JP5748314B1 - Cargo handling vehicle - Google Patents

Abstract

A cargo handling vehicle capable of improving the detection accuracy of an obstacle is provided. A forklift that is a cargo handling vehicle includes a vehicle main body, a camera that captures an obstacle and a road surface, and reference areas B1 to B4 and detection target areas F1 to F8, R1 based on camera images captured by the camera. A region extraction unit that extracts .about.R8, a reference gradation degree calculation unit that calculates a reference gradation degree for distinguishing between a road surface area and a non-road surface area based on the gradation degrees of the reference areas B1 to B4, and a reference gradation degree. And an obstacle detection unit that detects an obstacle by detecting non-road surface areas included in the detection target areas F1 to F8 and R1 to R8. The reference areas B1 to B4 are a plurality of areas provided apart from each other in the camera image, and the detection target areas F1 to F7 and R1 to R7 are arranged in the direction in which the vehicle main body 10 travels in the camera image and from the vehicle main body 10. It is a plurality of areas that become smaller as the distance increases. [Selection] Figure 3

Description

  The present invention relates to a cargo handling vehicle such as a forklift that detects an obstacle that hinders travel of a vehicle body.

  Forklifts equipped with obstacle detection sensors are known as cargo handling vehicles for detecting obstacles. Patent Documents 1 and 2 include an obstacle detection unit for detecting an obstacle present in front of the vehicle body. A provided forklift is described.

  The forklift disclosed in Patent Document 1 includes sensors provided at the lower part of the vehicle body and at the tip of the fork. Paragraph [0023] of Patent Document 1 describes that these sensors are configured by photoelectric sensors or ultrasonic sensors.

  In addition, the forklift disclosed in Patent Document 2 includes a front obstacle detector that detects an obstacle existing below the fork and in front of the vehicle body. In paragraph [0015] of Patent Document 2, it is described that a video camera is used as an obstacle detector, and it is suggested that an interval from the obstacle is detected by processing an image obtained by the video camera. Has been.

JP 2011-46451 A Japanese Patent Laid-Open No. 2004-1992

  By the way, since the photoelectric sensor and the ultrasonic sensor described in Patent Document 1 have a disadvantage that the detection range of the obstacle is narrow, it is preferable to use a camera in order to detect the obstacle from a wide range. In order to detect obstacles on the road surface using the camera, the road image area and obstacles in which the road surface is shown are included from the camera image, which is an image taken by the camera, based on a predetermined reference gradation level. It is necessary to detect an obstacle by distinguishing it from a non-road surface area. And in order to distinguish a road surface area | region and a non-road surface area | region accurately, it is necessary to calculate the reference | standard gradation degree suitable for the road surface where a cargo handling vehicle drive | works.

  However, when calculating the reference gradation level from the gradation level included in the camera image, the camera image may include a non-road surface area where the road surface is not reflected, and the reference gradation level suitable for the road surface cannot be calculated. There was a case. Therefore, there is a problem that the detection accuracy of the obstacle is lowered.

  The present invention has been made in view of the above circumstances, and an object of the present invention is to provide a cargo handling vehicle capable of increasing the detection accuracy of an obstacle in the detection of an obstacle using a camera.

  In order to solve the above-described problem, a cargo handling vehicle according to the first aspect of the present invention includes a vehicle body that can travel on a road surface, an obstacle that prevents the vehicle body from traveling, a camera that captures the road surface, and the camera A reference region that is a part of the camera image photographed by the camera image is extracted from the camera image, and a region extraction unit that extracts a detection target region that is a part of the camera image from the camera image; and a gradation level of the reference region A reference gradation degree calculating unit for calculating a reference gradation degree for distinguishing between a road surface area where the road surface is reflected and a non-road surface area where the road surface is not reflected, and the detection based on the reference gradation degree An obstacle detection unit that detects the obstacle by detecting the non-road surface region included in the target region, and the region extraction units are separated from each other in the camera image. A plurality of regions to be extracted from the camera image as the reference region, and a plurality of regions arranged in the camera image in a direction in which the vehicle main body travels and become smaller from the vehicle main body as the detection target region. It extracts from the said camera image, It is characterized by the above-mentioned.

