JP5947938B1 - Obstacle detection device and moving body equipped with the same - Google Patents

Abstract

An obstacle detection device that detects an obstacle on a running road with a reduced amount of calculation and a moving body including the obstacle detection device. A distance information detection unit that detects distance information to an object ahead of a vehicle, a parallax image generation unit that creates a parallax image in front of the vehicle, and a position of the vehicle from a start position So. The track position information storage unit 25 that stores track position information that identifies the track Tr on the parallax image, the position estimation unit 27 that estimates the position of the vehicle 1 on the track Tr, the estimated position and the track A parallax image extraction unit 66 for extracting a parallax image on the road Tr based on position information and an obstacle detection unit 31 ′ for detecting an obstacle based on the parallax image on the road are provided, and the road position information is distance information. And the obstacle detection device 56 specified by the coordinate information of the vehicle 1 in the horizontal direction. [Selection] Figure 12

Description

  The present invention relates to an obstacle detection device and a moving body including the same.

  Some golf carts that run on a predetermined course in a golf course have an autonomous running function. Autonomous traveling is performed by detecting a magnetic field generated from an electromagnetic induction wire. On the course of a golf course, there are work vehicles and people crossing the course as obstacles to the golf cart. When traveling autonomously without a driver, it is necessary to detect these obstacles. In view of this, some obstacle sensors using ultrasonic waves or the like are provided on the front surface of the golf cart, and an obstacle detection device that detects an obstacle in front of the golf cart is provided.

  Conventional obstacle detection devices detect an object that is not on the road as an obstacle, and as a result, the golf cart may be subjected to deceleration control. Therefore, in the technique of Patent Document 1, it is determined whether an obstacle exists on the road. As a result, stop control by an obstacle not on the runway is suppressed.

The technique of patent document 1 In patent document 1, the driving | running | working in a mobile body coordinate system is described from the coordinate position of the moving body obtained as a whole coordinate system, and the coordinate position data of each point along a running path obtained as a whole coordinate system. It describes that the coordinate position data of the road is acquired, the traveling road surface in the three-dimensional image in the moving body coordinate system is specified, and it is determined whether or not the detected obstacle is on the traveling road surface.

Japanese Patent Laid-Open No. 10-141954

  However, since the traveling road surface in the three-dimensional image is specified using the three-dimensional data, a calculation load is applied to the CPU of the obstacle detection device, and calculation time is required. If the calculation time increases, the moving body is traveling, so the distance from the obstacle becomes closer and the possibility of contact increases. Also, as a device for vehicles that pass just between trees like a golf course, if the estimated current position error is large, the trees around the runway are judged as obstacles in the runway and stopped. Will occur. Therefore, the current position must be able to be estimated with high accuracy. In Patent Document 1, three-dimensional position coordinates are used for the current position of the moving body. However, if the current position is to be estimated with high accuracy in three dimensions, expensive GPS, sensors, and the like are required. Required for the number of units.

  The present invention has been made in view of such circumstances, and an object of the present invention is to provide an obstacle detection device that detects an obstacle on a runway with a reduced amount of computation, and a moving body including the obstacle detection device. To do.

In order to achieve the above object, the present invention has the following configuration.
That is, according to a first aspect of the present invention, in an obstacle detection device mounted on a moving body traveling on a predetermined road, distance information from the obstacle detection device to an object ahead of the moving body is detected. A distance information detection unit that performs, a three-dimensional information generation unit that generates three-dimensional information in front of the mobile body based on the distance information, and track position information corresponding to the position of the mobile body from a predetermined position On the road that extracts 3D information on the road in the three-dimensional information based on the road position information, based on the road position information, the storage unit that stores the position, the position estimation unit that estimates the position of the moving body on the road and the three-dimensional information extracting unit, which is extracted, based on the three-dimensional information on the track and a obstacle detection unit for detecting an obstacle, the track position information in the horizontal direction of the moving body and the distance information Identified by coordinate information.

  According to this invention, the distance information detection unit detects distance information from the obstacle detection device to the object ahead of the moving body. The three-dimensional information generation unit generates three-dimensional information ahead of the moving object based on the detected distance information. The storage unit stores road position information that specifies a road on the three-dimensional information according to the position of the moving body from a predetermined position. The position estimation unit estimates the position of the moving body on the runway. The three-dimensional information extraction unit on the road extracts three-dimensional information on the road in the three-dimensional information based on the estimated position and the road position information. The obstacle detection unit detects the obstacle based on the three-dimensional information on the road.

  The track position information used to extract the three-dimensional information on the track is two-dimensional information of distance information and horizontal coordinate information of the moving body, and thus has been specified using a three-dimensional information amount. The amount of information handled can be reduced, and the calculation load can be reduced. Thereby, the calculation time can be shortened, it can be determined whether or not the obstacle is on the runway when the distance from the obstacle is secured, and the avoidance of contact with the obstacle can be improved. it can.

  Moreover, it is preferable to provide the person determination part which determines whether the said obstruction on a runway is a person. Or it is preferable to provide the person determination part which determines whether the said obstacle determined to be on a runway is a person. By narrowing down human detection to obstacles on the runway, processing time for human detection can be shortened. Moreover, since the person who does not have the possibility of contact such as being far away from the moving body is not detected, the moving body can travel smoothly.

  In addition, an extended region 3D information extraction unit that extracts three-dimensional information in an extended track region that is a predetermined region adjacent to the track based on the track position information is provided, and the obstacle detection unit includes the track It is preferable that an obstacle is also detected on the extension area, and the person determination unit determines whether or not the obstacle detected in the runway extension area is a person. Thereby, a detection rate can be improved with respect to the person who exists in the area | region of the width direction both ends of a runway which is easy to be confused with trees around a runway.

