Obstacle detection device

Abstract

Provided is an obstacle detection device that suppresses a decrease in positional accuracy of an obstacle. An obstacle detection device includes a camera and a control unit. The control unit acquires a captured image from the camera. The control unit calculates the position of the obstacle on the basis of the feature of the captured image and the angle of the optical axis. The control unit does not correct the angle of the optical axis when the specific condition is satisfied, and corrects the angle of the optical axis when the specific condition is not satisfied.

Description

Technical Field

[0001] This disclosure relates to obstacle detection devices. Background Technology

[0002] Patent Document 1 discloses an obstacle detection device comprising a camera and a detection unit. The detection unit acquires a straight line from the image captured by the camera. The detection unit estimates the position of the vanishing point based on the straight line that forms the vanishing point. The detection unit calculates the camera's tilt angle based on the distance between the vanishing point and the center of the image captured by the camera, as well as the camera's focal length. By taking the camera's tilt angle into account when estimating the position of the obstacle, errors caused by the camera's tilt angle can be corrected.

[0003] Existing technical documents

[0004] Patent documents

[0005] Patent Document 1: Japanese Patent Application Publication No. 2023-132153 Summary of the Invention

[0006] The problem the invention aims to solve

[0007] When an obstacle detection device is mounted on a moving body, if the tilt is corrected, the positional accuracy of the obstacle decreases, requiring further improvement of the obstacle detection device.

[0008] Solution for solving the problem

[0009] The obstacle detection device that solves the above problem is an obstacle detection device mounted on a moving body, which includes a camera and a control unit. The control unit acquires a video image from the camera, calculates the position of the obstacle based on the feature portion of the video image and the angle of the optical axis of the camera, and does not correct the angle of the optical axis when a specific condition is met, and corrects the angle of the optical axis when the specific condition is not met.

[0010] Under certain conditions, the angle of the optical axis is not corrected. Therefore, by correcting the angle of the optical axis under certain conditions, the decrease in the positional accuracy of obstacles can be suppressed.

[0011] Regarding the aforementioned obstacle detection device, the camera may also be a single-lens camera, and the optical axis may extend in a direction that intersects the ground on which the moving body is traveling.

[0012] Regarding the aforementioned obstacle detection device, the specific condition may also include: the moving body is in operation.

[0013] Regarding the aforementioned obstacle detection device, the moving body may also have an engine, and the specific condition includes: the vibration generated by the moving body is not in a stable state.

[0014] Invention Effects

[0015] According to the present invention, it is possible to suppress the decrease in the positional accuracy of obstacles. Attached Figure Description

[0016] Figure 1 This is a rough sketch of a forklift.

[0017] Figure 2 This is a schematic diagram of the forklift and obstacle detection device.

[0018] Figure 3 This is a diagram showing the optical axis of a camera.

[0019] Figure 4 It is a flowchart that shows the position of the obstacle and calculates the control.

[0020] Figure 5 This is a flowchart illustrating the correction control.

[0021] Explanation of reference numerals in the attached figures

[0022] 10… A forklift as an example of a moving body; 61… An obstacle detection device; 62… A camera; 63… An optical axis; 81… A control unit. Detailed Implementation

[0023] One embodiment of the obstacle detection device will be described.

[0024] like Figure 1 As shown, the forklift 10 includes a body 11, drive wheels 12, steering wheels 13, and a cargo loading / unloading device 21. The forklift 10 is an example of a mobile vehicle. The mobile vehicle can be an industrial vehicle that includes the forklift 10 or a trailer, or it can be a passenger vehicle.

[0025] The vehicle body 11 has a roof guard 14. The roof guard 14 is provided, for example, above the driver's seat.

[0026] The cargo loading and unloading device 21 includes a mast 22 and forks 23. The forks 23 move up and down by raising and lowering the mast 22. The forks 23 also move tilted by tilting the mast 22. Cargo is loaded onto the forks 23.

[0027] like Figure 2As shown, the forklift 10 includes an acceleration operation component 31, an acceleration sensor 32, a steering lever 33, a steering switch 34, a vehicle control device 41, an engine 51, a driving control device 52, a power transmission mechanism 53, a hydraulic pump 54, a hydraulic mechanism 55, a cargo loading / unloading lever 56, and a vehicle speed sensor 57.