  Further, the cargo handling vehicle according to claim 2 is the cargo handling vehicle according to claim 1, wherein the camera is configured by an omnidirectional camera capable of photographing a range of 360 ° when viewed from the vertical direction. And

  Further, the cargo handling vehicle according to claim 3 is the cargo handling vehicle according to claim 1 or 2, wherein the region extraction unit is a front region located in front of the vehicle main body and a rear located in the rear of the vehicle main body. An area is configured to be extracted from the camera image as the detection target area, and the obstacle detection unit is configured to be able to detect the obstacle existing in front of and behind the vehicle body. It is characterized by.

  In addition, the cargo handling vehicle according to claim 4 is the cargo handling vehicle according to claim 3, wherein the region extraction unit performs either one of the front region and the rear region according to a traveling mode of the vehicle body. The obstacle detection unit is extracted from the camera image as the detection target region, and the obstacle detection unit detects the obstacle present in front of or behind the vehicle main body according to a traveling mode of the vehicle main body. .

  Further, the cargo handling vehicle according to claim 5 is the cargo handling vehicle according to any one of claims 1 to 4, wherein the region extraction unit is located closer to the vehicle body than the detection target region. A region is extracted from the camera image as the reference region.

  In addition, the cargo handling vehicle according to claim 6 is the cargo handling vehicle according to any one of claims 1 to 5, wherein the obstacle detection unit is a pixel indicating the non-road surface area included in the detection target area. The non-road surface area is detected based on the fact that the number of the above is a predetermined ratio or more with respect to the total number of pixels of the detection target area.

  Further, the cargo handling vehicle according to claim 7 is the cargo handling vehicle according to any one of claims 1 to 6, wherein the detection target area in which the non-road surface area is detected in the camera image is set as a detection block, The obstacle detection unit detects the obstacle based on a change in the number of the detection target areas adjacent to each other constituting the detection block as the vehicle main body travels. To do.

  According to the present invention, it is possible to provide a cargo handling vehicle capable of increasing the accuracy of detecting an obstacle in the detection of an obstacle using a camera.

BRIEF DESCRIPTION OF THE DRAWINGS It is a figure which simplifies and shows the external appearance of the cargo handling vehicle which concerns on one Embodiment of this invention, Comprising: (A) is a side view, (B) is a top view. It is a block diagram which shows the structure of the cargo handling vehicle which concerns on the same embodiment. It is a schematic diagram which shows an example of the image image | photographed with the camera which concerns on the embodiment. (A) And (B) is a schematic diagram for demonstrating the detection of the obstruction based on the detection block which concerns on the same embodiment. It is a flowchart which shows the flow of the periphery monitoring process which the cargo handling vehicle which concerns on the embodiment performs.

  An embodiment of a cargo handling vehicle according to the present invention will be described with reference to the drawings. In the drawings, a front-rear direction X indicated by an arrow X, a left-right direction Y indicated by an arrow Y, and a vertical direction Z indicated by an arrow Z are directions orthogonal to each other.

As shown in FIG. 1, the reach-type forklift 1 is a cargo handling vehicle including a vehicle body 10, a traveling device 20, a cargo handling device 30, an operation device 40, a camera 50, and the like.
The vehicle main body 10 houses a battery (not shown) that is a power source of the traveling device 20 and the cargo handling device 30. The vehicle main body 10 is provided with a riding section 11 on which an operator (not shown) driving the forklift 1 rides, and a head guard 12 disposed above the riding section 11. The head guard 12 is constituted by a lattice-like structure, and is supported by a stay 12 </ b> A provided to stand on the vehicle main body 10. In FIG. 1B, the head guard 12 is not shown. The vehicle body 10 is configured to be able to travel in the front-rear direction X on the road surface S when the traveling device 20 operates.