  Further, based on the runway position information, an extended area 3D information extraction unit that extracts 3D information in an extended runway area that is a predetermined area adjacent to the runway, and the 3D information in the runway extension area A person detection unit for detecting a person based on the above may be provided. The human detection processing time can be shortened by performing human detection within a wide range of a predetermined area on the specified road. Furthermore, since a person near the runway can be detected, a person trying to enter the runway can also be detected. Thereby, contact avoidance with a person can be improved. In addition, since the obstacle detection unit and the human detection unit are provided as separate bodies, it is difficult to judge the possibility of contact with obstacles on the road with high possibility of contact, and it is confused with trees around the road. It is possible to perform different processing on people who are in areas near both ends in the width direction of the easy track.

  According to a second aspect of the present invention, in an obstacle detection device mounted on a moving body traveling on a predetermined road, distance information from the obstacle detection device to an object ahead of the moving body is detected. A distance information detection unit that performs the three-dimensional information generation unit that generates three-dimensional information ahead of the moving body based on the distance information, and the three-dimensional information according to the position of the moving body from a predetermined position. A three-dimensional information according to the storage unit storing the road position information for specifying the road on the information, the position estimation unit for estimating the position of the moving body on the road, and the estimated position and the road position information. A road identification unit that identifies a road in the vehicle, an obstacle detection unit that detects an obstacle based on the three-dimensional information, and determines whether the detected obstacle is on the identified road A determination unit, and the runway Location information is identified by the horizontal coordinate information of the movable body and the distance information.

  According to the present invention, the distance information detection unit detects distance information from the obstacle detection device to an object ahead of the moving body. The three-dimensional information generation unit generates three-dimensional information ahead of the moving object based on the detected distance information. The storage unit stores road position information that specifies a road on the three-dimensional information according to the position of the moving body from a predetermined position. The position estimation unit estimates the position of the moving body on the runway. The runway specifying unit specifies the runway in the three-dimensional information based on the estimated position and the runway position information. The obstacle detection unit detects an obstacle based on the three-dimensional information. The determination unit determines whether or not the detected obstacle is on the identified road.

  The track position information for specifying the track on the three-dimensional information is two-dimensional information of the distance information and the horizontal coordinate information of the moving body. The amount of information handled can be reduced, and the calculation load can be reduced. Thereby, the calculation time can be shortened, it can be determined whether or not the obstacle is on the runway when the distance from the obstacle is secured, and the avoidance of contact with the obstacle can be improved. it can.

  In addition, an area specifying unit that specifies an extended runway area that is a predetermined area adjacent to the specified runway is provided, and the determination unit includes the extended runway in which the detected obstacle is specified. It is preferable to determine whether or not the obstacle is on the area, and the person determination unit determines whether or not the obstacle determined to be on the extended track area is a person. Thereby, a detection rate can be improved with respect to the person who exists in the area | region of the width direction both ends of a runway which is easy to be confused with trees around a runway.

  In addition, an area specifying unit that specifies an extended runway area that is a predetermined area adjacent to the specified runway, and a person detection unit that detects a person in the specified extended runway area are provided. Good. The human detection processing time can be shortened by performing human detection within a wide range of a predetermined area on the specified road. Furthermore, since a person near the runway can be detected, a person trying to enter the runway can also be detected. Thereby, contact avoidance with a person can be improved. In addition, since the obstacle detection unit and the human detection unit are provided as separate bodies, it is difficult to determine the possibility of contact with obstacles on the road with high possibility of contact, and the trees around the road Different processing can be performed on people who are in the vicinity of both ends in the width direction of the runway that are easily confused.

  Moreover, it is preferable that the said position estimation part estimates the position of the said mobile body on the said runway from the movement distance from a predetermined position. Since the moving body travels on a predetermined traveling path, the position of the moving body on the traveling path can be specified by the moving distance from the predetermined position. Since the position estimation unit estimates the position of the moving body on the runway based on the moving distance from a predetermined position, the position of the moving body can be estimated using only the vehicle coordinate system without using the overall coordinate system. Furthermore, since the movement distance is a one-dimensional data amount, the amount of data handled can be significantly reduced.

  The distance information detection unit may include a stereo camera, and the three-dimensional information may be a parallax image. By using a stereo camera as the distance information detection unit, a parallax image can be used as three-dimensional information. Thereby, determination of the runway and detection of an obstacle in a parallax image can be implemented appropriately. In addition, in a one-dimensional scanning laser radar, it is easy to detect the road surface as an obstacle due to the up and down of the road surface, but by using a stereo camera, it is possible to detect the obstacle more unlikely to be affected by road surface undulations. .

  The distance information detection unit may have a radar. By using a radar as the distance information detection unit, three-dimensional information can be acquired appropriately. Rather than using a stereo camera, distance measurement can be performed with high accuracy, particularly at a distance, and it is possible to appropriately perform determination of a runway and detection of an obstacle in three-dimensional information.

  Moreover, the mobile body which concerns on this invention is a mobile body provided with the said obstacle detection apparatus. Since the obstacle detection device is provided, it can be determined whether or not there is an obstacle on the running road with a reduced amount of calculation.

  Moreover, it is preferable that a mobile body is provided with the alarm device which issues a warning, when the said obstruction is detected on the said runway. By issuing a warning to pedestrians, animals, drivers of other moving objects, etc. existing on the runway, contact avoidance with these obstacles can be improved. In addition, a warning can be issued to alert the driver of his / her moving body.

  Moreover, it is preferable that a mobile body is provided with the speed control part which decelerates or stops the said mobile body when the said obstruction is detected on the said runway. Even when the obstacle present on the runway is not a person or an animal, or when the vehicle is traveling autonomously without a driver of a moving body, contact avoidance with the obstacle can be improved.

