JP2019090711A - Wall detector - Google Patents

Abstract

To detect a wall.SOLUTION: A wall detector is mounted in a fork lift. The wall detector comprises two imaging devices and an image processing device. Each imaging device is a stereo camera. The image processing device calculates parallax from the image captured by each imaging device. The image processing device calculates coordinates of an object shown in the image from the calculated parallax, and plots the object to a world coordinate system. When there exists a point group P1 that represents a plurality of feature points P that are in a row on an X axis of a world coordinate system and there is no feature point P in the back from the point group P1, the image processing device determines that the point group P1 is the upper end of a wall.SELECTED DRAWING: Figure 8

Description

  The present invention relates to a wall detection device.

In order to prevent a mobile from colliding with an obstacle, an obstacle detection device for detecting an obstacle is mounted on the mobile (see, for example, Patent Document 1).
The obstacle detection device of Patent Document 1 detects an obstacle using parallax (parallax image) obtained by a stereo camera. A stereo camera is composed of a plurality of cameras. The parallax is a difference (difference in the number of pixels) in the image position of the object shown in the image taken by each camera when the same object is imaged by a plurality of cameras. The obstacle detection device determines that the object is a wall when it detects an object including a surface continuous from the height position corresponding to the ceiling downward to a predetermined length or more.

Unexamined-Japanese-Patent No. 2004-326264

  By the way, when detecting an obstacle using a stereo camera, the position of the obstacle is detected from the parallax. In order to obtain parallax, it is necessary to extract the same object from the image captured by each camera. When extracting the same object, a portion (edge) that is a feature is extracted from the image, and this feature portion is used as the object. However, when a wall having no feature such as a pattern or unevenness is imaged by a stereo camera, parallax may not be obtained and detection of the surface may not be performed. Then, in the obstacle detection device of Patent Document 1, there may occur a case where the wall can not be detected.

  An object of the present invention is to provide a wall detection device capable of detecting a wall.

  A wall detection device for solving the above-mentioned problems is a wall detection device which is mounted on a mobile body and detects a wall erected in the vertical direction, and is imaged by the stereo camera which images the periphery of the mobile body and the stereo camera Calculating means for calculating a parallax from the captured image, an axis extending in the traveling direction of the movable body as a first axis, an axis extending in the vertical direction as a second axis, the first axis, and the second axis The coordinate calculation means for calculating the coordinates in the real space about the feature point for which the parallax is obtained in a three-axis orthogonal coordinate system in which the axis orthogonal to the third axis is the third axis, and arranged on the third axis The point group which is the plurality of feature points is present at a position higher than a predetermined height in the vertical direction, and the feature point does not exist at the back of the moving body in the traveling direction than the point group If the point cloud is the top of the wall It comprises a constant determining means.

  Since the boundary between the ceiling and the space exists at the upper end of the wall, the parallax calculating means can calculate the parallax from the upper end of the wall shown in the image. When coordinates are calculated for the feature points for which parallax has been obtained, the feature points indicating the upper end of the wall are aligned on the third axis. In addition, since it is not possible to image the back of the wall in the traveling direction, when the wall is imaged, there is no feature point at the back of the point cloud. For this reason, when the point group exists at a position higher than a predetermined height in the vertical direction, and the feature point does not exist behind the point group, the determination unit determines that the point group is the upper end of the wall It can be determined. The wall detection device can detect the wall by judging that the wall is present from the upper end to the lower side.

  In the wall detection device, when the lower point group which is the plurality of the feature points arranged on the third axis is present below the point group, the determination means determines that the lower point group is the wall It may be determined that the lower end of.

  Similar to the top of the wall, there is a boundary with the floor at the bottom of the wall. When the lower point group exists below the point group, the determination means can determine that the lower point group is the lower end of the wall. The wall detection device can determine that a wall exists between the point cloud and the lower point cloud.