  The traveling device 20 includes a front wheel 21, a rear wheel 22, a traveling motor (not shown) that drives the rear wheel 22, a steering device (not shown) that changes the direction of the rear wheel 22, and the like. The cargo handling device 30 is provided in front of the vehicle main body 10 and includes a mast 31 that can be expanded and contracted in the vertical direction Z and a fork 32 that can be moved up and down along the mast 31. The traveling device 20 and the cargo handling device 30 are configured to operate based on instructions from an operator input using the operation device 40.

  The operating device 40 is provided in the vehicle body 10 and includes a steering handle 41, a travel lever 42, a brake pedal (not shown) that is a deadman brake, and the like. When the operator rotates the steering handle 41, an instruction to change the direction of the rear wheel 22 is input to the forklift 1. Further, when the operator tilts the traveling lever 42 forward, an instruction to drive the rear wheel 22 so that the vehicle body 10 advances forward is input to the forklift 1, and when the operator tilts the traveling lever 42 backward, An instruction to drive the rear wheel 22 so that the main body 10 moves rearward is input to the forklift 1. When the operator depresses the brake pedal, an instruction to release the braking of the travel of the vehicle body 10 is input to the forklift 1.

  The camera 50 is provided on the head guard 12 and is composed of an omnidirectional camera capable of photographing a range of 360 ° as viewed from the vertical direction. The camera 50 images the road surface S existing below the camera 50 from above the riding section 11 and also images obstacles that hinder the traveling of the vehicle body 10 existing on the road surface S. The camera 50 captures the road surface S and the like in color, and acquires a two-dimensional image composed of a plurality of pixels arranged along a first coordinate axis and a second coordinate axis that are orthogonal to each other. The camera 50 repeatedly acquires the above-described image showing the road surface S and the like at a minute time interval (for example, about 33 milliseconds).

With reference to FIG. 2, a configuration for detecting an obstacle from an image photographed by the camera 50 (hereinafter referred to as “camera image”) will be described.
As shown in FIG. 2, the forklift 1 includes a preprocessing unit 61 that performs predetermined preprocessing on a camera image, an area extraction unit 62 that extracts a predetermined area from the camera image, and a reference for calculating a reference gradation level. A gradation degree calculation unit 63 and an obstacle detection unit 64 that detects an obstacle are provided.

  The preprocessing unit 61 is configured by an integrated circuit that performs predetermined preprocessing on the camera image. Specifically, the preprocessing unit 61 performs a process of converting a color image into a grayscale image on the camera image. Each of the pixels constituting the gray scale image has a pixel value included in a range of 0 to 255, for example. The pixel value represents the gradation level.

  The region extraction unit 62 is configured by an integrated circuit that partially extracts image data relating to a camera image. The area extraction unit 62 extracts reference areas B1 to B4 (see FIG. 3) which are a part of the camera image from the camera image, and further, detection target areas F1 to F8 and R1 to R8 which are a part of the camera image (FIG. 3). Reference) is extracted from the camera image. The area extraction unit 62 extracts the detection target areas F1 to F8 from the camera image when the vehicle main body 10 travels forward among the detection target areas F1 to F8 and R1 to R8, and the vehicle main body 10 travels backward. Sometimes, the detection target areas R1 to R8 are extracted from the camera image. That is, the region extraction unit 62 extracts one of the detection target regions F1 to F8 or the detection target regions R1 to R8 from the camera image according to the traveling mode of the vehicle body 10. The reference areas B1 to B4 and the detection target areas F1 to F8 and R1 to R8 will be described later with reference to FIG.

  The reference gradation degree calculation unit 63 is configured by an integrated circuit that calculates the reference gradation degree based on the gradation degrees of the reference regions B1 to B4. The reference gradation level is a gradation level for distinguishing between a road surface area where the road surface S is shown and a non-road surface area where the road surface S is not shown (that may include an obstacle) in the camera image. Specifically, for example, the reference gradation degree calculation unit 63 calculates the average pixel value of the pixels included in the reference areas B1 to B4 as the gradation degree of the road surface area, and a predetermined range including the gradation degree of the road surface area. The gradation level is set as the reference gradation level. For example, if the road surface area has a gradation of 20, 15 to 25 is calculated as the reference gradation.