  In addition, the moving body is a speed control unit that decelerates or stops the moving body with respect to the detected obstacle on the running road, and an alarm device that issues a warning to the detected person on the extended running road area. It is preferable to provide. Thereby, when an obstacle including a person is on the runway, the moving body is decelerated or stopped by detecting this. Further, when there is a person in the extended runway area, a warning is issued from the alarm device by detecting this. By this warning, it is possible to prevent a person in the extended runway area from entering the runway, and as a result, unnecessary deceleration or stop of the moving body can be suppressed.

  ADVANTAGE OF THE INVENTION According to this invention, the obstruction detection apparatus which reduces the amount of calculations and detects the obstruction on a runway, and a mobile body provided with the same can be provided.

1 is a front view of a vehicle according to an embodiment. 1 is a block diagram illustrating a configuration of a vehicle according to a first embodiment. 1 is a block diagram illustrating a configuration of an obstacle detection device according to Embodiment 1. FIG. It is explanatory drawing which shows the relationship between the stereo camera which concerns on Example 1, and the direction of distance information. It is explanatory drawing which shows an example of the runway where the vehicle which concerns on an Example drive | works. It is explanatory drawing explaining specification of a runway in the parallax image which concerns on Example 1. FIG. It is explanatory drawing which shows the obstruction on a runway in the parallax image which concerns on Example 1. FIG. 6 is a flowchart illustrating a flow of obstacle detection according to the first embodiment. FIG. 6 is a block diagram illustrating a configuration of a vehicle according to a second embodiment. FIG. 6 is a block diagram illustrating a configuration of a vehicle according to a third embodiment. It is explanatory drawing which shows an extended runway area | region in the parallax image which concerns on Example 3. FIG. FIG. 10 is a block diagram illustrating a configuration of a vehicle according to a fourth embodiment. It is explanatory drawing which shows the obstruction on a runway in the parallax image which concerns on Example 4. FIG. It is explanatory drawing which shows an extended runway area | region in the parallax image which concerns on Example 4. FIG. 10 is a flowchart illustrating a flow of obstacle detection according to a fourth embodiment.

  Embodiment 1 of the present invention will be described below with reference to the drawings. As an embodiment of the moving body in the present invention, a golf cart that runs autonomously is given. The moving body in the present invention is not limited to a golf cart, and includes an automatic guided vehicle that runs in a factory or an orchard. Further, the moving body in the present invention is not limited to a four-wheeled vehicle, and may be a three-wheeled vehicle or a monorail type. In the following description, front and rear and left and right are based on the advancing direction of the moving body.

1. Schematic configuration of vehicle Referring to FIG. FIG. 1 is a front view of the vehicle 1 according to the embodiment. The vehicle 1 is a golf cart that autonomously travels within a golf course. The vehicle 1 can travel autonomously by being guided by an electromagnetic wave emitted from a guide wire embedded in the runway. A stereo camera 3 is provided in the center of the front surface of the vehicle 1.

  The vehicle 1 also includes a right front wheel 5 and a left front wheel 6 that are steered by the rotation of the handle 4. A reading unit 7 that reads a fixed point Fp (FIG. 5) embedded in advance on the runway Tr is provided in the lower portion of the vehicle 1. In addition to the start position So, a plurality of fixed points Fp are embedded on the runway Tr. The fixed point Fp is constituted by a combination of magnetic poles of a plurality of magnets. The reading unit 7 is a magnetic sensor that reads such magnetic field information from a fixed point Fp. The magnetic field information includes position information indicating a start position So or a specific position, and speed control information such as temporary stop and vehicle speed increase / decrease. Note that the fixed point Fp is not limited to a magnet, but may be an RF tag used in RFID (radio frequency identifier).

  The right front wheel 5 is provided with a rotation angle sensor 8 that detects the rotation angle of the right front wheel 5 and a wheel speed sensor 9 that detects the rotation speed of the right front wheel 5. A touch panel display 12 for displaying various information is installed in the vicinity of the handle 4. The rotation angle sensor 8 detects the rotation angle of the wheel and is, for example, a rotary encoder. The rotation angle sensor 8 and the wheel speed sensor 9 may be provided on the left front wheel 6 or the rear wheel instead of the right front wheel 5.

  Reference is now made to FIG. FIG. 2 is a functional block diagram showing the configuration of the vehicle 1. The vehicle 1 includes an obstacle detection device 14 that detects obstacles on the road, an autonomous traveling control unit 15 that controls the vehicle 1 autonomously traveling along the guide line, and autonomous traveling and obstacle detection. A vehicle speed control unit 16 that controls acceleration / deceleration of the vehicle speed, and an alarm device 17 that issues a warning to the front of the vehicle and the passenger by detecting an obstacle are provided. Since the autonomous traveling control unit 15 uses a conventional autonomous traveling control method, description thereof is omitted here. The alarm device 17 is provided on the front surface of the vehicle 1 and can notify the surroundings that the vehicle 1 is approaching by sound.

2. Configuration of Obstacle Detection Device Next, the configuration of the obstacle detection device 14 provided in the vehicle 1 will be described with reference to FIG. FIG. 3 is a block diagram illustrating a configuration of the obstacle detection device 14.

  The obstacle detection device 14 includes a distance information detection unit 21 that detects distance information to an object ahead of the vehicle 1, a parallax image generation unit 23 that generates a parallax image in front of the vehicle 1, and a vehicle from the start position So. A road position information storage unit 25 that stores road position information that identifies a road on the parallax image according to the position of 1, a position estimation unit 27 that estimates the position of the vehicle 1 on the road, and the estimated position Whether or not the detected obstacle is on the specified road, the road identification unit 29 for specifying the road in the parallax image based on the road position information, the obstacle detection unit 31 for detecting the obstacle based on the parallax image, and A determination unit 33 for determining the distance, a distance calculation unit 34 for calculating distance information from the parallax information, and a person determination unit 35 for determining whether or not the detected obstacle is a person.

