JP2020113134A - Unmanned moving object, unmanned moving object control method and unmanned moving program - Google Patents

Abstract

To provide an unmanned moving object enabling a prompt control on a moving direction without the placement of a marker in advance and without the use of positioning radio waves from a navigation satellite.SOLUTION: An unmanned moving object 1 capable of autonomously moving includes: image obtaining means for picking up the image of an exterior covering a goal position that is a position where the unmanned moving object is to move and reach, and for obtaining the image; feature point extracting means for extracting multiple feature points within the image; relative position defining means for defining the goal position within the image in accordance with the relative relation to the respective positions of the multiple feature points; predicting means for calculating a change in position of each of the multiple feature points in the image as a predicted change when the unmanned moving object moves toward the goal position through a predetermined route; change determining means for determining whether or not the actual change that is the actual change in position of each of the multiple feature points in the image is consistent with the predicted change while the unmanned moving body is moving toward the goal position; and moving direction changing means for changing the moving direction of the unmanned moving object so as to cancel the change when the determination is made that the actual change is inconsistent with the predicted change.SELECTED DRAWING: Figure 1

Description

The present invention relates to an unmanned moving body, a method for controlling the unmanned moving body, and an unmanned moving program.

Conventionally, a technique has been proposed in which an unmanned vehicle is automatically driven while detecting a marker arranged in the vicinity of a route in advance (for example, Patent Document 1).

Japanese Patent No. 5046310

In the above-mentioned technique, it is necessary to arrange the marker near the route along which the unmanned vehicle travels. Therefore, it is necessary to prepare the markers, and it takes time to arrange the markers. Also, it is difficult or impossible to place the marker depending on the situation. It is possible to use positioning radio waves from navigation satellites such as GPS (Global Positioning System) satellites, and to drive while using an electronic compass or acceleration sensor together, but it is suitable for positioning using radio waves for positioning. Therefore, it is difficult to quickly correct the traveling direction.

The present invention has attempted to solve such a problem, and an unmanned vehicle capable of rapidly controlling its moving direction without arranging markers in advance and using positioning radio waves from a navigation satellite. , An unmanned moving body control method and an unmanned moving program are provided.

A first aspect of the present invention is an unmanned mobile body capable of autonomous movement, which captures an image of a direction of a target position, which is a position where the unmanned mobile body moves and reaches, and an image capturing unit that captures an image. In the feature point extracting means for extracting a plurality of feature points, a relative position defining means for defining the direction of the target position in the image in a relative relationship with the positions of the plurality of feature points, and the unmanned moving body. Prediction means for calculating a change in the position of the plurality of characteristic points in the image as a predicted change under the condition of moving toward the target position on the planned route, and the unmanned moving body moves toward the target position. In the meantime, the actual change, which is the actual change of the positions of the plurality of feature points in the image, is changed by the change determination means for determining whether or not the actual change is different from the predicted change, and the actual change is performed by the change determination means. When it is determined that there is a difference between the change and the predicted change, the moving direction changing unit that changes the moving direction of the unmanned moving body so as to cancel the difference is an unmanned moving body.

According to the configuration of the first invention, the unmanned moving body defines the direction of the target position in a relative relationship with the plurality of feature points, and sets the plurality of feature points under the condition of moving toward the target position on the planned route. The change in position is calculated as a predicted change, and it is judged whether or not it is the same as the change in a plurality of feature points during the actual movement. If they are different, the moving direction is changed so as to cancel the difference. be able to. That is, the movement direction can be changed using a plurality of feature points as markers. For example, when there is only one feature point, the position in the image may not change even if the direction of the unmanned moving body changes, or it may be extremely small. On the other hand, when a plurality of feature points are used, such as two feature points or three feature points, the position of at least one feature point always changes if the moving direction of the unmanned moving body changes. Further, the change in the positions of the plurality of feature points in the image is easier to detect because the change amount of the pixels in the image is larger than the change in one feature point. Furthermore, when there is only one feature point, the feature point may not be displayed in the image due to some natural phenomenon or artificial phenomenon. That is, by using a plurality of feature points, the reliability of the judgment by the change judging means is improved. Moreover, since the time required for the image processing is significantly shorter than the time required for the positioning processing, it is possible to quickly control the moving direction. Thereby, the moving direction can be controlled without arranging the marker in advance and using the positioning radio wave from the navigation satellite.

A second invention is, in the configuration of the first invention, among the plurality of the feature points extracted by the feature point extracting means, the feature points indicating a movement different from the overall tendency of the movement of the plurality of the feature points, It is an unmanned moving body that has a means for eliminating it.

When using feature points as markers, feature points of moving objects such as moving vehicles and people walking (hereinafter referred to as "moving objects") are eliminated, and natural objects such as trees and buildings It is desirable to use the characteristic points of the structure (hereinafter, referred to as "animal"). By the way, it is general that there are more inanimate animals than moving objects in the range in which an image is to be acquired. Since the majority of the image is occupied by inanimate animals, the overall trend of movement of multiple feature points in the image is formed by the inanimate feature points. Then, the movement of the feature points of the inanimate object according to the movement mode of the unmanned moving body matches the overall tendency. For example, when the unmanned moving body is stopped, the tendency of the overall movement shows a zero velocity vector, and the movement of each feature point of the inanimate animal also shows a zero velocity vector. On the other hand, the movement of each feature point of the moving object is not a vector with zero velocity, and is different from the overall tendency. This is the same when the unmanned moving body is moving forward, for example, the feature points for the inanimate animal on the right side with respect to the traveling direction show the overall tendency to move to the right side, and The feature points for inanimate animals show an overall tendency to move to the left, but the movement of feature points for moving objects is different from the overall tendency. In this regard, according to the configuration of the second aspect of the present invention, since the feature point exhibiting a movement different from the overall tendency can be eliminated by the eliminating means, only the feature point regarding the inanimate object can be used. ..