  The obstacle detection unit 64 is configured by an integrated circuit that detects an obstacle by detecting non-road surface areas included in the detection target areas F1 to F8 and R1 to R8 based on the reference gradation. The obstacle detection unit 64 performs binarization processing, non-road surface area detection processing, and detection block comparison processing as the obstacle detection processing. The binarization process is a process for binarizing the pixel values constituting the detection target areas F1 to F8 and R1 to R8 that are part of the camera image based on the reference gradation. In the binarization process, for example, when the reference gradation is 15 to 25, the pixel values of 15 to 25 indicating the road surface area are converted to 0, and the other pixel values indicating the non-road surface area are set to 1. Converted. In the non-road surface area detection process, the number of pixels indicating the non-road surface area included in each of the detection target areas F1 to F8 and R1 to R8 is equal to the total number of pixels of the detection target areas F1 to F8 and R1 to R8. This is a process for detecting a non-road surface area on the basis of the fact that the ratio is not less than a predetermined ratio. In the non-road surface area detection process, for example, when the total number of pixels in the detection target area F1 is 100,000 pixels, the number of pixels having a pixel value of 1 is 80,000 pixels (80% of 100,000 pixels) or more. In some cases, a non-road surface area is detected from the detection target area F1. The detection of the non-road surface area performed in this way is similarly performed in each of the other detection target areas F2 to F8 and R1 to R8. The detection of the non-road surface area is performed except for the area where a part of the forklift 1 is shown and the area where the operator is highly likely to be reflected from the detection target areas F1 to F8 and R1 to R8. The detection block comparison process compares a detection target area (hereinafter referred to as “detection block”) in which a non-road surface area is detected in two camera images photographed at a minute time interval, and detects an obstacle based on a change in the detection block. It is a process to detect. The obstacle detection based on the change of the detection block will be described later with reference to FIG. In the non-road surface area detection process, the obstacle detection unit 64 detects the non-road surface area from the detection target areas F1 to F8 when the vehicle body 10 travels forward among the detection target areas F1 to F8 and R1 to R8. When the vehicle body 10 travels backward, the non-road surface area is detected from the detection target areas R1 to R8. That is, the obstacle detection unit 64 detects an obstacle from one of the detection target areas F1 to F8 or the detection target areas R1 to R8 according to the traveling mode of the vehicle body 10.

In addition, the forklift 1 includes a control unit 71 that controls a predetermined operation based on an obstacle detection result by the obstacle detection unit 64, and a notification unit 72 that performs a notification operation.
The control unit 71 is configured by an integrated circuit that controls the notification unit 72. When the obstacle detection unit 64 detects an obstacle, the control unit 71 causes the notification unit 72 to perform a notification operation.

  The notification unit 72 includes a speaker and a warning lamp. The notification unit 72 notifies the operator that an obstacle that hinders the traveling of the vehicle body 10 has been detected by emitting sound and light.

With reference to FIG. 3, camera images, reference regions B1 to B4, detection target regions F1 to F8, and R1 to R8 will be described.
As shown in FIG. 3, an operator who is on boarding section 11 located below head guard 12 is shown in the center of the camera image. The camera image shows a road surface S on which the vehicle body 10 travels, and a portion of the forklift 1 such as the head guard 12, the stay 12A, and the mast 31.

  The reference areas B1 to B4 are rectangular areas that are set in advance as highly likely areas where the road surface S is reflected in the camera image. The reference areas B1 to B4 are a plurality of areas provided apart from each other in the camera image. The reference areas B1 and B3 correspond to places on the road surface S that are located on the left side of the vehicle body 10 and are separated from each other in the front-rear direction X. Further, the reference areas B2 and B4 correspond to places on the road surface S that are located to the right of the vehicle body 10 and that are located apart from each other in the front-rear direction X. The reference areas B1 and B2 are areas closer to the vehicle main body 10 than the detection target areas F1 to F8, and the reference area B3 is closer to the vehicle main body 10 than the detection target areas R1 to R8. It is an area to do.