  The distance information detection unit 21 includes a stereo camera 3 that captures an image in front of the vehicle 1 and a buffer 41 that temporarily stores an image captured by the stereo camera 3.

  The stereo camera 3 is composed of two image sensors, a left image sensor 3a and a right image sensor 3b. The left image sensor 3a and the right image sensor 3b are general visible light sensors such as a CCD and a CMOS. The left image sensor 3a and the right image sensor 3b are installed in the vehicle 1 under predetermined geometric conditions. In the first embodiment, the left image sensor 3a and the right image sensor 3b are installed at a constant distance in the horizontal direction. That is, the left image sensor 3a and the right image sensor 3b are arranged in a parallel stereo positional relationship.

The left image sensor 3a and the right image sensor 3b are arranged so that the positions of the respective rows of the captured images are matched, that is, the epipolar lines are matched. An image photographed by the left image sensor 3a is a left image, and an image photographed by the right image sensor 3b is a right image. In the first embodiment, the left image sensor 3a is a reference camera, and the left image is a reference image. Note that the number of image sensors provided in the stereo camera 3 is not limited to two and may be three or more. The right image may be the reference image.

  The coordinate system of the image sensors 3a and 3b will be described with reference to FIG. The X axis is provided in the horizontal direction (left and right direction) of the stereo camera 3, the Y axis is provided in the vertical direction (up and down direction) of the stereo camera 3, and the Z axis is provided in the front direction (depth direction) of the stereo camera 3. ing. The X axis and the Y axis are the coordinate axes of the left image and the right image. The disparity information between the left image and the right image is distance information from the obstacle detection device 14 to an object ahead of the vehicle 1.

  Returning to FIG. The buffer 41 temporarily stores images sent from the stereo camera 3, that is, images sent from the image sensors 3a and 3b. The buffer 41 uses a memory, a flash memory, a hard disk (HDD), or the like. In FIG. 3, the buffer 41 is provided corresponding to each of the image sensors 3a and 3b, but may be a single buffer. Each stored image is preferably an image in which lens distortion, variation in focal length, and the like are corrected using correction parameters. By such correction, the texture of the plane perpendicular to the optical axis of the image sensor is similarly projected on the image projection plane.

  The parallax image generation unit 23 generates a parallax image from each input image. That is, a parallax image is generated based on the left image and the right image stored in the buffer 41. The parallax image can be generated by stereo matching such as SAD. Other stereo matching methods include the area correlation method and the Census transform. The parallax here indicates the amount of pixel shift between a plurality of images. In the first embodiment, the parallax is a pixel shift amount in the horizontal direction of the right image with respect to the left image. Note that the X and Y coordinates on the parallax image are the same as the X and Y coordinates on the reference image. The generated parallax image is output to the obstacle detection unit 31 and the runway identification unit 29.

  The runway position information storage unit 25 stores position information for specifying the runway based on the parallax value and the X coordinate value on the parallax image according to the distance from the start position So of the runway. As shown in FIG. 5, the parallax value and the X coordinate value of the runway on the parallax image at each of the points S1 to Sn are measured and stored in advance at, for example, equidistant intervals from the start point So on the runway Tr. Yes. In addition, in FIG. 5, the display of the point is abbreviate | omitted on the way.

This will be described with reference to FIG. FIG. 6 shows, for example, a parallax image P1 at a point Sa at a distance La from the start position So in FIG. In the runway position information storage unit 25, the X coordinate value of the runway Tr in the parallax image P1 is stored in advance in association with the parallax value. That is, as the runway position information of the point Sa, the X coordinate area X ₁ -X ₁₀ for the parallax value da, the X coordinate area X ₂ -X ₉ for the parallax value db, and the X coordinate area X for the parallax value dc _The X-coordinate region X ₄ -X ₇ for the _3- X ₈ parallax value dd, and the X-coordinate region X ₅ -X ₆ for the parallax value de are stored in the road position information storage unit 25. The relationship between the parallax values da to dg is da>db>dc>dd>de>df> dg.

  Thus, the runway position information storage unit 25 stores the X coordinate value and the parallax value of the runway Tr in association with each other in the parallax image at each point S1 to Sn. The track position information storage unit 25 uses a memory, a flash memory, a hard disk (HDD), or the like.

  Returning to FIG. The position estimation unit 27 includes a reading unit 7, a rotation angle sensor 8, and a position calculation unit 43. The detection values of the reading unit 7 and the rotation angle sensor 8 are output to the position calculation unit 43.

  The position calculation unit 43 calculates the current position of the vehicle 1 from the start position So using a simplified odometry method based on each input detection value. That is, the start position So or the position information of each fixed point Fp is input from the reading unit 7 to the position calculating unit 43. Further, the position calculation unit 43 can calculate the movement distance from the start position So or each fixed point Fp by counting the detection value of the rotation angle sensor 8. The calculated position information of the vehicle 1 is output to the road position information storage unit 25, and road position information corresponding to the position information of the vehicle 1 is read out. The read track position information is output to the track specifying unit 29.

  Since the vehicle 1 travels on the predetermined track Tr, the position of the vehicle 1 on the track Tr can be specified by the start position So that is a predetermined position or the movement distance from each fixed point Fp. Thus, since the position estimation unit 27 estimates the position of the vehicle 1 on the track Tr based on the start position So or the movement distance from each fixed point Fp, only the vehicle coordinate system can be used without using the overall coordinate system. The position of the vehicle 1 can be estimated. Furthermore, since the moving distance is a one-dimensional data amount, the data amount handled for specifying the vehicle position can be significantly reduced. Further, the position estimation unit 27 can correct the error of the calculated position by reading the magnetic field information of the fixed point Fp embedded in a position other than the start position So. Thereby, the current position of the vehicle 1 can be estimated with high accuracy.

