can measure only the three-dimensional shape of the obstacle within a certain measurement range, but can also measure the three-dimensional shape of the obstacle in all directions. It is cheap compared.
- the device that measures the obstacle and generates the three-dimensional shape information that can specify the three-dimensional shape of the obstacle is not limited to the three-dimensional distance sensor 2.
- a camera imaging device
- an obstacle in a certain range may be imaged by a camera, and image information indicating an image in the imaged range may be generated as three-dimensional shape information. That is, as long as information that can identify the three-dimensional shape of the obstacle in the measurement range can be obtained from the contents, the format of the three-dimensional shape information and the device for generating the three-dimensional shape information adopt any one can do.
- the three-dimensional shape information may be information indicating the three-dimensional shape of the obstacle as a distance like the distance information, and the three-dimensional shape of the obstacle is analyzed by analyzing the content like the image information. It may be information that can be derived.
- the omnidirectional sensor 3 measures obstacles around the autonomous mobile body 1 in all directions, and generates obstacle position information that can identify the positions of these obstacles as measurement results. Sensor.
- the omnidirectional sensor 3 is disposed on the upper part of the main body 4 so that the omnidirectional sensor 1 can measure all azimuths.
- the omnidirectional sensor 3 is, for example, a sensor that measures the distance to obstacles around the autonomous mobile body 1 in all directions and generates distance information indicating the distance to each obstacle in all directions as obstacle position information. Good. In this case, the omnidirectional sensor 3 is an inexpensive distance sensor that can measure the distance to the obstacle in all directions, but cannot measure the three-dimensional shape. Further, the omnidirectional sensor 3 may be a sensor that images obstacles around the autonomous mobile body 1 in all directions and generates image information indicating the captured omnidirectional images as obstacle position information. That is, as long as information that can identify the position of each obstacle in all directions can be obtained from the contents, any type of obstacle position information format and all-direction sensor can be adopted. That is, the obstacle position information may be information indicating the distance to the obstacle like distance information, and the distance to the obstacle can be derived by analyzing the content like image information. Information that can be used.
- the omnidirectional sensor 3 is inexpensive and cannot measure the three-dimensional shape of the surrounding obstacle, but can measure the position of the obstacle in a wide range. As described above, the omnidirectional sensor 3 cannot measure the three-dimensional shape of the obstacle, but the measurement range in the horizontal direction is wider than that of the three-dimensional distance sensor 2.
- the main body 4 is provided with wheels 5a and 5b at the lower part thereof.
- the autonomous mobile body 1 moves by driving the wheels 5a and 5b.
- the autonomous mobile body 1 moves autonomously avoiding obstacles around the autonomous mobile body 1 using the measurement results of the three-dimensional distance sensor 2 and the omnidirectional sensor 3. Moreover, even if the measurement range of the three-dimensional distance sensor 2 that measures the three-dimensional shape is limited by combining the three-dimensional distance sensor 2 and the omnidirectional sensor 3 with the autonomous mobile body 1 according to the present embodiment. This makes it possible to grasp the three-dimensional shape of surrounding obstacles over a wide range.
- FIG. 3 is an internal configuration diagram of the autonomous mobile body 1 according to the present embodiment.
- the autonomous mobile body 1 includes a control unit 10 and actuators 14a and 14b.
- the control unit 10 controls the autonomous mobile body 1 in an integrated manner.
- the control unit 10 has a CPU (Central Processing Unit) and a storage unit (not shown), and the CPU executes various programs as the control unit 10 by executing a program stored in the storage unit. To do. That is, this program includes code for causing the CPU of the control unit 10 to execute various processes as the control unit 10.
- the storage unit stores the program and various information used by the control unit 10.
- the storage unit includes at least one storage device capable of storing the program and various types of information. For example, at least one of a memory and a hard disk may be arbitrarily used as the storage device.
- the control unit 10 functions as a three-dimensional obstacle position storage unit 11, a route plan unit 12, and a moving body drive control unit 13.
- the three-dimensional obstacle position storage unit 11 Based on the obstacle position information output from the omnidirectional sensor 3, the three-dimensional obstacle position storage unit 11 recognizes obstacles around the autonomous mobile body 1 and their positions, and for each recognized obstacle. Identification information.
- the identification information is information that can uniquely identify each obstacle. For example, a unique integer may be assigned as identification information.
- the three-dimensional obstacle position storage unit 11 tracks the current position of each recognized obstacle based on the position of each obstacle specified by the obstacle position information. This makes it possible to track the current position of the obstacle regardless of whether the obstacle is within the measurement range of the three-dimensional distance sensor 2.
- the three-dimensional obstacle position storage unit 11 The obstacle position information indicating the position of the obstacle and the three-dimensional shape information indicating the three-dimensional shape of the obstacle specified based on the three-dimensional shape information output from the three-dimensional distance sensor 2 are stored in the storage unit in association with each other. . That is, for each obstacle, the obstacle position information and the three-dimensional shape information are associated with each other and stored in the storage unit. At this time, the format of the obstacle position information output from the omnidirectional sensor 3 and the format of the obstacle position information stored in the storage unit may be the same or different.