In the wall detection device, a plurality of cameras constituting the stereo camera may be arranged in the vertical direction.
When a plurality of cameras constituting a stereo camera are arranged in the vertical direction, the same object shown in an image captured by each camera is captured with a vertical shift. Then, the parallax appears as a difference of vertical pixels in each image. When the parallax is calculated from the edge extending in the horizontal direction of the image, such as the upper end of the wall, the parallax calculating unit can calculate the parallax more easily by the difference of the vertical pixels. Therefore, the wall detection device can easily detect the wall by arranging the cameras in the vertical direction.

  In the wall detection device, when there is a horizontal point group which is a plurality of the feature points arranged on the second axis, the determination means is an end portion in the left-right direction of the wall It may be determined that

  When imaging the left and right ends of the wall, the ends become feature points aligned on the second axis. Therefore, when the horizontal point group is present, the determination means can determine that the horizontal point group is the lateral end of the wall.

  According to the present invention, walls can be detected.

The side view of the forklift with which a wall detection device is carried. The schematic block diagram of a forklift and a wall detection apparatus. Schematic which shows a wall. (A) is a figure which shows a 1st image when a wall is imaged with a 1st camera, (b) is a figure which shows a 2nd image when a wall is imaged with a 2nd camera. The flowchart which shows wall detection processing. The flowchart which shows wall feature search processing. The flowchart which shows relative position calculation processing. FIG. 3 is a schematic diagram showing feature points plotted on a world coordinate system. The figure for demonstrating the state of a passage. FIG. 3 is a schematic diagram showing feature points plotted on a world coordinate system.

Hereinafter, one embodiment of a wall detection device is described.
As shown in FIGS. 1 and 2, the forklift 10 as a moving body includes a vehicle body 11 and a cargo handling device 12 provided at the front of the vehicle body 11. The forklift 10 includes a drive wheel 13, a steered wheel 14, and a traveling motor M 1 that drives the drive wheel 13. The forklift 10 may be one in which a traveling operation and a cargo handling operation are automatically performed, or may be one in which a traveling operation and a cargo handling operation are performed by an operation by a passenger (operator).

  The forklift 10 includes a main controller 20, a traveling control device 23 that controls the traveling motor M1, and a vehicle speed sensor 24. The main controller 20 performs control relating to the traveling operation and the cargo handling operation. The main controller 20 includes a CPU (arithmetic unit) 21 and a memory 22 in which programs for performing various controls are stored.

  The CPU 21 of the main controller 20 gives the travel control device 23 a command for the rotational speed of the traveling motor M1 so that the vehicle speed of the forklift 10 becomes the target speed. The main controller 20 controls the traveling direction (forward or reverse) of the forklift 10 by giving the traveling control device 23 an instruction for rotating the traveling motor M1.

  The travel control device 23 of the present embodiment is a motor driver. The vehicle speed sensor 24 of the present embodiment is a rotation speed sensor that detects the rotation speed of the traveling motor M1. The vehicle speed sensor 24 outputs the rotation speed of the traveling motor M1 to the traveling control device 23. The traveling control device 23 controls the traveling motor M1 based on the command from the main controller 20 so that the rotational speed of the traveling motor M1 matches the command. The traveling control device 23 controls the traveling motor M1 such that the rotation direction of the traveling motor M1 matches the command from the main controller 20.

  A wall detection device 30 is mounted on the forklift 10. The wall detection device 30 includes two imaging devices 31 and 41. One of the two imaging devices 31 and 41 is the front imaging device 31, and the other is the rear imaging device 41. The front imaging device 31 is a stereo camera that images the front of the forklift 10. The rear imaging device 41 is a stereo camera that shoots the rear of the forklift 10.

  Each of the imaging devices 31 and 41 is disposed in the upper part of the vehicle body 11. The front imaging device 31 includes two cameras 32 and 33. As the cameras 32 and 33, for example, a CCD image sensor or a CMOS image sensor is used. The cameras 32 and 33 are arranged such that their optical axes are parallel to each other. The optical axes of the cameras 32 and 33 may or may not be horizontal. In the present embodiment, the two cameras 32 and 33 are arranged side by side in the vertical direction. Of the two cameras 32 and 33, the camera 32 located at the upper side is the first camera 32, and the camera 33 located at the lower side is the second camera 33. The rear imaging device 41 has the same configuration as the front imaging device 31, and includes a first camera 42 and a second camera 43.