  The detection target areas F1 to F8 and R1 to R8 are rectangular areas that are set in advance as highly probable areas in which obstacles that obstruct the traveling of the vehicle main body 10 appear in the camera image. The detection target areas F1 to F8 and R1 to R8 are areas arranged in the front-rear direction X in which the vehicle body 10 travels. The detection target areas F <b> 1 to F <b> 8 are front areas located in front of the vehicle main body 10, and the detection target areas R <b> 1 to R <b> 8 are rear areas located behind the vehicle main body 10. The detection target areas F <b> 1 to F <b> 7 are a plurality of areas that become smaller in the camera image as the distance from the vehicle body 10 (that is, toward the front). Similarly, the detection target areas R1 to R7 are a plurality of areas that become smaller in the camera image as they move away from the vehicle body 10 (that is, toward the rear). That is, when two areas adjacent to each other among the detection target areas F1 to F7 or the detection target areas R1 to R7 are defined as the areas A1 and A2, the size in the y direction of the area A1 located away from the vehicle body 10 is The area A2 located near the vehicle body 10 is smaller than the area A1 in the y direction, and the total number of pixels constituting the area A1 is smaller than the total number of pixels constituting the area A2. . Specifically, for example, the total number of pixels in each of the detection target regions F1 to F7 or the detection target regions R1 to R7 is about 104,000 pixels, about 58,000 pixels, and about 31,000 pixels. About 14,000 pixels, about 90,000 pixels, about 77,000 pixels, and about 55,000 pixels. Note that each of the detection target regions F1 to F7 and R1 to R7 has the same dimension on the road surface S and corresponds to a place positioned at equal intervals in the front-rear direction X.

Next, detection of an obstacle based on the change of the detection block will be described with reference to FIG.
For example, as shown in FIG. 4A, when an obstacle is present at a location corresponding to the detection target region R3, the obstacle has a three-dimensional shape. Therefore, the camera image detects not only the detection target region R3 but also the detection target region R3. A non-road surface area is also detected from the target areas R4 and R5. That is, in FIG. 4A, the detection target regions R3 to R5 constitute a detection block. When the vehicle main body 10 travels backward from the state shown in FIG. 4A, the obstacle moves relatively to a location corresponding to the detection target region R2, as shown in FIG. 4B. At this time, a non-road surface area is detected from the detection target areas R2 and R3 in the camera image. That is, in FIG. 4B, the detection target regions R2 and R3 constitute a detection block. The obstacle detection unit 64 detects the obstacles existing on the road surface S by detecting the change in the position and the number of the areas constituting the detection block in this way, and detects a three-dimensional obstacle and a two-dimensional obstacle. Distinguish from obstacles.

  Next, with reference to FIG. 5, the flow of the surrounding monitoring process performed by the forklift 1 will be described. The surrounding monitoring process is started when an instruction to drive the vehicle body 10 forward or backward is input to the forklift 1 by the operator operating the traveling lever 42.

  In the surrounding monitoring process, first, the camera 50 acquires two images including the road surface S photographed at a minute time interval (step S1), and then the preprocessing unit 61 converts each color camera image into a grayscale image. Preprocessing including processing for conversion into a camera image is performed (step S2).

  Next, the area extraction unit 62 extracts the reference areas B1 to B4 from each grayscale camera image (step S3), and further, the detection target areas F1 to F8 or the detection target areas positioned in the direction in which the vehicle body 10 travels. R1 to R8 are extracted (step S4). In step S4, when an instruction to travel the vehicle body 10 forward is input to the forklift 1, the detection target areas F1 to F8 are extracted, and an instruction to travel the vehicle body 10 backward is input to the forklift 1. If so, detection target regions R1 to R8 are extracted.