- the information format may be converted and stored as in the case of specifying the position and three-dimensional shape from the image information.
- the 3D obstacle position storage unit 11 outputs the 3D shape information of the obstacle from the 3D distance sensor 2 while the position of the obstacle is within the measurement range of the 3D distance sensor 2. It is updated by replacing it to indicate the latest three-dimensional shape specified based on the three-dimensional shape information.
- the 3D obstacle position storage unit 11 stores, for each obstacle, a 3D shape in which all of the 3D shapes recognized so far are stored separately from the 3D shape information indicating the latest 3D shape. Is also stored in the storage unit.
- the three-dimensional obstacle position storage unit 11 accumulates the three-dimensional shape of the obstacle with reference to the obstacle position. Thereby, when the obstacle operates like a human and the shape of the obstacle changes, the range in which the obstacle can move can be complemented sequentially as its three-dimensional shape.
- the three-dimensional obstacle position storage unit 11 stores the storage indicated by the three-dimensional shape storage information of the obstacle stored in the storage unit. Based on the three-dimensional shape, the three-dimensional shape estimated as the range in which the obstacle can move is calculated according to a predetermined calculation method. The calculation method will be described in detail later.
- the three-dimensional obstacle position storage unit 11 stores the three-dimensional shape information indicating the calculated three-dimensional shape as three-dimensional shape information indicating the latest three-dimensional shape of the obstacle. In this way, when an obstacle has gone out of the measurement range of the three-dimensional distance sensor 2, the three-dimensional shape of the obstacle is expanded to a three-dimensional shape that is estimated to be a range in which the obstacle can move. By recognizing the shape, the autonomous mobile body 1 and the obstacle are prevented from contacting each other.
- the three-dimensional obstacle position storage unit 11 is based on the three-dimensional shape information output from the three-dimensional distance sensor 2. The updating of the 3D shape information indicating the latest 3D shape of the obstacle is resumed.
- the obstacle is tracked based on the measurement result of the omnidirectional sensor 3, and the three-dimensional of the obstacle is detected. It is possible to keep track of the shape.
- the route planning unit 12 plans a route along which the autonomous mobile body 1 moves so that the autonomous mobile body 1 does not contact the obstacle based on the obstacle position information and the three-dimensional shape information stored in the storage unit.
- the mobile body drive control unit 13 calculates drive command values to the actuators 14a and 14b so that the autonomous mobile body 1 moves along the route planned by the route planning unit 12. Then, the moving body drive control unit 13 outputs the calculated drive command value to the actuators 14a and 14b.
- Actuators 14a and 14b drive the wheels 5a and 5b of the autonomous mobile body 1.
- the actuator 14a rotates the wheel 5a at a rotation speed corresponding to the drive command value from the moving body drive control unit 13.
- the actuator 14b rotates the wheel 5b at a rotation speed according to the drive command value from the moving body drive control unit 13.
- the wheels 5a and 5b correspond to the wheels attached to the lower part of the main body 4 described above.
- the moving body drive control unit 13 controls the actuators 14a and 14b to rotate the wheels 5a and 5b, thereby moving the moving path planned so that the autonomous moving body 1 does not contact the obstacle. To do.
- FIG. 4 is a flowchart showing processing of the autonomous mobile body 1 according to the present embodiment.
- FIG. 5 is a flowchart showing the three-dimensional shape information providing process in step S13 of FIG.
- FIG. 6 is a diagram illustrating a flow of data handled by the autonomous mobile body 1 according to the embodiment.
- FIG. 7 is a diagram illustrating an example of correspondence between an obstacle position and a three-dimensional shape.
- FIG. 8 is a diagram illustrating an example of accumulating the three-dimensional shape of the obstacle in the measurement range of the three-dimensional distance sensor.
- FIG. 9 is a diagram illustrating an example of generating a three-dimensional shape of an obstacle outside the measurement range of the three-dimensional distance sensor. Note that S1 to S5 in FIGS. 4 and 5 correspond to S1 to S5 in FIG. Further, S2 and S3 in FIG. 5 correspond to FIGS. 7 and 8, and S4 and S5 in FIG. 5 correspond to FIG.
- the autonomous mobile body 1 has started autonomous movement (S10).
- the autonomous movement to the destination set by the user is started.
- the autonomous mobile body 1 is not limited to the one for which the destination is set, and may be one that repeatedly moves and stops for an arbitrary purpose.
- the autonomous mobile body 1 acquires information by the three-dimensional distance sensor 2 and the omnidirectional sensor 3 (S11). Specifically, as described above, the three-dimensional distance sensor 2 measures obstacles within a certain measurement range among obstacles around the autonomous mobile body 1, generates three-dimensional shape information, and performs tertiary The data is output to the original obstacle position storage unit 11. Further, as described above, the omnidirectional sensor 3 measures obstacles around the autonomous mobile body 1 in all directions, generates obstacle position information, and outputs the obstacle position information to the three-dimensional obstacle position storage unit 11. The three-dimensional obstacle position storage unit 11 acquires the three-dimensional shape information output from the three-dimensional distance sensor 2 and the obstacle position information output from the omnidirectional sensor 3.
