JPH07152434A - Own position detecting method for autonomously traveling work wagon - Google Patents

Abstract

PURPOSE:To surely recognize a light reflecting mark even at an uneven ground or an area where there is undulation so that correct own position can be detected by recognizing the light reflecting mark by using a picture when light is projected from a projecting means and the picture when the light is not projected. CONSTITUTION:A stroboscope 8 and a CCD camera 9 are fitted coaxially to a rotating mechanism part 10 erected vertically at the front side of a work wagon 1. The stroboscope 8 and the CCD camera 9 are made to face the light reflecting mark 20, and the picture at the time of the non-light emission of the stroboscope and the picture at the time of the light emission of the stroboscope are image-picked up. The difference of luminance is determined for every corresponding picture element between these pictures, and a luminance difference picture is generated. Then, a binarized picture is generated by extracting the picture element exceeding a slice level set beforehand from the generated luminance difference picture. A picture element group to form a mass on this binarized picture is extracted, and the number of the picture elements to form this picture element group is calculated. Then, the picture element group in which the number of the picture elements is over a set value is extracted, and this picture element group is judged to be the mark 20.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention recognizes a plurality of signs installed around a specific work area to detect its own position, and performs self-positioning of an autonomous traveling work vehicle for carrying out work traveling in grass cutting, lawn mowing, etc. It relates to a detection method.

[0002]

2. Description of the Related Art Conventionally, for an autonomous vehicle that is autonomously autonomously traveling, there is a technique in which an electric wire is buried underground and a magnetic sensor detects a magnetic field generated by the electric wire as self-position detection for autonomous traveling. Although it has been proposed, if the autonomous driving area is large, such as an autonomous vehicle that performs unmanned grass cutting, lawn cutting, etc. in various fields such as golf courses, river banks, parks, etc. It is difficult to embed the electric wire, and the installation cost becomes large.

For this reason, a technique has been developed in which a plurality of signs are installed in the vicinity of an autonomously traveling area and the position of the vehicle is detected by recognizing the signs.
In Japanese Patent No. 5606, a plurality of retroreflective poles (light reflective markers) standing upright in a specific area are projected, the reflected light is received by an optical sensor, and an angle facing adjacent reflective poles is detected and detected. There is disclosed a technique for detecting the self-position by the three-point angle method (triangulation method) from the angle thus obtained and the known coordinates of the sign.

[0004]

However, there are many cases where an autonomously traveling work vehicle autonomously travels on an uneven terrain or swell. In such an area, the vehicle tilts and is reflected from a light reflecting sign. Light may not be detected by the optical sensor.

Further, in an area with uneven terrain or swell, the light reflecting sign cannot be installed so that the reflected light is reflected in the horizontal direction, so that the reflected light has an elevation angle and a depression angle, and the sign by the optical sensor. Becomes difficult to detect.

The present invention has been made in view of the above circumstances, and the self-propelled vehicle is capable of accurately recognizing the light reflecting sign even in an uneven terrain or a swelling area and accurately detecting the self-position. It is intended to provide a position detection method.

[0007]

According to a first aspect of the present invention, there is provided a self-position detecting method for an autonomous traveling work vehicle, which recognizes a plurality of light reflecting signs installed around a specific work area and detects the self-position by triangulation. The autonomous mobile work vehicle includes a light projecting unit for projecting light onto the light reflective sign, and an image capturing unit for imaging the light reflective sign. After capturing the image when the light is projected and the image when the light is not projected by the image pickup means to obtain a pair of images, a feature amount at a corresponding position of the pair of images is obtained, and the light amount is calculated from the feature amount. Characterized by recognizing a reflective sign.

In a second aspect based on the first aspect, a brightness difference exceeding a set level is obtained as the characteristic amount for each pixel at corresponding positions of the pair of images, and the brightness difference exceeding the set level is obtained. Is extracted to generate a binarized image, and a set of pixels equal to or more than a set number in the binarized image is recognized as the light reflection sign.

[0009]

According to the first aspect of the invention, a pair of images are obtained by picking up an image when the light is projected from the light projecting unit and an image when the light is not projected by the image sensing unit, and the feature at the corresponding position of the pair of images is obtained. A plurality of light-reflecting signs installed around a specific work area are recognized to find the quantity, and the self-position is detected by triangulation.

