JP6876484B2 - Data processing equipment, data processing methods, and programs - Google Patents

Description

The present invention relates to processing of point cloud data acquired by an external sensor.

In recent years, map data has been generated based on 3D point cloud data generated by a 3D lidar (LiDAR: Light Detection and Ringing) or the like. Generally, point cloud data such as the same road is repeatedly acquired to improve the accuracy of map data, but the measured point cloud data contains errors due to sensor performance limits, etc., so it is different. If the measurement conditions are different, such as the schedule and time of day, the point cloud data acquired at the same point may be out of sync.

In this case, a method of associating points corresponding to the same point in each point cloud data and then moving / rotating / enlarging / reducing the point cloud data to eliminate the deviation between the point cloud data. However, this has the following problems. That is, it is difficult to determine an arbitrary one point from the 3D space, and a sufficient number of corresponding points cannot be obtained. In addition, automatic extraction does not provide a sufficiently accurate response. Furthermore, if all points are targeted, the processing cost becomes enormous.

In relation to this, Patent Document 1 describes a method of voting three-dimensional point cloud data in a two-dimensional grid to generate a road surface image and estimating a deformation amount representing a relative positional relationship of a plurality of road surface images. ..

JP-A-2017-10393

However, in the method of Patent Document 1, since the cost calculation based on the matching cost is performed in order to estimate the deformation amount indicating the relative positional relationship of the road surface image, there is a problem that the calculation amount becomes enormous and the cost becomes high. ..

As an example of the problem to be solved by the present invention, the above is given as an example. An object of the present invention is to provide a data processing apparatus capable of adjusting a positional deviation between a plurality of 3D point cloud data while suppressing a calculation amount.

The invention according to claim 1 is a data processing apparatus, wherein an acquisition unit that acquires first and second three-dimensional point cloud data and the first and second three-dimensional point cloud data are first. And the conversion unit that converts to the second two-dimensional image, and the first pixel in the first two-dimensional image and the second pixel in the second two-dimensional image corresponding to the same feature point are extracted. Then, the first generation unit that generates the first information indicating the positional relationship of the representative points in each of the first pixel and the second pixel, and the first three-dimensional point cloud data based on the first information. A second generation unit that generates second information indicating the positional relationship between the second three-dimensional point cloud and the second three-dimensional point cloud data is provided , and the first generation unit is the point cloud included in the first pixel. The point having the brightness closest to the brightness of the first pixel is a representative point in the first pixel, and the point having the brightness closest to the brightness of the second pixel among the point cloud included in the second pixel is the second point. It is a representative point in the pixel .
The invention according to claim 2 is a data processing apparatus, wherein an acquisition unit that acquires first and second three-dimensional point cloud data and the first and second three-dimensional point cloud data are first. And the conversion unit that converts to the second two-dimensional image, and the first pixel in the first two-dimensional image and the second pixel in the second two-dimensional image corresponding to the same feature point are extracted. Then, the first generation unit that generates the first information indicating the positional relationship of the representative points in each of the first pixel and the second pixel, and the first three-dimensional point cloud data based on the first information. A second generation unit that generates second information indicating the positional relationship between the second three-dimensional point cloud and the second three-dimensional point cloud data is provided, and the first generation unit is the center of gravity of the point cloud included in the first pixel. The point closest to the position is designated as the representative point in the first pixel, and the point closest to the position of the center of gravity in the point cloud included in the second pixel is designated as the representative point in the second pixel.
The invention according to claim 3 is a data processing apparatus, wherein an acquisition unit that acquires first and second three-dimensional point cloud data and the first and second three-dimensional point cloud data are first. And the conversion unit that converts to the second two-dimensional image, and the first pixel in the first two-dimensional image and the second pixel in the second two-dimensional image corresponding to the same feature point are extracted. Then, the first generation unit that generates the first information indicating the positional relationship of the representative points in each of the first pixel and the second pixel, and the first three-dimensional point cloud data based on the first information. A second generation unit that generates second information indicating the positional relationship between the second three-dimensional point cloud and the second three-dimensional point cloud data is provided, and the first generation unit is one of the point clouds included in the first pixel. The point closest to the position of the center of gravity of the feature region included in the first pixel is set as a representative point in the first pixel, and the position of the center of gravity of the feature region included in the second pixel among the point cloud included in the second pixel is set. The point closest to is designated as a representative point in the second pixel.