- the autonomous mobile body 1 identifies and tracks obstacles (S12). Specifically, the three-dimensional obstacle position storage unit 11 recognizes surrounding obstacles and their positions based on the obstacle position information output from the omnidirectional sensor 3, and identifies and tracks the obstacles. . The three-dimensional obstacle position storage unit 11 generates and assigns identification information for an obstacle recognized for the first time. Then, the three-dimensional obstacle position storage unit 11 associates the generated identification information with the obstacle position information indicating the position of the recognized obstacle and stores them in the storage unit (S1). Further, the three-dimensional obstacle position storage unit 11 tracks the position of an already recognized obstacle.
- the three-dimensional obstacle position storage unit 11 determines the position indicated by the obstacle position information of the obstacle stored in the storage unit as the current position (omnidirectional sensor).
- the position of the obstacle specified based on the obstacle position information output from 3 is updated (S1).
- the position indicated by the obstacle position information of a certain obstacle is the position of the certain obstacle specified from the obstacle position information from the omnidirectional sensor 3 most. If they are close, it may be determined that these obstacles are the same obstacle.
- the autonomous mobile body 1 gives the three-dimensional shape information of the obstacle to the recognized obstacle (S13). That is, for the obstacle recognized for the first time, its three-dimensional shape information is newly given, and for the recognized obstacle, the three-dimensional shape newly recognized is indicated as the three-dimensional shape information. By replacing, the shape recognized by the autonomous mobile body 1 as the three-dimensional shape of the obstacle is updated (S3). Thereby, as shown in FIG. 7, identification information, three-dimensional shape information, and obstacle position information are linked and stored in the storage unit for the identified obstacle. The process in step S13 will be described later in detail with reference to FIG.
- the autonomous mobile body 1 plans a route along which the autonomous mobile body 1 moves, calculates a drive command value for driving the wheels 5a and 5b so as to move along the planned route, and drives the wheels 5a and 5b (S14). .
- the route planning unit 12 determines that the autonomous mobile body 1 is an obstacle based on the obstacle position information and the three-dimensional shape information of each obstacle stored in the storage unit by the three-dimensional obstacle position storage unit 11. Plan a route to move so as not to touch the door.
- information indicating the three-dimensional shape of the autonomous mobile body 1 is stored in advance in the storage unit, and when the autonomous mobile body 1 moves based on the information, the three-dimensional shape of the autonomous mobile body 1 is stored. And a route that does not overlap with the three-dimensional shape of each obstacle may be derived.
- the moving body drive control unit 13 drives the actuators 14a and 14b by giving drive command values to the actuators 14a and 14b so that the autonomous moving body 1 moves along the route planned by the route planning unit 12. Thereby, the wheels 5a and 5b are driven, and the autonomous mobile body 1 moves along the route planned by the route planning unit 12.
- the autonomous mobile body 1 determines whether or not the autonomous mobile body 1 has reached the goal of autonomous movement (S15). For example, the autonomous mobile body 1 determines that the goal has been reached when the destination set by the user is reached.
- the autonomous mobile body 1 When it is determined that the autonomous mobile body 1 has reached the goal of autonomous movement (S15: Yes), the autonomous mobile body 1 ends the autonomous movement. When it is determined that the autonomous mobile body 1 has not reached the goal of autonomous movement (S15: No), the autonomous mobile body 1 returns to step S11.
- FIG. 5 is a flowchart for explaining step S13 of FIG. 6 in detail.
- the three-dimensional obstacle position storage unit 11 determines whether or not the identified obstacle is within the measurement range of the three-dimensional distance sensor 2 (S20).
- the three-dimensional obstacle position storage unit 11 determines the obstacle from the measurement result of the three-dimensional distance sensor 2. Are acquired and stored (S21). Specifically, the three-dimensional obstacle position storage unit 11 cuts out a certain range of three-dimensional shape information including the obstacle from the three-dimensional shape information output from the three-dimensional distance sensor 2 to obtain three-dimensional shape accumulation information. (S2).
- the three-dimensional shape information is accumulated at regular intervals. Specifically, as shown in FIG. 8, when the measured three-dimensional shape has already been accumulated as the three-dimensional shape of the obstacle, the three-dimensional shape newly acquired around the position of the obstacle. Are updated as the three-dimensional shape of the obstacle. In other words, the three-dimensional shape that is the range in which the obstacle can move is updated by complementing the shape newly found by measurement. According to this, for example, when the obstacle moves like a human, the posture change range (the range in which the arm or foot moves) can be recorded.
