JP2009198382A - Environment map acquiring device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To accurately measure a shape, without being affected by the irregularities of the road surface or by installation errors. <P>SOLUTION: An environment map acquiring device is provided with: a low-precision measuring domain specification part 13 for specifying a measuring domain of point group data having low measurement precision; and a scan procedure changing part 15 for controlling the low angle of a movable stand 4 so that the scanning angle range of laser light includes the measuring domain specified by the low-precision measuring domain specifying part 13, and collecting additional point group data measured by a laser scanner 5. In the device, a shape-specifying part 17 synthesizes point group data collected by a scan processing part 11, with the additional point group data collected by the scan procedure changing part 15, and specifies the shape of a reflection object from the composited point group data. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to an environmental map (in the case of outdoor, for example, the shape of a structure such as the ground, surrounding buildings, guardrails, etc.) corresponds to the environmental map, and in the case of indoor, for example, a wall, ceiling, desk, partition, table, etc. The present invention relates to an environmental map acquisition device that acquires the shape of furniture or the shape of a bottle or can placed on a floor corresponds to an environmental map.

A conventional environment map acquisition device measures a shape by bringing it to a place where a person wants to measure, and uses a device called a laser scanner as a measurement device (see, for example, Non-Patent Document 1).
This laser scanner is a distance measuring device that uses a laser. By irradiating the laser beam to the left and right in the horizontal direction, it irradiates the laser beam at various directional angles and determines the distance to the reflector present at the directional angle. To get.
The conventional environmental map acquisition device acquires the shape of the inside of a building by fixing the laser scanner to the carriage and obtaining the distance of the horizontal horizontal direction angle in the traveling direction of the carriage.

Hiraoka et al., “Robust Superposition of Time-Series Laser Scanner Data for Indoor Mapping” Photo Measurement and Remote Sensing, P37-41, VOL. 46, NO. 2,2007

  Since the conventional environmental map acquisition device is configured as described above, the measurement of the laser scanner occurs when there is an installation error of the laser scanner with respect to the carriage or the carriage tilts due to the unevenness of the road surface on which the carriage is placed. When the direction is not horizontal with respect to the ground, there is a problem that the shape cannot be measured accurately.

  The present invention has been made to solve the above-described problems, and an object of the present invention is to obtain an environmental map acquisition device that can accurately measure the shape without being affected by road surface unevenness and installation errors. To do.

  The environmental map acquisition device according to the present invention collects distance data indicating the distance measured by the laser distance measuring device while controlling the pitch angle of the movable table and changing the irradiation direction of the laser light by the laser distance measuring device. First distance data collecting means for analyzing, distance data collected by the first distance data collecting means, low-precision measurement area specifying means for specifying a measurement area of distance data having low measurement accuracy, laser measurement Distance data indicating the distance measured by the laser distance measuring device by controlling the low angle of the movable base so that the scanning angle range of the laser beam by the distance device includes the measurement area specified by the low precision measurement area specifying means. Second distance data collecting means for collecting the distance data, and the shape specifying means is collected by the distance data collected by the first distance data collecting means and the second distance data collecting means. The distance data is synthesized, is obtained so as to identify the shape of the reflector from the distance data after the synthesis.

  According to the present invention, the distance data indicating the distance measured by the laser distance measuring device is collected while changing the irradiation direction of the laser light by the laser distance measuring device by controlling the pitch angle of the movable base. Distance data collecting means, low-precision measurement area specifying means for analyzing distance data collected by the first distance data collecting means, and specifying a measurement area of distance data with low measurement accuracy, and laser by a laser distance measuring device First, collect distance data indicating the distance measured by the laser distance measuring device by controlling the low angle of the movable base so that the scanning angle range of the light includes the measurement area specified by the low-precision measurement area specifying means. 2 distance data collecting means, and the shape specifying means combines the distance data collected by the first distance data collecting means and the distance data collected by the second distance data collecting means. And, since it is configured to identify the shape of the reflector from the distance data after synthesis, without being affected by the road surface irregularities and installation error, there is an effect that can be measured accurately shape.

Embodiment 1 FIG.
FIG. 1 is a perspective view showing a state where an environmental map acquisition apparatus according to Embodiment 1 of the present invention is mounted on a moving table. FIG. 2 is a block diagram showing an environment map acquisition apparatus according to Embodiment 1 of the present invention.
1 and 2, the moving base 1 has two front wheels 2 attached to the left and right sides of the front, and one rear wheel 3 attached to the center of the rear.
A stepping motor is mounted on the two front wheels 2 and rotates in an arbitrary direction by a specified angular displacement amount according to a control signal output from the computer 6. The front wheel 2 has a tachometer attached thereto, and transmits the rotation angle detected by the tachometer to the computer 6.
The rear wheel 3 is a wheel that can be bent left and right.

The movable table 4 is configured using, for example, an automatic precision stage or a device called an electric pan head, and is fixed to the upper surface of the movable table 1. The movable table 4 can rotate in the vertical direction (pitch direction) with respect to the upper surface portion of the movable table 1 and in the left-right direction (low direction) centered on the axis that goes in the traveling direction.
That is, the movable table 4 rotates in the vertical direction (pitch direction) or the left-right rotation direction (low direction) with the traveling direction as an axis in accordance with a control signal output from the computer 6. Further, in order for the computer 6 to tilt the movable table 4 at an arbitrary pitch angle and rotation angle, an angle in the pitch direction and an angle from the horizontal in the low direction are detected at the center of the movable table 4. An angle meter detected by the inclinometer is transmitted to the computer 6.
Note that the azimuth direction can be rotated by an arbitrary displacement angle under the instruction of the computer 6 as in the pitch direction and the low direction. In this case, the arbitrary displacement angle is a displacement angle from the front angle.

The laser scanner 5 is installed on the upper surface of the movable table 4 and rotates horizontally from side to side about a light source 5a that irradiates laser light, thereby moving the laser light to a scanning angle range (the laser scanner 5 emits laser light). This is a laser distance measuring device that measures the distance from the time it takes for the reflected light of the laser beam (the laser beam reflected by the object) to return to the reflecting object.
The computer 6 is fixed to the upper surface portion of the moving table 1 and controls the front wheels 2 and the rear wheels 3 to carry out a process of moving the moving table 1 to any position and direction such as front and rear, left and right, and changing the orientation. At the same time, a process of moving the movable table 4 in an arbitrary pitch direction and a low direction is performed.
Further, the computer 6 collects distance data indicating the distance measured by the laser scanner 5 and performs a process of specifying the shape of the reflector from the distance data.

FIG. 3 is a block diagram showing details of the computer 6 of the environmental map acquisition apparatus according to Embodiment 1 of the present invention.
In FIG. 3, the scan processing unit 11 controls the pitch angle of the movable table 4 and changes the irradiation direction of the laser light by the laser scanner 5, and is a point group that is distance data indicating the distance measured by the laser scanner 5. Implement a process to collect data. The scan processing unit 11 constitutes a first distance data collection unit.
The point cloud data storage unit 12 is a memory that stores the point cloud data collected by the scan processing unit 11.