[0028] Acceleration control component 31 is, for example, an accelerator pedal. Acceleration control component 31 may also be, for example, a lever. Acceleration sensor 32 detects the amount of operation of acceleration control component 31.

[0029] The steering lever 33 determines the direction of travel of the forklift 10. The steering lever 33 is operated by the rider of the forklift 10. The steering lever 33 is operated from a neutral position to either a forward position indicating forward movement or a reverse position indicating reverse movement. For example, the forward position is where the steering lever 33 is tilted forward from the neutral position. The reverse position is where the steering lever 33 is tilted backward from the neutral position.

[0030] The direction switch 34 switches according to the operating direction of the direction lever 33. The direction switch 34, for example, has three contacts. The direction lever 33 switches the contacts of the direction switch 34 depending on whether it is in a neutral, forward, or backward position.

[0031] The vehicle control device 41 includes a processor 42 and a storage unit 43. The processor 42 may be, for example, a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), or a DSP (Digital Signal Processor). The storage unit 43 includes RAM (Random Access Memory) and ROM (Read Only Memory). The storage unit 43 stores program code or instructions configured to cause the processor 42 to perform processing. The storage unit 43, i.e., computer-readable medium, includes any usable medium accessible to a general-purpose or special-purpose computer. The vehicle control device 41 may also be constructed from hardware circuits such as ASICs (Application Specific Integrated Circuits) or FPGAs (Field Programmable Gate Arrays). The vehicle control device 41, as a processing circuit, may include one or more processors, ASICs, or FPGAs, or combinations thereof, that operate according to a computer program.

[0032] Engine 51 is the driving source for the forklift 10's driving and loading / unloading operations.

[0033] The driving control device 52 is an engine control unit that controls the engine 51. For example, the driving control device 52 adjusts the throttle opening. By adjusting the throttle opening, the driving force of the engine 51 is adjusted.

[0034] The power transmission mechanism 53 transmits the driving force of the engine 51 to the drive wheels 12. The power transmission mechanism 53 includes, for example, a torque converter and a transmission.

[0035] Hydraulic pump 54 is driven by engine 51. Hydraulic pump 54 draws working oil from oil tank.

[0036] The hydraulic mechanism 55 distributes the working oil drawn by the hydraulic pump 54. For example, the hydraulic mechanism 55 distributes working oil to hydraulic cylinders such as tilt cylinders or lifting cylinders provided in the cargo loading and unloading device 21. A tilt cylinder is a hydraulic cylinder that tilts the mast 22. A lifting cylinder is a hydraulic cylinder that raises or lowers the mast 22.

[0037] The loading boom 56 is operated by the rider of the forklift 10. The loading boom 56 includes a lifting lever that is operated when the forks 23 are moved up and down, and a tilting lever that is operated when the mast 22 is tilted. The loading device 21 is operated by supplying working oil from the hydraulic mechanism 55 to the loading device 21 according to the operation of the loading boom 56.

[0038] The vehicle speed sensor 57 is used to detect the speed of the forklift 10. The vehicle speed sensor 57 outputs a pulse signal corresponding to the speed of the forklift 10. The detection result of the vehicle speed sensor 57 is acquired by the travel control device 52. Thus, the travel control device 52 can obtain the speed of the forklift 10.

[0039] The vehicle control device 41 is configured to communicate with the travel control device 52. The vehicle control device 41 can obtain the speed of the forklift 10 from the travel control device 52.

[0040] The vehicle control unit 41 obtains the operation amount of the acceleration operation member 31 from the acceleration sensor 32. The vehicle control unit 41 issues a command to the travel control unit 52 in such a way that the forklift 10 travels at a speed corresponding to the operation amount of the acceleration operation member 31.

[0041] The vehicle control unit 41 identifies the position of the steering lever 33 by identifying which contact of the steering switch 34 is being identified.

[0042] The vehicle control unit 41 identifies the loading and unloading status of the forklift 10. The loading and unloading status of the forklift 10 is, for example, whether the loading and unloading device 21 is in operation. The operation of the loading and unloading device 21 includes the tilting movement of the mast 22 and the raising and lowering of the forks 23.