- the three-dimensional obstacle position storage unit 11 stores the three-dimensional shape of the obstacle in association with the obstacle position from the measurement result of the three-dimensional distance sensor 2 (S22). Specifically, the three-dimensional obstacle position storage unit 11 stores the cut-out three-dimensional shape information as three-dimensional shape information indicating the latest three-dimensional shape in association with the obstacle position information of the identified obstacle. Save to the department.
- the 3D shape information stored in the storage unit is not limited to the format of the cut out 3D shape information itself, and arbitrary format conversion may be performed. Thereby, as shown in FIG. 7, identification information, three-dimensional shape information, and obstacle position information are linked and stored in the storage unit for the identified obstacle (S3).
- the obstacle position information and the three-dimensional shape information of the obstacle are displayed.
- the position of the obstacle is tracked while being updated. If the obstacle is out of the measurement range of the three-dimensional distance sensor 2, the position of the obstacle can be traced. Cannot be measured or updated.
- the three-dimensional obstacle position storage unit 11 Based on the shape recognized as the three-dimensional shape, the posture change range of the obstacle (the range in which the three-dimensional shape of the obstacle changes) is calculated (S23). Specifically, as shown in FIG. 9, the three-dimensional obstacle position storage unit 11 rotates the three-dimensional shape accumulated for the obstacle around an axis extending vertically upward from the position of the obstacle. A three-dimensional shape is calculated (S4). That is, a three-dimensional shape obtained by rotating the accumulated three-dimensional shape indicated by the three-dimensional shape accumulation information about the obstacle position axis is calculated.
- the three-dimensional obstacle position storage unit 11 converts the calculated three-dimensional shape into the three-dimensional shape information indicating the three-dimensional shape, and the third order indicating the latest three-dimensional shape of the obstacle outside the measurement range.
- the original shape information is stored in the storage unit in association with the obstacle position information of the obstacle (S24, S25, S5).
- the three-dimensional shape information is continuously stored in the storage unit until the obstacle enters the measurement range of the three-dimensional distance sensor 2 again. Then, when the obstacle enters the measurement range of the three-dimensional distance sensor 2 again, the three-dimensional obstacle position storage unit 11 restores the original three-dimensional shape information. Or you may make it update 3D shape information by the process in said step S22, without forcibly returning to original 3D shape information.
- an arbitrary position in the accumulated three-dimensional shape may be determined as the rotation axis of the three-dimensional shape.
- the center of gravity of the three-dimensional shape may be determined as the accumulated rotation axis of the three-dimensional shape.
- the center position of a predetermined characteristic part of the three-dimensional shape may be determined as the accumulated three-dimensional rotation axis.
- the center position of the head may be determined as the rotation axis.
- the position of the obstacle and the rotation axis are made to coincide. To do.
- the three-dimensional shape obtained by rotating the accumulated three-dimensional shape becomes a three-dimensional shape in which the range in which the obstacle can move by its own posture change (for example, rotational motion) is estimated.
- the obstacle is a human being
- the three-dimensional distance sensor 2 can measure a three-dimensional shape with one arm raised, after the person goes out of the measurement range of the three-dimensional distance sensor 2,
- the autonomous mobile body 1 does not recognize the protrusion of the arm as a three-dimensional shape, so autonomous movement
- the body 1 may come into contact with humans.
- the three-dimensional shape obtained by rotating the three-dimensional shape of the recognized obstacle is recognized as the three-dimensional shape of the obstacle, avoid contact even in such a case. Is possible.
- the three-dimensional shape information calculated and stored as described above is used.
- the three-dimensional shape including the posture change range is estimated and recognized as the three-dimensional shape of the obstacle.
- the latest three-dimensional shape or the accumulated three-dimensional shape may be used as it is without estimating the three-dimensional shape. Even in such a case, it is possible to track an obstacle by the omnidirectional sensor 3 and perform interference determination in consideration of the three-dimensional shape of the obstacle. Therefore, the three-dimensional shape of the surrounding obstacle can be grasped in a wide range.
- the obstacle operates and the three-dimensional shape is changed. Even if it changes, contact with the obstacle can be avoided.
- the control unit 10 performs the tertiary when the obstacle (object) is within the measurement range of the three-dimensional distance sensor 2 (three-dimensional measurement device). Based on the three-dimensional shape information generated by the original distance sensor 2, the shape recognized as the three-dimensional shape of the obstacle is updated. In addition, the control unit 10 tracks the obstacle based on the obstacle position information generated by the omnidirectional sensor 3 even when the obstacle is out of the measurement range of the three-dimensional distance sensor 2. Yes.
- the three-dimensional distance sensor 2 cannot measure a three-dimensional shape at a low cost. Although the measurement range is narrow, the blind spot of the three-dimensional distance sensor 2 that can measure a three-dimensional shape with high accuracy can be compensated. Therefore, according to the present embodiment, it is possible to continuously grasp a three-dimensional shape of surrounding obstacles while suppressing costs.
- the control part 10 is based on the shape acquired as a three-dimensional shape of an obstacle about the obstacle which came out of the measurement range of the three-dimensional distance sensor 2.