The low-accuracy measurement region specifying unit 13 analyzes the point cloud data collected by the scan processing unit 11 and performs a process of specifying the measurement region of the point cloud data with low measurement accuracy. The low accuracy measurement area specifying unit 13 constitutes a low accuracy measurement area specifying means.
The low-accuracy range data memory 14 is a memory for storing low-accuracy range data indicating a measurement region of point cloud data with low measurement accuracy specified by the low-accuracy measurement region specifying unit 13.

The scanning method changing unit 15 includes the low angle of the movable table 4 so that the scanning angle range of the laser beam by the laser scanner 5 includes the measurement region of the point cloud data with the low measurement accuracy specified by the low accuracy measurement region specifying unit 13. And a process of collecting additional point cloud data that is distance data indicating the distance measured by the laser scanner 5 is performed. The scanning method changing unit 15 constitutes second distance data collection means.
The additional point cloud data storage unit 16 is a memory for storing the additional point cloud data collected by the scan method changing unit 15.

  The shape specifying unit 17 combines the point cloud data stored in the point cloud data memory 12 and the additional point cloud data stored in the additional point cloud data storage unit 16, and reflects the reflected object from the combined point cloud data. The process of specifying the shape of In addition, the shape identification part 17 comprises the shape identification means.

First, the overall operation will be described.
The movable platform may be automatically advanced or may be pushed forward by a person. First, the scan processing unit 11 automatically acquires point cloud data at regular time intervals. Immediately after this, the low-accuracy range data is accumulated by the operation of the low-accuracy measurement region specifying unit 13. At this time, if the low-accuracy region data exists, the scanning method changing unit 15 operates and sends an instruction to the movable table and the laser scanner 5 to change the laser acquisition position and the acquisition method. Collect group data.
Next, the operation of each unit will be described.
First, the scan processing unit 11 of the computer 6 controls the pitch angle and low angle of the movable table 4 to adjust, for example, the movable table 4 to be horizontal and adjust the measurement direction.
Next, the scan processing unit 11 collects distance data indicating the distance measured by the laser scanner 5 while controlling the pitch angle of the movable table 4 and changing the irradiation direction of the laser light from the laser scanner 5.

That is, the scanning processing unit 11 moves the pitch angle of the movable table 4 in increments of 1 degree from, for example, minus 20 degrees to plus 40 degrees, and causes the laser scanner 5 to perform one scan at each pitch angle. Instruct.
The laser scanner 5 oscillates the laser light in each scanning angle range within the scanning angle range, and measures the distance from the time until the reflected light of the laser light (laser light reflected by the object) returns to the reflecting object. .
When the scanning angle range of the laser scanner 5 is, for example, a range of minus 90 degrees to plus 90 degrees, and the laser scanner 5 measures the distance with a displacement of 1 degree, 181 distance data are used as point group data as a scan processing unit. 11 is output.
As described above, the scan processing unit 11 collects 61 sets of point cloud data from the laser scanner 5 when the pitch angle of the movable table 4 is movable in increments of 1 degree from, for example, minus 20 degrees to plus 40 degrees, 61 sets of point cloud data are stored in the point cloud data storage unit 12.

Here, FIG. 4 is an explanatory diagram showing an example of point cloud data.
In the example of FIG. 4, the position of the center of the laser scanner 5 is represented by orthogonal coordinates X, Y, and Z as the measurement position.
The reference of the measurement position is, for example, Y as the north direction, X as the east direction and Z as the height direction, with the starting point of the environment map acquisition device as the origin, and the unit is 1.0 per meter.
The measurement azimuth indicates the azimuth on which the laser scanner 5 is directed on a horizontal plane, with north being 0 and a clockwise angle.
The case where the azimuth is accurately known is a case where a geomagnetic sensor is added to the inclinometer. In this case, when the geomagnetic sensor detects the direction of the movable platform 4 as a deviation of the angle from the north direction and sends it to the computer 6, the computer 6 moves the wheels 2 and 3 and moves by a certain number of revolutions. The absolute position is calculated from the distance (obtained from the rotation speed and the wheel diameter). If the orientation cannot be obtained without having a geomagnetic sensor, the pitch and low angle should be recorded accurately, and the data should be recorded assuming that the orientation is the north direction and the absolute position is the starting point. And

The measurement results are pitch angles P1, P2,..., Low angles R1, R2,..., Measurement directions θ-m, θ-m + 1,. −m) to D1 (θ + m), D2 (θ−m) to D2 (θ + m),.
For example, D1 (θ-m) is a distance measurement value of the laser scanner 5 in the direction of the horizontal angle θ-m, and has a centimeter accuracy, for example.
The angles are 2m + 1 directions up to θ−m, θ−m + 1,. The angular resolution varies depending on the accuracy of the laser scanner 5, but is, for example, 1 degree.

When the scan processing unit 11 collects point cloud data, the low accuracy measurement region specifying unit 13 of the computer 6 analyzes the point cloud data and specifies a measurement region of point cloud data with low measurement accuracy.
That is, the low-precision measurement area specifying unit 13 specifies an area where it is difficult to accurately measure the shape of the reflector. For example, the shape near the edge of the reflector (distance data near the edge) is accurate. Therefore, it is mainly intended to specify the position of the edge.

Hereinafter, the processing content of the low-accuracy measurement region specifying unit 13 will be specifically described.
As shown in FIG. 6, the error range of the displacement in the scanning direction of the laser scanner 5 is kept constant depending on the product specifications of the laser scanner 5.
On the other hand, the posture of the movable platform 1 may not be horizontal depending on the road surface condition.
In addition, since the level of the movable table 4 is not completely guaranteed, the distance measurement of the horizontal surface cannot be performed only by adjusting the pitch angle and the low angle of the movable table 4.
Furthermore, when an inexpensive inclinometer with low accuracy is used, the orientation of the movable table 4 itself cannot be accurately taken, and the direction of the measurement surface of the laser scanner 5 becomes inaccurate.

When the scanning surface of the laser scanner 5 (the surface on which the laser scanner 5 oscillates the laser beam) is moved up and down, as shown in FIG. 7, the laser beam may penetrate through the object due to imbalance in the shape of the object (B , D, and E), and laser light may strike the side of the object (see A and C). In such a case, a certain scanning surface may be perceived as an uneven shape.
Also, if the scan plane that was intended to be adjusted horizontally is not horizontal, as shown in FIG. 8, the laser beam penetrates the top of the box (see D and E), and the shape cannot be determined. There is.
In addition, as shown in FIG. 9, when the pitch of the scan surface is gradually increased, on one surface, all the laser beams pass beside the box, and on one surface, all the laser beams pass over the box. There is something wrong. In such a case, the angular displacement during this period is not an accurate value with the accuracy of moving the movable table 4. In other words, it is not an accurate value such as the exact number of displacements or the absolute angle with respect to the horizontal.