[0043] Obstacle Detection Device

[0044] The forklift 10 is equipped with an obstacle detection device 61. The forklift 10 may be equipped with one obstacle detection device 61 or multiple obstacle detection devices 61.

[0045] The obstacle detection device 61 includes a camera 62, an inertial measurement unit 71, and a control unit 81. The camera 62 and the inertial measurement unit 71 are modular. Therefore, when the camera 62 is tilted, the inertial measurement unit 71 also tilts in the same way as the camera 62. The control unit 81 can also be modular with the camera 62 and the inertial measurement unit 71.

[0046] Camera 62 is a digital camera. Camera 62 has an image-capturing element. The image-capturing element is, for example, a CCD (Charge Coupled Device) image sensor or a CMOS (Complementary Metal Oxide Semiconductor) image sensor. Camera 62 is a single-lens reflex camera.

[0047] like Figure 3 As shown, the optical axis 63 of camera 62 extends in a direction intersecting the ground G on which the forklift 10 travels. Camera 62 is mounted, for example, on the overhead guard 14. Camera 62 is mounted facing downwards. Camera 62 can be mounted simply in the direction in which obstacle detection is desired. If it is desired to detect obstacles behind, camera 62 can also be mounted facing backwards. If it is desired to detect obstacles to the left or right, camera 62 can also be mounted facing left or right.

[0048] like Figure 2 As shown, the inertial measurement device 71 includes an accelerometer 72 and a gyroscope 73. The accelerometer 72 is a 3-axis accelerometer with three mutually orthogonal axes. The accelerometer 72 detects the acceleration acting on each axis. The gyroscope 73 is a 3-axis gyroscope with three mutually orthogonal axes. The gyroscope 73 detects the angular velocity acting on each axis.

[0049] The control unit 81, for example, has the same hardware configuration as the vehicle control device 41. The control unit 81 includes a processor 82 and a storage unit 83. The control unit 81 is configured to communicate with the vehicle control device 41 via a vehicle communication protocol. Therefore, the control unit 81 can obtain various information from the vehicle control device 41.

[0050] The control unit 81 calculates the position of the obstacle. The position of the obstacle is based on the position of the forklift 10, that is, the relative position of the forklift 10 and the obstacle. The control for calculating the position of the obstacle is explained.

[0051] <Obstacle Position Calculation Control>

[0052] like Figure 4 As shown, in step S1, the control unit 81 acquires a video image from the camera 62.

[0053] Next, in step S2, the control unit 81 calculates the position of the obstacle based on the feature portion of the camera image and the angle of the optical axis 63. The feature portion is the part used to extract objects from the camera image. The feature portion can be detected, for example, based on the brightness gradient of the camera image. The control unit 81 can also distinguish and extract people from other objects within the obstacle.

[0054] The control unit 81 converts the coordinates of obstacles in the camera image into the coordinates of obstacles in actual space based on the angle of the optical axis 63. This conversion can be performed using known techniques. The angle of the optical axis 63 is the angle between the optical axis 63 and the ground G on which the forklift 10 travels. In other words, it is the angle between the plane parallel to the ground G and the optical axis 63. The coordinates in actual space can be represented by orthogonal coordinates or polar coordinates. The coordinates in actual space can be two-dimensional or three-dimensional.

[0055] As described above, the control unit 81 calculates the position of the obstacle using the angle of the optical axis 63. The angle of the optical axis 63 can be predetermined. Therefore, by storing the angle of the optical axis 63 in advance in a storage medium such as the storage unit 83, the position of the obstacle can be calculated. Figure 3 As shown, when the angle of the optical axis 63 changes from a known angle θ1 to an unknown angle θ2, the actual position of the obstacle will differ from the position calculated by the control based on the obstacle's position. That is, the accuracy of the calculated obstacle position may decrease.

[0056] The control unit 81 corrects the angle of the optical axis 63 when the angle of the optical axis 63 changes from the known angle θ1. Therefore, even when the angle of the optical axis 63 changes from the known angle θ1, the error in the position of the obstacle can be reduced.