  Next, the reference gradation degree calculation unit 63 calculates the reference gradation degree of each camera image based on the gradation degrees of the reference regions B1 to B4 of each camera image (step S5). Then, the obstacle detection unit 64 performs the above-described obstacle detection process (step S6).

  Based on the result of the obstacle detection process in step S6, the controller 71 determines whether or not an obstacle has been detected by the obstacle detector 64 (step S7). If it is not determined in step S7 that an obstacle has been detected, that is, if no obstacle has been detected, the processing from step S1 onward is repeated.

  On the other hand, if it is determined in step S7 that an obstacle has been detected, that is, if an obstacle has been detected, the control unit 71 causes the notification unit 72 to start a notification operation to indicate that the obstacle has been detected. To the operator (step S8).

In the forklift 1 of the present embodiment, the following effects are obtained.
(1) The forklift 1 includes a vehicle main body 10, a camera 50, a region extraction unit 62 that extracts the reference regions B1 to B4 and the detection target regions F1 to F8 and R1 to R8 from the camera images taken by the camera 50, A reference gradation degree calculation unit 63 and an obstacle detection unit 64 are provided. The region extraction unit 62 extracts a plurality of regions provided apart from each other in the camera image as reference regions B1 to B4 from the camera image, and further aligns in the direction in which the vehicle main body 10 travels and moves away from the vehicle main body 10 in the camera image. A plurality of regions F1 to F7 and R1 to R7 that become smaller as the detection target regions are extracted from the camera image. According to the above configuration, since the obstacle is detected by detecting the non-road surface area, the obstacle detection accuracy is accurately detected from the camera image with high probability of the non-road surface area in which the obstacle is reflected. Will increase. The reference gradation level for detecting the non-road surface area is calculated based on the gradation levels of a plurality of areas (reference areas B1 to B4) provided apart from each other in the camera image. For this reason, even when the road surface is not shown in one area constituting the reference areas B1 to B4, it is possible to calculate the reference gradation that can distinguish the road area from the non-road area with high accuracy. Therefore, the non-road surface area can be detected from the camera image with high accuracy based on the reference gradation, and the obstacle detection accuracy can be increased. Further, the non-road surface area is detected from a plurality of areas (detection target areas F1 to F7, R1 to R7) that are arranged in the direction in which the vehicle main body 10 travels in the camera image and become smaller as the distance from the vehicle main body 10 increases. Therefore, since the obstacle can be detected from the detection target areas F1 to F7 and R1 to R7 that are small with respect to the entire camera image, the load for detecting the obstacle can be reduced.

  (2) The camera 50 is configured by an omnidirectional camera that captures a 360 ° range as viewed from the vertical direction. For this reason, it becomes possible to detect an obstacle from a wide range without using a plurality of cameras.

  (3) The region extracting unit 62 is configured to extract from the camera image, as detection target regions, the front regions F1 to F8 located in front of the vehicle body 10 and the rear regions R1 to R8 located behind the vehicle body 10. The obstacle detection unit 64 is configured to be able to detect obstacles existing in front and rear of the vehicle body 10. According to this configuration, it is possible to detect an obstacle from a wide range compared to a configuration in which an obstacle existing in either the front or rear of the vehicle body 10 is detected.

  (4) The region extraction unit 62 extracts any one of the front regions F1 to F8 and the rear regions R1 to R8 from the camera image as a detection target region according to the traveling mode of the vehicle body 10, and the obstacle detection unit 64 Detects an obstacle present in front of or behind the vehicle main body 10 according to the traveling mode of the vehicle main body 10. According to this configuration, when the vehicle body 10 travels forward by detecting an obstacle according to the switching of the travel mode of the vehicle body 10, the obstacle present in front of the vehicle body 10 is detected. When the vehicle body 10 travels backward, an obstacle present behind the vehicle body 10 can be detected. Since the non-road surface area is detected from the front area or the rear area extracted according to the traveling mode of the vehicle body 10, both the front areas F1 to F8 and the rear areas R1 to R8 are irrespective of the traveling mode of the vehicle body 10. As compared with the configuration in which the non-road surface area is detected, the load for detecting the obstacle can be reduced.