The invention according to claim 7 is a data processing method executed by a data processing apparatus, which includes an acquisition step of acquiring first and second three-dimensional point cloud data, and the first and second three-dimensional aspects. In the conversion step of converting the point cloud data into the first and second two-dimensional images, and in the first pixel and the second two-dimensional image in the first two-dimensional image corresponding to the same feature points. The first generation step of extracting the second pixel of the above and generating the first information indicating the positional relationship of the representative points in each of the first pixel and the second pixel, and the first generation step based on the first information. The first generation step includes a second generation step of generating a second information indicating a positional relationship between the three-dimensional point cloud data of 1 and the second three-dimensional point cloud data, and the first generation step is performed on the first pixel. The point having the brightness closest to the brightness of the first pixel in the included point cloud is set as the representative point in the first pixel, and the brightness closest to the brightness of the second pixel in the point group included in the second pixel. Is a representative point in the second pixel .
The invention according to claim 8 is a data processing method executed by a data processing apparatus, which includes an acquisition step of acquiring first and second three-dimensional point group data, and the first and second three-dimensional data. In the conversion step of converting the point group data into the first and second two-dimensional images, and in the first pixel and the second two-dimensional image in the first two-dimensional image corresponding to the same feature points. The first generation step of extracting the second pixel of the above and generating the first information indicating the positional relationship of the representative points in each of the first pixel and the second pixel, and the first generation step based on the first information. The first generation step includes a second generation step of generating a second information indicating a positional relationship between the three-dimensional point group data of 1 and the second three-dimensional point group data, and the first generation step is performed on the first pixel. The point closest to the center of gravity position among the included point groups is designated as the representative point in the first pixel, and the point closest to the center of gravity position among the point groups included in the second pixel is designated as the representative point in the second pixel.
The invention according to claim 9 is a data processing method executed by a data processing apparatus, which includes an acquisition step of acquiring first and second three-dimensional point group data, and the first and second three-dimensional data. In the conversion step of converting the point group data into the first and second two-dimensional images, and in the first pixel and the second two-dimensional image in the first two-dimensional image corresponding to the same feature points. The first generation step of extracting the second pixel of the above and generating the first information indicating the positional relationship of the representative points in each of the first pixel and the second pixel, and the first generation step based on the first information. The first generation step includes a second generation step of generating a second information indicating a positional relationship between the three-dimensional point group data of 1 and the second three-dimensional point group data, and the first generation step is performed on the first pixel. Among the included point groups, the point closest to the position of the center of gravity of the feature region included in the first pixel is set as a representative point in the first pixel, and among the point groups included in the second pixel, the second pixel The point closest to the position of the center of gravity of the included feature region is set as a representative point in the second pixel .