[0057] <Correction Control>

[0058] This describes the correction control for correcting the angle of the optical axis 63.

[0059] like Figure 5As shown, in step S11, the control unit 81 determines whether the forklift 10 is in motion. Whether the forklift 10 is in motion is determined solely by its speed. If the forklift 10's speed is above the motion determination threshold, the control unit 81 determines that the forklift 10 is in motion. If the forklift 10's speed is below the motion determination threshold, the control unit 81 determines that the forklift 10 is stationary. The forklift 10's speed is obtained from the vehicle control device 41 based on the detection result of the speed sensor 57. The motion determination threshold is set, for example, in the range of 0 [km / h] to 0.5 [km / h]. If the determination result in step S11 is negative, the control unit 81 proceeds to step S12. If the determination result in step S11 is positive, the control unit 81 proceeds to step S17.

[0060] In step S12, the control unit 81 determines whether the forklift 10 is loading or unloading goods. Whether the forklift 10 is loading or unloading goods is determined solely by the loading / unloading status. The loading / unloading status can be obtained from the vehicle control device 41. If the loading / unloading device 21 is operating, the vehicle control device 41 determines that the forklift 10 is loading or unloading goods. If the loading / unloading device 21 is not operating, the vehicle control device 41 determines that the forklift 10 is not loading or unloading goods. If the determination result in step S12 is negative, the control unit 81 proceeds to step S13. If the determination result in step S12 is positive, the control unit 81 proceeds to step S17.

[0061] In step S13, the control unit 81 determines whether the vibration generated by the forklift 10 is in a stable state. Vibration generated even when the moving body is not moving is considered a stable vibration. If the moving body, such as the forklift 10, is equipped with a cargo loading / unloading device 21, then vibration generated even when the moving body is neither moving nor loading / unloading cargo is considered a stable vibration. In contrast, vibration generated when an impact or vibration is applied to the forklift 10 from the outside, and when the resulting vibration does not converge, is considered an unstable vibration. When the vibration is not in a stable state, it is considered an unstable vibration. A stable vibration is, for example, vibration generated when the engine 51 is idling. Whether the vibration is in a stable state can be determined based on the detection results of the inertial measurement device 71.

[0062] If the amplitude of the vibration detected by the inertial measurement device 71 within the most recent specified frame is continuously below the vibration threshold within the specified frame, the control unit 81 determines that the vibration is in a stable state. If the amplitude of the vibration detected by the inertial measurement device 71 within the most recent specified frame is greater than the vibration threshold, the control unit 81 determines that the vibration is in an unstable state. The number of specified frames can be arbitrarily set. For example, the number of specified frames can be set such that when vibration or impact is applied to the forklift 10 from the outside, it can be determined whether the resulting vibration has converged.

[0063] The amplitude of the vibration detected by the inertial measurement device 71 can be determined by using at least one of the amplitudes of the acceleration detected by each axis of the accelerometer sensor 72 and the amplitudes of the angular velocity detected by each axis of the gyroscope sensor 73.

[0064] If the determination result in step S13 is positive, the control unit 81 proceeds to step S14. If the determination result in step S13 is negative, the control unit 81 proceeds to step S17.

[0065] In step S14, the control unit 81 estimates the angle of the optical axis 63. Gravitational acceleration acts on each axis of the accelerometer 72. Each axis of the accelerometer 72 detects the component of gravitational acceleration based on its tilt relative to the direction of gravity. Since the camera 62 and the inertial measurement device 71 are tilted together, if the angle of the optical axis 63 changes, the magnitude of the gravitational acceleration detected by each axis of the accelerometer 72 will change. When the forklift 10 is neither moving nor loading / unloading goods, the acceleration acting on the accelerometer 72 can be considered as gravitational acceleration. The control unit 81 estimates the angle of the optical axis 63 from the magnitude of the gravitational acceleration detected by each axis of the accelerometer 72.