Therefore, in the low-accuracy measurement region specifying unit 13, as shown in FIG. 7, in the same scan data (point group data collected by the scan processing unit 11 at a certain pitch angle), the direction angle at which the measurement points are adjacent to each other. If there is more than a certain variation in (for example, if the dispersion of N distance values at successive measurement azimuths is more than a certain value), the variation is as low-accuracy range data (data indicating a certain range of point cloud data). Are recorded in the low accuracy range data storage unit 14 in a certain range of pitch angles, low angles, and measurement points in the measurement direction.
Here, FIG. 5 is an explanatory diagram showing an example of the low accuracy range data.
The format of the low accuracy range data is the same as the format of the point cloud data, but only data relating to the pitch angle, low angle, and measurement direction of the low accuracy range is recorded.

  Further, the low-accuracy measurement region specifying unit 13 also has a variation range as the low-accuracy range data even when there is a variation of a certain distance or more in continuous measurement azimuth angles as in C and D of FIG. The pitch angle, the low angle, and the measurement points in the measurement direction are recorded in the low accuracy range data storage unit 14.

Further, as shown in FIG. 9, the low-accuracy measurement region specifying unit 13 varies as low-accuracy range data even when the distance measurement error of the same measurement azimuth angle exceeds a certain value at successive pitch angles. Are recorded in the low accuracy range data storage unit 14 in a certain range of pitch angles, low angles, and measurement points in the measurement direction.
In FIG. 9, it is assumed that the measurement points of Y, A, B, C, D and Z, E, F, G, H are the same measurement direction with respect to the measurement points of X, P, Q, R, S. A portion (P, Q, R, S or A, B, C, D) where the distance is changed at a constant increase / decrease rate is defined as one object surface.
At this time, since A and E, A and F, A and G, and A and H each have a certain distance displacement, the pitch angle, the low angle, and the measurement direction in which Y, A, B, C, and D are measured. And the measurement points in the pitch angle, low angle and measurement direction in which Z, E, F, G, and H are measured are recorded as low-accuracy range data.
7, 8, and 9 are all convex corners of a box or the like, but even a concave corner specifies a portion that can be measured only with low accuracy by the same processing.

When the low-accuracy measurement region specifying unit 13 records the low-accuracy range data in the low-accuracy range data storage unit 14, the scanning method changing unit 15 of the computer 6 changes the scanning angle range of the laser beam by the laser scanner 5 to the low-accuracy range. The low angle of the movable table 4 is controlled so as to include the measurement region indicated by the data, and distance data indicating the distance measured by the laser scanner 5 is collected.
Hereinafter, the processing content of the scanning method change part 15 is demonstrated concretely.

The scan technique changing unit 15 tilts the new scan surface of the laser scanner 5 perpendicularly to the edge direction so that the shape of the reflector in the measurement region of the point cloud data with low measurement accuracy can be accurately measured. Determine the direction.
For example, the scan surface of the laser scanner 5 is determined as follows.
(1) In the example of FIG. 7, since the angle of the edge is unknown, a plane perpendicular to the original scan plane is determined as a new scan plane (plane indicated by a dotted line).
(2) In the example of FIG. 8, by referring to the scan data of a plurality of pitch angles, there is an edge between 2 and 4 having a difference in measurement distance in the same measurement direction as shown in FIG. It is determined that there is an edge between 3 and 5, and a direction perpendicular to the directions 2 and 3 or the directions 4 and 5 is determined as a new scan plane (a plane indicated by a dotted line in FIG. 10).
(3) In the example of FIG. 9, a plane perpendicular to A, B, C, and D is determined as a new scan plane (a plane indicated by a dotted line).

When the scanning method changing unit 15 determines a new scanning surface of the laser scanner 5, as shown in FIG. 11, the scanning method changing unit 15 changes the orientation of the movable table 4 so that the scanning surface of the laser scanner 5 coincides with the determined scanning surface. Control is performed (the direction of the movable table 4 is controlled so that the scan surface is perpendicular to the edge to be measured), and a scan execution command is output to the laser scanner 5.
Thereby, the range including the low-accuracy measurement region is scanned by the laser scanner 5 shaking the laser beam.
As shown in FIG. 11, when a range including the low-precision measurement region is scanned and another edge is newly detected, data in the range of the edge newly detected from the measurement target edge (FIG. 11). In the example of FIG. 6, the additional point cloud data storage unit 16 stores the additional point cloud data as additional point cloud data.
The format of the additional point cloud data is the same as the format of the low-accuracy range data, and the content of the data is a newly added measurement position, measurement orientation, and measurement result (measurement orientation and its distance).

When the scan method changing unit 15 stores the additional point cloud data in the additional point cloud data storage unit 16, the shape specifying unit 17 of the computer 6 stores the point cloud data and the additional point cloud data stored in the point cloud data memory 12. The additional point cloud data stored in the storage unit 16 is combined, and the shape of the reflector is specified from the combined point cloud data.
That is, the shape specifying unit 17 specifies the shape of the reflector as follows.

Since the center position of the light source 5a in the laser scanner 5 does not move, the shape specifying unit 17 uses the position of the light source 5a in the laser scanner 5 as an origin and the unit vector to the measurement direction of the point cloud data (see FIG. 4). Then, the angle conversion of the measurement direction angle, the low rotation angle, the azimuth rotation angle, and the pitch rotation angle is added, and the vector after the angle conversion is obtained for all the measurement results.
Similarly, the unit vector to the measurement direction of the additional point cloud data is subjected to angle conversion of the measurement direction angle, low rotation angle, azimuth rotation angle, and pitch rotation angle, and all the vectors after angle conversion are measured. Ask for results.

Next, the shape specifying unit 17 compares the angle of the vector after the angle conversion obtained from the measurement result of the point cloud data and the angle of the vector after the angle conversion obtained from the measurement result of the additional point cloud data.
If the angle difference between the two vectors is less than a certain angle, the point cloud data measurement result related to the vector is deleted, and the remaining point cloud data measurement result and additional point cloud data measurement result (additional) The shape of the reflector is specified from the measurement result of the point cloud data in the same manner as the measurement result of the point cloud data.
An example of a method for identifying the shape of a reflector from the point cloud data measurement results is as follows. The process of generating a vector that connects close points is repeated. At this time, the vector is generated only when the distance is within a certain distance.
Next, the connected components connected by the vectors are grouped into one group, and a vector that connects the points closest to each other in the points in one group and the points in the other group is generated. Similar vector generation is repeated for other groups except for groups connected by vectors. At this time, the vector is generated only when the distance is within a certain distance.
As a result of repeating such repetition until no vector can be generated, the shape having the generated vector group as a plane can be reproduced as data representing the shape of the surface of the reflector.