- FIG. the three-dimensional shape including the range in which the obstacle changes is estimated, and the estimated three-dimensional shape is recognized as the three-dimensional shape of the object outside the measurement range of the three-dimensional distance sensor 2.
- the three-dimensional shape that the obstacle can take can be grasped more appropriately.
- the three-dimensional shape of the obstacle can be suitably recognized in the autonomous mobile body 1 that moves while avoiding the obstacle.
- the three-dimensional shape of the posture change range of the obstacle is estimated by rotating the accumulated three-dimensional shape, but is not limited thereto. For example, do not accumulate a separate 3D shape, and estimate the 3D shape of the obstacle's posture change range by rotating the latest 3D shape of the obstacle around the position of the obstacle. May be. However, preferably, the three-dimensional shape including all the obstacle posture change ranges is estimated by rotating the accumulated three-dimensional shape to estimate the three-dimensional shape of the obstacle posture change range. Therefore, contact with an obstacle can be avoided more.
- the identified obstacle is categorized by analyzing the measurement result of at least one of the two sensors, and a detailed three-dimensional shape of the posture change range is designated according to the category of the obstacle. You may do it.
- the failure of that category prepared in advance
- the three-dimensional shape of the posture change range of the object may be used as the three-dimensional shape of the posture change range of the obstacle.
- 3D shape information indicating the three-dimensional shape of the posture change range of the human stored in advance in the storage unit is stored in the obstacle.
- the latest three-dimensional shape may be stored in the storage unit as three-dimensional shape information.
- measurement pattern information indicating a measurement pattern possessed by the obstacle of the category and three-dimensional shape information of the obstacle of the category are stored in the storage unit in advance.
- the three-dimensional obstacle position storage unit 11 stores in the storage unit at least one of the measurement results (obstacle position, latest three-dimensional shape, and accumulated three-dimensional shape) of the three-dimensional distance sensor 2 and the omnidirectional sensor 3.
- the three-dimensional shape information corresponding to the measurement pattern information is used as the three-dimensional shape information indicating the three-dimensional shape of the obstacle. Identify.
- the three-dimensional shape of the obstacle can be recognized more appropriately. Moreover, thereby, more accurate three-dimensional contact determination can be performed in the route plan, and autonomous movement with higher reliability and higher traveling performance is possible.
- the latest three-dimensional shape is recognized by the three-dimensional distance sensor 2 as the latest three-dimensional shape, but is not limited thereto.
- the accumulated three-dimensional shape may be recognized as the latest three-dimensional shape.
- the three-dimensional shape recognition device is not limited to the autonomous mobile body 1 and is applied to various devices that perform arbitrary control using the three-dimensional shape of a surrounding obstacle (object) recognized by the three-dimensional shape recognition device. can do.
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Aviation & Aerospace Engineering (AREA)
