JP3015138B2 - Automatic vehicle track correction device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an automatic vehicle trajectory correcting apparatus for a production line such as a factory. 2. Description of the Related Art An automatic traveling vehicle (hereinafter, referred to as a self-propelled vehicle) that transports parts and tools on a production line or the like of a factory travels on a predetermined track. The conventional self-propelled vehicle was stopped when there was an obstacle on the track. However, recently, a self-propelled vehicle equipped with a TV camera has also developed a self-propelled vehicle that, once it finds an obstacle on the track, leaves the track once, avoids the obstacle and returns to the original track again to continue running. Have been.

Conventionally, in such a self-propelled vehicle, the trajectory is corrected based on the image of the trajectory drawn on the floor. Did not. Also, it has been developed a method of automatically traveling seeking track traveling by ultrasonic or the like, on which require wall that reflects ultrasonic waves on both sides of the track, was not reliable. There is also a method of returning to an original correct trajectory (hereinafter simply referred to as a trajectory) while counting the number of rotations of the motor after the start of obstacle avoidance, but it does not operate accurately.

[0003] The present invention can be used even if the orbit line is not drawn.
An object of the present invention is to provide a self-propelled vehicle trajectory correction device that avoids an obstacle based on an image captured by a TV camera, and reliably returns to a trajectory and continues running.

[0004]

2. Description of the Related Art A conventional method for avoiding obstacles in a self-propelled vehicle will be described with reference to FIG. Fig. 7 (a), (b), (c), (d)
, 100 is an obstacle, 105 is a trajectory of a self-propelled vehicle, 1
07 and 108 are wall surfaces.

FIG. 1A is an explanatory diagram of obstacle avoidance. In the figure, 110 is a self-propelled vehicle trying to avoid the obstacle 100, and 111 is a self-propelled vehicle 110 avoiding the obstacle 100
Represents Self-propelled vehicle 11 that has traveled on the correct track 105
0 finds the obstacle 100, and the self-propelled vehicle 1 in the figure
It is necessary to avoid the position 11 and return to the track 105 to travel.

As a method of avoiding such obstacles, a method using ultrasonic waves, a method in which two TV cameras are mounted on a self-propelled vehicle, and a trajectory line is detected and returned by binocular stereovision,
Measures the number of rotations of the motor driven by the self-propelled vehicle since the start of the avoidance operation, and the rotation angle of the self-propelled vehicle to avoid obstacles. There is a method of returning.

FIG. 1B shows a method of returning by ultrasonic waves.
In the figure, reference numeral 120 denotes a self-propelled vehicle, and reference numerals 121 and 122 denote wavefronts of ultrasonic waves. The self-propelled vehicle 120 has sonars installed on both sides of the self-propelled vehicle, and the ultrasonic wave causes ultrasonic waves to be applied to the wall 10 from both sides.
Measure the distance to 7,108. The traveling position of the self-propelled vehicle 120 is corrected so that this distance becomes the same as when traveling on the track, and the vehicle returns to the track.

FIG. 1C shows a method based on binocular stereovision. In the figure, 130 is a self-propelled vehicle, and 131 is 2 attached to a self-propelled vehicle
One camera. The self-propelled vehicle 130 has two cameras 13
1 detects the track line (track 105) drawn on the floor, measures the distance between the self-propelled vehicle 130 and the track line (track 105), corrects the position, and returns to the track line (track 105). I do.

FIG. 4D shows a method of returning by measuring the movement. In the figure, reference numeral 140 denotes a self-propelled vehicle. Obstacle 1
In order to avoid 00, the physical quantity such as the angle θ at which the self-propelled vehicle has rotated and the number of rotations of the motor since the start of avoidance is measured and stored. For example, the number of rotations and the angle θ of the motor from the point P to the point Q in the figure are measured, and at the point Q, the direction is changed by rotating the angle θ in the opposite direction, and the self-propelled vehicle 140 is moved by the same motor speed By doing so, it returns to the orbit 105.