[0066] The control unit 81 can also correct the estimated angle of the optical axis 63 by using the detection results of the gyroscope sensor 73. The control unit 81 can also estimate the change in the angle of the optical axis 63 from the acceleration detected by each axis of the accelerometer sensor 72. For example, the change in the angle of the optical axis 63 can be estimated from the difference between the acceleration detected by each axis of the accelerometer sensor 72 when the tilt of the optical axis 63 has not changed and the acceleration detected by each axis of the accelerometer sensor 72 when the tilt of the optical axis 63 has changed.

[0067] When estimating the angle of the optical axis 63, the influence of vibration can be eliminated using hardware or software filters. Sometimes, a steady-state vibration is applied to the forklift 10 when estimating the angle of the optical axis 63. In this case, the acceleration detected by the accelerometer 72 or the angular velocity detected by the gyroscope sensor 73 may contain noise. This noise can also be removed using hardware or software filters. Furthermore, the angle of the optical axis 63 can be estimated from the noise-removed acceleration or angular velocity.

[0068] Next, in step S15, the control unit 81 determines whether the forklift 10 is located on the ramp. Whether the forklift 10 is on the ramp can be determined based on the amount of change in the angle of the optical axis 63.

[0069] The reasons for the change in the angle of the optical axis 63 can be thought of as follows.

[0070] Changes in the mounting angle of camera 62 caused by: loosening of the bracket or bolts securing camera 62 due to years of deterioration; changes in the mounting angle of camera 62 caused by loads applied to camera 62 such as when a passenger of forklift 10 bumps into it; changes in the tilt angle of vehicle body 11 caused by tire wear of drive wheel 12 and steering wheel 13; changes in the tilt angle of vehicle body 11 caused by loading goods onto fork 23; changes in the tilt angle of vehicle body 11 caused by forklift 10 driving onto a step; and swaying of at least one of camera 62 and vehicle body 11 due to impact or vibration.

[0071] When the forklift 10 is on a ramp, the angle between the ground G and the optical axis 63 does not change from the known angle θ1. However, as the forklift 10 tilts with the ramp, the magnitude of the gravitational acceleration detected by each axis of the acceleration sensor 72 changes. Therefore, although the angle of the optical axis 63 relative to the ground G does not change, the angle of the optical axis 63 relative to the direction of gravity changes. Thus, when the angle of the optical axis 63 is estimated in step S14, a value different from the known angle θ1 is estimated.

[0072] If the change in the angle of the optical axis 63 exceeds a threshold, the control unit 81 determines that the forklift 10 is located on a ramp. The change in the angle of the optical axis 63 is the difference between the known angle θ1 and the estimated angle of the optical axis 63. The change in the angle of the optical axis 63 when the forklift 10 is on a ramp is greater than the change in the angle of the optical axis 63 caused by the aforementioned reasons. Therefore, the threshold is a value obtained by accumulating the change in the angle of the optical axis 63 caused by the aforementioned reasons using the square and square root.

[0073] If the determination result in step S15 is negative, the control unit 81 proceeds to step S16. If the determination result in step S15 is positive, the control unit 81 proceeds to step S17.

[0074] In step S16, the control unit 81 corrects the angle of the optical axis 63. For example, the control unit 81 calculates the position of the obstacle using the angle of the optical axis 63 estimated in step S14 in the obstacle position calculation control.

[0075] In step S17, the control unit 81 does not correct the angle of the optical axis 63. In this case, the control unit 81 can also calculate the position of the obstacle using the known angle θ1 of the optical axis 63 in the control. The control unit 81 can also calculate the position of the obstacle using the angle of the optical axis 63 that was recently corrected in step S16.

[0076] When the processing of step S16 or step S17 is completed, the control unit 81 returns to step S11. The calibration control is repeated during the start-up of the forklift 10.

[0077] As described above, the control unit 81 does not correct the angle of the optical axis 63 when certain conditions are met, and corrects the angle of the optical axis 63 when certain conditions are not met. The forklift 10 is considered to be in operation when it is traveling or when it is loading / unloading goods. Specific conditions include: the forklift 10 is in operation. Specific conditions include: the vibration generated by the forklift 10 is not in a stable state. Specific conditions include: the forklift 10 is located on a ramp. It is sufficient for the specific conditions to include at least one of these.