- Automation & Control Theory (AREA)
- Electromagnetism (AREA)
- Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)
- Computer Networks & Wireless Communication (AREA)
- Image Analysis (AREA)
- Length Measuring Devices By Optical Means (AREA)
- Image Processing (AREA)
- Optical Radar Systems And Details Thereof (AREA)
Abstract
Description
まず、本実施の形態に係る自律移動体1について説明する。図1及び図2を参照して、本実施の形態に係る自律移動体1の外部構成について説明する。図1は、本実施の形態に係る自律移動体1の外部構成図である。図2は、実施の形態に係る自律移動体1による物体の測定状況の一例を示す図である。
上記の実施の形態では、障害物の姿勢変化範囲が未知であると仮定して、障害物の三次元形状を計測結果に基づいて随時更新するとともに、2つのセンサ(三次元距離センサ2、全方位センサ3)の共通計測範囲外に出た障害物については、その三次元形状を回転させることで姿勢変化範囲を推定し、利用するようにしていた。
2 三次元距離センサ
3 全方位センサ
4 本体
5a、5b 車輪
10 制御部
11 三次元障害物位置保存部
12 経路計画部
13 移動体駆動制御部
14a、14b アクチュエータ
Claims (7)
- 全方位について周辺の物体を測定し、測定結果として前記物体の位置を特定することができる位置情報を生成する全方位センサと、
前記周辺の物体のうち、一定の測定範囲内の物体を測定し、測定結果として前記物体の三次元形状を特定することができる三次元形状情報を生成する三次元測定装置と、
前記物体が前記三次元測定装置の測定範囲内にある場合、前記三次元測定装置が生成した三次元形状情報に基づいて、前記物体の三次元形状として認識する形状を更新する制御部と、を備え、
前記制御部は、前記物体が前記三次元測定装置の測定範囲外に出た場合であっても、前記全方位センサが生成した位置情報に基づいて前記物体を追跡する、
三次元物体認識装置。 - 前記制御部は、前記三次元測定装置の測定範囲外に出た物体については、前記物体の三次元形状として認識した形状に基づいて前記物体が変化する範囲を含む三次元形状を推定し、推定した三次元形状を前記三次元測定装置の測定範囲外に出た物体の三次元形状として認識する、
請求項1に記載の三次元物体認識装置。 - 前記制御部は、前記物体が変化する範囲を含む形状として、前記物体の形状として認識した三次元形状を回転させた形状を算出する、
請求項2に記載の三次元物体認識装置。 - 前記物体の形状として認識した三次元形状は、前記三次元測定装置が生成した三次元形状情報に基づいて特定した前記物体の三次元形状を蓄積した形状である、
請求項2又は3に記載の三次元物体認識装置。 - 前記三次元物体認識装置は、さらに、前記物体の三次元形状として認識する形状を示す三次元形状情報と、前記物体の位置を示す位置情報とが対応付けて格納される記憶部を備え、
前記制御部は、前記物体の三次元形状として認識する形状の更新として、前記三次元測定装置が生成した三次元形状情報に基づいて前記記憶部に格納された三次元形状情報が示す三次元形状を更新するとともに、前記物体の追跡として、前記全方位センサが生成した位置情報に基づいて前記記憶部に格納された位置情報が示す前記物体の位置を更新する、
請求項1乃至4のいずれか1項に記載の三次元物体認識装置。 - 三次元測定装置によって、周辺の物体のうち、前記三次元測定装置の一定の測定範囲内における物体について三次元形状を測定し、測定結果に基づいて前記三次元測定装置の測定範囲内にある物体の三次元形状として認識する形状を更新するステップと、
全方位センサによって、全方位について周辺の物体の位置を測定し、測定結果に基づいて前記三次元測定装置の測定範囲外に出た物体を追跡するステップと、
を備えた三次元物体認識方法。 - 全方位について周辺の物体を測定し、測定結果として前記物体の位置を特定することができる位置情報を生成する全方位センサと、
前記周辺の物体のうち、一定の測定範囲内の物体を測定し、測定結果として前記物体の三次元形状を特定することができる三次元形状情報を生成する三次元測定装置と、
前記物体が前記三次元測定装置の測定範囲内にある場合、前記三次元測定装置が生成した三次元形状情報に基づいて、前記物体の三次元形状として認識する形状を更新するとともに、前記認識している物体の三次元形状に基づいて、前記物体と接触しないように移動経路を計画し、計画した移動経路を移動するように制御する制御部と、を備え、
前記制御部は、前記物体が前記三次元測定装置の測定範囲外に出た場合であっても、前記全方位センサが生成した位置情報に基づいて前記物体を追跡する、
移動体。
Priority Applications (3)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
DE112014004163.7T DE112014004163B4 (de) | 2013-09-11 | 2014-09-09 | Vorrichtung zur Erkennung eines dreidimensionalen Objekts, Verfahren zur Erkennung eines dreidimensionalen Objekts und Fahrzeug |
KR1020157030923A KR101971317B1 (ko) | 2013-09-11 | 2014-09-09 | 삼차원 물체 인식 장치, 삼차원 물체 인식 방법 및, 이동체 |
US14/888,184 US9633438B2 (en) | 2013-09-11 | 2014-09-09 | Three-dimensional object recognition apparatus, three-dimensional object recognition method, and vehicle |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2013188267A JP6032163B2 (ja) | 2013-09-11 | 2013-09-11 | 三次元物体認識装置、三次元物体認識方法、及び移動体 |
JP2013-188267 | 2013-09-11 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2015037222A1 true WO2015037222A1 (ja) | 2015-03-19 |
Family
ID=52665353
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2014/004618 WO2015037222A1 (ja) | 2013-09-11 | 2014-09-09 | 三次元物体認識装置、三次元物体認識方法、及び移動体 |
Country Status (5)
Country | Link |
---|---|
US (1) | US9633438B2 (ja) |
JP (1) | JP6032163B2 (ja) |
KR (1) | KR101971317B1 (ja) |
DE (1) | DE112014004163B4 (ja) |
WO (1) | WO2015037222A1 (ja) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2021504715A (ja) * | 2017-11-28 | 2021-02-15 | 深▲せん▼市杉川机器人有限公司Shenzhen 3Irobotix Co., Ltd. | 領域輪郭作成の方法、装置及びコンピュータ読取可能な記憶媒体 |