[0010]

The method of measuring the distance to the wall surface by the ultrasonic wave shown in FIG. 7B and returning to the orbit is as follows.
Walls that reflect ultrasonic waves are required on both sides of the orbit, and such an environment is difficult to realize in an actual site.

FIG. 7 (c) shows a method based on binocular stereopsis, in which it is necessary to draw a trajectory line for determining the trajectory on the floor, it takes time to measure the distance, and the accuracy is not good. FIG.
In the method based on the motion measurement in (d), if there is unevenness on the floor avoiding the obstacle, the rotation speed of the motor and the moving distance cannot be correlated, resulting in poor accuracy.

The present invention provides a trajectory correcting device for a self-propelled vehicle that can reliably return to the trajectory by using an image of a TV camera mounted on the self-propelled vehicle even if no trajectory line is drawn on the running surface of the self-propelled vehicle. The purpose is to do.

[0013]

According to the present invention, straight lines parallel to each other in a three-dimensional space are converted to a two-dimensional perspective image by utilizing an infinite point where an extension of the straight line intersects one point. Corrected the trajectory of running vehicles.

The means for solving the problem will be described with reference to FIGS. 8, 9 and 10. FIG. 8 is an explanatory view (1) of the means for solving the problem. In the figure, 160 is a self-propelled vehicle, 161 is an imaging device, 162 is the track of the self-propelled vehicle, 163
Is the optical axis of the imaging device.

According to the present invention, the shift amount from the infinity point and the trajectory of the self-propelled vehicle is calculated based on the image of the parallel line of the subject taken by the image pickup device 161. Figure (b) shows a state in which the self-propelled vehicle is traveling correctly on the track. Figure (c)
FIG. 4 is a diagram showing a case where the direction of the self-propelled vehicle is correct but the vehicle is shifted from the track.

In FIGS. 2 (b) and 2 (c), 170 is located on a track, a self-propelled vehicle in the correct direction, 171 is a track, 172 is an optical axis of an image pickup device (TV camera), 175 and 176 are wall surfaces, 180 is a self-propelled vehicle shifted from the orbit in the correct direction, and 182 is the optical axis of the TV camera.

FIG. 9 is an explanatory view of means for solving the problem.
(2). Figure (a) is an image when it is on the orbit (image at the position in Figure 8 (b)). In the figure, reference numeral 200 denotes an image for making a TV camera. The lower left corner of the image is the coordinate origin, the horizontal scanning direction is the X axis, and the vertical scanning direction is the Y axis. 210 is a point at infinity, which is a line parallel to each other in the subject (for example, parallel lines W1, W
2, W3, W4, etc.). 211 is a reference line, which is a vertical line passing through the point at infinity. 215, 21
Numerals 6 are parallel wall surfaces (however, the wall is not necessarily required in the present invention). W1 and W2 are wall surfaces 2
15 parallel lower and upper ends. W3 and W4 are a lower end and an upper end of the wall surface 216 that are parallel to each other. S is the center line between the wall surfaces (however, in the present invention, the center line representing the trajectory does not have to be drawn). (Gx, gy) is the coordinates of the point at infinity, A is the intersection between the reference line 211 and the X axis, B is the shift amount calculation point, the intersection between W1 and the X axis, and C is the shift amount calculation point. Is the intersection of W3 and the X axis. Δq
Is the distance between AB and Δr is the distance between AC.

FIG. 8B shows an image when the direction of the self-propelled vehicle is correct but the vehicle is shifted from the track (the image at the position shown in FIG. 8C). 230 is an image, 240 is an infinity point, 2
41 is a reference line, and 245 and 246 are parallel wall surfaces. W1 and W2 are a lower end and an upper end of the wall surface 245 parallel to each other. W3 and W4 are a lower end and an upper end of the wall surface 246 which are parallel to each other. S is the center line between the walls, (gx, gy) is the coordinates of the point at infinity, A is the intersection of the reference line 241 with the X axis, B
Is a shift amount calculation point, which is an intersection between W1 and the X axis, and C is a shift amount calculation point, which is an intersection between W3 and the X axis.
Δa ₁ is the distance between AB and Δa ₂ is the distance between AC.
As shown in FIGS. 9 (a) and 9 (b), the coordinates of the point at infinity in the image do not change when the self-propelled vehicle is oriented in the correct direction and shifted parallel to the trajectory. However, the point at which a straight line such as W1 or W3 passing through the point at infinity intersects the X axis (shift amount calculation point) varies.