  Assuming that the image captured by the first camera 32, 42 is a first image, and the image captured by the second camera 33, 43 is a second image, the same object is vertically oriented in the first image and the second image. The image is shifted in the direction (Y axis of the image).

  As shown in FIG. 3, it is assumed that the wall W erected in the vertical direction is imaged by the front imaging device 31. In this case, the upper end E1 of the wall W and the lower end E2 of the wall W appear in the image captured by each of the cameras 32 and 33. The wall W erected in the vertical direction also includes a wall erected in an inclined manner with respect to the vertical direction. That is, it may be a wall provided with the upper end E1 and the lower end E2.

  As shown in FIGS. 4A and 4B, in the first image and the second image, the upper end E1 and the lower end E2 are shifted in the vertical direction (Y-axis of the image). In the present embodiment, the number of pixels of the image captured by each of the cameras 32, 33, 42, 43 is the same. For example, VGA (640 × 480 pixels) is used as the image.

  As shown in FIG. 2, the wall detection device 30 includes an image processing device 50 that detects the wall W from the image captured by each of the imaging devices 31 and 41. The image processing apparatus 50 is configured of a CPU, a RAM, a ROM, and the like. Although the image processing apparatus 50 detects an object that can be an obstacle other than the wall W, for convenience of description, only the case of detecting the wall W will be described. The image processing device 50 can also be said to be an obstacle detection device that detects an obstacle including the wall W from the images captured by the imaging devices 31 and 41.

Hereinafter, the wall detection process performed by the image processing device 50 will be described together with the operation of the wall detection device 30.
As shown in FIG. 5, the image processing device 50 calculates parallax from the image captured by the imaging devices 31 and 41 (step S <b> 10). Parallax compares the first image and the second image for each of the imaging devices 31 and 41, and calculates the difference between the image position (pixel) of the first image and the second image for the same object shown in each image It is obtained by doing. The difference between the image positions is the difference [px] in the number of pixels between the object shown in the first image and the object shown in the second image. In the present embodiment, the parallax is obtained as a parallax image. The parallax image is an image in which an object appearing in an image captured by the imaging devices 31 and 41 is associated with parallax of the object. In the parallax image, parallax is associated with each pixel. The number of pixels of the parallax image is the same as the number of pixels of the first image (second image). Note that the parallax image does not necessarily need to be displayed, but indicates data in which parallax is associated with each pixel in the parallax image (X coordinate and Y coordinate in parallax image). In the present embodiment, the image processing device 50 functions as a parallax calculation unit.

  As shown in FIGS. 4A and 4B, when pixels of the same object shown in each image are compared, the horizontal pixels in the first image and the second image become the same, while The pixels are different. That is, parallax appears as a difference between vertical pixels. When calculating the parallax between the first image and the second image, the same object in each image is extracted by scanning each image in the vertical direction (Y axis). The object is a characteristic portion included in each image, and is, for example, a portion that can be recognized as a boundary, such as an edge of an object. The object can be detected from luminance information or the like, and the image processing apparatus 50 can calculate parallax from the object. On the other hand, parallax can not be calculated for a portion (pixel) having no feature in the image.

  As shown in FIG. 5, the image processing device 50 calculates the position of an object present around the forklift 10 from the parallax (step S <b> 11). First, the image processing device 50 calculates the position of the object in the camera coordinate system from the parallax images obtained by the imaging devices 31 and 41. The camera coordinate system is a three-axis orthogonal coordinate system in which an optical axis is taken as a Z axis, and each of two axes orthogonal to the optical axis is taken as an X axis and a Y axis. The position of the object in the camera coordinate system can be represented by Z coordinate Zc, X coordinate Xc, and Y coordinate Yc in the camera coordinate system. The Z coordinate Zc, the X coordinate Xc, and the Y coordinate Yc can be calculated using the following equations (1) to (3), respectively.