The runway specifying unit 29 specifies the runway Tr on the parallax image generated by the parallax image generating unit 23 based on the input runway position information. The runway is specified by a parallax value corresponding to an X coordinate value stored in advance. The parallax image P2 shown in FIG. 7 is the same spot as the parallax image P1 shown in FIG. In FIG. 7, the X coordinate area of the parallax value de in the road position information is X ₅ -X ₆ , but there is no area of the parallax value de having the X coordinate of X ₅ due to the obstacle Ob4. In this case, X coordinate area in the range of disparity values de is identified as track between _X 5 -X _6. The identified road area is output to the determination unit 33.

  In the parallax image, the obstacle detection unit 31 touches different parallax values in the X direction and has the same parallax value area having a number of pixels in the Y direction equal to or larger than a predetermined value, in other words, the same parallax value. A region having a predetermined height is detected as an obstacle. In FIG. 6 and FIG. 7, the regions Ob1, Ob2, Ob3, Ob4 are detected as obstacles. The detected obstacle areas Ob <b> 1 to Ob <b> 4 are output to the determination unit 33. In addition, it is not restricted to the same parallax value as an area | region of an obstacle, The area | region of the obstruction which has a several parallax value by processing the area | region of the parallax value of the range defined beforehand on the basis of a certain parallax value as a group May be specified.

  The determination unit 33 determines whether or not the detected obstacle exists on the track Tr. The determination unit 33 determines whether or not the input X coordinate value range at the lower end of the obstacle area is included in the X coordinate range of each parallax of the runway. Thereby, for example, it is determined that only the obstacle Ob4 is on the track Tr out of the obstacles Ob1 to Ob4 shown in FIG.

  The determination unit 33 determines whether the rectangular area surrounding the obstacle area overlaps the runway area, and then performs the above-described determination for the obstacle having the rectangular area overlapping the runway area. You may implement. By this two-stage determination, it is possible to determine whether or not an obstacle exists on the runway at a higher speed. In addition, when the Y coordinate value (height) of the lower end of the rectangle is larger than a predetermined height, it is determined that it is not an obstacle as a part of a tree such as a bridge or a branch over the runway. The accuracy of object determination can also be improved.

  When it is determined that there is an obstacle on the road Tr, the distance calculation unit 34 calculates a distance in the real space to the obstacle based on the parallax information of the area of the obstacle. As a method for calculating the distance from the parallax image, a conventionally used parallel stereo method is used. The calculated distance is output to the vehicle speed control unit 16. The vehicle speed control unit 16 decelerates or stops the vehicle speed of the vehicle 1 according to the distance to the obstacle. The closer the distance between the vehicle 1 and the obstacle, the greater the deceleration. Thereby, the contact avoidance with the vehicle 1 and an obstruction can be improved.

  The person determination unit 35 determines whether or not the area on the reference image corresponding to the area of the obstacle existing on the runway in the parallax image is a person. As a human determination method, a method in which an image local feature amount (HOG feature amount) and a statistical learning method are combined is used. In addition, person determination may be performed by matching using a template prepared in advance. If it is determined in the reference image that the area of the obstacle is a person, the touch panel display 12 alerts the passenger that the person is in front of the vehicle 1. Further, a warning whistle is sounded from the alarm device 17 to warn a person in front of the vehicle 1. When it is determined in the reference image that the area of the obstacle is not a person, the touch panel display 12 alerts the passenger that there is an obstacle in front of the runway.

  The parallax image generation unit 23, the runway identification unit 29, the obstacle detection unit 31, the determination unit 33, the distance calculation unit 34, the person determination unit 35, and the position calculation unit 43 are a microprocessor or an FPGA (Field Programmable Gate Array; reconfiguration) Possible gate array).

  Next, referring to FIG. 8, the operation of obstacle detection in the first embodiment will be described. FIG. 8 is a flowchart showing a processing procedure for distance estimation.

  When the vehicle 1 starts moving from the start point So of the track Tr, the position estimation unit 27 estimates the current position of the vehicle 1 on the track Tr (step S01). In addition, the stereo camera 3 captures an image in front of the vehicle 1 at a predetermined time interval. Thus, parallax information (distance information) is acquired by photographing a plurality of images. Moreover, the parallax image generation part 23 produces | generates 3D information by producing | generating a parallax image based on a some picked-up image (step S02). Next, the runway specifying unit 29 specifies the runway Tr in the parallax image based on the runway position information corresponding to the current position of the vehicle 1 (step S03). In addition, the obstacle detection unit 31 detects an obstacle in the generated parallax image (step S04). Next, the determination unit 33 determines whether or not the detected obstacle is on the track Tr (step S05).

  In the obstacle determination, when it is determined that there is no obstacle on the track Tr (No in step S06), the obstacle detection operation is terminated, and the processing from step S01 is started. In the obstacle determination, when it is determined that there is an obstacle on the runway Tr (Yes in step S06), the person determination unit 35 determines whether the obstacle is a person (step S07). If the obstacle is not a person (No in step S08), the vehicle speed control unit 16 performs deceleration or stop control including braking by the vehicle 1 according to the distance from the obstacle (step S10). When it is determined that the obstacle is a person (step S08), a warning sound is generated from the alarm 17 of the vehicle 1 that the vehicle 1 is approaching in front of the vehicle (step S09). In addition, the vehicle speed control unit 16 controls the vehicle 1 to be decelerated or stopped including brake braking according to the distance from the person (step S10).