The magnitude of the shift of the self-propelled vehicle from the track can be determined based on the difference between the coordinate values of the shift amount calculation point B when the vehicle is on the track and the shift amount calculation point B at the shifted position. It is possible (the point C may be compared). That is, considering the screen on the left side of the reference line, in order to return the self-propelled vehicle to the trajectory, the difference Δs = Δq−Δa1 between the x-coordinate of the shift amount calculation point and that on the trajectory
Is calculated and the difference Δs =
The self-propelled vehicle may be controlled so as to be 0 (hereinafter, Δs is referred to as a shift amount). When the shift amount Δs = 0, the self-propelled vehicle has returned to the track.

Referring to FIG. 10 which is an explanatory view (3) of the means for solving the problem, the movement of the infinity point when the self-propelled vehicle rotates from the correct direction will be described. Figure (a) shows a state in which the self-propelled vehicle has rotated. In the figure, 175 and 176 are wall surfaces, and 171 is a track. Reference numeral 190 denotes a self-propelled vehicle rotated from a correct direction, and reference numeral 192 denotes an optical axis of the self-propelled vehicle 190.

FIG. 2B shows an image of a TV camera of a self-propelled vehicle rotated from a correct direction. In the figure, 250 is an image, 251 is an infinity point, 252 is a reference line, 253, 2
54 are parallel wall surfaces. W1 and W2 are wall surfaces 25
3 are a lower end and an upper end parallel to each other. W3 and W4 are a lower end and an upper end of the wall surface 254 that are parallel to each other. S is a center line between the wall surfaces, and 255 is an infinity point in the correct direction of the vehicle (infinity point in FIG. 8B or 8C).
t is the difference in the x direction between the point at infinity before rotation and the point at infinity after rotation.

Assuming that the focal length of the lens of the TV camera is F, the rotation angle is θ, and the difference in the x direction between the infinity point in the correct direction and the infinity point when rotated is t, θ = tan ⁻¹ (t /
F). Therefore, in order to return the self-propelled vehicle to the correct direction, the amount of change t at the point at infinity is calculated, θ is obtained by the above relational expression, and the self-propelled vehicle may be rotated by the angle θ.

According to the present invention, based on the above principle, the return to the trajectory after the self-propelled vehicle avoids the obstacle is performed based on the fluctuation amount and the shift amount of the point at infinity on the image.
The principle of the present invention will be described with reference to FIG. In the figure, 1
Denotes an image input unit for inputting an image created by a TV camera provided in a self-propelled vehicle and converting the image into a binary image. Reference numeral 2 denotes a rotation amount calculation unit that calculates coordinates of an infinity point from a ridgeline of a subject and calculates a variation amount of the infinity point based on coordinates of the infinity point before and after avoiding an obstacle. It is. 3
A shift amount calculation unit calculates a shift amount based on the coordinates of the shift amount calculation point when the self-propelled vehicle is on the track and when the self-propelled vehicle is shifted from the track. Numeral 4 denotes a traveling trajectory correction unit which corrects the trajectory of the self-propelled vehicle based on the variation amount of the infinity point calculated by the rotation amount calculation unit 2 and the shift amount calculated by the shift amount calculation unit 3.

The infinity point judging unit 10 judges an infinity point by extracting a parallel straight line of the object and finding an intersection, and then finds its coordinates. The infinity point comparison unit 11 compares the coordinates of the infinity point before and after avoiding the obstacle and calculates the amount of change of the infinity point. Reference numeral 12 denotes a shift amount calculation point determination unit that determines a shift amount calculation point from a straight line passing through the point at infinity. Numeral 13 compares the coordinates of the shift amount calculation points before and after avoiding the obstacle and calculates the shift amount.

