JP7498151B2 - Shape Recognition Device - Google Patents

Description

This disclosure relates to a shape recognition device that recognizes the shape of an object by irradiating it with laser light.

Several methods have been proposed for recognizing the three-dimensional shape of objects with complex shapes. Typical methods include the TOF method and triangulation method. With the TOF method, a laser beam is irradiated onto the object, and the distance to the object is recognized by measuring the time it takes for the reflected light to return from the object. With the triangulation method, a laser beam is irradiated onto the object, and the distance to the object is recognized from the position where the reflected light is focused.

All of these methods recognize the distance to an object by emitting a laser beam from a certain point and measuring the time or angle it takes for the laser beam to hit the object and the reflected light to return. Therefore, with these methods, the position at which the laser beam is emitted is moved little by little to collect a large amount of distance data to the object, making it possible to recognize the three-dimensional shape of the object.

However, with these methods, in order to recognize the distance to an object, the object is basically irradiated with laser light from a single point. As a result, the laser light cannot reach the shadowed parts of the object, making it difficult to recognize the shape of these parts as blind spots.

Figure 8 is an explanatory diagram showing the problem of blind spots in conventional technology. In order to reduce the blind spot area, it is possible to adopt methods such as increasing the distance L between the laser scanner and the target object (see Figure 8(a)) or using multiple laser scanners (see Figure 8(b)).

However, neither the method shown in FIG. 8(a) nor the method shown in FIG. 8(b) can completely eliminate blind spots. In addition, when these methods are used, the recognition system becomes large-scale and expensive, which is problematic, and they are not a complete solution to the blind spots.

There is a conventional method for solving this problem by eliminating blind spots (see, for example, Patent Document 1). The device in Patent Document 1 is configured to emit detection light radially while rotating, and to emit it toward the target via two mirrors (reflective members). As a result, the problem of blind spots is solved by using detection light emitted toward the target via each mirror in addition to detection light emitted directly to the target.

JP 2018-84489 A

However, the conventional techniques have the following problems.
9 to 11 are diagrams for explaining the problems of the conventional technology. As in the method disclosed in Patent Document 1, measuring the distance by reflecting laser light on each of two mirrors is the same as arranging two laser scanners at the plane symmetrical portions of the respective mirrors as shown in FIG.

Therefore, in order to recognize the overall shape of the object, it becomes necessary to measure the distance by distinguishing between the reflection measurement data obtained when the object is illuminated through one of the two mirrors and the direct measurement data obtained when the object is illuminated directly without passing through a mirror. However, when the shape of the object is unknown and the uneven shape is also unknown, a problem arises in that it is not possible to distinguish between the reflection measurement data and the direct measurement data.

Furthermore, in the case of an object as shown in FIG. 10, reflection measurement data reflected by mirror A is used to recognize faces A1 to A4. On the other hand, reflection measurement data reflected by mirror B is used to recognize faces B1 to B4. Therefore, in order to accurately recognize the shape of an object as shown in FIG. 10, it is necessary to combine the reflection measurement data reflected by mirror A and the reflection measurement data reflected by mirror B, making data processing extremely difficult.

In addition, the detection light that is irradiated to the object via the mirror is irradiated from an oblique direction to the object. Therefore, as shown in Figure 11, in an object that has a part surrounded by two tall protrusions, there will still be a blind spot.

As such, the method described in Patent Document 1 is effective when the general shape of the target object is known in advance, but it is difficult to accurately measure the shape of an object with an unknown shape.

The present disclosure has been made to solve the problems described above, and aims to provide a shape recognition device that can recognize the shape of an object while suppressing the occurrence of blind spots that occur when using a single laser scanner.