  As described above, according to the first embodiment, the track position information for specifying the track Tr on the three-dimensional information is a two-dimensional map of a parallax value that is distance information and an X coordinate value that is coordinate information in the horizontal direction of the vehicle 1. Therefore, it is possible to reduce the amount of information handled as compared with the conventional method specified by using the three-dimensional information amount, and to reduce the calculation load. Thereby, the calculation time can be shortened, it can be determined whether or not the obstacle is on the runway when the distance from the obstacle is secured, and the avoidance of contact with the obstacle can be improved. it can.

  In addition, by limiting the determination of the obstacle to the road Tr, a three-dimensional object such as a tree on the side of the road is not determined as an obstacle, so that the vehicle 1 can be prevented from being erroneously stopped and can be smoothly driven. In this way, the vehicle 1 can be surely decelerated and stopped with respect to an obstacle on the traveling road Tr having a high possibility of contact.

  Further, since the distance information detection unit 21 includes the stereo camera 3, the parallax image can be used as three-dimensional information. Thereby, determination of the runway Tr in a parallax image and the detection of an obstruction can be implemented appropriately. In addition, when a one-dimensional scan laser radar is used, it is easy to detect the road surface as an obstacle by up and down the road surface. However, by using the stereo camera 3, the obstacle can be detected less easily by the influence of the road surface. Can do.

  In order to further improve the accuracy of the current position estimation, the error correction interval may be shortened. For example, it is only necessary to increase the number of fixed points Fp embedded in the runway. By doing in this way, since it is not necessary to mount expensive GPS and a sensor for every vehicle, cheap and highly accurate present position estimation is attained.

  Next, an obstacle detection apparatus according to the second embodiment will be described with reference to FIG. FIG. 9 is a block diagram illustrating a configuration of the obstacle detection apparatus according to the second embodiment. In the second embodiment, the parts indicated by the same reference numerals as those in the first embodiment have the same configuration as in the first embodiment, and thus the description thereof is omitted here. The configurations of the vehicle and the obstacle detection device other than those described below are the same as those in the first embodiment.

  The feature of the obstacle detection device 14 ′ of the second embodiment is that a distance measuring device 21 ′ is used instead of the stereo camera 3 as the distance information detection unit 21 of the obstacle detection device 14 of the first embodiment. In the first embodiment, distance information is detected by obtaining parallax information using two image sensors of the stereo camera 3. In the second embodiment, distance information from the obstacle detection device 14 ′ to an object ahead of the vehicle 1 is directly detected by using the distance measuring device 21 ′. Therefore, in the second embodiment, the measured value of the distance information is used instead of the parallax value of the first embodiment.

  The obstacle detection device 14 according to the second embodiment has a configuration in which a distance measuring device 21 ′ is provided instead of the stereo camera 3. The distance measuring device 21 'is, for example, a distance measuring radar or a laser radar (LIDAR). Distance information in polar coordinates can be obtained by radar scanning. Therefore, in the second embodiment, the scan angle θ is used instead of the X coordinate value of the first embodiment. Further, in place of the parallax image generation unit 23 of the first embodiment, a distance image generation unit 23 ′ is provided in the second embodiment. The distance image generation unit 23 ′ generates a distance image in which the horizontal axis is the scan angle θ, the vertical axis is the Y axis, and the pixel value is distance information. The Y coordinate value is preferably at least two points. In addition, the road position information storage unit 25 ′ according to the second embodiment stores the θ value of the road Tr and the distance value r in association with each other in the distance image at each point S1 to Sn.

  As described above, according to the second embodiment, the distance information r with respect to the scan angle θ can be appropriately acquired by using the distance measurement radar as the distance information detection unit. Rather than using the stereo camera 3, distance measurement can be performed with accuracy particularly at a distance, and the determination of the track Tr and the detection of an obstacle in the three-dimensional information can be appropriately performed.

  Next, an obstacle detection apparatus according to the third embodiment will be described with reference to FIG. FIG. 10 is a block diagram illustrating a configuration of the obstacle detection apparatus according to the third embodiment. In the third embodiment, the parts indicated by the same reference numerals as those shown in the first embodiment have the same configuration as that of the first embodiment, and thus the description thereof is omitted here. The configurations of the vehicle and the obstacle detection device other than those described below are the same as those in the first embodiment.

  In the first embodiment, a person is detected by determining whether or not the obstacle is determined to be on the track Tr. On the other hand, the feature of the obstacle detection device 54 according to the third embodiment is that it is determined whether or not a person is an obstacle present in a wider range than the runway Tr including the runway Tr. It is a point to detect. Thus, when detecting a person, it is preferable to detect a person in the range extended about 1 m by actual distance in the track width direction rather than the specified track Tr area | region. Thereby, since it is possible to detect a person who has only a part of the body in the runway Tr region, an avoidance action can be taken early.

  The obstacle detection device 54 according to the third embodiment is that the configuration of the obstacle detection device 14 according to the first embodiment includes a runway extension area specifying unit 64. In the first embodiment, the determination unit 33 determines whether or not there is an obstacle in the runway identified by the runway identification unit 29. In the third embodiment, in addition to this, as shown in FIG. 11, the road extension area specifying unit 64 specifies an area extended from the road Tr by a predetermined range outward in the width direction as the road extension area Tr ′. . In this runway extension region Tr ', obstacle determination and person determination are performed. The predetermined range to be expanded is a range expanded by about 1 m, for example, in the actual road width direction. That is, the number of pixels corresponding to the parallax corresponding to 1 m is extended outward from the specified traveling width shown in FIG.