The invention according to claim 10 is a program executed by a data processing device including a computer, and is an acquisition unit for acquiring first and second three-dimensional point cloud data, the first and second three dimensions. A conversion unit that converts point cloud data into first and second two-dimensional images, a first pixel in the first two-dimensional image and a second two-dimensional image corresponding to the same feature point. The first generation unit that extracts the second pixel and generates the first information indicating the positional relationship of the representative points in each of the first pixel and the second pixel, the first generation unit based on the first information. The computer is operated as a second generation unit that generates second information indicating the positional relationship between the three-dimensional point cloud data and the second three-dimensional point cloud data, and the first generation unit is the first pixel. The point having the brightness closest to the brightness of the first pixel among the point cloud included in the first pixel is set as the representative point in the first pixel, and the point cloud included in the second pixel is closest to the brightness of the second pixel. A point having brightness is designated as a representative point in the second pixel .
The invention according to claim 11 is a program executed by a data processing device including a computer, and is an acquisition unit for acquiring first and second three-dimensional point cloud data, the first and second three dimensions. A conversion unit that converts point cloud data into first and second two-dimensional images, a first pixel in the first two-dimensional image and a second two-dimensional image corresponding to the same feature point. The first generation unit that extracts the second pixel and generates the first information indicating the positional relationship of the representative points in each of the first pixel and the second pixel, the first generation unit based on the first information. The computer is operated as a second generation unit that generates second information indicating the positional relationship between the three-dimensional point cloud data and the second three-dimensional point cloud data, and the first generation unit is the first pixel. The point closest to the center of gravity position in the point cloud included in the first pixel is the representative point in the first pixel, and the point closest to the center of gravity position in the second pixel is the representative point in the second pixel. ..
The invention according to claim 12 is a program executed by a data processing device including a computer, and is an acquisition unit for acquiring first and second three-dimensional point cloud data, the first and second three dimensions. A conversion unit that converts point cloud data into first and second two-dimensional images, a first pixel in the first two-dimensional image and a second two-dimensional image corresponding to the same feature point. The first generation unit that extracts the second pixel and generates the first information indicating the positional relationship of the representative points in each of the first pixel and the second pixel, the first generation unit based on the first information. The computer is operated as a second generation unit that generates second information indicating the positional relationship between the three-dimensional point cloud data and the second three-dimensional point cloud data, and the first generation unit is the first pixel. Of the point cloud included in the first pixel, the point closest to the position of the center of gravity of the feature region included in the first pixel is set as a representative point in the first pixel, and the second pixel among the point cloud included in the second pixel. The point closest to the position of the center of gravity of the feature region included in the second pixel is set as a representative point in the second pixel .

The data processing apparatus which concerns on Example is shown. The outline of the position adjustment process is shown. It is a flowchart of a position adjustment process. The relationship between the ortho image and the point cloud data is shown. An example of a representative point in a pixel of an ortho image is shown. An example of associative information between pixels of an ortho image is shown.

In a preferred embodiment of the present invention, the data processing apparatus obtains the first and second three-dimensional point group data and the first and second three-dimensional point group data. The conversion unit that converts the two two-dimensional images and the first pixel in the first two-dimensional image and the second pixel in the second two-dimensional image corresponding to the same feature points are extracted. The first generation unit that generates the first information indicating the positional relationship of the representative points in each of the first pixel and the second pixel, and the first three-dimensional point group data and the said based on the first information. It includes a second generation unit that generates second information indicating a positional relationship with the second three-dimensional point group data.

The above data processing apparatus acquires the first and second three-dimensional point cloud data, and converts the first and second three-dimensional point cloud data into the first and second two-dimensional images. Next, the first pixel in the first two-dimensional image and the second pixel in the second two-dimensional image corresponding to the same feature point are extracted, and in each of the first pixel and the second pixel. The first information indicating the positional relationship of the representative points is generated. Further, based on the first information, second information indicating the positional relationship between the first three-dimensional point cloud data and the second three-dimensional point cloud data is generated.

As described above, the above-mentioned data processing device converts the three-dimensional point cloud data into a two-dimensional image, detects the positional relationship between the pixels in the two-dimensional image corresponding to the same feature point, and based on this, 3 Detect the positional relationship between the 3D point cloud data. Therefore, it is possible to adjust the position between the point cloud data while suppressing the amount of calculation.

In one aspect of the above data processing apparatus, the first generation unit uses a point having a brightness closest to the brightness of the first pixel among the point cloud included in the first pixel as a representative point in the first pixel. A point cloud having a brightness closest to the brightness of the second pixel among the point cloud included in the second pixel is set as a representative point in the second pixel.