A third aspect of the invention is the configuration of the first aspect or the second aspect of the invention, in the case where the movement path of the unmanned vehicle is a straight line, the unmanned vehicle moves the direction of the target position to the center of the image. The predictive means is an unmanned moving body which is set at a position and which calculates the predicted change as a change in a manner in which a change in the shape formed by the plurality of feature points maintains a similar shape.

According to the configuration of the third invention, when the movement path of the unmanned moving body is a straight line, the direction of the target position is set to the center position of the image, and the predicting means forms a shape composed of a plurality of feature points. Calculates the predicted change that expands while maintaining a similar shape. Then, for example, when the similar shape shifts to the left or right in the actual image, it is an unexpected change in position, and therefore the unmanned air vehicle changes the moving direction so as to cancel the change.

A fourth aspect of the present invention is the configuration of any one of the first to third aspects of the invention, wherein the unmanned vehicle is an unmanned vehicle traveling on the ground, and the change determining means includes the actual change and the prediction. When the difference from the change is only the difference in the vertical direction, the unmanned moving body determines that the actual change and the predicted change are not different.

According to the configuration of the fourth invention, the unmanned vehicle traveling on the uneven ground can maintain the horizontal traveling direction even if the traveling direction changes in the vertical direction.

A fifth invention is the configuration of any one of the first invention to the third invention, wherein the unmanned vehicle is an unmanned aerial vehicle, and the moving direction changing means is a horizontal direction of the unmanned aerial vehicle and An unmanned vehicle configured to change the moving direction in the vertical direction.

According to the configuration of the fifth aspect, the unmanned moving body can change the moving direction in the horizontal direction and the vertical direction by changing the positions of the plurality of feature points in the image.

A sixth invention is an autonomously movable unmanned moving body, capturing an image of a direction of a target position, which is a position where the unmanned moving body moves and reaches, and an image acquiring step of acquiring an image, in the image, A characteristic point extraction step of extracting a plurality of characteristic points; and a characteristic point showing a movement different from the overall tendency of movement of the plurality of characteristic points among the plurality of characteristic points extracted in the characteristic point extraction step, An exclusion step of eliminating, a relative position defining step of defining the direction of the target position in the image in a relative relationship with the positions of the plurality of characteristic points, and the unmanned moving body moving toward the target position on a planned route. In the condition of moving the plurality of feature points, a prediction step of calculating a change in the position of the plurality of feature points in the image as a predicted change, and a plurality of the plurality of feature points in the image during the movement of the unmanned moving body toward the target position. Between the change determination step of determining whether the actual change, which is the actual change of the position of the feature point, is different from the predicted change, and the change determination step, between the actual change and the predicted change. When it is determined that there is a difference, the moving direction changing step of changing the moving direction of the unmanned moving body so as to cancel the difference is a control method of the unmanned moving body.

A seventh aspect of the invention is an image acquisition unit that captures an image of a computer that controls an autonomously movable unmanned moving body in a direction of a target position, which is a position where the unmanned moving body moves and reaches, and the image. Among the plurality of characteristic points extracted by the characteristic point extracting means for extracting a plurality of characteristic points, the characteristic indicating a movement different from the overall tendency of the movement of the plurality of characteristic points Exclusion means for excluding points, relative position defining means for defining the direction of the target position in the image in relation to the positions of the plurality of characteristic points, and the unmanned vehicle moving toward the target position on a planned route. In a condition of moving the plurality of feature points, a prediction unit that calculates a change in position of the plurality of feature points in the image as a predicted change, and a plurality of the plurality of feature points in the image while the unmanned moving body moves toward the target position. An actual change, which is an actual change in the position of the feature point, is a change determination unit that determines whether or not the predicted change is the same, and the change determination unit determines whether the actual change is between the predicted change and the actual change. When it is determined that there is a difference between the two, it is an unmanned movement program for functioning as movement direction changing means for changing the movement direction of the unmanned moving body so as to cancel the difference.

According to the present invention, it is possible to quickly control the moving direction without arranging markers in advance and using positioning radio waves from navigation satellites.

It is a schematic diagram showing an operation of the unmanned mobile object concerning a first embodiment of the present invention. It is a schematic diagram showing an unmanned mobile. It is a schematic diagram showing the functional composition of an unmanned mobile. It is a conceptual diagram which shows a feature point. It is a conceptual diagram which shows a feature point. It is a conceptual diagram which shows the change of the position of the feature point in an image accompanying the movement of an unmanned mobile body. It is a conceptual diagram which shows the change of the position of the feature point in an image accompanying the movement of an unmanned mobile body. It is a figure for demonstrating the problem when there is one characteristic point. It is a clear diagram for explaining an advantage when there are a plurality of feature points. It is a flowchart which shows operation|movement of an unmanned mobile body. It is a conceptual diagram which shows the change of the position of the feature point in an image accompanying the movement of the unmanned mobile body which concerns on 2nd embodiment of this invention. It is the schematic which shows the effect|action of the unmanned air vehicle which concerns on 3rd embodiment of this invention. It is a schematic diagram showing an unmanned aerial vehicle. It is a schematic diagram showing the functional composition of an unmanned aerial vehicle. It is a flowchart which shows operation|movement of an unmanned air vehicle.