The shape recognition device according to the present disclosure includes a parabolic mirror having a parabolic surface that reflects light, a laser scanner that is installed at a focal position of the parabolic mirror and irradiates a laser light toward the parabolic surface, thereby irradiating the laser light that has been made parallel via the parabolic mirror from directly above the object, and a control unit that executes scanning control of the laser scanner, wherein the control unit executes scanning control of the laser scanner so that the laser light that is perpendicular to the parallel laser light and is parallel to the object when the surface on which the laser scanner is installed is set as a reference plane, thereby recognizing the shape of the object from the measurement result of the distance from the reference plane to the object, and the parabolic mirror is configured by an area portion obtained by slicing a part of the parabolic reflecting surface of the parabolic mirror, with both ends being in contact with the reference plane, so as to obtain a path along which the laser light emitted when the laser scanner is scanned in the first axial direction can be reflected so as to be irradiated to the object from directly above without passing through the laser scanner, and the control unit controls the scanning of the laser scanner in the first axial direction, and collects point cloud data for one slice of the object by irradiating the object with the parallel laser light along the path.

According to the present disclosure, it is possible to obtain a shape recognition device that can recognize the shape of an object while suppressing the occurrence of blind spots when using a single laser scanner.

FIG. 1 is a diagram for explaining a configuration of a shape recognition device according to a first embodiment of the present disclosure. 1 is a diagram for explaining a shape recognition method using a shape recognition device according to a first embodiment of the present disclosure. FIG. 1 is a cross-sectional view for explaining the configuration of a parabolic mirror according to a first embodiment of the present disclosure, and is a cross-sectional view taken along the irradiated laser light. FIG. 2 is a cross-sectional view for explaining the configuration of the parabolic mirror according to the first embodiment of the present disclosure, showing the state when the parabolic mirror side is viewed from the reference plane. 5A to 5C are diagrams for explaining a method for recognizing a three-dimensional shape using the parabolic mirror described in FIGS. 3 and 4 according to the first embodiment of the present disclosure. FIG. 1 is an explanatory diagram relating to a helical scan method according to a first embodiment of the present disclosure. FIG. 2 is an explanatory diagram relating to a corrected helical scan method according to the first embodiment of the present disclosure. FIG. 1 is an explanatory diagram showing a blind spot problem in the prior art. FIG. 1 is a diagram for explaining problems with the conventional technology. FIG. 1 is a diagram for explaining problems with the conventional technology. FIG. 1 is a diagram for explaining problems with the conventional technology.

A preferred embodiment of the shape recognition device of the present disclosure will be described below with reference to the drawings. The shape recognition device of the present disclosure has a technical feature in that it is configured to irradiate an object with laser light from directly above using a parabolic mirror. With such a configuration, it is possible to realize the effect of suppressing blind spots caused by the object and enabling shape recognition over the entire irradiation range of the laser light, even for objects whose shapes are not known in advance.

Embodiment 1.
1 is a diagram for explaining the configuration of a shape recognition device according to a first embodiment of the present disclosure. The shape recognition device according to the first embodiment is configured to include a parabolic mirror 10 and a laser scanner 20. The parabolic mirror 10 has a parabolic surface for reflecting light.

The laser scanner 20 is installed at the focal position of the parabolic mirror 10 and emits the laser light necessary to recognize the shape of the object 1. Here, the laser scanner 20 according to the first embodiment is characterized in that it does not directly irradiate the object 1 with laser light, but irradiates the laser light toward the parabolic surface and indirectly irradiates the object with the collimated laser light via the parabolic mirror 10.

In other words, the laser light in this disclosure is indirectly irradiated onto the object 1 via the parabolic mirror 10, so that it is irradiated in parallel from directly above the object 1. As a result, a simple configuration using a single laser scanner makes it possible to suppress shadows (blind spots) caused by the object.

Figure 2 is a diagram for explaining a shape recognition method using a shape recognition device according to embodiment 1 of the present disclosure. Figure 2 shows a cross-sectional view of an object 1 placed on an installation surface P1, in which laser light is irradiated from a laser scanner 20 via a parabolic mirror 10.

The laser scanner 20 installed at the focal position of the parabolic mirror 10 is disposed on the reference plane P2 as shown in FIG. 2. In other words, the reference plane P2 is a plane perpendicular to the laser light collimated by the parabolic mirror 10, and corresponds to a plane including the focal position where the laser scanner is installed.