[0025]

The operation of the configuration shown in FIG. 1 for explaining the principle of the present invention will be described. Reference is made to FIGS. 9 and 10 as necessary. The image input unit 1 inputs images before and after obstacle avoidance (hereinafter, before and after avoidance, respectively) created by a TV camera (having a known focal length) attached to the self-propelled vehicle. And convert it to a binary image. Infinity point judgment unit 10
Calculates a ridge line of a subject in a binary image by a pixel tracking method or the like, converts the ridge line into a broken line by approximation of a straight line, and calculates coordinates of a broken point, a function of a broken line, and the like. Then, the intersection of the straight lines is determined, and the maximum cumulative point of the intersection is determined as the point at infinity. The infinity point comparison unit 11 compares the x-coordinates of the infinity point before and after the avoidance and calculates the amount of variation t of the infinity point. Then, θ = tan ⁻¹ (t / F) is calculated as an angle at which the self-propelled vehicle should be rotated based on the focal length F of the lens.

The shift amount calculation point judging section 12 searches for a straight line passing through the infinity point from the infinity point and the straight line obtained by the infinity point judging section 10, finds an intersection with the x axis, and sets it as a shift amount calculation point. . If there are a plurality of points, the point closest to the x coordinate of the point at infinity is set as the shift amount calculation point (one point may be obtained on the right and left sides of the image with the reference line as a boundary).

The shift amount calculation point comparison unit 13 calculates the distance (Δq, Δa1, see FIG. 9) from the point A based on the x coordinate of the shift amount calculation point before and after the avoidance, and calculates the shift amount (Δs
= Δa1−Δq). The traveling trajectory correction unit 4 corrects the traveling angle of the self-propelled vehicle based on the variation t of the point at infinity. The correction is repeated until t = 0. When the traveling angle is corrected, the traveling position of the self-propelled vehicle is corrected based on the shift amount Δs. The correction is repeated until Δs = 0, and the self-propelled vehicle is returned to the track.

In the above description, the case where there are walls on both sides of the track has been described. However, the wall does not need to be specially installed on the track, and the ridgeline for finding the point at infinity is the wall and ceiling of the building, Anything at the site where the self-propelled vehicle is used, such as a ridgeline of an obstacle, may be used.

The center line indicating the correct trajectory does not need to be drawn on the floor, but if it is drawn, the point at infinity can be easily obtained from the trajectory line and the shift amount can be calculated. The above values are also easily obtained based on the center line, which facilitates the implementation of the present invention. The imaging device may be mounted not only on the front part of the self-propelled vehicle but also on the rear part, and the infinity point may be obtained from the scene behind the self-propelled vehicle.

[0030]

FIG. 2 shows an embodiment of the present invention. Reference is made to FIGS. 9 and 10 as necessary. In the figure, reference numeral 30 denotes an image input unit which performs A / D conversion of an image photographed by a TV camera mounted on a self-propelled vehicle. For example, the image input unit 30 has 512 pixels × 512 pixels × 8
The image is stored in the internal memory as a 256-bit digital image, and a portion (edge) having sharp undulations is extracted from the image, and a pixel value larger than a predetermined threshold value is 1,
Small pixel values are binarized by giving a value of zero.
Further, the image input unit 30 extracts a ridge line of the object in the image, and transfers the binarized image to the infinity point determination unit.

Reference numeral 31 denotes a rotation amount calculation unit, and 32 denotes a shift amount calculation unit. Reference numeral 33 denotes an infinity point determination unit which obtains the position (coordinates (x, y)) of the infinity point by performing image processing on the binarized image created by the image input unit 30. The algorithm for finding the point at infinity will be described later (see FIGS. 4 and 5).