Hereinafter, modes for carrying out the present invention (hereinafter, embodiments) will be described in detail. In the following description, the same components are designated by the same reference numerals, and the description thereof will be omitted or simplified. It should be noted that description of configurations that can be appropriately performed by those skilled in the art will be omitted, and only the basic configuration of the present invention will be described.

<First embodiment>
The unmanned vehicle 1 (hereinafter, referred to as "unmanned vehicle 1") illustrated in FIG. 1 is an example of an unmanned vehicle that can autonomously move along a predetermined route. The unmanned aerial vehicle 1 is also an example of an unmanned vehicle traveling on the ground. The unmanned aerial vehicle 1 is designed to operate in response to an instruction by a radio signal from a control device (also called “propo”) that controls the unmanned aerial vehicle 1.

It is assumed that the mission of the unmanned aerial vehicle 1 is to automatically draw a straight line in the soccer stadium 100 (hereinafter referred to as "stadium 100"). Specifically, a line connecting the positions P1 and P2, a line connecting the positions P2 and P3, a line connecting the positions P3 and P4, and a line connecting the positions P4 and P1 are drawn as lines showing the outline of the soccer field. The position P1 and the like are corners of the soccer field. It is assumed that the coordinates (latitude, longitude and altitude) of the positions P1 to P4 are known.

Outside the stadium 100, there are various structures and natural objects such as buildings 102A to 102C, elevated expressways 104, and trees 106A to 106C. When these structures and natural objects are photographed and an image is acquired, a large number of feature points can be extracted from the image. Here, a characteristic point is a characteristic point such as a point or a corner of an object compared with the surroundings, and is a portion that is easy to distinguish from other portions because the brightness is changed in the image. The feature point is specified (extracted) by a well-known algorithm after the image is captured. This algorithm is called, for example, SIFT (Scale lnvariant Feature Transform), and is a technique for finding invariant feature quantity in each pixel in an image in rotation, scale, and change in illumination environment.

For example, the positions SP1, SP2, and SP3 on the contours of the trees 106C and 106B, the side portion SP4 of the highway 104, and the corner portion S5 of the building 102C are examples of feature points. The unmanned aerial vehicle 1 uses the plurality of feature points to control the moving direction.

As shown in FIG. 2, the drone 1 has a vehicle body 10. Tires 12A to 12D are arranged in the vehicle body 10. A motor is arranged in the vehicle body 10, and the unmanned aerial vehicle 1 can run by rotating the tires 12C and 12D of the rear wheels by the rotation of the motor. The traveling direction of the unmanned aerial vehicle 1 can be changed by rotating the front tires 12A and 12B.

The vehicle body 10 includes a housing 14 in which a computer that controls each unit of the unmanned aerial vehicle 1, an autonomous traveling device, a wireless communication device, a positioning device using GPS (Global Positioning System), an inertial sensor, an atmospheric pressure sensor, and the like are stored. It is arranged. Further, the vehicle body 10 includes a battery 16 for supplying electric power to each part of the unmanned aerial vehicle 1, a camera 18, an antenna 20 for wireless communication, and an antenna for receiving positioning radio waves from a navigation satellite. 22 are arranged. Although the direction of the optical axis of the camera 18 can be changed, while the unmanned aerial vehicle 1 is traveling, it is fixed to the traveling direction when the unmanned aerial vehicle 1 goes straight. That is, the traveling direction of the drone 1 and the direction of the optical axis of the camera 18 match.

In the vehicle body 10, a discharge device 24 for discharging the powder for drawing a line is arranged in front of the front wheels on the side where the tires 12A and 12B are located.

FIG. 3 is a diagram showing a functional configuration of the unmanned aerial vehicle 1. The unmanned aerial vehicle 1 includes a CPU (Central Processing Unit) 50, a storage unit 52, a wireless communication unit 54, a satellite positioning unit 56, an inertial sensor unit 58, a drive control unit 60, an image processing unit 62, and a power supply unit 64.

The unmanned aerial vehicle 1 can communicate with the outside through the wireless communication unit 54. The wireless communication unit 54 of the unmanned aerial vehicle 1 receives an instruction such as starting from a control device (propo) 70 operated by a user.

The unmanned aerial vehicle 1 can measure the position of the unmanned aerial vehicle 1 itself by the satellite positioning unit 56. The satellite positioning unit 56 basically receives positioning radio waves from four or more navigation satellites and measures the position of the unmanned aerial vehicle 1. The satellite positioning unit 56 is an example of a positioning unit that measures the current position. The inertial sensor unit 58 detects the movement of the unmanned aerial vehicle 1 by using, for example, an acceleration sensor and a gyro sensor. The position information of the unmanned aerial vehicle 1 itself is used not only for determining the moving route of the unmanned aerial vehicle 1 and for autonomous movement, but also for associating the image data captured by the camera 18 with the coordinates (position).

The drone 1 controls the rotation of the motor connected to the rear tires 12C and 12D and the rotation of the front tires 12A and 12C by the drive control unit 60 to control the traveling speed and the traveling direction. ing.

The drone 1 operates the camera 18 (see FIG. 2) by the image processing unit 62 to acquire an external image, and processes the acquired shooting data.