Here, B is the interocular distance [m] of each imaging device 31, 41, f is the focal distance [m], and d is the parallax [px]. xp is an arbitrary X coordinate (horizontal pixel of an arbitrary pixel) in the parallax image, and x 'is a central X coordinate in the parallax image. yp is an arbitrary Y coordinate (longitudinal pixel of any pixel) in the parallax image, and y ′ is a central Y coordinate in the parallax image. The calculations of the equations (1) to (3) described above are performed for all the objects in the parallax image.

  Here, an axis extending in the traveling direction of the forklift 10 is a Y axis as a first axis, and an axis extending in a vertical direction is a third axis which is orthogonal to the Z axis as a second axis, Y axis, and Z axis. Coordinates in a three-axis orthogonal coordinate system (world coordinate system), which is an X-axis as a coordinate in (world coordinate) in real space, are used as The position of the object in the real space can be represented by the X coordinate Xw, the Y coordinate Yw, and the Z coordinate Zw in the real space.

  When the imaging devices 31 and 41 are installed such that the optical axes of the cameras 32, 33, 42, and 43 are horizontal (orthogonal to the vertical direction), Xc = Xw, Yc = Zw, and Zc = Yw. That is, the position of the object in the camera coordinate system is the position of the object in real space.

  On the other hand, when the optical axis of each camera 32, 33, 42, 43 is not horizontal (perpendicular to the vertical direction), the image processing apparatus 50 performs coordinate conversion (world coordinate conversion) using the following equation (4) Do.

Here, H is the mounting height of the imaging devices 31 and 41 in the world coordinate system, and θ is an angle of + 90 ° between the optical axes of the cameras 32, 33, 42 and 43 and the horizontal plane. In the present embodiment, the image processing apparatus 50 functions as coordinate calculation means.

  Next, the image processing apparatus 50 determines the traveling direction (step S12). The traveling direction (forward or reverse) can be determined from, for example, the rotation direction of the traveling motor M1 detected by the vehicle speed sensor 24.

The image processing device 50 performs wall feature search processing for searching whether the wall feature appears from the position of the object obtained in step S11 (step S13).
As shown in FIG. 8, the search for the wall feature is performed for each area A present in the traveling direction among a plurality of areas A obtained by dividing the XY plane in the world coordinate system into an arbitrary size. If the traveling direction is before, search for wall features is performed for each area A before the forklift 10, and if the traveling direction is after, search for wall features is performed for each area A after the forklift 10.

  The search for the wall feature is performed by plotting the position of the object obtained in step S11 in the world coordinate system. Hereinafter, the position of the object plotted in the world coordinate system will be described as the feature point P. The feature point P can be said to be the coordinates at which the parallax (the position of the object) is obtained in the world coordinate system.

  As shown in FIG. 6, the image processing apparatus 50 determines whether or not the point group P1, which is a plurality of feature points P aligned on the X-axis, exists at a position higher than a predetermined height in the vertical direction. (Step S21). That is, it is determined whether or not a plurality of feature points P having similar values of the Z coordinate Zw and the Y coordinate Yw exist in the area A to be searched along the X axis. In addition, with "the same value" said here, a slight shift | offset | difference is accept | permitted.

  When the wall W is imaged by the imaging devices 31 and 41, the upper end E1 of the wall W has a boundary with a ceiling, a space, and the like, so that parallax can be obtained from the upper end E1 captured in the image. Since the upper end E1 extends in the horizontal direction, when the feature points P indicating the upper end E1 are plotted, a plurality of feature points P are arranged on the X-axis to form a point group P1. Note that “a position higher than a predetermined height in the vertical direction” is a height at which it is assumed that the upper end E1 of the wall exists, and for example, at a position higher than the height (highest point) of the forklift 10 is there. Whether or not the upper end E1 of the wall W is at a position higher than a predetermined height in the vertical direction can be determined from the Z coordinate Zw of the feature point P constituting the point group P1.