Families Citing this family (15)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US9625582B2 (en) * | 2015-03-25 | 2017-04-18 | Google Inc. | Vehicle with multiple light detection and ranging devices (LIDARs) |
US9888174B2 (en) | 2015-10-15 | 2018-02-06 | Microsoft Technology Licensing, Llc | Omnidirectional camera with movement detection |
US10277858B2 (en) * | 2015-10-29 | 2019-04-30 | Microsoft Technology Licensing, Llc | Tracking object of interest in an omnidirectional video |
JP6473685B2 (ja) * | 2015-11-19 | 2019-02-20 | 日立建機株式会社 | 車両制御装置及び作業機械 |
US10204274B2 (en) * | 2016-06-29 | 2019-02-12 | Cellular South, Inc. | Video to data |
JP2018045482A (ja) * | 2016-09-15 | 2018-03-22 | ソニー株式会社 | 撮像装置、信号処理装置、及び、車両制御システム |
JP6773503B2 (ja) * | 2016-09-27 | 2020-10-21 | 株式会社トプコン | レーザスキャナシステム及び点群データのレジストレーション方法 |
US10386856B2 (en) | 2017-06-29 | 2019-08-20 | Uber Technologies, Inc. | Autonomous vehicle collision mitigation systems and methods |
US10065638B1 (en) | 2017-08-03 | 2018-09-04 | Uber Technologies, Inc. | Multi-model switching on a collision mitigation system |
DE102018114293A1 (de) * | 2018-06-14 | 2019-12-19 | Helga Sommer | Computerimplementiertes Verfahren für die Erstellung einer Route für Transport und Durchführung von Transport |
JP7199984B2 (ja) * | 2019-02-01 | 2023-01-06 | 株式会社小松製作所 | 作業車両の制御システム及び作業車両の制御方法 |
DE102019202511A1 (de) * | 2019-02-25 | 2020-08-27 | Kuka Deutschland Gmbh | Autonomes Fahrzeug mit einer Sensoreinrichtung |
CN110108231B (zh) * | 2019-05-08 | 2020-10-27 | 广东嘉腾机器人自动化有限公司 | 基于角点检测的钜体箱体三维尺寸测量方法及存储介质 |
EP3748393A1 (en) * | 2019-06-04 | 2020-12-09 | Jabil Optics Germany GmbH | Surround-view imaging system |
US20240139945A1 (en) * | 2022-10-31 | 2024-05-02 | Omron Corporation | Method and apparatus for dynamically modifying a representation of an autonomous robot |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH07248817A (ja) * | 1994-03-12 | 1995-09-26 | Nec Corp | 水中走行機 |
JP2006039760A (ja) * | 2004-07-23 | 2006-02-09 | Victor Co Of Japan Ltd | 移動ロボット |
JP2013057541A (ja) * | 2011-09-07 | 2013-03-28 | Ihi Corp | 対象物との相対位置計測方法と装置 |
Family Cites Families (17)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH09331520A (ja) * | 1996-06-13 | 1997-12-22 | Nippon Telegr & Teleph Corp <Ntt> | 自動追跡システム |
JPH11296229A (ja) * | 1998-02-13 | 1999-10-29 | Komatsu Ltd | 車両の誘導装置 |
DE10242852A1 (de) * | 2002-09-14 | 2004-03-25 | Technische Universität Ilmenau Abteilung Forschungsförderung und Technologietransfer | Verfahren zur Minimierung des Einflusses von Störsignalen bei der Formelementeberechnung aus Koordinatenpunkten |
JP3760918B2 (ja) * | 2003-01-21 | 2006-03-29 | 株式会社日立製作所 | セキュリティシステム |
JP2006220521A (ja) * | 2005-02-10 | 2006-08-24 | Hitachi Ltd | 自己位置計測装置及び自己位置計測方法を実行するためのプログラム |
JP2009092383A (ja) * | 2007-10-03 | 2009-04-30 | Toyota Motor Corp | 障害物検知センサ搭載シミュレーション装置、障害物検知センサ搭載シミュレーション方法及び障害物検知センサ搭載シミュレーションプログラム |
KR20090062881A (ko) * | 2007-12-13 | 2009-06-17 | 삼성전자주식회사 | 이동 로봇 및 그 이동 객체 검출 방법 |
JP5178276B2 (ja) | 2008-03-27 | 2013-04-10 | ダイハツ工業株式会社 | 画像認識装置 |
US9740921B2 (en) * | 2009-02-26 | 2017-08-22 | Tko Enterprises, Inc. | Image processing sensor systems |
JP5462597B2 (ja) | 2009-11-12 | 2014-04-02 | 株式会社日立情報通信エンジニアリング | 物体測定システム及びその方法 |
JP5430456B2 (ja) * | 2010-03-16 | 2014-02-26 | キヤノン株式会社 | 幾何特徴抽出装置、幾何特徴抽出方法、及びプログラム、三次元計測装置、物体認識装置 |
JP2011209019A (ja) * | 2010-03-29 | 2011-10-20 | Sony Corp | ロボット装置及びロボット装置の制御方法 |
JP5872818B2 (ja) * | 2010-08-20 | 2016-03-01 | パナソニック株式会社 | 測位処理装置、測位処理方法、および画像処理装置 |
JP5770486B2 (ja) * | 2011-02-21 | 2015-08-26 | 株式会社トプコン | 全周画像計測装置 |
JP5810296B2 (ja) * | 2011-06-07 | 2015-11-11 | パナソニックIpマネジメント株式会社 | 画像表示装置及び画像表示方法 |