The power supply unit 64 is, for example, a replaceable rechargeable battery, and supplies power to each unit of the drone 1.

The storage unit 52 stores the following programs in addition to various data and programs necessary for autonomous movement such as data indicating a movement plan for autonomous movement from a starting point to a target position.

The storage unit 102 stores an image capturing program, a feature point extraction program, an exclusion program, a relative position defining program, a prediction program, a change determination program, and a moving direction changing program. The CPU 50 and the image capturing program are an example of an image acquisition unit. The CPU 50 and the feature point extraction program are an example of feature point extraction means. The CPU 50 and the exclusion program are examples of exclusion means. The CPU 50 and the relative position defining program are an example of a relative position defining unit. The CPU 50 and the prediction program are examples of prediction means. The CPU 50 and the change determination program are examples of change determining means. The CPU 50 and the moving direction changing program are examples of moving direction changing means.

The unmanned aerial vehicle 1 captures an image by capturing an image of a direction of a target position, which is a position where the unmanned aerial vehicle 1 moves and reaches, by an image capturing program. As described above, the direction of the optical axis of the camera 18 is fixed in the straight traveling direction. For example, as shown in FIG. 4A, the drone 1 captures an image at the position P1 in the direction of the target position P2 before starting traveling. It is assumed that the image includes feature points SP1 to SP5. The unmanned aerial vehicle 1 continuously captures images, captures multiple images in the direction of the target position P2, and acquires multiple images with different capture timings.

The unmanned aerial vehicle 1 extracts a plurality of feature points in an image (hereinafter referred to as “camera image”) acquired by the camera 18 by using the feature point extraction program. FIG. 4B is a conceptual diagram showing the acquired image D1. The unmanned aerial vehicle 1 extracts the feature points SP1 to SP5 in the image D1. The position P2, which is the target position, may be a feature point, but may not be a feature point. The larger the number of feature points to be extracted, the better, and it is preferably 30 or more.

The unmanned aerial vehicle 1 uses the plurality of feature points as markers for controlling the moving direction, as described above. When the movement of the feature points shown in the plurality of images continuously captured by the camera 18 is not the movement of the drone 1 but the movement of the feature points themselves, the movement direction control is performed. It is difficult to use as a marker of. For example, it is difficult to use a characteristic point of a walking person or a running vehicle as a marker for controlling the moving direction. A characteristic point of a stationary object is desirable as a marker for controlling the moving direction. In this regard, the unmanned aerial vehicle 1 eliminates a feature point showing a movement different from the overall tendency of the movement of the plurality of feature points from the plurality of feature points extracted by the feature point extraction program by the elimination program. That is, the unmanned aerial vehicle 1 determines the overall tendency of the movements of the plurality of characteristic points based on the plurality of characteristic points, and excludes the characteristic points exhibiting a movement different from the overall tendency. As a result, feature points of moving objects such as vehicles running on the highway 104 are eliminated, and only feature points of inanimate objects such as the buildings 102A to 102C, the elevated highway 104, and the trees 106A to 106C are used. can do. Generally, there are more non-animals than moving objects in the range from which images are to be acquired. Since the majority of the image is occupied by inanimate animals, the overall trend of movement of multiple feature points in the image is formed by the inanimate feature points. Then, the movement of the feature points of the inanimate object according to the movement mode of the drone 1 matches the overall tendency. For example, when the unmanned aerial vehicle 1 is stopped, the tendency of the overall movement shows a zero velocity vector, and the movement of each feature point of the inanimate animal also shows a zero velocity vector. On the other hand, the movement of each feature point of the moving object is not a vector with zero velocity, and is different from the overall tendency. This is the same when the drone 1 is moving forward, for example, the feature points for the inanimate object on the right side with respect to the traveling direction show the general tendency to move to the right side, and The feature points for inanimate animals show an overall tendency to move to the left, but the movement of feature points for moving objects is different from the overall tendency. The unmanned aerial vehicle 1 eliminates feature points that do not match the overall tendency. Since a plurality of feature points are extracted by the feature point extraction program, a plurality of feature points that can be used as markers remain even if some feature points are excluded by the exclusion program.

The unmanned aerial vehicle 1 defines the direction of the target position in the image in the relative relationship with the positions of the plurality of feature points by the relative position defining program. For example, in the image D1, the direction of the target position P2 is defined in relation to the characteristic points SP1 to SP5. When the movement path from the position P1 to the target position P2 is a straight line, the drone 1 adjusts its posture and direction so that the target position P2 is at the center of the image D1.

After reaching the position P2 from the position P1, the unmanned aerial vehicle 1 displays the image at the position P2 at the position P3, which is the next target position, in the same manner as described above, as shown in FIGS. 5A and 5B. Is obtained, a plurality of characteristic points in the image are extracted, and the target position P3 is defined in the relative relationship with the plurality of characteristic points. After reaching the position P3, similar processing is performed on the position P4 which is the next target position, and after reaching the position P4, similar processing is performed on the position P1 which is the next target position.

The unmanned aerial vehicle 1 acquires an image by the above-described image capturing program not only before starting traveling but also during traveling (moving). Then, the feature point extraction program is configured to extract a plurality of feature points in the image. Furthermore, the exclusion program eliminates feature points that do not match the overall tendency.