According to a second aspect of the present invention, in the first aspect of the present invention, the luminance exceeding the set level is set for each pixel at a position corresponding to the image projected by the light projecting means and the image captured without projecting the light. The difference is obtained as the feature amount, and the pixel having the brightness difference exceeding the set level is extracted to generate the binarized image. Then, a set of pixels equal to or more than the set number in this binarized image is recognized as a light reflection marker.

[0011]

Embodiments of the present invention will be described below with reference to the drawings. 1 and 2 relate to an embodiment of the present invention.
6 is a flow chart of a self-position detection routine, FIG. 3 is a flow chart of a main control routine, FIG. 4 is a schematic explanatory view showing the appearance of a lawnmower vehicle, FIG. 5 is an explanatory view showing marker imaging by a CCD camera of the lawnmower vehicle, and FIG. Is a block diagram of a control system, FIG. 7 is a circuit configuration diagram of a vehicle position detection unit, FIG. 8 is an explanatory diagram showing a travel route and a work area, and FIG. 9 is an explanatory diagram of self-position detection by triangulation.

In FIG. 4, reference numeral 1 denotes an autonomous traveling work vehicle which is capable of autonomous driving by an unmanned vehicle. In this embodiment, it is a lawn mowing work vehicle for performing grass and lawn mowing work on a golf course or the like. This lawnmower vehicle 1 is driven by an engine, and the steering angles of the front and rear wheels can be independently controlled.
In the lower part of the vehicle body, a cutting blade mechanism 2 such as a mower for performing grass / lawn mowing work is provided, a dead reckoning sensor for measuring the current position based on the traveling history, and a traveling obstacle for detecting an obstacle. A sensor for obstacle detection, a sensor for profile travel for performing profile travel along the cutting boundary in the grass / lawn mowing work area, etc. are installed, and a plurality of lights installed upright around the grass / lawn mowing work area. It is provided with a light projecting means and an imaging means for imaging the reflective sign, recognizing the sign from the picked-up image, and detecting the self-position by triangulation.

As the dead reckoning sensor, a geomagnetic sensor 3 and a wheel encoder 4 are provided in the lawnmower work vehicle 1, and the obstacle detecting sensor is a non-contact type including an ultrasonic sensor or an optical sensor. Sensors 5a, 5b
Are attached to the front and rear parts of the lawn mowing work vehicle 1, and a contact type sensor 6a using a micro switch or the like,
6b are attached to the front and rear ends of the lawnmower working vehicle 1.

As the copying traveling sensor, for example, a sled plate which moves up and down according to the length of grass and grass and a rotary encoder which detects a rotation angle of a member supporting the sled plate are combined. In addition, a cut mark boundary detection sensor 7 for detecting a cut mark boundary between the already cut portion and the uncut portion is provided at the lower portion of the vehicle body of the lawnmower work vehicle 1.

As the light projecting means and the image capturing means for capturing the image of the light reflecting sign, a strobe 8 is used as the light projecting means, and a CCD camera 9 using a solid-state image sensor (CCD) is used as the image capturing means. The lawn mowing work vehicle 1 is coaxially attached to a rotating mechanism portion 10 provided upright on the front side in the same direction. Then, as shown in FIG. 5, the strobe 8 and the CCD camera 9 are provided.
Toward a light reflection sign 20 having a retroreflector having a property of constantly reflecting light in the incident direction, such as a corner cube type light reflection sheet, for a predetermined length, and strobe non-light emission. A time image and an image at the time of strobe light emission are picked up.

Further, as shown in FIG. 6, the lawnmower working vehicle 1 is equipped with a control device 30 using a plurality of microcomputers, and the control device 30 receives signals from respective sensors and actuators. Each of the detection units for processing, that is, the vehicle position detection unit 31, the cut mark boundary detection unit 35,
Further, an obstacle detection unit 36 is provided, and a traveling control unit 37 for performing traveling control based on signals from these detection units 31, 35, 36, and work referred to by the traveling control unit 37. A work data storage unit 38 in which a data map is stored, a vehicle control unit 39 for performing vehicle control according to an instruction from the traveling control unit 37, and a lawnmower work vehicle based on an output from the vehicle control unit 39 are provided. 1 to drive each mechanism unit, the drive control unit 40,
A steering control unit 41 and a cutting blade control unit 42 are provided.