The symbols shown in FIG. 2 are defined as follows.
θ1: the angle of the laser light emitted from the laser scanner 20 with respect to the reference plane P1.
L1: the distance from the laser scanner 20 to the parabolic mirror 10, which is a known value as a function of θ at the time of designing the parabolic mirror 10.
N1: the distance from the parabolic mirror 10 to the reference surface, which is a known value as a function of θ at the time of designing the parabolic mirror 10.
K: the distance between the reference plane P2 and the installation plane P1, which is a known value depending on the installation position of the parabolic mirror 10.
H1: The distance from the reference plane P2 to the surface of the object 1.
M1: Height of object 1.

For example, when the laser beam is irradiated at an angle θ1, the measurement distance by the laser scanner 20 is given by the following formula (1).
Measurement distance (θ1)=L1+N1+H1 (1)

Here, L1 and N1 are values that are known as functions of θ1 at the time of designing the parabolic mirror 10, so the distance H1 at angle θ1 can be determined by subtracting L1 and N1 from the measurement result of the measured distance (θ1) according to the above formula (1). Therefore, the height M1 of the object 1 can be found by subtracting the value of H1 from the known value of K. That is, M1 can be found according to the following formula (2).
M1 = L1 + N1 + K - measured distance (θ1) (2)

By gradually changing the irradiation angle θ of the laser scanner and calculating the height M at the position of the object 1 corresponding to each irradiation angle, the two-dimensional shape of the object 1 can be obtained.

For two-dimensional shape recognition, only one cross-sectional portion of the parabolic mirror 10 needs to be used as a reflecting mirror, and the specific configuration of the parabolic mirror 10 will be described in detail using Figures 3 to 5.

Figure 3 is a cross-sectional view for explaining the configuration of the parabolic mirror 10 according to the first embodiment of the present disclosure, and is a cross-sectional view along the irradiated laser light. Also, Figure 4 is a cross-sectional view for explaining the configuration of the parabolic mirror 10 according to the first embodiment of the present disclosure, and is a cross-sectional view showing the state when the parabolic mirror 10 side is viewed from the reference plane P2.

To obtain the two-dimensional shape of the target object 1, it is sufficient to use only a portion of the parabolic mirror shown by the solid line in Fig. 3 and Fig. 4 as the parabolic mirror 10. In other words, the parabolic mirror 10 can be configured with an area portion that can reflect the laser light emitted when the laser scanner 20 is scanned in the first axial direction.

Figure 5 is a diagram for explaining a method for recognizing a three-dimensional shape using the parabolic mirror 10 described in Figures 3 and 4 according to the first embodiment of the present disclosure. At a certain position, the laser scanner 20 irradiates the parabolic mirror 10 with laser light while sequentially changing the angle θ, thereby making it possible to measure the two-dimensional shape at one cross section.

As shown in FIG. 5, the traveling device 30 on which the parabolic mirror 10 and the laser scanner 20 are mounted is configured to be movable in a second axis direction perpendicular to the first axis direction, which is the scanning direction of the laser scanner 20. Therefore, the three-dimensional shape of the target object 1 can be recognized by controlling the traveling device 30 to move and repeatedly measure the two-dimensional shape at different cross-sectional positions.

There are four possible scanning methods for recognizing a three-dimensional shape:
<Case 1: Method for measuring the two-dimensional shape of one cross section while the traveling device 30 is stopped>
In the method of Case 1, the laser light is irradiated onto the object 1 while the traveling device 30 is stopped. After performing shape recognition for one slice, the traveling device 30 is moved by a predetermined pitch, and shape recognition for the next slice is performed. By sequentially repeating this pitch feed, it becomes possible to recognize a three-dimensional shape.

<Case 2: Helical scan method>
In the method of Case 2, the traveling device 30 is moved at a constant speed while continuously laser scanning the object 1. Fig. 6 is an explanatory diagram of the helical scanning method in the first embodiment of the present disclosure. In the helical scanning method of Case 2, three-dimensional shape recognition is performed while obliquely cutting the object 1 into slices at an angle obtained by combining the scanning speed of the laser scanner 20 and the scanning speed of the traveling device 30, as shown in Fig. 6.