  If the determination result in step S21 is negative, the image processing apparatus 50 determines that the wall W does not exist in the corresponding area A (step S24). If the determination result of step S21 is affirmative, the image processing apparatus 50 proceeds to step S22.

  The image processing apparatus 50 determines whether the lower point group P2 which is a point group different from the point group P1 is present at a position lower than a predetermined height in the vertical direction (step S22). Similar to the upper end E1 of the wall W, the lower edge E2 of the wall W also has a boundary with the floor. When the feature points P indicating the lower end E2 are plotted, the plurality of feature points P are arranged on the X-axis to become the lower point group P2. The “position lower than the point group P1 in the vertical direction” in step S22 is a position lower than the point group P1, and is, for example, a position at the same height as the floor surface on which the forklift 10 is traveling.

  If the determination result in step S22 is negative, the image processing apparatus 50 determines that the wall W does not exist in the corresponding area A (step S24). If the determination result in step S22 is affirmative, the image processing apparatus 50 proceeds to step S23.

  The image processing apparatus 50 determines whether or not the feature point P exists in the back of the point group P1 (lower point group P2) (step S23). The point group P1 is more forward than the point group P1 when the forklift 10 is moving forward, and is more backward than the point group P1 when the forklift 10 is moving backward. If the determination result in step S23 is negative, the image processing apparatus 50 determines that the wall W is present in the corresponding area A (step S25). If the determination result in step S23 is affirmative, the image processing apparatus 50 determines that the wall W does not exist in the corresponding area A (step S24).

  As shown in FIG. 9, even when there is no wall W in the direction of travel, the case where two passages F1 and F2 of different colors are connected or there is a joint between two passages F1 and F2, etc. 2 Parallax (position information of an object) may be calculated from the boundary between two passages F1 and F2. In this case, the point group P1 is detected even though the wall W is not present at the boundary between the two paths F1 and F2. If the wall W does not exist, an object O behind the boundary between the passages F1 and F2 appears in the image, and parallax is obtained by this object. Therefore, as shown in FIG. 10, even if the point group P1 is present, it can be determined that the wall W does not exist if the feature point P is present behind the point group P1.

  On the other hand, since the back can not be imaged behind the wall W in the traveling direction, when the wall W is imaged, the feature point P does not exist behind the point group P1. When there is no feature point P behind the point group P1, it can be determined that the point group P1 is the upper end E1 of the wall W and the lower point group P2 is the lower end E2 of the wall W. Then, the image processing apparatus 50 can detect that the wall W is present between the point group P1 and the lower point group P2 in the area where the point group P1 and the lower point group P2 exist. In the present embodiment, the image processing apparatus 50 functions as a determination unit.

  As shown in FIG. 6, when the search of the area A is finished, that is, when the process of step S24 or step S25 is performed, the image processing apparatus 50 increments (adds) the number of searches (step S26). The feature search process ends.

  As shown in FIG. 5, when the wall feature search process is completed, the image processing apparatus 50 compares the number of areas A present in the traveling direction with the number of searches, and the number of areas A present in the traveling direction is searched It is determined whether the number is larger than the number (step S14). That is, it is determined whether or not wall feature search processing has been performed for all of the areas A present in the traveling direction. If the determination result in step S14 is affirmative, the image processing apparatus 50 performs wall feature search processing on the area A in which the search for wall features has not ended. On the other hand, if the determination result of step S14 is negative, the image processing apparatus 50 proceeds to step S15.

  The image processing device 50 determines whether the wall W is present in the traveling direction of the forklift 10 (step S15). That is, in the wall feature searching process in step S13, it is determined whether the wall W is detected. If the determination result in step S15 is negative, the image processing apparatus 50 ends the process. On the other hand, when the determination result of step S15 is affirmative, the image processing device 50 performs relative position calculation processing (step S16).

  As shown in FIG. 7, in the relative position calculation process, the image processing device 50 calculates the relative position of the forklift 10 and the wall W (step S31). The relative position can be calculated from the Y coordinate of the feature point P constituting the point group P1 or the lower point group P2. The image processing device 50 outputs the calculated relative position to the main controller 20.