  In addition to the function of the determination unit 33 according to the first embodiment, the determination unit 33 ′ according to the third embodiment is configured such that the obstacle detected by the obstacle detection unit 31 is determined by the road extension region specifying unit 64. Determine if it exists in '. Furthermore, in addition to the function of the human determination unit 35 of the first embodiment, the human determination unit 35 ′ of the third embodiment includes an obstacle determined to be present in the lane expansion area Tr ′ specified by the lane expansion area specification section 64. Determine if you are a person.

  With the above configuration, in the determination unit 33 ′ and the person determination unit 35 ′, it is possible to change the processing frequency of obstacle determination and person determination on the track Tr and the track extension region Tr ′. That is, processing time is required, and the processing frequency of human detection around the road where the detection result is not used for speed control can be reduced. Therefore, the frame rate of the obstacle detection process on the runway Tr can be ensured. As a result, erroneous detection of an obstacle can be avoided, and the vehicle 1 can be stably stopped. Since using the detection results of a plurality of frames can suppress erroneous stop due to erroneous detection rather than using only the detection result of one frame, it is better that the frame rate of the obstacle detection process on the track Tr is higher.

  In this way, the vehicle 1 can be surely decelerated and stopped with respect to an obstacle on the traveling road Tr having a high possibility of contact. Also, people near the end of the runway in the width direction are easily confused with stationary objects such as trees around the runway. Furthermore, it is difficult to determine the possibility of contact that a part of the body such as a human hand comes into contact with the vehicle 1. Even for such people, by detecting the person, an alarm can be issued, and by moving to a position away from the runway, both reduction of contact rate and avoidance of unpleasant false stop are achieved. Can be made possible.

  Next, an obstacle detection apparatus according to the fourth embodiment will be described with reference to FIG. FIG. 12 is a block diagram illustrating a configuration of the obstacle detection apparatus according to the fourth embodiment. In the fourth embodiment, the parts indicated by the same reference numerals as those shown in the first embodiment have the same configuration as that of the first embodiment, and thus the description thereof is omitted here. The configurations of the vehicle and the obstacle detection device other than those described below are the same as those in the first embodiment.

  In the first embodiment, an obstacle is detected from the entire generated parallax image, and it is determined whether or not the detected obstacle is on the road. On the other hand, in Example 4, a parallax image on the road is extracted from the generated parallax image, and an obstacle is detected from the extracted parallax image on the road. Thereby, since the area | region which detects an obstruction becomes small, a calculation load can be reduced more.

  The parallax image extraction unit 66 on the road according to the fourth exemplary embodiment uses the parallax image generated by the parallax image generation unit 23 based on the road position information stored in the road position information storage unit 25 and the parallax on the road. Extract only images. For example, it is assumed that the parallax image generating unit 23 generates the parallax image illustrated in FIG. The on-road parallax image extraction unit 66 extracts, from the road position information, a parallax image in the Y direction from the parallax image of the road Tr area and the road Tr having the same parallax value for each parallax area. FIG. 13 shows the extracted on-road parallax image P3.

The region Ob4 has a parallax value dc. The region Ob4 is extracted as a parallax image in the Y-direction upward direction from the road Tr between the X coordinate areas X _{3 to} X ₈ having the parallax value dc on the road Tr. In addition, the region Ob1 ′ is extracted as a parallax image that is located in the Y direction upward from the road Tr between the X coordinate areas X _{5 and} X ₆ having a parallax value de on the road Tr. Since only the parallax image on the runway Tr is extracted, not only the runway Tr is specified, but Ob2 and Ob3 are deleted as compared with the parallax image of FIG. Need not be detected.

  The extracted on-road parallax image is sent to the obstacle detection unit 31 '. The obstacle detection unit 31 ′ detects an obstacle from the on-road parallax image P <b> 3 as in the first embodiment. As a result, the region Ob4 is detected as an obstacle. The area Ob1 'is not detected as an obstacle because the minimum value of the Y coordinate value is away from the runway Tr. The person determination unit 35 determines whether or not the detected obstacle Ob4 is a person. The obstacle detection device 56 may include an extended area extraction unit 64. As shown in FIG. 14, the extension area extraction unit 64 extracts disparity information of the road extension area Tr ′ extended from the extracted road Tr by a predetermined range outward in the width direction. FIG. 14 shows, for example, the parallax image P4 on the runway Tr and the runway extension region Tr ′ extracted when the parallax image generation unit 23 generates the parallax image P1 shown in FIG. The obstacle detection unit 31 ′ detects an obstacle also on the extracted runway extension region Tr ′. When an obstacle is detected in the runway extension region Tr ′, the person determination unit 35 determines whether or not the detected obstacle is a person. Thereby, the effect similar to the case where the runway expansion area specific | specification part 64 in Example 3 is provided can be acquired.

  Next, with reference to FIG. 15, the operation of obstacle detection in the fourth embodiment will be described. FIG. 15 is a flowchart showing a processing procedure for obstacle detection. Since steps S01 and S02 are the same as those in the first embodiment, the description thereof is omitted.

  The on-road parallax image extraction unit 66 extracts a parallax image on the road from the parallax image based on the road position information corresponding to the current position of the vehicle 1 (step S03). Next, the obstacle detection unit 31 'starts obstacle detection in the extracted parallax image on the road (step S04). If no obstacle is detected in the parallax image on the road (No in step S06), the obstacle detection operation is terminated, and the processing from step S01 is started. In the obstacle detection, when an obstacle is detected on the track Tr (Yes in step S06), the person determination unit 35 determines whether the obstacle is a person (step S07). In addition, since the process after step S07 is the same as that of Example 1, description is abbreviate | omitted.