The infinity point comparison unit 34 includes an infinity point determination unit 3
The coordinates (x, y) of the infinity point after avoidance calculated in 3 are compared with the coordinates (gx, gy) of the infinity point before avoidance. At this time, the y coordinate is regarded as having no difference and is excluded from comparison. If gx is equal to x, it is determined that the self-propelled vehicle does not rotate, and the processing is shifted to the shift amount calculation unit 32. If not equal, obtains the amount of change of the focal length F and point at infinity of the TV camera lens (x-gx) a self-propelled vehicle of rotational adjustment angle ^{θ = tan -1 ((x-} gx) / F) , To the traveling trajectory correction unit 38.

The shift amount calculation point determination unit 36 determines a shift amount calculation point when the self-propelled vehicle is correctly oriented in the traveling direction. The shift amount calculation points are determined as points B and C where a straight line passing through the point at infinity intersects the x-axis, and the length of the line segment AB is determined as Δa1, and the length of the line segment AC is determined as Δa2.

The shift amount calculation point 37 determines the shift amount calculation points B and C before and after the avoidance, and calculates the shift amount Δs = Δa1
−Δq and Δs = Δa2−Δr are calculated. Reference numeral 38 denotes a traveling trajectory correction unit. The traveling trajectory correction unit 38 calculates the rotation trajectory θ from the infinity point comparison unit 34 based on the rotation adjustment angle θ.
Adjust the steering of the self-propelled vehicle. If θ is positive, clockwise; if negative, counterclockwise. Then, the self-propelled vehicle moves in parallel based on the difference in the shift amount sent from the shift amount calculation point 37 after the self-propelled vehicle faces the correct direction.

35 is data (Δq, Δg) based on the infinity point (gx, gy) calculated by the infinity point judgment unit 33 before avoidance and the shift amount calculation point obtained by the shift amount calculation point judgment unit 36.
r). FIG. 3 is a flowchart of the configuration of the embodiment.

A description will be given according to the steps in the figure. S1 An obstacle avoidance operation is started, and an image is input. S2: The initial value of the point at infinity is obtained from the image at the start of traveling or the image immediately before avoiding the obstacle.

S3: The initial value of the shift amount is obtained from the image at the time of starting running or the image immediately before avoidance. S4 The infinity point determination unit calculates the infinity point. S5: The infinity point comparison unit compares the infinity point calculated by the infinity point determination unit with the initial value of the infinity point. If they do not match,
The rotation angle of the self-propelled vehicle is calculated, and the traveling direction of the self-propelled vehicle is adjusted. Further, an image is input, and the processing of S4 to S6 is repeated.

If the infinity point calculated by the infinity point determination unit matches the initial value of the infinity point, the process proceeds to S7. S7 The image is input, and the shift amount calculation point determination unit obtains the shift amount calculation point. S8: The shift amount calculation point comparison unit calculates the distance difference between the shift amount calculation point before and after the avoidance and the point A (see FIG. 9) (Δs
= Δa1-Δq, Δs = Δa2-Δr).

If the difference between the shift amounts is not 0, the data is sent to the traveling trajectory correction unit, and the self-propelled vehicle is translated. Further, an image is input and the processing from S7 to S9 is repeated. If the shift amounts match and the difference from the shift amount initial value becomes 0, the self-propelled vehicle has returned to the correct track, and the process is terminated.

FIG. 4 is an algorithm of an infinity point calculation method in the infinity point determination unit. The figure shows an example using a pixel tracking method as an embodiment of an algorithm for obtaining an infinite point. The infinity point determination unit includes a pixel tracking unit 41,
Broken line approximation unit 42, drawing unit 43, maximum accumulated point search unit 44
It consists of.