  The main controller 20 determines whether the relative distance between the forklift 10 and the wall W is equal to or less than a predetermined value (step S32). The predetermined value is, for example, a value determined based on the maximum speed of the forklift 10, braking performance, and the like.

  If the determination result in step S32 is negative, the main controller 20 ends the process. If the determination result in step S32 is affirmative, the main controller 20 performs a deceleration process (step S33). As the deceleration processing, processing to stop the forklift 10, processing to gradually reduce the speed of the forklift 10 from the current speed, processing to restrict the speed of the forklift 10, and control traveling at a speed exceeding the speed limit And various treatments.

Therefore, according to the said embodiment, the following effects can be acquired.
(1) When the point group P1 exists and there is no feature point P behind the point group P1, the image processing device 50 determines that the point group P1 is the upper end E1 of the wall W. The image processing apparatus 50 can detect that the wall W exists downward from the point cloud P1. The upper end E1 of the wall W is present even in the case where the wall W has no feature such as a pattern or unevenness. Therefore, by detecting the wall W based on the parallax obtained from the upper end E1, the wall W can be detected even when the wall W has no feature.

  (2) When the lower point group P2 exists below the point group P1, the image processing apparatus 50 determines that the lower point group P2 is the lower end E2 of the wall W. The image processing device 50 can detect that the wall W exists between the point group P1 and the lower point group P2. Therefore, the wall W can be detected more accurately.

  (3) If the cameras 32, 33, 42, 43 are arranged in the lateral direction, the first image and the second image become images shifted in the lateral direction (X coordinate in the image). Since the upper end E1 of the wall W extends in the horizontal direction, when the first image and the second image shift in the lateral direction, the same object as the object shown in the first image is any pixel of the second image It is difficult to determine if it corresponds to On the other hand, by arranging the cameras 32, 33, 42, 43 in the vertical direction, the images captured by the cameras 32, 33 become images shifted in the vertical direction. Then, the parallax appears as a difference between vertical pixels in each image (difference in Y coordinate in the image). As a result, it becomes easy to extract the same object from each image, and it becomes easy to calculate the parallax. Therefore, the image processing apparatus 50 can easily detect the wall W.

The embodiment may be modified as follows.
○ As indicated by a broken line in FIG. 8, when there is a vertical point group P3 that is a plurality of feature points P arranged on the Z axis in the world coordinate system, the image processing device 50 determines that the vertical point group P3 is a wall W It may be determined that it is the end in the left-right direction in. In this case, the image processing apparatus 50 detects that the wall W is present in a portion surrounded by the upper end E1, the lower end E2, and the end in the left-right direction. The image processing apparatus 50 may also detect that the wall W is present in a portion between the two ends, when the two ends in the left-right direction are detected.

  The image processing device 50 may perform the process of detecting the wall W, and may not calculate the relative position. For example, the image processing apparatus 50 may perform the processing up to step S15 and output the detection result as to whether or not the wall W is detected to the main controller 20. In this case, the main controller 20 may notify the display unit that the wall W is present or sound a buzzer. That is, the process performed when the wall W is detected may be arbitrarily changed.

  The cameras 32, 33, 42 and 43 constituting the imaging devices 31 and 41 may be provided side by side in the horizontal direction (left and right direction). In this case, the parallax can be easily calculated from the edge extending in the vertical direction.

  ○ In the image processing apparatus 50, when the point group P1 exists at a position higher than a predetermined height in the vertical direction and the feature point P does not exist behind the point group P1, downward from the point group P1 It may be determined that the wall W is present. That is, step S22 which is a judgment of whether lower end E2 of wall W exists may be omitted.

  The image processing apparatus 50 may determine that the wall W exists in the area A where the point group P1 exists, when the feature point P exists between the point group P1 and the lower point group P2. In this case, the wall W is a wall having a feature such as a pattern or unevenness.

As the imaging devices 31 and 41, one having three or more cameras may be used.
The mobile object may be a construction machine, an automatic carrier, a vehicle such as a truck, or a biped robot. In addition, it is preferable that it is what moves indoors as a moving body.