  As apparent from the above, according to the first embodiment, the point angle measured by the laser scanner 5 while controlling the pitch angle of the movable table 4 and changing the irradiation direction of the laser light by the laser scanner 5. A scan processing unit 11 that collects data, a point cloud data collected by the scan processing unit 11, a low accuracy measurement region specifying unit 13 that specifies a measurement region of point cloud data with low measurement accuracy, and a laser scanner The additional point cloud data measured by the laser scanner 5 by controlling the low angle of the movable table 4 so that the scanning angle range of the laser beam by 5 includes the measurement region specified by the low-precision measurement region specifying unit 13. And a scan method changing unit 15 that collects the point cloud data collected by the scan processing unit 11 and the additional information collected by the scan method changing unit 15. Since the point cloud data is synthesized and the shape of the reflector is specified from the synthesized point cloud data, the shape can be accurately measured without being affected by road surface irregularities and installation errors. There is an effect.

That is, even when the resolution of the movable angle of the movable table 4 is low or the laser scanner 5 itself is tilted, edge portions such as desks, partitions, and wall edges can be measured with high accuracy.
In the example of FIG. 11, the distance from edge to edge becomes accurate by oscillating the laser beam vertically (in the example of FIG. 9, the pitch angle is left to the operation of the movable table 4, so the pitch can be accurately adjusted. I don't know if there is a difference, and I can't measure it accurately).

Embodiment 2. FIG.
12 is a block diagram showing details of the computer 6 of the environmental map acquisition apparatus according to the second embodiment of the present invention. In FIG. 12, the same reference numerals as those in FIG.
Similarly to the scan processing unit 11 in FIG. 3, the scan processing unit 21 controls the pitch angle of the movable table 4 to change the irradiation direction of the laser light by the laser scanner 5 and change the distance measured by the laser scanner 5. A process of collecting point cloud data, which is distance data shown, is performed.
However, the scan processing unit 21 collects point cloud data at each installation position each time the environment map acquisition device is moved and the movable platform 4 is installed at a different position. The scan processing unit 21 constitutes a distance data collection unit.
The point cloud data storage unit 22 is a memory that stores the point cloud data collected by the scan processing unit 21.

The feature shape extraction unit 23 reads a plurality of point group data acquired from different positions of the movable platform 4 from the point group data storage unit 22, and performs a process of extracting characteristic shapes from the plurality of point group data. Note that the feature shape extraction unit 23 constitutes a feature shape extraction unit.
The feature data storage unit 24 is a memory that stores feature data indicating the characteristic shape extracted by the feature shape extraction unit 23.

The shape associating unit 25 compares a plurality of characteristic shapes extracted by the characteristic shape extracting unit 23 with each other, and performs a process of associating characteristic shapes indicating the same space. The shape association unit 25 constitutes a shape association unit.
The shape specifying unit 26 synthesizes the point cloud data from which the characteristic shapes associated with each other by the shape association unit 25 are combined, and performs the process of specifying the shape of the reflector from the combined point cloud data. The shape specifying unit 26 constitutes a shape specifying means.

First, the overall operation will be described.
The movable table 4 may be automatically advanced or may be pushed forward by a person. First, the scan processing unit 21 automatically acquires point cloud data at regular time intervals. For example, a person may move the movable table 4 and perform a scanning process at various places. A plurality of movable platforms 4 may be moved separately, and the point cloud data acquired by each movable platform 4 may be collected into one point cloud data.
Next, the feature shape extraction unit 23 extracts features of individual point cloud data and records them as feature data. Next, the shape associating unit 25 extracts the feature data to be associated. In accordance with this, the shape specifying unit 26 performs processing for connecting the point cloud data.
Next, each operation will be described.
First, the scan processing unit 21 of the computer 6 controls the pitch angle and the low angle of the movable table 4 to adjust the movable table 4 to be horizontal, for example, in the same manner as the scan processing unit 11 of FIG. , Adjust the measurement direction.
Next, as with the scan processing unit 11 in FIG. 3, the scan processing unit 21 controls the pitch angle of the movable table 4 and changes the laser beam irradiation direction by the laser scanner 5 while measuring with the laser scanner 5. Collect distance data indicating the distances that have been placed.

In other words, the scan processing unit 21 moves the movable platform 4 with a pitch angle of, for example, minus 20 degrees to plus 40 degrees in increments of 1 degree, and causes the laser scanner 5 to perform one scan at each pitch angle. Instruct.
The laser scanner 5 oscillates the laser light in each scanning angle range within the scanning angle range, and measures the distance from the time until the reflected light of the laser light (laser light reflected by the object) returns to the reflecting object. .
When the scanning angle range of the laser scanner 5 is, for example, a range of minus 90 degrees to plus 90 degrees, and the laser scanner 5 measures the distance with a displacement of 1 degree, 181 distance data are used as point group data as a scan processing unit. 21 is output.
As described above, the scan processing unit 21 collects 61 sets of point cloud data from the laser scanner 5 when the pitch angle of the movable table 4 is movable in increments of 1 degree from, for example, minus 20 degrees to plus 40 degrees, 61 sets of point cloud data are stored in the point cloud data storage unit 22.

However, unlike the scan processing unit 11 in FIG. 3, the scan processing unit 21 has 61 sets of point groups at each installation position each time the environment map acquisition device is moved and the movable platform 4 is installed at a different position. Collect data.
In addition, when the installation position of the movable platform 4 changes due to the movement of the movable platform 1 of the environment map acquisition device, the measurement position is recorded with the departure point as the origin.
On the other hand, when the user lifts the environment map acquisition device and moves it from an arbitrary point to another point, the measurement position and the measurement direction are not recorded under the instruction of the operating user.
The format of the point cloud data is the same as that in the first embodiment (see FIG. 4).

When the scan processing unit 21 stores the point cloud data of each installation position in the point cloud data storage unit 22, the feature shape extraction unit 23 reads the point cloud data of each installation position from the point cloud data storage unit 22, and each installation Each characteristic shape is extracted from the point cloud data of the position.
That is, the feature shape extraction unit 23 extracts shapes that serve as marks, such as desk corners and partition edges, from the point cloud data at each installation position.
Then, the feature shape extraction unit 23 uses the relative position of the shape to be a mark, the value representing the linearity of the feature part, the value representing the rectangle, and the value representing the inclusion relation of the shape as the feature data, and the feature data storage unit 24. Output to.

As a technique for extracting a characteristic shape from point cloud data, a known technique may be used. Here, an example in which a straight line is extracted as a characteristic shape from point cloud data will be described.
FIG. 14 is an explanatory diagram showing a method of extracting a straight line as a characteristic shape. In the figure, the upper four laser trajectories pass through the top of the box shape, and of the four left and right including point 1, the left two hit the wall and the right two extend to the part without the wall. The left and right four of 2 and the four left and right of point 3 are the same as point 1.
First, the feature shape extraction unit 23 pays attention to the fact that the adjacent pitch angles and the distances in the adjacent measurement directions are abruptly changed as indicated by points 1, 2, and 3 in FIG. 14.
In the example of FIG. 14, although it is a convex part, if it is a recessed part, a distance does not change suddenly, but becomes a part where the distance becomes the longest (this is a point reaching the farthest part of the concave part. ).