[0078] [The function of this implementation method]

[0079] When the control unit 81 corrects the angle of the optical axis 63, the positional accuracy of the obstacle may sometimes decrease due to the correction performed by the control unit 81. For example, when the forklift 10 is in motion or loading / unloading, the vibration generated by the forklift 10 is large. This vibration becomes an external disturbance, resulting in significant noise in the acceleration detected by the accelerometer 72 or the angular velocity detected by the gyroscope sensor 73. Therefore, even if the angle of the optical axis 63 is estimated using the accelerometer 72 or the gyroscope sensor 73, the deviation may become larger. The same applies when the vibration is not in a stable state.

[0080] like Figure 1As shown, when the forklift 10 is on the ramp, the angle between the ground G and the optical axis 63 remains unchanged. Therefore, by controlling the obstacle position calculation without correcting the angle of the optical axis 63, the position of the obstacle O on the ground G where the forklift 10 is located can be accurately calculated. When the angle of the optical axis 63 is corrected, the correction is performed based on gravitational acceleration. That is, when the angle of the optical axis 63 is corrected, the obstacle Oi is determined to be located on the plane Gi, which is orthogonal to gravitational acceleration.

[0081] As described above, specific conditions are defined as the presence of large noise in the detection results of the accelerometer 72 or the gyroscope 73, or the change in the angle between the optical axis 63 and the ground G, even if the angle between the optical axis 63 and the ground G remains unchanged. Furthermore, since the positional accuracy of the obstacle would decrease if the optical axis 63 were corrected under these specific conditions, it is assumed that the angle of the optical axis 63 will not be corrected under these conditions.

[0082] [Effects of this implementation method]

[0083] (1) Under certain conditions, the angle of the optical axis 63 is not corrected. Therefore, by correcting the angle of the optical axis 63 under certain conditions, the decrease in the positional accuracy of the obstacle can be suppressed.

[0084] (2) Camera 62 is a single-lens camera. Compared with using a stereo camera as camera 62, it can reduce manufacturing costs.

[0085] (3) Specific conditions include: the forklift 10 is in operation. When the forklift 10 is in operation, the detection results of the accelerometer 72 or the gyroscope sensor 73 contain a lot of noise. Therefore, by not correcting the optical axis 63 when the forklift 10 is in operation, it is possible to suppress the decrease in the positional accuracy of the obstacle caused by correcting the angle of the optical axis 63.

[0086] (4) Specific conditions include: the vibration generated by the forklift 10 is not in a stable state. When the vibration generated by the forklift 10 is not in a stable state, the detection results of the accelerometer 72 or the gyroscope sensor 73 contain a lot of noise. Therefore, by not correcting the optical axis 63 when the vibration generated by the forklift 10 is not in a stable state, it is possible to suppress the deterioration of the positional accuracy of the obstacle caused by correcting the angle of the optical axis 63.

[0087] (5) Specific conditions include: Forklift 10 is located on a ramp. When forklift 10 is located on a ramp, the position of the obstacle on the ground G where forklift 10 is located can be accurately calculated without correcting the angle of optical axis 63. Therefore, by not correcting optical axis 63 when forklift 10 is located on a ramp, the decrease in the positional accuracy of the obstacle caused by correcting the angle of optical axis 63 can be suppressed.

[0088] [Example of Change]

[0089] The implementation method can be modified as follows. The implementation method and the following variations can be combined with each other to the extent that they are not technically contradictory.

[0090] In step S11, the control unit 81 can also determine whether the forklift 10 is in motion based on the amount of operation of the acceleration operation member 31. The amount of operation of the acceleration operation member 31 can be obtained from the vehicle control device 41. If the amount of operation of the acceleration operation member 31 is above the acceleration operation amount threshold, the control unit 81 determines that the forklift 10 is in motion. The acceleration operation amount threshold is set to exclude minute operation amounts. The control unit 81 can also use both the vehicle speed and the amount of operation of the acceleration operation member 31 to make the determination in step S11.

[0091] When the acceleration operation member 31 is operated with the direction lever 33 in the neutral position, the forklift 10 will still vibrate even though it is not moving. If the angle of the optical axis 63 is corrected in this situation, the positional accuracy of obstacles may sometimes decrease due to the vibration. By determining the amount of operation of the acceleration operation member 31 in step S11, it is possible to prevent the angle of the optical axis 63 from being corrected when the acceleration operation member 31 is operated with the direction lever 33 in the neutral position.