Further, the vehicle position detecting section 31 integrates the vehicle speed detected by the wheel encoder 4 to obtain a traveling distance, and makes the traveling distance correspond to a change in traveling direction detected by the geomagnetic sensor 3. The dead reckoning position detection unit 32 that calculates the traveling history from the reference point and measures the current position of the vehicle by accumulating the accumulated data, the strobe 8, the CCD camera 9, and the rotation mechanism unit 1.
A sign imaging control unit 33 that outputs each drive signal for sign imaging to the step motor 11 for driving 0.
And a video signal from the CCD camera 9 is processed to recognize a plurality of signs, and a rotation angle signal from a rotation angle sensor 12 attached to the rotation mechanism section 10 and positioning from the dead reckoning position detection section 32. A triangulation position detection unit 3 that compares the data and the position data of each sign that has been stored in advance to identify each sign and detects the self-position based on the triangulation principle from the position data of the sign that is stored in advance.
4 and.

On the other hand, the cut boundary detection unit 35 processes the signal from the cut boundary detection sensor 7 according to the grass / turf height to detect the grass / turf boundary position. As described above, the cut boundary detecting sensor 7 is composed of a sled-shaped plate that moves up and down according to the length of grass / turf, and a rotary encoder that detects a rotation angle of a member that supports the sled-shaped plate. If so, grass
When the vertical displacement of the sled-shaped plate corresponding to the lawn height is converted into a rotation angle and a detection signal is input from the rotary encoder, the cut boundary detection unit 35 uses the detection signal to detect grass and lawn mowing work. The cut area boundary is detected by distinguishing the completed area and the unworked area, and the position data is output to the traveling control unit 37.

Further, the obstacle detecting section 36 detects an unpredictable obstacle by the non-contact type sensors 5a, 5b and the contact type sensors 6a, 6b, and outputs a detection signal to the traveling control section 37.

The traveling control unit 37 stores work data based on the information from the dead reckoning position detecting unit 32, the triangulation position detecting unit 34, the cut mark boundary detecting unit 35, and the obstacle detecting unit 36. The amount of error between the current position of the user and the target position is calculated by referring to the map of the section 38 (the work content and the map for traveling such as the terrain data of the work area where grass and lawn mowing work are performed), and this error amount is calculated. A travel route instruction and a vehicle control instruction are determined to be corrected and output to the vehicle control unit 39.

In the vehicle control unit 39, the instruction from the traveling control unit 37 is converted into a specific control instruction amount, and this control instruction amount is drive control unit 40, steering control unit 41, cutting blade control unit 42. Output to. Accordingly, the drive control unit 4
At 0, the hydraulic motor 13 is driven to generate hydraulic pressure for driving each mechanism section of the lawnmower working vehicle 1, and the steering control section 41 is operated.
Then, the front wheel steering mechanism and the rear wheel steering mechanism (not shown) are servo-controlled via the front wheel steering hydraulic control valve 14 and the rear wheel steering hydraulic control valve 15, respectively, and the cutting blade control section 42 The cutting blade mechanism 2 is servo-controlled via the cutting blade control hydraulic control valve 16.

FIG. 7 shows a concrete circuit configuration of the vehicle position detecting section 31. The CPU 50 has a RAM 51 for holding work data, and a light reflection marker installed around a specific area to be worked. Position data and other fixed data for control, and a ROM 52 in which a control program is stored, an I / O interface 53 for inputting / outputting various data and control signals, a data bus 54 and an address bus 55. A microcomputer mainly connected via the electronic flash controls the imaging by the strobe 8 and the CCD camera 9 and processes the captured image stored in the video memory 65 to detect the vehicle position by sign recognition and triangulation. It is like this.

That is, the CPU 50 executes the program of the self-position detecting routine shown in FIGS.
The strobe 8 and the CC coaxially attached to the rotation mechanism unit 10 by outputting a control signal from the / O interface 53 to the step motor drive circuit 56 and the strobe drive circuit 57.
After pointing the D camera 9 in a predetermined direction, an electronic shutter control signal is output to the CCD camera 9 to capture an image when the strobe is on and an image when the strobe is not lit.