The unmanned aerial vehicle 1 calculates a change in the positions of a plurality of feature points as a predicted change under the condition that the unmanned aerial vehicle 1 moves toward the target position on the planned route by the prediction program. When the planned route of the unmanned aerial vehicle 1 is a straight line, a change in the positions of the plurality of feature points is calculated as a predicted change, which is a change in which the shape formed by the plurality of feature points expands while maintaining a similar shape. .. The predicted change is, for example, a change mode in which the polygon formed by the feature points SP1 to SP5 shown in FIG. 4B changes while maintaining a similar shape.

The prediction program will be specifically described with reference to FIGS. 6 and 7. 6 and 7, only two feature points SP2 and SP4 are used as feature points for convenience of description, but it is actually preferable to use a larger number of feature points.

As shown in FIG. 6, positions midway between the current position P1 and the target position P2 are positions PM2, PM4, and PM6. If the unmanned aerial vehicle 1 moves on a straight line connecting the position P1 and the position P2 toward the position P2, a line segment connecting the characteristic points SP2 and SP4 is obtained as the unmanned aerial vehicle 1 approaches the position P2. Since the center position should be long toward both sides while maintaining the center position, the unmanned aerial vehicle 1 calculates such a change as a predicted change as a change in the positions of a plurality of feature points.

For example, if the image acquired at position P1 is image D10, the image acquired at position PM2 is image D12, the image acquired at position PM4 is image D14, and the image acquired at position PM6 is image D16, the predicted change is unmanned. As the aircraft 1 approaches the position P2, the line segment connecting the characteristic points SP2 and SP4 gradually changes toward both sides while maintaining its center position.

According to the change determination program, the unmanned aerial vehicle 1 determines whether the actual change, which is the actual change in the positions of the plurality of feature points in the image, is different from the predicted change while the unmanned aerial vehicle 1 moves toward the target position. To judge. The unmanned aerial vehicle 1 determines that there is no difference between the actual change and the predicted change when the only difference between the actual change and the predicted change is the difference in the vertical direction. Accordingly, when the unmanned aerial vehicle 1 travels on an uneven ground, even if the optical axis of the camera 18 shifts in the vertical direction and a difference in the vertical direction occurs between the actual change and the predicted change, the unmanned aerial vehicle 1 Can move toward the target position.

When the unmanned aerial vehicle 1 acquires a front image at the position PM2 shown in FIG. 6, for example, the line segment connecting the feature points SP2 and SP4 in the image has its center compared to the line segment at the position P1. If the length is increasing toward both sides while maintaining, it is as predicted by the above-mentioned prediction program. Therefore, the drone 1 determines that the actual change is not different from the predicted change. In this case, the unmanned aerial vehicle 1 moves in a straight line connecting the position P1 and the position P2 toward the position P2.

On the other hand, as shown in FIG. 7, at the position PM3, when the direction of the unmanned aerial vehicle 1 changes to the direction of the arrow R1 and faces the characteristic point SP4 side, the characteristic point in the image D13 acquired at the position PM3. The center position of the line segment connecting SP2 and SP4 is not maintained and is moving to the right, which is different from the change predicted by the above prediction program. Therefore, the drone 1 determines that the actual change is different from the predicted change. Further, at the position PM5, when the direction of the drone 1 changes to the direction of the arrow R2 and faces the feature point SP2 side, the central position of the line segment connecting the feature points SP2 and SP4 in the image D15 acquired at the position PM5. Is not maintained and is moving to the left, which is different from the change predicted by the above prediction program. Therefore, the drone 1 determines that the actual change is different from the predicted change.

When the unmanned aerial vehicle 1 determines that there is a difference between the actual change and the predicted change by the change determination program, the moving direction change program changes the moving direction of the unmanned aerial vehicle 1 so as to cancel the difference. Is configured to.

For example, at the position PM3 in FIG. 7, the moving direction of the drone 1 changes in the arrow R1 direction, and as a result, a difference between the actual change and the predicted change occurs. Therefore, in order to cancel the change in the arrow R1 direction, the arrow The moving direction is changed to the arrow R2 direction which is the opposite direction to R1. Similarly, at the position PM5 in FIG. 7, the moving direction of the drone 1 changes in the direction of arrow R2, and as a result, a difference between the actual change and the predicted change occurs. Therefore, in order to cancel the change in the direction of arrow R2, The movement direction is changed to the arrow R1 direction which is the opposite direction to the arrow R2.

Here, the technical significance of using a plurality of feature points will be described. For example, when there is only one feature point, the position in the image may not change even if the direction of the drone 1 changes, or it may be extremely small. As shown in FIG. 8A, it is assumed that there is a feature point SPa and the initial position in the image Dx1 acquired by the camera 18 is the position Da1. Depending on how the direction of the unmanned aerial vehicle 1 changes, as shown in FIG. 8B, even if the direction of the unmanned aerial vehicle 1 changes, the position Da1 is maintained as the position of the feature point SPa in the image Dx2.

On the other hand, when using a plurality of feature points, such as two feature points or three feature points, the position of at least one feature point always changes if the moving direction of the drone 1 changes. As shown in FIG. 9A, it is assumed that there are characteristic points SPa and SPb and the initial positions in the image Dx1 acquired by the camera 18 are positions Da1 and Db1, respectively. Depending on how the direction of the drone 1 changes, as shown in FIG. 9B, when the direction of the drone 1 changes, the position of the feature point SPa in the image Dx2 may maintain the position Da1. The position of the indication point SPb does not maintain the position Db1 but changes to the position Db2. Further, when using a plurality of feature points, the amount of difference (the number of different pixels) is larger when a difference occurs between the actual change and the predicted change than when using one feature point, Differences are easy to detect. That is, by using a plurality of feature points, the reliability of the determination of the direction change of the drone 1 is improved.