Then, the feature shape extraction unit 23 obtains a straight line that minimizes the sum of the distances from the coordinates of the points 1, 2 and 3 (the coordinates of the light source 5a are set to “0”). As a general method, the least square method is used. Or, for example, two points are extracted from three points, and a unit vector of inclination connecting the two points is obtained and combined, and a straight line passing through one of the three points is determined using the combined vector as an inclination vector. .
Then, among the points 1, 2 and 3, the laser trajectory having a continuous pitch angle, or both ends of the continuous measurement direction (in the example of FIG. 14, the pitch angle which is continuous with the points 1, 2, and 3) Among the arrival points of the laser trajectory, point 1 and point 3 are both ends) and recorded as feature data.
A plurality of feature data may be generated from one point cloud data. In this case, a number is assigned so that each original feature data can be identified.

FIG. 13 is an explanatory diagram showing an example of feature data.
In the example of FIG. 13, feature points are points at both ends that are features and the slope of a straight line connecting the points at both ends.
Note that when there are a large number of points representing one feature, inclination data obtained from a point group representing the feature is used. Also, a unique number is assigned to all feature data.

When the feature shape extraction unit 23 extracts characteristic shapes from the point cloud data at the respective installation positions, the shape association unit 25 compares the plurality of characteristic shapes with each other and displays the same space. The shape is associated and the corresponding part information is output to the shape specifying unit 26.
Here, the corresponding part information is information indicating which part of certain point cloud data corresponds to which part of other point cloud data among the point cloud data at each installation position.
Hereinafter, the processing content of the shape matching part 25 is demonstrated concretely.

The shape association unit 25 searches for feature data satisfying the following conditions from among a plurality of feature data stored in the feature data storage unit 24, and the number assigned to the feature data and corresponding part information The feature data (tilt, distance) and the like are output to the shape specifying unit 26.
[Case 1] Search for number pairs in which measurement position and measurement direction are recorded (part 1)
Condition (1) Two feature data conditions in which the inclination angle is constant or less (2) Two feature data in which the distance of the line segment is constant or less.
Condition (3) Compare feature data of point cloud data acquired from different measurement positions

[Case 2] Search for number pairs in which measurement position and measurement direction are recorded (part 2)
Condition (1) Condition for comparing feature data of point cloud data acquired from the same measurement position (2) Two feature data conditions with an inclination angle greater than or equal to (3) Two feature data conditions with a line segment distance less than or equal to (4) At least one of the points with the closest distance between the line segments of the two feature data satisfies the endpoint condition (5) Condition (1) to Condition (4) as a group.
(There may be multiple groups of feature data from the same measurement position)
For each set of feature data and feature data measured from other measurement positions, feature data matching the conditions (1) to (3) of case 1 is obtained.

[Case 3] Search for a pair of (A: feature data measured at the same measurement position and measurement direction) and (B: feature data measured at the same measurement position and measurement direction but different from A) (the coordinates of the measurement position are May or may not be fixed)
Condition (1) A and B feature data are compared, and feature data whose inclination angle is within a certain range are extracted. (2) Among feature data satisfying condition (1),
Ratio of (distance between feature data (two) of A) and (distance of feature data (two) of B)
Compared to the feature data satisfying the selection condition (3) condition (2)
(Difference in inclination between feature data of A) and (Difference in inclination between feature data of B)
And selecting a pair in which the difference between these differences is equal to or less than a certain value, the relative position between the feature data of A that simultaneously satisfies the conditions (4), (2), and (3)
), Those having the same relative position between feature data of B satisfying (3), A,
Select from B

Here, the processing contents of the shape association unit 25 will be described more specifically.
The angle of the arrangement of the point cloud data is such that the front direction of the laser scanner 5 is always treated as north, but when comparing two point cloud data with different installation positions and measurement directions, the angle is rotated to It is checked whether the length and direction of the straight line portion that is the same.
For example, as shown in FIG. 15, a curve and a straight line are detected from the point group, and further converted into a one-dimensional array for comparison. As an example, an example in which the shape of an intersection line between a horizontal portion such as a floor (or ceiling) and a wall can be acquired as a result of extracting a concave portion between the floor (or ceiling) and a wall as a feature amount. The end points of a curve or a straight line are expressed by the types shown in FIG.
As an example, L is a straight section, S is a blank section, E is a point where the end of the straight line is not a corner, and it is understood what the destination of the laser reaches, such as the limit of the reach of the laser and the maximum value of the scan range. The point that does not exist is expressed as U.
As for the end points of the straight line, A is a point that is bent to the back at the left end, and B is a point that is bent to the back at the right end. Let C be the point that is bent toward the front at the left end, and D be the point that is bent toward the front at the right end.
Further, a crank with the left side closer to the front than the right is G, and a crank with the right side closer to the front than the left is referred to as H.
Note that this expression example is a description example of the shape feature about the horizontal plane, and therefore, for the upper and lower surfaces and the oblique surface, the expression “left and right” may be applied to the vertical and oblique angles.

At the time of conversion to a one-dimensional array, the length and angle are rounded to several levels. By performing such rounding processing, for example, when there is a power cord or the like near the wall, depending on the timing and direction, it reflects on one of the point cloud data acquired from two locations, it does not reflect on the other, Different wall data.
About a continuation of a curve, you may make it recognize as one big curve with the direction of a curve start and a curve end. In this case, the error can be reduced.
In addition, when the point cloud data is regarded as a plane instead of a horizontal plane, a small cluster of points deviating from a certain plane may be removed (see FIG. 16). In this case, the error can be reduced.

In matching between point cloud data in which the sensed position and the direction of the laser scanner 5 are different, there are cases where only partial matching is performed.
For example, if it hits a convex corner, the corner can be taken as shown in FIG. 23, but the shape behind it may not be well understood.
Even in the case of comparison using a one-dimensional array, it is possible that partial match is possible instead of complete match.
As a specific example, an example in which the shape of the intersection of the horizontal plane and the wall such as the floor (or ceiling) can be acquired as a result of extracting the concave portion between the floor (or ceiling) and the wall as a feature amount. In this example, since the tangent line between the floor (or ceiling) and the wall can be extracted, it is not affected by the inclination of the sensor. As shown in FIG. 17, when two point group data having different acquisition positions and directions are matched, the features 1 to 8 and A to M are extracted separately. At this time, a portion where no data is taken is characterized as a space. The length of the straight line in contact with the space is not fixed (assuming it may be longer).
When these are compared, there is a matching part as shown in FIG. At this time, the length of the straight line in which both ends are in contact with the bent portion is determined. However, when only one of the straight lines is in contact with the bent portion, the length of the straight line is not determined, so the length is not compared. For example, the 1 straight line, the L straight line, the 3 straight line, and the A straight line may be longer, so the lengths are not compared.