  As described above, according to the fourth embodiment, a stationary object outside the road is not detected as an obstacle, so that the calculation load can be reduced. For example, when detecting a person leaning on a kite off the runway, the kite and the person are detected as an obstacle together, and it has been difficult to detect this single object as a person. However, according to the fourth embodiment, the wrinkles outside the road are removed on the parallax image, so that only people on the road can be detected as obstacles, and the detection accuracy of the obstacles can be improved. In addition, when using a radar instead of a stereo camera, what is necessary is just to mask the data outside a runway range.

  The present invention is not limited to the above embodiment, and can be modified as follows.

  (1) In Example 3, a person was detected by determining whether or not the obstacle was determined to be on the runway. Not only this but in parallel with the detection of an obstacle, you may provide the person detection part which detects a person within the range wider than the track Tr including the track Tr. Thereby, since it is possible to detect a person who has only a part of the body in the runway Tr region, an avoidance action can be taken early. Also, by providing a separate human detection unit from the obstacle detection unit 31, different processes can be performed in parallel. In other words, it is difficult to judge obstacles on the runway with high contact possibility, and it is difficult to judge the possibility of contact, and it is easy to be confused with trees around the runway. Can do.

  (2) In Examples 1 to 4, an obstacle detection device obtained by a combination of the configurations may be configured.

  (3) Although the obstacle detection device 14 is provided in the vehicle 1 in the above embodiment, the present invention is not limited to this. In addition, for example, it may be adopted in a vision system for a robot that travels autonomously or a support system for a visually impaired person.

DESCRIPTION OF SYMBOLS 1 ... Vehicle 3 ... Stereo camera 3a ... Left image sensor 3b ... Right image sensor 14, 56 ... Obstacle detection device 21 ... Distance information detection part 21 '... Distance measuring device 23 ... Three-dimensional information generation part 25 ... Track position information storage Unit 27 ... Position estimation unit 29 ... Runway identification unit 31, 31 '... Obstacle detection unit 33 ... Determination unit 35 ... Person determination unit 66 ... Runway parallax image extraction unit

Claims (15)

In the obstacle detection device mounted on a moving body that runs on a predetermined track,
A distance information detection unit that detects distance information from the obstacle detection device to an object in front of the moving body;
A three-dimensional information generation unit that generates three-dimensional information ahead of the moving body based on the distance information;
A storage unit in which track position information corresponding to the position of the moving body from a predetermined position is stored;
A position estimator for estimating the position of the moving object on the runway;
A three-dimensional information extraction unit on the road for extracting three-dimensional information on the road in the three-dimensional information based on the road position information;
Extracted, and a obstacle detection unit for detecting an obstacle based on the three-dimensional information on the track,
The travel path position information is specified by the distance information and horizontal coordinate information of the moving object. The obstacle detection device according to claim 1,
An obstacle detection device comprising a person determination unit that determines whether or not the obstacle on the runway is a person. In the obstacle detection device according to claim 2,
Based on the road position information, an extended area 3D information extraction unit that extracts 3D information in an extended road area that is a predetermined area adjacent to the road,
The obstacle detection unit detects an obstacle also on the extended runway area,
The obstacle determination device is configured to determine whether the obstacle detected in the extended runway region is a person. The obstacle detection device according to claim 1,
Based on the runway position information, an extended area 3D information extraction unit that extracts 3D information in an extended runway area that is a predetermined area adjacent to the runway;
An obstacle detection apparatus comprising a human detection unit that detects a person based on the three-dimensional information in the extended runway region. In the obstacle detection device mounted on a moving body that runs on a predetermined track,
A distance information detection unit that detects distance information from the obstacle detection device to an object in front of the moving body;
A three-dimensional information generation unit that generates three-dimensional information ahead of the moving body based on the distance information;
A storage unit that stores road position information for specifying a road on the three-dimensional information according to the position of the moving body from a predetermined position;
A position estimator for estimating the position of the moving object on the runway;
A road identification unit that identifies a road in the three-dimensional information based on the estimated position and the road position information;
An obstacle detection unit for detecting an obstacle based on the three-dimensional information;
A determination unit that determines whether or not the detected obstacle is on the identified road,
The travel path position information is specified by the distance information and horizontal coordinate information of the moving object. In the obstacle detection device according to claim 5,
An obstacle detection device comprising a person determination unit that determines whether or not the obstacle determined to be on a runway is a person. The obstacle detection device according to claim 6,
An area specifying unit for specifying an extended road area that is a predetermined area adjacent to the specified road;
The determination unit determines whether or not the detected obstacle is on the specified extended track area,
The said person determination part determines whether the said obstacle determined to be on an extended runway area | region is a person The obstacle detection apparatus. In the obstacle detection device according to claim 5,
An area specifying unit for specifying an extended road area that is a predetermined area adjacent to the specified road;
An obstacle detection device comprising a person detection unit for detecting a person in the specified extended track area. In the obstacle detection device according to any one of claims 1 to 8,
The said position estimation part estimates the position of the said mobile body on the said runway from the movement distance from the predetermined position The obstruction detection apparatus. In the obstacle detection device according to any one of claims 1 to 9,
The distance information detection unit has a stereo camera,
The obstacle detection apparatus, wherein the three-dimensional information is a parallax image. In the obstacle detection device according to any one of claims 1 to 10,
The distance information detection unit includes an radar.   A moving body comprising the obstacle detection device according to any one of claims 1 to 11. The moving body according to claim 12,
A moving body comprising an alarm device that issues a warning when the obstacle is detected on the runway. The moving body according to claim 12 or 13,
A moving body comprising a speed control unit that decelerates or stops the moving body when the obstacle is detected on the running path. In a moving object provided with the obstacle detection device according to any one of claims 3, 4, 7, and 8.
A speed control unit for decelerating or stopping the moving body with respect to the detected obstacle on the runway;
A moving body comprising: an alarm device that issues a warning to a detected person on the extended runway area.

Images