In addition, when there is only one feature point, the feature point may not be displayed in the image due to some natural phenomenon or an artificial phenomenon. Therefore, by using the plurality of feature points, the reliability of the determination of the direction change of the drone 1 is also improved. In addition, it is possible to judge the overall tendency of the movement of the feature points by using the plurality of feature points, eliminate the feature points of the moving object, and leave a plurality of feature points that can be used as markers. ..

Further, while the unmanned aerial vehicle 1 is moving, the time required for processing (hereinafter, referred to as “judgment processing”) from external image acquisition to determination of the identity of the actual change and the predicted change is the positioning radio wave from the navigation satellite. Since it is much shorter than the time required for positioning using, the direction of movement can be quickly controlled. For example, if the time required for positioning using the positioning radio wave is 1/10 second (s), the time required for the determination process is 1/1000 second to 1/100 second.

The operation of the unmanned aerial vehicle 1 will be described below with reference to the flowchart of FIG. It is assumed that the unmanned aerial vehicle 1 travels from the current position P1 to the target position P2 with the planned route as a straight route. When the unmanned aerial vehicle 1 starts its operation, it adjusts the direction (posture) so that the target position P2 is located at the center of the image at the current position P1 before starting traveling (step ST1 in FIG. 10). Subsequently, the unmanned aerial vehicle 1 acquires an image in the direction of the target position P2 (step ST2), and extracts a plurality of feature points around the direction of the position P2 (step ST3). In step ST2, the unmanned aerial vehicle 1 acquires a plurality of images, and in step ST3, excludes feature points that show a movement different from the overall tendency of the movement of the plurality of feature points. Step ST3 is an example of a feature point extraction step, and is also an example of an exclusion step. Subsequently, the unmanned aerial vehicle 1 predicts changes in the positions of a plurality of feature points in the image acquired during the movement of the planned route (step ST4).

Subsequently, the unmanned aerial vehicle 1 starts traveling toward the target position P2 and acquires an external image (step ST5). The unmanned aerial vehicle 1 determines whether or not the change (actual change) in the positions of the plurality of feature points in the image is different from the predicted change while traveling (step ST6). Is determined to be the target position P2 (step ST7), and if the target position P2, the process is stopped. Whether or not the current position is the target position P2 is determined by the positioning by the satellite positioning unit 56 (see FIG. 3) depending on whether or not the coordinates of the current position and the coordinates of the target position P2 match.

When the unmanned aerial vehicle 1 determines in step ST6 that the actual change is different from the predicted change, it changes the moving direction so as to cancel the difference (step ST8).

Note that, unlike the present embodiment, the ejection direction of the ejection device 24 may be configured to be changeable, and both the posture control of the unmanned aerial vehicle 1 and the ejection direction control may be used. For example, when the actual change and the predicted change are different, the unmanned aerial vehicle 1 may be configured to correct the attitude of the unmanned aerial vehicle 1 while changing the ejection direction so as to cancel the difference. This makes it possible to draw a line that is closer to the planned line.

<Second embodiment>
Next, a second embodiment will be described with reference to FIG. It should be noted that description of items that are common to the first embodiment will be omitted, and the description will focus on the parts that differ from the first embodiment.

The configuration of the unmanned aerial vehicle 1 of the second embodiment is the same as that of the first embodiment, but in the second embodiment, the planned route of the unmanned aerial vehicle 1 is a curve.

As shown in FIG. 11, it is assumed that the unmanned aerial vehicle 1 moves from the current position P1 to the target position P5 along a curved path. The unmanned aerial vehicle 1 has an optical axis of the camera 18 according to a planned moving route so that the direction of the target position P5 is projected on the image acquired by the camera 18 from the current position P1 to the target position P5. Set conditions such as the direction of. Therefore, for example, the drone defines a curved path connecting the current position P1 and the target position P5 in the map data, and if the width is the width W1, the center of the image is located at the center position of the width W1. The above-mentioned conditions of the camera 18 are set so that is positioned. In the example of FIG. 11, at the starting position P1, the target position P5 is located near the right end of the image D10. Then, in the vicinity of the target position P5, it comes to be located at the center of the image D13.

As the unmanned aerial vehicle 1 progresses from the current position P1 to the positions P11, P12, P13, the directions of the unmanned aerial vehicle 1 and the optical axis of the camera 18 in the curved path are directions V10, V11, V12, V13. Changes to. If the unmanned aerial vehicle 1 travels along a curved path that turns to the right, the feature points SP1 and SP2 should gradually move to the left as the unmanned aerial vehicle 1 travels along a planned route.

The unmanned aerial vehicle 1 predicts the above-described change in the positions of the characteristic points SP1 and SP2 in the image during the movement of the curved path by the prediction program. For example, in the map data, by defining a curved path connecting the current position P1 and the target position P5, and calculating the tangential direction in the curved path, the characteristic points SP1 and SP2 in the image D10 at the current position P1 are curved. The change in the image acquired at each position in the route, that is, the predicted change can be calculated. Since this calculation can be performed by using a well-known technique, detailed description will be omitted.

The drone 1 uses the change determination program to determine whether the actual change is different from the predicted change while the drone 1 is moving toward the target position. , The moving direction of the drone 1 is changed so as to cancel the difference.