<Case 3: Corrected helical scan method>
In the method of Case 3, the laser light scanning trajectory of the laser scanner 20 is inclined by an angle that cancels the scanning speed of the traveling device 30. Fig. 7 is an explanatory diagram of the corrected helical scan method in the first embodiment of the present disclosure. As shown in Fig. 7, by inclining the scanning direction by the laser scanner 20, the target object 1 can be sliced in a direction perpendicular to the traveling direction of the traveling device 30, and three-dimensional shape recognition can be performed.

In addition, in the above-mentioned cases 1 to 3, the parabolic mirror 10 and the laser scanner 20 are moved using the travel device 30, but the present disclosure is not limited to such a configuration. Three-dimensional shape recognition is also possible by disposing the parabolic mirror 10 and the laser scanner 20 in a fixed position and using a travel device 30 that moves the target object 1 side relatively.

In other words, the traveling device 30 needs to be able to move in one axial direction while carrying the parabolic mirror 10 and the laser scanner 20, or the target object 1.

<Case 4: Stereoscopic Scanning Method>
3 to 5, instead of using only one cross-sectional portion of the parabolic mirror as the parabolic mirror 10, the entire parabolic mirror is used as the parabolic mirror 10. In this case 4, a three-dimensional laser scanner is used as the laser scanner 20, and the laser light is scanned over the entire parabolic mirror, without moving the parabolic mirror 10 and the laser scanner 20, or the target object 1, to collect three-dimensional point cloud data.

As described above, according to the first embodiment, a shape recognition device can be constructed that is configured to irradiate an object with laser light from directly above using a parabolic mirror. As a result, it is possible to suppress blind spots where the object cannot be irradiated with laser light, thereby realizing the effect of enabling shape recognition over the entire irradiation range of the laser light.

In addition, by using only one cross-sectional portion of the parabolic mirror, it is possible to recognize the two-dimensional shape of one cross-section of the object. Furthermore, by repeating two-dimensional shape recognition while moving the parabolic mirror and the laser scanner, or while moving the object, it is possible to easily recognize the three-dimensional shape of the object with a simple configuration.

Although not shown in the figure, scanning control of the laser scanner 20 and travel control of the travel device are performed by a controller equivalent to the control unit, and the control unit can realize two-dimensional shape recognition and three-dimensional shape recognition depending on the application.

1 Object, 10 Parabolic mirror, 20 Laser scanner, 30 Traveling device.

Claims (2)

a parabolic mirror having a parabolic surface that reflects light;
a laser scanner that is installed at a focal position of the parabolic mirror and irradiates a laser beam toward the parabolic surface, thereby irradiating a target object with the laser beam that has been made parallel via the parabolic mirror from directly above;
A control unit that executes scanning control of the laser scanner,
the control unit executes scanning control of the laser scanner so that the parallel laser light is irradiated onto the object when a surface perpendicular to the parallel laser light and on which the laser scanner is installed is taken as a reference surface, thereby recognizing a shape of the object from a measurement result of a distance from the reference surface to the object;
the parabolic mirror is configured by a region obtained by slicing a part of a parabolic reflecting surface of the parabolic mirror, with portions in contact with the reference surface as both ends , so as to obtain a path along which the laser light emitted when the laser scanner is scanned in a first axial direction can be reflected so as to be irradiated from directly above the target object without passing through the laser scanner,
The control unit controls the laser scanner to scan in the first axis direction, and collects point cloud data for one slice of the object by irradiating the collimated laser light along the path to the object. The parabolic mirror and the laser scanner are mounted, or the object is mounted, and the traveling device is movable in a second axis direction perpendicular to the first axis direction;
The shape recognition device according to claim 1 , wherein the control unit is configured to move the traveling device along the second axial direction to thereby enable three-dimensional shape recognition of the object.