  (5) The region extraction unit 62 extracts regions located closer to the vehicle body 10 than the detection target regions F1 to F8 as reference regions B1 and B2 from the camera image. In addition, the region extraction unit 62 extracts a region located closer to the vehicle body 10 than the detection target regions R1 to R8 as a reference region B3 from the camera image. Considering that the forklift 1 is operated so that the vehicle main body 10 does not come into contact with an obstacle, the probability that no obstacle exists near the vehicle main body 10 is higher than the position away from the vehicle main body 10. . Therefore, according to the above configuration, the road region is included in the reference region as compared to the configuration in which the region located farther from the vehicle body 10 than the detection target regions F1 to F8 and R1 to R8 is extracted as the reference region. Therefore, it is possible to calculate the reference gradation that enables the road surface area and the non-road surface area to be accurately distinguished.

  (6) The obstacle detection unit 64 is configured such that the number of pixels indicating the non-road surface area included in the detection target area (for example, the detection target area F1) is a predetermined ratio or more with respect to the total number of pixels in the detection target area. Based on a certain thing, a non-road surface area is detected. According to this configuration, it is possible to detect an obstacle without performing pattern recognition, and the load for detecting the obstacle can be further reduced.

  (7) The detection target area in which the non-road surface area is detected in the camera image is set as a detection block, and the obstacle detection unit 64 detects the detection target areas adjacent to each other constituting the detection block as the vehicle body 10 travels. An obstacle is detected based on the change in position and number. According to this configuration, an obstacle having a three-dimensional shape approaching the vehicle main body 10 can be accurately detected based on a comparison between two camera images taken while the vehicle main body 10 is traveling.

  The present invention is not limited to the above-described embodiment, and the configuration of the above-described embodiment can be changed as appropriate. For example, the above configuration can be changed as follows, and the following changes can be combined.

  As long as an obstacle can be detected from the detection target region, the obstacle detection process may be performed by a method other than the process described in the above embodiment. For example, the obstacle detection unit 64 detects the non-road surface area from a plurality of detection target areas based on the detection of the non-road surface area from the detection target areas without being based on the change in the position and number of detection blocks. An obstacle can also be detected based on what has been done. Further, for example, the obstacle detection unit 64 can detect an obstacle by performing pattern recognition.

  The region extraction unit 62 can extract only one of the front region located in front of the vehicle main body 10 and the rear region located behind the vehicle main body 10 from the camera image as a detection target region. That is, the obstacle detection unit 64 can also be configured to detect an obstacle present in either the front or the rear of the vehicle body 10.

  If the obstacle can be detected, the number, arrangement, and size of the detection target areas extracted from the camera image may be changed as appropriate. That is, a plurality of detection target areas F1 to F7 and R1 to R7 excluding a part of the detection target areas F8 and R8 can be made smaller areas as they move away from the vehicle main body 10, and the vehicle main body 10 travels in the camera image. It is also possible to make all the detection target areas arranged in the direction to be smaller as the distance from the vehicle main body 10 decreases.

  If the non-road surface area can be detected with high accuracy, the number, arrangement, and size of the reference areas extracted from the camera image may be appropriately changed. For example, a configuration in which two, three, or five or more regions are extracted from the camera image as the reference region may be employed. In addition, the calculation method of the reference gradation may be changed to a method other than that described in the above embodiment. For example, the degree of gradation of each of the plurality of reference areas is calculated, and based on these degrees of gradation, a reference area having a higher probability of showing the road surface is determined, and the probability that the road surface is reflected among the plurality of reference areas. It is also possible to calculate the reference gradation level based on the gradation level of a higher region.

  -As long as the road surface S and an obstruction can be image | photographed, you may change the structure which concerns on the camera 50 suitably. For example, the arrangement and the number of cameras 50 can be changed. Moreover, the camera 50 can also be comprised by cameras other than an omnidirectional camera.