<Third embodiment>
Next, a third embodiment will be described with reference to FIGS. It should be noted that description of items common to the first embodiment and the second embodiment will be omitted, and the description will focus on parts different from the first embodiment and the second embodiment.

As shown in FIG. 12, the unmanned aerial vehicle 200 of the third embodiment is an unmanned aerial vehicle. It is assumed that the unmanned aerial vehicle 200 is to fly over the field 120 so as to connect the positions P101 to P107 as shown by the arrows and spray fertilizer. Positions P101 to P107 are positions above the field 120.

For example, the unmanned aerial vehicle 200 acquires an image in the direction of the next target position P102 at the starting position P101 and extracts a plurality of feature points SP101, SP102, SP103, and SP104. It is assumed that the characteristic points SP101 and the like have variations in the horizontal direction and the vertical direction with respect to the sky above the target position P102.

If the unmanned aerial vehicle 200 flies from the sky above the position SP101 to the sky above the position SP102 in a straight line at a constant altitude, if the actual change in the image of the feature points that vary in the horizontal direction differs from the predicted change. , The moving direction in the horizontal direction is changed so as to cancel the difference. If the actual change in the image of the characteristic points that are scattered in the vertical direction is different from the predicted change, the moving direction in the vertical direction is changed so as to cancel the difference.

As shown in FIG. 13, the drone 200 has a housing 202. A computer that controls each unit of the unmanned aerial vehicle 200, an autonomous flight device, a wireless communication device, a positioning device that uses positioning radio waves from a navigation satellite system such as GPS, and a battery are arranged in the housing 202. An inertial sensor and an atmospheric pressure sensor are arranged in the autonomous flight device. A camera 214 is arranged in the housing 202 via a fixing device 212.

The fixing device 212 is a triaxial fixing device (so-called gimbal) that can minimize the blurring of the image captured by the camera 214 and control the optical axis of the camera 214 in any direction. However, in this embodiment of the unmanned aerial vehicle 200, the direction of the optical axis of the camera 214 is fixed to the traveling direction.

A round bar-shaped arm 204 is connected to the housing 202. A motor 206 is connected to each arm 204, and a propeller 208 is connected to each motor 206. Each motor 206 is a DC motor (brushless DC motor). Each motor 206 is independently controlled by an autonomous flight device in the housing 202, so that the unmanned aerial vehicle 200 can be freely moved in the vertical and horizontal directions, stopped in the air (hovering), and attitude control. Has become.

A protective frame 210 is connected to the arm 204 to prevent the propeller 208 from directly contacting an external object. The arm 204 and the protective frame 210 are made of, for example, carbon fiber reinforced plastic, and are lightweight while maintaining strength.

FIG. 14 is a diagram showing a functional configuration of the unmanned aerial vehicle 200. The unmanned aerial vehicle 200 includes a CPU 250, a storage unit 252, a wireless communication unit 254, a satellite positioning unit 256, an inertial sensor unit 258, a drive control unit 260, an image processing unit 262, and a power supply unit 264.

The unmanned aerial vehicle 200 can communicate with the outside through the wireless communication unit 254. The unmanned aerial vehicle 200 receives, via the wireless communication unit 254, an instruction such as starting from a control device (propo) 270 operated by the user.

The unmanned aerial vehicle 200 can measure the position of the unmanned aerial vehicle 200 itself by the satellite positioning unit 256. The satellite positioning unit 256 basically receives positioning radio waves from four or more navigation satellites and measures the position of the unmanned aerial vehicle 200. The satellite positioning unit 256 is an example of a positioning unit that measures the current position. The inertial sensor unit 258 detects the movement of the unmanned aerial vehicle 200 by using, for example, an acceleration sensor and a gyro sensor. The position information of the unmanned aerial vehicle 200 itself is used not only for determining the movement route of the unmanned aerial vehicle 200 and for autonomous movement, but also for associating the image data captured by the camera 214 with the coordinates (position).

The drive control unit 260 controls the unmanned aerial vehicle 200 to control the rotation of the motor 206 to control the flight speed and the flight direction.

The image processing unit 262 causes the unmanned aerial vehicle 200 to operate the camera 214 to acquire an external image and process the acquired shooting data.

The power supply unit 264 is, for example, a replaceable rechargeable battery and supplies power to each unit of the unmanned aerial vehicle 200.

The storage unit 252 stores various data and programs necessary for autonomous movement, such as data indicating a movement plan for autonomous movement from a starting point to a target position, as well as the image capturing program as in the first and second embodiments. , A feature point extraction program, an exclusion program, a relative position definition program, a prediction program, a change determination program, and a moving direction change program are stored.

In the unmanned aerial vehicle 200, the traveling direction is controlled not only in the horizontal direction but also in the vertical direction. That is, the unmanned aerial vehicle 200 determines by the change determination means that there is no difference between the actual change and the predicted change not only in the horizontal direction but also in the vertical direction.

The operation of unmanned aerial vehicle 200 will be described below with reference to the flowchart of FIG. It is assumed that the unmanned aerial vehicle 200 heads from the current position P101 to the target position P102 by a straight path.

When the unmanned aerial vehicle 200 starts its operation, it reaches the current position P101, adjusts its posture so as to include the target position P102 in the image captured by the camera 214 (step ST101 in FIG. 15), and acquires an image in the direction of the target position 102 ( In step ST102), a plurality of feature points are extracted (step ST103). In step ST102, the unmanned aerial vehicle 200 acquires a plurality of images, and in step ST103, excludes feature points exhibiting a motion different from the overall tendency of the motion of the feature points. Subsequently, the unmanned aerial vehicle 200 predicts changes in the positions of a plurality of feature points in the image acquired while the planned route is moving, and calculates predicted changes (step ST104).