First, the pixel tracking unit 41 scans the binary image in the horizontal direction from the upper left corner, and searches for a pixel having a value of 1. A in the figure shows an example of 8-neighbor search in the pixel tracking unit 41. In FIG. A, reference numeral 51 denotes a pixel having a value of 1 found as a result of the horizontal scanning. Then, starting from the found pixel 51 having the value 1 as a starting point, a search is made for 8 neighborhoods, and among the 8 neighboring pixels, the pixel 5 having the value 1 is searched.
Look for 2. At this time, the already found points are deleted from the image, and the coordinate values of the found points are stored in the memory. This search is repeated until there is no point having a pixel value of 1 near 8. Scanning is continued and the same process is repeated until the lower right corner of the image is reached. Next, the polygonal line approximation unit 42 performs a polygonal line approximation for each of the curves obtained as a result of the pixel tracking by the pixel tracking unit 41. The polygonal line approximation will be described later (see FIG. 5). As a result of the polygonal line approximation, the coordinates of the polygonal point and the function value of the polygonal line are obtained. The drawing unit 43 writes a straight line into the frame memory using the coordinates of the broken point calculated by the broken line approximating unit 42 and the function value. Then, a point at each coordinate of the straight line has a pixel value of 1, and 1 is added to a pixel in which one straight line shares the same pixel as another straight line (one is added for each shared straight line).

The maximum cumulative point searching section 44 scans the frame memory on which the straight line is drawn, and obtains a coordinate point having the maximum pixel value (the value of the pixel sharing the straight line (cumulative degree)). The obtained point is set as an infinity point 45. The infinity point is sent to the infinity point comparison unit and the shift amount adjustment unit.

FIG. 5 shows an embodiment of the broken line approximation in the broken line approximation section. Figure (a) Let the end points a and b of the traced curve k. (B) A point c on the curve k farthest from the line segment ab connecting the end points a and b is obtained. Then, assuming that the intersection point of the perpendicular drawn from the point c to the line segment ab and the line segment ab is P, the point c
The distance D between the point and the point P is obtained.

(C) Similar processing is performed on a line segment ac connecting the point c and the end point a and a curve k1 connecting the end point a and the point c.
A point c ′ having the maximum distance from the line segment ac is obtained on the curve k1. Then, a distance D1 between the point c 'and the line segment ac is obtained.
Similarly, for a line segment bc connecting the point c and the end point b and a curve k2 connecting the point c and the end point b, a point c ″ is obtained on the curve k2,
The distance D2 between the point c "and the line segment bc is obtained.

The above processing is repeated, and when the distances D1 and D2 between the point obtained on the curve and the line segment become equal to or less than a predetermined value, the polygonal line approximation for the curve is terminated. FIG. 6 shows an embodiment of the shift amount calculation point determination unit. The lower left end of the image is the origin of the coordinate axes, the horizontal direction is the X axis, and the vertical direction is the Y axis.

70 is an image, 71 is a reference line, 7
2 is a point at infinity. A is the intersection of the reference line and the X axis, B and C
Is a shift amount calculation point. Figure (a) shows the relationship between the point at infinity and the reference line in the image. To calculate the shift amount,
In the present invention, an image is divided into a left half and a right half with respect to a reference line, and each is stored in a memory.

FIG. 8B shows the left half image, and FIG. 8C shows the right half image. (D) shows the shift amount calculation for the left half image, and (e) shows the shift amount calculation for the right half image. Position of infinity point and broken line approximation data of each curve (coordinates of end points of broken line, inclination of straight line, y intercept)
Enter Then, it is determined whether the polygonal line is included on the right or left side of the image based on the coordinates of the end point of the polygonal line. At the same time, check whether the straight line passes through the point at infinity. Then, a straight line that does not pass through the point at infinity is discarded, and only a straight line that passes through the point at infinity is written in an area including the end point. This writing is performed as a pixel value 1 of a point on a straight line. After writing the straight line,
In the image on the right, pixels adjacent to the X axis (that is, 1
) Is scanned rightward from the point at infinity, the first non-zero pixel C (pixel having a value of 1) is set as a shift amount calculation point, and the number of pixels from point A is set as Δa2. Similarly, in the left image, pixels adjacent to the X axis are scanned leftward starting from the X coordinate of the point at infinity, and the first non-zero pixel B (pixel having a value of 1) found is shifted to the shift amount calculation point. And the number of pixels from point A is Δa1.

The shift amount calculation point comparing section calculates the difference between the shift amount calculated by the shift amount calculation point determination section and the initial value Δq and Δr of the shift amount before obstacle avoidance. That is, the shift amount Δs1 = Δa1−Δq is calculated for the left image, and the shift amount Δs2 = Δa2−Δr is calculated for the right image, and is transferred to the traveling trajectory correction unit.