  -The operation | movement which the alerting | reporting part 72 performs can also be changed suitably. For example, the notification unit 72 can notify the operator that either an obstacle or a sound that hinders the traveling of the vehicle body 10 has been detected.

  -The operation | movement which the control part 71 controls based on the detection result of an obstruction can also be changed suitably. For example, the control unit 71 may control the operation of the traveling device 20, and when an obstacle is detected by the obstacle detection unit 64, the control unit 71 sets the vehicle body to the obstacle regardless of the operation of the brake pedal. The traveling device 20 can also be controlled so that 10 does not collide.

  -The structure which concerns on the vehicle main body 10, the traveling apparatus 20, the cargo handling apparatus 30, and the operating device 40 of the forklift 1 can also be changed suitably. For example, the present invention can be applied to a counterbalance type forklift, and the present invention can also be applied to a cargo handling vehicle that does not include the fork 32.

1 Forklift (handling vehicle)
DESCRIPTION OF SYMBOLS 10 Vehicle main body 11 Boarding part 12 Head guard 12A Stay 20 Traveling apparatus 21 Front wheel 22 Rear wheel 30 Cargo handling apparatus 31 Mast 32 Fork 40 Operation apparatus 41 Steering handle 42 Traveling lever 50 Camera 61 Preprocessing part 62 Area extraction part 63 Reference gradation degree Calculation unit 64 Obstacle detection unit 71 Control unit 72 Notification unit B1 to B4 Reference region F1 to F8 Detection target region (front region)
R1 to R8 detection target area (rear area)
S road surface X longitudinal direction Y lateral direction Z vertical direction

Claims (7)

A vehicle body capable of traveling on the road surface;
A camera that captures obstacles that obstruct travel of the vehicle body and the road surface;
A region extraction unit that extracts a reference region that is a part of a camera image photographed by the camera from the camera image, and further extracts a detection target region that is a part of the camera image from the camera image;
A reference gradation degree calculation unit for calculating a reference gradation degree for distinguishing between a road surface area where the road surface is reflected and a non-road surface area where the road surface is not reflected, based on the gradation degree of the reference area;
An obstacle detection unit that detects the obstacle by detecting the non-road surface area included in the detection target area based on the reference gradation degree;
The region extraction unit extracts a plurality of regions provided apart from each other in the camera image as the reference region from the camera image, and further aligns in the direction in which the vehicle body travels in the camera image and from the vehicle body. A cargo handling vehicle characterized in that a plurality of areas that become smaller as they move away are extracted from the camera image as the detection target area. The cargo handling vehicle according to claim 1, wherein the camera is configured by an omnidirectional camera capable of photographing a range of 360 ° when viewed from the vertical direction. The region extraction unit is configured to extract a front region located in front of the vehicle body and a rear region located behind the vehicle body as the detection target region from the camera image,
The cargo handling vehicle according to claim 1 or 2, wherein the obstacle detection unit is configured to be able to detect the obstacle existing in front and rear of the vehicle main body. The region extraction unit extracts one of the front region and the rear region from the camera image as the detection target region according to a traveling mode of the vehicle body,
The cargo handling vehicle according to claim 3, wherein the obstacle detection unit detects the obstacle present in front of or behind the vehicle main body according to a traveling mode of the vehicle main body. The said area extraction part extracts the area | region located near the said vehicle main body compared with the said detection object area | region from the said camera image as said reference | standard area. The Claim 1 characterized by the above-mentioned. Cargo handling vehicle. The obstacle detection unit, based on the fact that the number of pixels indicating the non-road surface area included in the detection target area is a predetermined ratio or more with respect to the total number of pixels of the detection target area, An area is detected. The cargo handling vehicle according to any one of claims 1 to 5. The detection target area where the non-road surface area is detected in the camera image is a detection block,
The obstacle detection unit detects the obstacle based on a change in the position or number of the detection target areas adjacent to each other constituting the detection block as the vehicle main body travels. The cargo handling vehicle according to claim 1, wherein the cargo handling vehicle is a vehicle.

Images