An NTSC video signal from the CCD camera 9 is amplified by an amplifier 58 and supplied to a synchronizing circuit 59 and an AD converter 60, respectively. The synchronizing circuit 59 separates the synchronizing signal from the video signal to generate a timing signal and supplies it to the AD converter 60 and the address control circuit 61.

The AD converter 60 converts the video signal from the amplifier 58 into digital image data in synchronization with the timing signal and outputs it to the switching circuit 64 via the data bus 62. Further, the address control circuit 61 generates address data in synchronization with the timing signal and supplies it to the switching circuit 64 via the address bus 63.

The switching circuit 64 selectively connects either one of the data bus 54 and address bus 55 on the CPU 50 side and the data bus 62 and address bus 63 on the AD converter 60 side to the video memory 65. Yes, while the address data is being supplied from the address control circuit 61 to the switching circuit 64, the data bus 6 on the AD converter 60 side is provided.
2 is connected to the video memory 65 so that the image data can be written.
Access to 5 is prohibited.

When the supply of the video signal from the CCD camera 9 is stopped and the video memory 65 of the CPU 50 can be accessed, the image data (the image when the strobe is lit and the image when the strobe is not lit) is output from the video memory 65. )
Is read out, and after this image data is processed to recognize a plurality of markers around the work area, each of the geomagnetic sensor 3, the wheel encoder 4, and the rotation angle sensor 12 read via the I / O interface 53. Each sign is identified based on the signal, the position of the vehicle is calculated by triangulation, and the vehicle position information is output from the I / O interface 53 to the traveling control unit 37.

Below, with respect to the work area as shown in FIG.
Explain the case where unmanned grass and lawn mowing work is performed. In this case, the lawn mowing vehicle 1 is assumed to have been moved in advance by manned operation from an arbitrary preparation position 70 to the grass / lawn mowing start point 74 in the work area 72 before starting work. The lawn mowing work is autonomously performed according to the programs shown in the flowcharts of FIGS. After the work is completed, the work is moved from the work end point 75 to the standby position 78 again by manned driving.

In FIG. 8, reference numerals PA, PB, PC
Is a light-reflective sign previously installed around the work area 72,
It is similar to the light reflection sign 20 described above, and will be simply referred to as signs PA, PB, and PC hereinafter.

First, in the main control routine shown in FIG. 3, in step S101, a self-position detecting routine shown in FIGS. 1 and 2 which will be described later is executed to recognize the marks PA, PB and PC and to detect the self-position by triangulation. Measure and start the grass / lawn mowing work 7
When it is confirmed that the lawn mowing work vehicle 1 has reached step 4, in step S102, the cutting blade control hydraulic control valve 16 is opened to supply hydraulic pressure to the cutting blade mechanism section 2 to operate the cutting blade. Start grass and lawn mowing work.

Next, in step S103, the lawn mowing work is performed 1
It is checked whether or not it is the first time, and if the work is the first stroke (first column), the process proceeds to step S104, and if it is the second or later second column, the process branches to step S109.

This is the first operation, and when the process proceeds from step S103 to step S104, the self-position detecting routine is executed again, and after detecting the vehicle position and traveling direction of the own vehicle by triangulation and dead reckoning, the following steps are performed. In S105, the work data storage unit 38 is referred to, and the amount of error from the current position is obtained with the end point position of the first time (first column) in the work area 72 as the target position.

Next, when the process proceeds to step S106,
Each target steering amount of the front and rear wheels is determined according to the error amount obtained in S105, and in the subsequent step S107, the front wheel steering mechanism is operated via the front wheel steering hydraulic control valve 14 and the rear wheel steering hydraulic control valve 15. The rear wheel steering mechanism is driven to control the target steering angle.

Then, in step S108, it is checked whether or not the lawnmower work vehicle 1 has reached the end point position of the first time (first row),
When the end point position is not reached, the above-mentioned step S104
Return to and continue grass and lawn mowing work, and when the end point is reached,
In step S116, it is checked whether or not all the work has been completed. In this case, since it is the first work, the above steps
Return to S103, check again whether it is the first work operation, work 2
After the first time, the process branches to step S109 to follow the working path 73 along the cut boundary.

That is, in step S109, when the vehicle body is laterally shifted by the width of the cutting blade and moved to the next work position, the cut mark boundary detection sensor 7 detects the cut mark boundary of the previous work in step S110, and in step S111. , The amount of error for realizing the set lawn mowing overlap amount is calculated by comparing the cut mark boundary with the vehicle position.