The front imaging device 31 and the rear imaging device 41 may not have the same configuration.
The determination of the traveling direction of the forklift 10 may be performed based on a result other than the detection result of the vehicle speed sensor 24. For example, when the traveling direction of the forklift 10 is changed by the direction lever, it may be performed based on the detection result of the direction sensor which detects the operation position of the direction lever. Further, the rotation direction of the traveling motor M1 may be acquired from the main controller 20.

  The image processing apparatus 50 is provided as an apparatus for performing image processing of an image captured by the imaging devices 31 and 41. However, the image processing may be performed by the main controller 20 or the like. That is, image processing may be performed as one function of the main controller 20 by performing control related to the traveling operation and the cargo handling operation.

The image processing apparatus 50 functions as a parallax calculation unit, a coordinate calculation unit, and a determination unit. However, individual devices functioning as the respective units may be provided.
In each embodiment, an engine may be used as a traveling drive device. That is, the forklift 10 may travel by driving the engine. In this case, the travel control device is a device that controls the amount of fuel injection to the engine and the like.

In the case of a moving object that can perform only forward movement, the rear imaging device 41 may not be provided.
In the case where the wall detection device 30 is mounted on a movable body that can be carried by a passenger, only the rear wall W in which visual recognition by the passenger is not easily performed may be detected. In this case, the front imaging device 31 may not be provided.

  The front imaging device 31 may be disposed at any position as long as the front of the forklift 10 can be imaged. The rear imaging device 41 may be disposed at any position as long as the rear of the forklift 10 can be imaged. For example, the front imaging device 31 may be disposed in the cargo handling device 12, and the rear imaging device 41 may be disposed in a counterweight provided behind the vehicle body 11.

  ○ Coordinate conversion may be performed by table data instead of equation (4). The table data is table data in which Y coordinate Yw corresponds to a combination of Y coordinate Yc and Z coordinate Zc, and Z data Zw corresponds to a combination of Y coordinate Yc and Z coordinate Zc. By storing these table data in the ROM of the image processing apparatus 50, the Y coordinate Yw and the Z coordinate Zw in the world coordinate system can be obtained from the Y coordinate Yc and the Z coordinate Zc in the camera coordinate system. In the embodiment, since the X coordinate Xc in the camera coordinate system matches the X coordinate Xw in the world coordinate system, the table data for obtaining the X coordinate Xw is not stored.

  P: characteristic point, P1: point group, P2: lower point group, P3: vertical point group, 10: forklift (moving object), 30: wall detection device, 31, 41: imaging device (stereo camera), 32, 33, 42, 43 ... camera, 50 ... image processing apparatus (parallax calculation means, coordinate calculation means, determination means).

Claims (4)

A wall detection device mounted on a mobile body and detecting a wall standing upright in the vertical direction,
A stereo camera for imaging the periphery of the moving body;
Disparity calculating means for calculating disparity from an image captured by the stereo camera;
An axis extending in the direction of movement of the movable body is a first axis, an axis extending in the vertical direction is a second axis, and an axis orthogonal to the first axis and the second axis is a third axis Coordinate calculation means for calculating coordinates in real space about a feature point for which the parallax has been obtained in an orthogonal coordinate system;
A plurality of point groups which are the plurality of feature points arranged on the third axis are present at positions higher than a predetermined height in the vertical direction, and in the traveling direction of the moving body than the point groups And a determination unit that determines that the point cloud is the upper end of the wall when the feature point does not exist at the back.   The determination means determines that the lower point group is the lower end of the wall, when lower point groups which are the plurality of feature points arranged on the third axis exist below the point group. The wall detection device according to claim 1.   The wall detection device according to claim 1, wherein the plurality of cameras constituting the stereo camera are arranged in line in the vertical direction.   The judging means judges that the horizontal point group is a lateral end of the wall, when the horizontal point group which is the plurality of feature points arranged on the second axis is present. A wall detection device according to any one of the preceding claims.