[0092] In step S12, the control unit 81 can also determine whether the forklift 10 is loading or unloading goods based on the amount of operation of the loading / unloading lever 56. In this case, the forklift 10 is equipped with a sensor that detects the amount of operation of the loading / unloading lever 56. If the amount of operation of the loading / unloading lever 56 is above or equal to a loading / unloading operation amount threshold, the control unit 81 determines that the forklift 10 is loading or unloading goods. The loading / unloading operation amount threshold is set to exclude minute operations. The control unit 81 can also use both the loading / unloading status and the amount of operation of the loading / unloading lever 56 to make the determination in step S12.

[0093] In step S13, the control unit 81 can also estimate the angle of the optical axis 63 based on the detection results of the inertial measurement device 71, and determine that the vibration is in a stable state if the estimated angle of the optical axis 63 is below the angle threshold for a specified frame. The angle threshold is a predetermined value. The angle threshold is set so that it is possible to determine whether the vibration is in a stable or unstable state. If the noise caused by the vibration is included in the detection results of the accelerometer sensor 72 or the gyroscope sensor 73, the estimated angle of the optical axis 63 becomes larger, and thus the angle of the optical axis 63 exceeds the angle threshold. Therefore, by using the angle threshold, it is possible to determine whether the vibration is in a stable state.

[0094] Alternatively, step S13 can be omitted by setting a low-pass filter to remove large vibrations or shocks. In this case, the processing speed of the correction control can be improved. Furthermore, installation becomes easier.

[0095] In step S15, the control unit 81 can also determine whether the forklift 10 is located on the ramp based on the result of its own position estimation. For example, the control unit 81 can simply compare the map data recording the ramp location in the area where the forklift 10 is used with its own position on the map data estimated by its own position estimation. The self-position estimation can also be performed by the vehicle control device 41. The map data is stored in a storage device that can be read by the control unit 81 or the vehicle control device 41.

[0096] In step S15, the control unit 81 can also determine whether the forklift 10 is located on the ramp based on the detection results of the ground distance sensors. The ground distance sensors are, for example, installed at two locations separated in the forward and backward direction of the forklift 10. When moving from flat ground to the ramp, a difference arises in the distance to the ground G detected by the two ground distance sensors. The control unit 81 can determine whether the forklift 10 is located on the ramp simply based on the difference in distance to the ground G detected by the two ground distance sensors.

[0097] ○ The angle of the optical axis 63 can also be estimated when the forklift 10 is shipped from the factory, with the forklift 10 positioned on a flat platform. Furthermore, this angle of the optical axis 63 can be stored as a known angle θ1 in the storage unit 83. Compared to the case where the operator measures the angle of the optical axis 63, the efficiency of the operation can be improved.

[0098] ○ Camera 62 can also be a stereo camera.

[0099] ○ Forklift 10 can also be a forklift that is driven by a motor.

[0100] [definition]

[0101] The term "at least one" as used in this specification means "more than one" of the desired options. As an example, if the number of options is two, then "at least one" as used in this specification means "only one option" or "both options". As another example, if the number of options is three or more, then "at least one" as used in this specification means "only one option" or "any combination of two or more options".

Claims

1. An obstacle detection device, which is an obstacle detection device mounted on a moving body, characterized in that, have: Camera; and Control Department The control unit Acquire video images from the camera. The position of the obstacle is calculated based on the feature portion of the captured image and the angle of the camera's optical axis. The angle of the optical axis is not corrected when certain conditions are met, and the angle of the optical axis is corrected when the specific conditions are not met.

2. The obstacle detection device according to claim 1, wherein, The camera in question is a single-lens reflex camera. The optical axis extends in a direction that intersects the ground on which the moving body is traveling.

3. The obstacle detection device according to claim 1, wherein, The specific condition includes: the moving body is in operation.

4. The obstacle detection device according to claim 1, wherein, The mobile body is equipped with an engine. The specific condition includes: the vibration generated by the moving body is not in a stable state.

Images