In the following step S112, the above step S111
Each target steering amount of the front and rear wheels is determined according to the error amount obtained in step S113, and in step S113, the front wheel steering hydraulic control valve 1
4, front wheel steering mechanism via the rear wheel steering hydraulic control valve 15,
The rear wheel steering mechanism is driven to control the target steering angle.

Then, in step S114, a self-position detecting routine, which will be described later, is executed to detect the vehicle position and traveling direction of the own vehicle by triangulation and dead reckoning.
Then, it is checked whether or not the lawn mowing vehicle 1 has reached the end point position after the second time (after the second row), and if it has not reached the end point position, the process returns to step S110 described above and follows the cut boundary. When the end point position of the second time or later (after the second row) is reached, the process proceeds from step S115 to step S116, and it is determined whether or not the grass / lawn mowing work in the work area 72 is completed.

If the grass / lawn mowing work in the work area 72 is not completed, the process returns from step S116 to step S103 to continue the work, and when the grass / lawn mowing work in the work area 72 is completed, this main control routine is executed. Stop grass and lawn mowing work and stop the vehicle.

Next, self-position detection by sign recognition / triangulation by the self-position detection routine of FIGS. 1 and 2 will be described.

In this self-position detection routine, first, in step S201, the step motor 11 is driven through the step motor drive circuit 56 to rotate the rotating mechanism 10 to a set angle to set the strobe 8 and the CCD camera 9. Turn in the direction.

Next, in step S202, the strobe 8 is operated via the strobe drive circuit 57. In step S203, the CCD camera 9 is synchronized with the light emission of the strobe 8.
The electronic shutter is triggered to capture an image when the strobe is lit, and in step S204, the image when the strobe is lit is input. That is, the AD converter 60 converts the analog video signal from the CCD camera 9 into a digital signal in synchronization with the timing signal from the synchronizing circuit 59 at a sample rate of, for example, about 8 MHz, and for example, 512 pixels × 1 frame.
It is written in the video memory 65 as image data of 512 pixels.

Thereafter, when the operation proceeds to step S205, the driving of the strobe 8 is stopped, and in step S206, the electronic shutter of the CCD camera 9 is triggered to pick up an image when the strobe is not lit. , Step S2
In step 07, the same is written in the video memory 65 in step S208.
Go to.

In step S208, when the writing of the two images to the video memory 65 with and without the strobe lighting is completed, the switching circuit 64 enables the CPU 50 to access the video memory 65. The image data and the image data when the strobe is not lit are read, the difference in brightness is obtained for each corresponding pixel between these images, and a brightness difference image is generated.

Next, in step S209, pixels exceeding the preset slice level are extracted from the luminance difference image generated in step S208 to generate a binarized image. Then, in step S210, a pixel group that is a block on the binarized image is extracted, and the number of pixels forming this pixel group is calculated. In step S211, the number of pixels is equal to or greater than the set value. A group is extracted, and this pixel group is determined as a marker.

That is, if any of the markers PA, PB, and PC is within the angle of view of the CCD camera 9, there is a large difference in brightness between the pixels with the marker in the image when the strobe is lit and the image when the strobe is not lit. Occurs, and on the binarized image in which this brightness difference is extracted, the marker is a cluster of pixels exceeding the set slice level. Therefore, if the number of pixels forming the cluster of pixels is equal to or larger than the set value, it can be determined as a marker.

Next, in step S212, the center of gravity (X0, Y0) of the pixel group on the image is calculated.
Then, the direction of the sign is calculated from the X-axis position X0 of the center of gravity on the image and the camera angle θ with respect to the lawn mower 1 obtained from the rotation angle signal from the rotation angle sensor 12 attached to the rotation mechanism section 10.

In this case, the image obtained from the CCD camera 9 has an angle of view in the horizontal and vertical directions.
Even if the work area has uneven terrain or swell and the vehicle is tilted, or the signs PA, PB, PC are installed tilted and the reflected light from the sign has an elevation angle and a depression angle, Can recognize the sign. Further, since the CCD camera 9 captures an image when light is projected from the strobe 8 to the markers PA, PB, and PC, and an image when light is not projected, the marker can be recognized regardless of day or night.