In the example of FIG. 19, the point cloud is viewed from directly above, but the point cloud is regarded as a surface, normals at various positions on the surface are obtained, and the inclination of the normal and the connection relation of the surfaces May be connected as a character arrangement.
In this case, when the surface is connected to the surface A (for example, the surface B, the surface C), and the surface is connected to the surface S (for example, the surface T, the surface U, the surface V), the normal line of A and the B The angle difference between the normal line and the angle difference between the normal line of A and the normal line of C is, for example, the angle difference between the normal lines of two surfaces in contact with the surface S among the surfaces S, T, U, and V. If they are equal and the direction difference between the tangent lines is equal, they are matched.
Use multiple devices to obtain environmental maps by combining point cloud data acquired at distant points, or point cloud data acquired from different directions by moving back and forth in a corridor. An environmental map of the location where the device has passed can be created (see FIG. 20).
Further, as a specific means for comparing one-dimensional arrays, for example, a DP (dynamic programming) matching method may be applied.

When the shape associating unit 25 associates the characteristic shapes indicating the same space, the shape specifying unit 26 synthesizes the point cloud data from which the characteristic features are associated with each other. The shape of the reflector is specified from the point cloud data.
That is, the shape specifying unit 26 refers to the corresponding part information output from the shape associating unit 25, and moves one point cloud data in parallel or rotates the other point having the corresponding relationship. Combine with group data.
That is, the average value of the difference between the numbers and inclinations of both feature data is obtained, and the entire point cloud data measured from one measurement position is tilted about the light source as much as this average value. After tilting the entire point cloud data, the distance of the corresponding line segment is calculated again to obtain an average value of the difference in distance, and the entire point cloud data is moved in parallel by the distance indicated by the average value.
In the example of FIG. 17, the difference between the inclinations of the 5th and C straight lines, the difference between the 3rd and A line inclinations, the difference between the 1st line and the L line inclination, and the difference between the 7th line and the J line inclination are within a certain range. If it is within, the point group on the left side of FIG. 17 is rotated by the inclination that is the average value of the difference, and the position of 4 bend and the position of B bend, the position of 8 bend and the position of K bend By moving the point group in parallel to the position where they overlap, the point group can be connected as shown at the right end of FIG.

  As apparent from the above, according to the second embodiment, the point group measured by the laser scanner 5 while controlling the pitch angle of the movable table 4 and changing the irradiation direction of the laser light by the laser scanner 5. A scan processing unit 21 that collects data, a feature shape extraction unit 23 that is installed at different positions of the movable table 4 and extracts characteristic shapes from a plurality of point cloud data collected by the scan processing unit 21; A plurality of characteristic shapes extracted by the characteristic shape extraction unit 23 are compared with each other, and a shape association unit 25 that associates characteristic shapes indicating the same space with each other is provided. Since the point cloud data from which the characteristic shapes associated by 5 are extracted are synthesized, and the shape of the reflector is specified from the synthesized point cloud data, it is measured at different positions. It enables joining together the point cloud data, a result, an effect that can be obtained a wide range of environmental map.

Note that the feature shape extraction unit 23 according to the second embodiment uses a method of extracting a straight line as a characteristic shape by paying attention to the fact that the adjacent pitch angle and the distance in the adjacent measurement direction change rapidly. As described above, a straight line may be extracted as a characteristic shape by performing a detection process using a generally known Hough transform.
Specifically, it is as follows.

First, before carrying out the detection process using the Hough transform, the distance between points when the point cloud data (laser acquisition point sequence) measured by the laser scanner 5 is projected on a plane parallel to the ground is within a threshold value. Point sequences are grouped for each connected component.
That is, if detection processing using the Hough transform is performed on all of the laser acquisition point sequences, since various linear components are included, straight line overdetection or detection omission may occur. For this reason, the point sequence is grouped before performing the detection process using the Hough transform.

By grouping the point sequence, it is divided into parts where point clouds are not continuous, such as the left wall, right wall, and front wall from the environmental map acquisition device, and only the linear direction suitable for each wall is divided. Since it can be detected, straight line overdetection and omission of detection can be eliminated.
FIG. 29 is an explanatory diagram showing an example in which a point group having a short distance is divided into three. Since it is divided like an ellipse, a long straight line can be extracted for each.

If a straight line is detected without dividing every point where the point group is not continuous, it is not known how many straight lines should be detected as a whole. Also, a small number of point sequences (point sequences indicating short walls) may not be detected.
FIG. 22 shows an example of excessive detection of a straight line, and FIG. 23 is an explanatory diagram showing an example of detection failure of a straight line.
If the point cloud is divided for each non-continuous portion, the connected component of only a short line is processed as one lump so that it can be detected.

Here, FIG. 24 is an explanatory diagram showing a portion where the distance between the straight line and the point changes suddenly.
For example, it is determined from the magnitude of the distance between the detected straight line and the point sequence whether the point sequence is a straight line or includes a curve.
In a sequence of points arranged in order, it is determined that a portion where the change in the distance between the straight line and the point is greater than or equal to a threshold in a certain section is a corner.
Thereby, it is possible to extract a minute depression (for example, a portion recessed from the lower surface of the hallway by a door or the like).