Subsequently, the unmanned aerial vehicle 200 starts flying toward the target position P102 (step ST105). The unmanned aerial vehicle 200 determines whether or not the actual changes in the positions of the plurality of feature points in the image are different from the predicted changes during the flight (step ST106), and if not, the current position is the target position P102. It is determined whether or not there is, and if the target position is P102, preparation is made for moving to the next target position P103. Whether or not the current position is the target position P102 is determined by whether or not the coordinates of the current position and the coordinates of the target position P102 match by the positioning by the satellite positioning unit 256.

When the unmanned aerial vehicle 200 determines in step ST106 that the actual change is different from the predicted change, it changes the moving direction so as to cancel the change (step ST108).

It should be noted that the present invention is not limited to the above-described embodiments, and modifications, improvements, etc. within the scope of achieving the object of the present invention are included in the present invention.

1 Unmanned Aerial Vehicle 10 Vehicle Main Body 12A-12D Tire 14 Housing 16 Battery 18 Camera 20 Antenna 22 Antenna 24 Discharge Device 200 Unmanned Aerial Vehicle 202 Housing 204 Arm 206 Motor 208 Propeller 210 Protective Frame 212 Fixing Device 214 Camera

Claims (7)

An unmanned vehicle capable of autonomous movement,
An image acquisition unit that captures an image by capturing a direction of a target position, which is a position where the unmanned vehicle moves and reaches.
In the image, feature point extraction means for extracting a plurality of feature points,
Relative position defining means for defining the direction of the target position in the image in a relative relationship with the positions of the plurality of characteristic points,
Prediction means for calculating a change in position of the plurality of feature points in the image as a predicted change under the condition that the unmanned moving body moves toward the target position on a planned route,
A change that determines whether the actual change, which is an actual change in the positions of the plurality of feature points in the image, is different from the predicted change while the unmanned moving body moves toward the target position. Judgment means,
When it is determined by the change determination means that there is a difference between the actual change and the predicted change, a moving direction changing means that changes the moving direction of the unmanned moving body so as to cancel the difference,
An unmanned vehicle having Of the plurality of characteristic points extracted by the characteristic point extracting means, there is an excluding means for excluding the characteristic points exhibiting a movement different from the overall tendency of the movement of the plurality of characteristic points.
The unmanned vehicle according to claim 1. When the movement path of the unmanned moving body is a straight line, the unmanned moving body sets the direction of the target position to the center position of the image,
The unmanned moving body according to claim 1 or 2, wherein the prediction unit calculates the predicted change as a change in a manner in which a change in a shape formed of the plurality of feature points maintains a similar shape. The unmanned vehicle is an unmanned vehicle traveling on the ground,
The change determination means determines that the actual change and the predicted change are not different when the difference between the actual change and the predicted change is only a difference in the vertical direction,
The unmanned vehicle according to any one of claims 1 to 3. The unmanned vehicle is an unmanned aerial vehicle,
The moving direction changing means is configured to change a moving direction of the unmanned aerial vehicle in a horizontal direction and a vertical direction,
The unmanned vehicle according to any one of claims 1 to 3. An unmanned vehicle that can move autonomously
An image acquisition step of capturing an image by capturing a direction of a target position, which is a position where the unmanned vehicle moves and reaches,
A feature point extracting step of extracting a plurality of feature points in the image;
An exclusion step of excluding the feature points that show a movement different from the overall tendency of the movement of the plurality of feature points among the plurality of the feature points extracted in the feature point extraction step,
A relative position defining step that defines the direction of the target position in the image in a relative relationship with the positions of the plurality of characteristic points;
A predicting step of calculating a change in the position of the plurality of feature points in the image as a predictive change under the condition that the unmanned moving body moves toward the target position in a planned route,
A change that determines whether the actual change, which is an actual change in the positions of the plurality of feature points in the image, is different from the predicted change while the unmanned moving body moves toward the target position. Decision step,
In the change determining step, when it is determined that the actual change is different from the predicted change, a moving direction changing step of changing a moving direction of the unmanned moving body so as to cancel the difference,
A method for controlling an unmanned moving body for carrying out. A computer that controls an unmanned vehicle that can move autonomously,
An image acquisition unit that captures an image by capturing an image of a direction of a target position, which is a position where the unmanned vehicle moves and reaches.
Feature point extracting means for extracting a plurality of feature points in the image,
Exclusion means for excluding the characteristic points showing a movement different from the overall tendency of the movement of the plurality of characteristic points among the plurality of characteristic points extracted by the characteristic point extraction means,
Relative position defining means for defining the direction of the target position in the image in a relative relationship with the positions of the plurality of characteristic points,
Prediction means for calculating a change in the position of the plurality of feature points in the image as a predicted change under the condition that the unmanned moving body moves toward the target position on a planned route,
While the unmanned moving body moves toward the target position, it is determined whether or not an actual change, which is an actual change in the positions of the plurality of feature points in the image, is the same as the predicted change. Change determination means, and
When the change determining unit determines that the actual change is different from the predicted change, a moving direction changing unit that changes the moving direction of the unmanned moving body so as to cancel the difference.
Unmanned traveling program to function as.