Then, in step S214, one rotation (36
It is judged whether the processing of 0 degree) is completed, and if the processing for one rotation is not completed, the procedure returns to step S201 and the strobe 8 and the CCD camera 9 are further rotated by the set angle and the same processing is performed. When all the markers PA, PB and PC are recognized after the processing for one rotation is completed, step S2
Proceed to 15.

In step S215, the current vehicle position and direction is estimated by dead reckoning from the position and direction of the vehicle calculated up to the previous time, and in step S216, the detected sign and the position of the stored sign are compared. Then, it is determined which sign the detected sign corresponds to, and the angle α formed between the sign PA and the sign PB and the angle β between the signs PB and PC centering on the vehicle are detected, and the process proceeds to step S217. move on.

In step S217, the current vehicle position is calculated by the known triangulation method from the positions of the markers PA, PB and PC stored in advance in the ROM 52.

That is, as shown in FIG. 9, in the X, Y coordinate system with the marker PA as the origin, the distance from the marker PA (coordinate origin) to the center position of the vehicle is based on each data stored in advance in the ROM 52. The distance L and the angle θ1 formed by the line connecting the center of the vehicle from the marker PA (coordinate origin) and the line connecting the markers PA and PB are calculated by the following equations (1) and (2), and the vehicle position ( X1, Y1) can be calculated by the equations (3) and (4). L = L1 × sin (α + θ1) / sinα (1) θ1 = atn (-L2 × sinα × sin (θB + α + β) / (L1 × sinβ + L2 × sinα × cos (θB + α + β))) (2) L1; Label PA, Distance between PBs L2; Distance between signs PB and PC θB; Angle formed by line connecting signs PA and PB and line connecting signs PB and PC θA; Angle of line connecting signs PA and PB with respect to X axis X1 = L × cos (θA−θ1) (3) Y1 = L × sin (θA−θ1) (4)

[0052]

As described above, according to the present invention, the light reflection marker is obtained by obtaining the feature amount at the corresponding position from the imaged image when the light is projected from the light projecting means and the imaged image when the light is not projected. Therefore, even if the vehicle is tilted due to uneven terrain or swell in the autonomous driving area or the light reflected from the light reflective sign has an elevation angle and a depression angle, the light reflective sign is reliably recognized. It is possible to obtain an excellent effect such that the self position can be accurately detected.

[Brief description of drawings]

FIG. 1 is a flowchart of a self-position detection routine.

FIG. 2 is a flowchart of a self-position detection routine (continued)

FIG. 3 is a flowchart of a main control routine.

FIG. 4 is a schematic explanatory view showing the appearance of a lawnmower work vehicle.

FIG. 5 is an explanatory diagram showing image capturing of a marker by a CCD camera of a lawnmower work vehicle.

FIG. 6 is a block diagram of a control system

FIG. 7 is a circuit configuration diagram of a vehicle position detection unit.

FIG. 8 is an explanatory diagram showing a travel route and a work area.

FIG. 9 is an explanatory diagram of self-position detection by triangulation.

[Explanation of symbols]

 1 Lawn mowing work vehicle (autonomous traveling work vehicle) 8 Strobe (light projecting means) 9 CCD camera (imaging means) PA, PB, PC Light reflective sign

Claims (2)

[Claims] 1. A self-position detecting method for an autonomous traveling work vehicle, which recognizes a plurality of light reflecting signs installed around a specific work area and detects the self-position by triangulation, comprising: A light projecting means (8) for projecting light on a reflective sign, and an imaging means (9) for imaging the light reflective sign are provided, and the light projecting means (8) projects the light reflective sign. The image capturing means displays an image when light is emitted and an image when light is not emitted.
After obtaining a pair of images by imaging in (9), obtain the feature amount at the corresponding positions of the pair of images,
A self-position detecting method for an autonomously traveling work vehicle, characterized in that the light reflection sign is recognized from the characteristic amount. 2. A binarized image in which a brightness difference exceeding a set level is obtained as the feature amount for each pixel at corresponding positions of the pair of images, and pixels having a brightness difference exceeding the set level are extracted. The self-position detecting method for an autonomously traveling work vehicle according to claim 1, wherein a set of pixels equal to or more than a set number in the binarized image is recognized as the light reflection sign.