A structure such as a building is extracted with a feature that there are more vertical and horizontal surfaces orthogonal to a curved surface or an oblique surface.
Thereby, it is possible to extract a minute depression (for example, a portion recessed from the lower surface of the hallway by a door or the like).
An example in which the feature shape extraction unit 23 according to the second embodiment can acquire another three-dimensional feature amount will be described. Using a technique called 3D Hough, not only a straight line part is extracted from a point group on a horizontal plane, but also a surface such as a wall or ceiling from a three-dimensional laser point group consisting of laser reach distances in various pitch directions. A part can be extracted.
It is also possible to find a plane from the entire point group and obtain the features from the extracted plane with the end portion of the plane as a straight line.
The explanation of the principle of 3D Hough transform is detailed in the following two papers. Takayuki Nakata, et al., “Spatial rolling angle detection in 3D Hough transform”, Journal of the Institute of Image Electronics Engineers of Japan, Vol.31, No.5, pp.794--799, Sept. 2002) and Motoki Yoshikawa, et al., “Extraction of glasses using 3D Hough transform”, IEICE Technical Report, Pattern Recognition / Media Understanding, Vol.101, No.568 (20020110) pp.17-24.
When the X1, Y1, and Z1 coordinates of a point in the three-dimensional space (in this patent, the coordinates of the point group) are on a plane that can be expressed from the three parameters ρ1, θ1, and φ1 by the three-dimensional Hough transform, Vote for ρ1, θ1, φ1 in the ρ, θ, φ space. For all points in the point cloud, vote on the calculation whether or not to place on a plane that can be expressed by various values of ρ, θ, φ, and on a plane that can be expressed by ρ, θ, φ with a large number of votes It can be seen that there are many point cloud data.
Suppose that the plane direction is represented by φ as the angle from the north or the front direction of the laser scanner, the angle from directly above is represented by θ, and the distance from the laser light source to the object is represented by ρ. In this case, if voting is performed with various ρ, θ, and φ for all of the point groups, there are many useless calculations. First, among the point groups, a point group that is horizontal to the ground (in the example of FIG. 4, the measurement direction A is For the horizontal point group, the straight line portion extraction process is performed by the above-described two-dimensional Hough transform.
Next, assuming that there is a plane obtained by extending the extracted straight line directly above, the values of ρ, θ, and φ that are perpendicular to the extracted straight line are obtained, and only the values near this value are converted into the three-dimensional Hough transform. The voting process will be performed.
Thereby, it is possible to find a plane on which a large number of point clouds are placed with a small amount of calculation. A point cloud line (for example, a boundary line between a wall and a ceiling or a corner of a wall) at the edge of the plane can be extracted as a feature line.
Note that a method for acquiring shape features with a small amount of calculation in the feature shape extraction unit 23 according to the second embodiment will be described. Install the laser scanner and camera in the same orientation. Thereby, the inclination of the laser scanner and the inclination of the camera are made the same. Also, the front direction of the laser scanner and the line-of-sight direction of the camera are made to coincide.
The feature shape extraction unit 23 extracts a straight line portion that becomes a shaded edge in an image captured by the camera. Next, calculate the viewing angle of the camera from the focal length of the lens attached to the camera, and determine from the coordinate position in the image which angle of the azimuth and pitch the shaded part is from the direction of the line of sight of the camera. Estimating that there is a straight edge part in the point group in the direction of the pitch angle, and by checking whether there is a straight line part in the range of this azimuth plus or minus several degrees and pitch plus or minus several degrees, the straight line part from the whole Time to extract can be reduced.

Embodiment 3 FIG.
FIG. 25 is a block diagram showing details of the computer 6 of the environmental map acquisition apparatus according to Embodiment 3 of the present invention. In FIG. 25, the same reference numerals as those in FIG.
The shielding area estimation unit 31 analyzes the point cloud data stored in the point cloud data storage unit 12 and performs a process of estimating a shielding area whose distance is not measured by the laser scanner 5. In addition, the shielding area estimation part 31 comprises the shielding area estimation means.

In order to acquire the point cloud data of the occlusion area estimated by the occlusion area estimation unit 31, the computer position designation unit 32 performs a process of determining a position where the environment map acquisition device should move.
The movement control unit 33 controls the movable table 1 and the movable table 4 under the instruction of the computer position specifying unit 32, and moves the environment map acquisition device to a position where the shielding region estimated by the shielding region estimation unit 31 can be measured. Move.
The computer position specifying unit 32 and the movement control unit 33 constitute a movement control means.

After the environmental map acquisition device is moved by the movement control unit 33, the scan processing unit 34 controls the pitch angle of the movable table 4, and changes the laser beam irradiation direction by the laser scanner 5, and measures by the laser scanner 5. A process of collecting point cloud data, which is distance data indicating the measured distance, is performed. The scan processing unit 34 constitutes second distance data collection means.
The point cloud data storage unit 35 is a memory that stores the point cloud data collected by the scan processing unit 34.
The shape specifying unit 36 combines the point cloud data stored in the point cloud data memory 12 and the point cloud data stored in the point cloud data storage unit 35, and determines the shape of the reflector from the combined point cloud data. Perform the specified process. The shape specifying unit 36 forms a shape specifying means.

Next, the operation will be described.
Similarly to the first embodiment, the scan processing unit 11 controls the pitch angle of the movable table 4 to change the laser light irradiation direction by the laser scanner 5 and indicates the distance measured by the laser scanner 5. Point cloud data that is distance data is collected, and the point cloud data is stored in the point cloud data storage unit 12.

When the scan processing unit 11 stores the point cloud data in the point cloud data storage unit 12, the occlusion area estimation unit 31 analyzes the point cloud data stored in the point cloud data storage unit 12 and uses the laser scanner 5 to measure the distance. Estimate the occluded area where is not measured.
In the example of FIG. 26, the wall portion located behind the object is a shielding area due to the influence of the object in front.
For example, when the shielding area estimation unit 31 performs scanning while leveling the scan plane and moving the pitch angle up and down, the measurement direction in which the pitch angle is continuous, or the measurement direction in which the azimuth angle of the scan plane is continuous, When the difference in distance, which is a measured value, is greater than or equal to a certain value, it is determined that there is a shielding part between the parts.
In the example of FIG. 27, it is determined that there is a shielding part between A and B.

When the shielding area estimation unit 31 estimates the shielding area, the computer position designation unit 32 determines a position where the environment map acquisition device should move in order to acquire point cloud data of the shielding area.
In the example of FIG. 27, on the extension line of OA, the position on the left side of B in the figure is determined as the movement position.
Also, the orientation of the environment map acquisition device is determined in such a direction that A and B can be observed simultaneously.

The movement control unit 33 controls the moving table 1 and the movable table 4 under the instruction of the computer position specifying unit 32, and obtains an environment map acquisition device up to a position where the shielding area estimated by the shielding area estimation unit 31 can be measured. Move.
That is, the movement control unit 33 controls the moving table 1 so as to reach the moving position determined by the computer position specifying unit 32, and moves the moving table 1 and the moving table 1 so as to be in the direction determined by the computer position specifying unit 32. The movable table 4 is controlled.
In addition, in order to be able to move the position of the laser scanner 5 up and down or to extend forward, the shape of the movable table 4 may be the following shape.
・ Crane arm (Laser scanner 5 can be extended upward)
・ Magic hand (The laser scanner 5 can be extended to a thin shadow.)
-Forklift (Laser scanner 5 can be extended upward)

  The scan processing unit 34 controls the pitch angle of the movable base 4 after the environmental map acquisition device is moved by the movement control unit 33, and controls the irradiation direction of the laser light from the laser scanner 5, similarly to the scan processing unit 11. While changing, point cloud data, which is distance data indicating the distance measured by the laser scanner 5, is collected, and the point cloud data is stored in the point cloud data storage unit 35.

When the scan processing unit 34 stores the point cloud data in the point cloud data storage unit 35, the shape specification unit 36 stores the point cloud data stored in the point cloud data memory 12 and the point cloud data storage unit 35. The point cloud data is synthesized, and the shape of the reflector is specified from the synthesized point cloud data.
For example, the point cloud data combining process in the shape specifying unit 36 is performed after the shape matching process like the shape matching unit 25 in FIG. 12 is performed.
Moreover, the process which specifies the shape of a reflector from the point cloud data after a synthesis | combination is performed like the shape specific | specification part 26 of FIG. 12, for example.