If the value of Δa1−Δq is positive, the traveling trajectory correction unit determines that the vehicle is deviating to the right from the correct trajectory, and if negative, it is deviating to the left, and shifts the self-propelled vehicle. In the above description, the shift amount is calculated for both the right half and the left half of the image, but only one of them may be calculated.

[0050]

According to the present invention, in order to return to a correct trajectory after avoiding an obstacle, it is not necessary to draw a line indicating the trajectory on the floor surface. Since the self-propelled vehicle is guided based on this, it is possible to reliably return to the track. Also, there is no need for the walls on both sides of the trajectory, which is necessary when guiding with ultrasonic waves. Works correctly even on surfaces.

Therefore, according to the present invention, since the self-propelled vehicle can be accurately guided without requiring a special environment for guiding the self-propelled vehicle, the usability at a manufacturing site such as a factory is greatly improved.

[Brief description of the drawings]

FIG. 1 is a diagram illustrating the principle of the present invention.

FIG. 2 is a diagram showing a configuration of an embodiment of the present invention.

FIG. 3 is a diagram showing a flow of an embodiment of the present invention.

FIG. 4 is a diagram illustrating an algorithm of an infinity point determination unit according to the present invention.

FIG. 5 is a diagram showing an example of broken line approximation in the present invention.

FIG. 6 is a diagram illustrating an embodiment of a shift amount calculation point determination unit according to the present invention.

FIG. 7 is a diagram showing a conventional obstacle avoidance method for a self-propelled vehicle.

FIG. 8 is a diagram showing a description (1) of means for solving the problem.

FIG. 9 is a diagram showing a description (2) of a means for solving the problem.

FIG. 10 is a diagram showing a description (3) of a means for solving the problem.

[Explanation of symbols]

 1: Image input unit 2: Rotation amount calculation unit 3: Shift amount calculation unit 4: Traveling trajectory correction unit 10: Infinity point judgment unit 11: Infinity point comparison unit 12: Shift amount calculation point judgment unit 13: Shift amount calculation Point comparison section

──────────────────────────────────────────────────続 き Continuation of the front page (56) References JP-A-3-139706 (JP, A) JP-A-3-90917 (JP, A) JP-A-64-26913 (JP, A) JP-A-63-63 79005 (JP, A) JP-A-63-314615 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name) G05D 1/02

Claims (3)

(57) [Claims] 1. An automatic traveling vehicle trajectory correction device including an imaging device for correcting a traveling trajectory based on a captured image during traveling, calculates an infinity point where a parallel line of a subject intersects in the captured image, and calculates a different position. A rotation amount calculator (2) that calculates the amount of change at the point at infinity at the time and calculates the amount of rotation of the autonomous vehicle in the traveling direction based on the amount of change, and the autonomous vehicle moves in the correct direction. If you have
A shift amount calculating unit (3) that calculates a shift amount representing a variation amount of a straight line position passing through an infinite point in the image at different times, and calculates a relative position with respect to the trajectory based on the shift amount; The traveling trajectory correction unit (4) rotates the autonomous vehicle based on the amount of change at the infinity point calculated by (2) and corrects the traveling position of the autonomous vehicle based on the shift amount calculated by the shift amount calculation unit (3). ), And calculates the amount of shift at infinity and the amount of shift before and after the autonomous vehicle leaves the track, corrects the running angle of the autonomous vehicle based on the amount of change at infinity, and calculates the shift amount. An automatic traveling vehicle trajectory correction device that corrects a traveling position based on the traveling position. 2. The automatic vehicle trajectory correction device according to claim 1, wherein the amount of change and the amount of shift at the infinity point are calculated before and after obstacle avoidance. 3. An apparatus according to claim 1 or 2, the shift amount calculating unit, the automatic vehicle track straight line passing through the point at infinity is and calculates the shift amount based on the coordinates of the intersection point with the lower end of the image Correction device.