JP5906028B2 (ja) * | 2011-06-22 | 2016-04-20 | キヤノン株式会社 | 画像処理装置、画像処理方法 |
JP5636014B2 (ja) | 2012-03-12 | 2014-12-03 | 株式会社三共 | 遊技機 |
-
2013
- 2013-09-11 JP JP2013188267A patent/JP6032163B2/ja active Active
-
2014
- 2014-09-09 DE DE112014004163.7T patent/DE112014004163B4/de not_active Expired - Fee Related
- 2014-09-09 KR KR1020157030923A patent/KR101971317B1/ko active IP Right Grant
- 2014-09-09 WO PCT/JP2014/004618 patent/WO2015037222A1/ja active Application Filing
- 2014-09-09 US US14/888,184 patent/US9633438B2/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH07248817A (ja) * | 1994-03-12 | 1995-09-26 | Nec Corp | 水中走行機 |
JP2006039760A (ja) * | 2004-07-23 | 2006-02-09 | Victor Co Of Japan Ltd | 移動ロボット |
JP2013057541A (ja) * | 2011-09-07 | 2013-03-28 | Ihi Corp | 対象物との相対位置計測方法と装置 |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2021504715A (ja) * | 2017-11-28 | 2021-02-15 | 深▲せん▼市杉川机器人有限公司Shenzhen 3Irobotix Co., Ltd. | 領域輪郭作成の方法、装置及びコンピュータ読取可能な記憶媒体 |
US11275178B2 (en) | 2017-11-28 | 2022-03-15 | Shenzhen 3Irobotix Co., Ltd. | Method and device for drawing region outline and computer readable storage medium |
Also Published As
Publication number | Publication date |
---|---|
JP6032163B2 (ja) | 2016-11-24 |
KR101971317B1 (ko) | 2019-04-22 |
DE112014004163B4 (de) | 2021-10-07 |
DE112014004163T5 (de) | 2016-05-25 |
JP2015055974A (ja) | 2015-03-23 |
US20160063710A1 (en) | 2016-03-03 |
KR20150135505A (ko) | 2015-12-02 |
US9633438B2 (en) | 2017-04-25 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP6032163B2 (ja) | 三次元物体認識装置、三次元物体認識方法、及び移動体 | |
JP6849330B2 (ja) | 地図生成方法、自己位置推定方法、ロボットシステム、およびロボット | |
JP6762148B2 (ja) | 地図生成方法、移動ロボット及び地図生成システム | |
JP5112666B2 (ja) | 移動装置 | |
JP4467534B2 (ja) | 障害物のある環境下で自律移動する移動ロボットおよび移動ロボットの制御方法。 | |
US8515612B2 (en) | Route planning method, route planning device and autonomous mobile device | |
KR101439921B1 (ko) | 비젼 센서 정보와 모션 센서 정보를 융합한 모바일 로봇용 slam 시스템 | |
JP6539125B2 (ja) | 移動ロボットシステム及び移動ロボットの制御方法 | |
KR20170088228A (ko) | 다중로봇의 자기위치인식에 기반한 지도작성 시스템 및 그 방법 | |
US7873438B2 (en) | Mobile apparatus and control program therefor | |
KR101990878B1 (ko) | 시간상태 영역에서의 온라인 양방향 경로 계획 방법, 이를 구현하기 위한 프로그램이 저장된 기록매체 및 이를 구현하기 위해 매체에 저장된 컴퓨터프로그램 | |
JP5276931B2 (ja) | 移動体および移動体の位置推定誤り状態からの復帰方法 | |
JP2009093308A (ja) | ロボットシステム | |
KR101341204B1 (ko) | 레이저 스캐너 및 구조물을 이용한 모바일 로봇의 위치추정장치 및 방법 | |
JP2008059218A (ja) | 自律走行ロボットの自己位置回復方法 | |
KR20150020900A (ko) | 레이저 스캐너를 이용한 위치인식 및 맵핑 시스템을 이용한 위치인식 방법 | |
JP5287060B2 (ja) | 経路計画装置及び自律移動装置 | |
JP2009136933A (ja) | ロボットおよびタスク実行システム | |
JP2018185767A (ja) | 環境整備ロボットおよびその制御プログラム | |
WO2017038012A1 (ja) | 地図生成方法、自己位置推定方法、ロボットシステム、およびロボット | |
JP5287050B2 (ja) | 経路計画方法、経路計画装置、及び自律移動装置 | |
US20220019224A1 (en) | Mobile body, method of controlling mobile body, and program | |
JP2011238104A (ja) | 移動ロボット | |
JP5803769B2 (ja) | 移動ロボット | |
JP5439552B2 (ja) | ロボットシステム |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 14844799 Country of ref document: EP Kind code of ref document: A1 |
|
ENP | Entry into the national phase |
Ref document number: 20157030923 Country of ref document: KR Kind code of ref document: A |
|
WWE | Wipo information: entry into national phase |
Ref document number: 14888184 Country of ref document: US |
|
WWE | Wipo information: entry into national phase |
Ref document number: 1120140041637 Country of ref document: DE Ref document number: 112014004163 Country of ref document: DE |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 14844799 Country of ref document: EP Kind code of ref document: A1 |