  As can be seen from the above, according to the third embodiment, the point cloud data collected by the scan processing unit 11 is analyzed, and the shielding region estimation for estimating the shielding region whose distance is not measured by the laser scanner 5 After the environmental map acquisition device is moved by the movement control unit 32, the movement control unit 32 that moves the environmental map acquisition device to a position where the measurement of the occlusion region estimated by the unit 31, the shielding region estimation unit 31 is possible, A scan processing unit that collects point cloud data measured by the laser scanner 5 while changing the irradiation direction of the laser beam by the laser scanner 5 by controlling the pitch angle of the movable table 4 is provided. The point cloud data collected by the scan processing unit 11 and the point cloud data collected by the scan processing unit 34 are combined, and the shape of the reflector is determined from the combined point cloud data. And then, it is constant, an effect that can be created throughout the environment map also shielding region environmental map acquiring device is not visible to be moved.

It is a perspective view which shows the state by which the environmental map acquisition apparatus by Embodiment 1 of this invention is mounted in the moving stand. It is a block diagram which shows the environment map acquisition apparatus by Embodiment 1 of this invention. It is a block diagram which shows the detail of the computer of the environmental map acquisition apparatus by Embodiment 1 of this invention. It is explanatory drawing which shows an example of point cloud data. It is explanatory drawing which shows an example of low precision range data. It is explanatory drawing which shows that the displacement of the scanning direction of a laser scanner is restrained uniformly. It is explanatory drawing which shows the example from which the measurement result of a laser scanner becomes inaccurate. It is explanatory drawing which shows the example from which the measurement result of a laser scanner becomes inaccurate. It is explanatory drawing which shows the example from which the measurement result of a laser scanner becomes inaccurate. It is explanatory drawing explaining the determination of an oblique edge. It is explanatory drawing which shows the example of a change of a scanning method. It is a block diagram which shows the detail of the computer of the environmental map acquisition apparatus by Embodiment 2 of this invention. It is explanatory drawing which shows an example of feature data. It is explanatory drawing which shows the method of extracting a straight line as a characteristic shape. It is explanatory drawing which shows the processing content of the shape matching part of a computer. It is explanatory drawing which shows the processing content of the shape matching part of a computer. It is explanatory drawing which shows the processing content of the shape matching part of a computer. It is explanatory drawing which shows the processing content of the shape matching part of a computer. It is explanatory drawing which shows the processing content of the shape matching part of a computer. It is explanatory drawing which shows the processing content of the shape matching part of a computer. It is explanatory drawing which shows the example which has divided into 3 what has a short distance. It is explanatory drawing which shows the example of overdetection of a straight line. It is explanatory drawing which shows the detection omission example of a straight line. It is explanatory drawing which shows the part from which the distance of a straight line and a point is changing rapidly. It is a block diagram which shows the detail of the computer of the environment map acquisition apparatus by Embodiment 3 of this invention. It is explanatory drawing which shows the shielding area | region where the distance is not measured by the laser scanner. It is explanatory drawing explaining the determination method of the shielding part by the shielding area estimation part. It is explanatory drawing explaining the shape of a linear end point, and its symbol representation. It is explanatory drawing explaining the result of having extracted the straight line which the division | segmentation unit of a point cloud and the divided point cloud ride on.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 Mobile stand, 2 front wheel, 3 back wheel, 4 movable stand, 5 laser scanner, 5a light source, 6 computer, 11 scan processing part (1st distance data collection means), 12, 22, 35 Point cloud data storage part , 13 Low-precision measurement region specifying unit (low-precision measurement region specifying unit), 14 Low-precision range data memory, 15 Scan method changing unit (second distance data collecting unit), 16 Additional point cloud data storage unit, 17, 26 , 36 Shape identification unit (shape identification unit), 21 Scan processing unit (distance data collection unit), 23 Feature shape extraction unit (feature shape extraction unit), 24 Feature data storage unit, 25 Shape association unit (shape association unit) ), 31 Shielding area estimation unit (shielding area estimation unit), 32 Computer position designation unit (movement control unit), 33 Movement control unit (movement control unit), 34 Scan processing (Second range data acquisition means).

Claims (3)

  A laser range finder that oscillates a laser beam in a scanning angle range and measures a distance from the time until the reflected light of the laser beam returns to a reflecting object, and a movable base equipped with the laser range finder And a first distance for collecting distance data indicating a distance measured by the laser distance measuring device while changing a laser beam irradiation direction by the laser distance measuring device by controlling a pitch angle of the movable table. By the data collection means, the low-precision measurement area specifying means for analyzing the distance data collected by the first distance data collection means and specifying the measurement area of the distance data having low measurement accuracy, and the laser distance measuring device The low angle of the movable base is controlled so that the scanning angle range of the laser beam includes the measurement area specified by the low-precision measurement area specifying means, and the distance measured by the laser distance measuring device is determined. A second distance data collecting means for collecting the distance data; a distance data collected by the first distance data collecting means; and a distance data collected by the second distance data collecting means. An environment map acquisition apparatus comprising shape specifying means for specifying the shape of the reflector from the distance data.   A laser range finder that oscillates a laser beam in a scanning angle range and measures a distance from the time until the reflected light of the laser beam returns to a reflecting object, and a movable base equipped with the laser range finder And distance data collection means for collecting distance data indicating the distance measured by the laser distance measuring device while changing the irradiation direction of the laser light by the laser distance measuring device by controlling the pitch angle of the movable table And the movable platform is installed at different positions, and the feature shape extraction means for extracting characteristic shapes from the plurality of distance data collected by the distance data collection means, respectively, and the feature shape extraction means A shape association unit that compares a plurality of characteristic shapes with each other and associates characteristic shapes indicating the same space with each other, and a characteristic shape associated with the shape association unit The extract the distance data between synthesized, environmental map acquisition apparatus and a shape specifying means for specifying the shape of the reflector from the distance data after the synthesis.   A laser range finder that oscillates a laser beam in a scanning angle range and measures a distance from the time until the reflected light of the laser beam returns to a reflecting object, and a movable base equipped with the laser range finder And a first distance for collecting distance data indicating a distance measured by the laser distance measuring device while changing a laser beam irradiation direction by the laser distance measuring device by controlling a pitch angle of the movable table. A data collection means, a shielding area estimation means for analyzing the distance data collected by the first distance data collection means, and estimating a shielding area whose distance is not measured by the laser distance measuring device; and the shielding area A movement control means for moving the movable table to a position where the shielding area estimated by the estimation means can be measured; and after the movable table is moved by the movement control means, A second distance data collecting means for collecting distance data indicating a distance measured by the laser distance measuring device while controlling an angle and changing an irradiation direction of the laser light by the laser distance measuring device; Shape specifying means for combining the distance data collected by the first distance data collecting means and the distance data collected by the second distance data collecting means and for specifying the shape of the reflector from the combined distance data; Environmental map acquisition device provided.

Images