In another aspect of the above data processing apparatus, the first generation unit uses the point closest to the position of the center of gravity of the point cloud included in the first pixel as a representative point in the first pixel, and the second pixel. The point closest to the position of the center of gravity among the point cloud included in the second pixel is set as a representative point in the second pixel.

In another aspect of the data processing apparatus, the first generation unit sets the point closest to the center of gravity of the feature region included in the first pixel among the point cloud included in the first pixel. The representative point in one pixel is defined as the representative point in the second pixel, and the point closest to the position of the center of gravity of the feature region included in the second pixel is defined as the representative point in the second pixel.

In the above data processing apparatus, preferably, the three-dimensional point cloud data is data detected on the road surface by an external sensor, and the feature area is a paint area on the road surface. Also, preferably, the feature point is a point corresponding to the corner of the paint.

In another aspect of the above data processing device, a position adjusting unit that adjusts the positional deviation between the first three-dimensional point cloud data and the second three-dimensional point cloud data based on the second information is provided. Be prepared. In this aspect, the positional deviation between the three-dimensional point cloud data can be adjusted based on the second information.

In another preferred embodiment of the present invention, the data processing method executed by the data processing apparatus includes an acquisition step of acquiring first and second three-dimensional point group data and the first and second three-dimensional data. In the conversion step of converting the point group data into the first and second two-dimensional images, and in the first pixel and the second two-dimensional image in the first two-dimensional image corresponding to the same feature points. The first generation step of extracting the second pixel of the above and generating the first information indicating the positional relationship of the representative points in each of the first pixel and the second pixel, and the first generation step based on the first information. It includes a second generation step of generating second information indicating the positional relationship between the three-dimensional point group data of 1 and the second three-dimensional point group data. This method also makes it possible to adjust the position between the three-dimensional point cloud data while suppressing the amount of calculation.

In another preferred embodiment of the invention, the program executed by the data processing apparatus including the computer is an acquisition unit that acquires first and second three-dimensional point group data, the first and second three dimensions. A conversion unit that converts point group data into first and second two-dimensional images, a first pixel in the first two-dimensional image and a second two-dimensional image corresponding to the same feature point. The first generation unit that extracts the second pixel and generates the first information indicating the positional relationship of the representative points in each of the first pixel and the second pixel, the first generation unit based on the first information. The computer functions as a second generation unit that generates second information indicating the positional relationship between the three-dimensional point group data and the second three-dimensional point group data. By executing this program on a computer, the above data processing device can be realized. This program can be stored and handled in a storage medium.

Hereinafter, preferred embodiments of the present invention will be described with reference to the drawings.
[System configuration]
FIG. 1 shows a data processing apparatus according to an embodiment. The data processing device 10 processes 3D point cloud data (hereinafter, simply referred to as "point cloud data") measured by a 3D rider mounted on a vehicle or the like. Specifically, the data processing device 10 performs a process of adjusting the positional deviation between the point cloud data measured at different dates and times. The data processing device 10 is composed of a computer device including a CPU, a ROM, a RAM, and the like, and executes data processing by executing a program prepared in advance.

Now, as shown in FIG. 1, there are point cloud data A and point cloud data B measured at different dates and times, and a point cloud database (hereinafter, “database” is also referred to as “DB”) 5A and a point cloud, respectively. It is assumed that it is stored in DB5B. The data processing device 10 acquires the point cloud data A from the point cloud DB 5A, acquires the point cloud data B from the point cloud DB 5B, and performs processing.

[Overview of position adjustment processing]
Next, the position adjustment process between the point cloud data performed by the data processing device 10 will be described. FIG. 2 schematically shows an outline of the position adjustment process. The point cloud data A is stored in the point cloud DB5A, and the point cloud data B is stored in the point cloud DB5B. Here, the point cloud data is a set of measurement point data measured by the 3D rider in the three-dimensional space, and each of the measurement point data is in the three-dimensional space, generally in the world coordinate system. X, Y, Z coordinates) and the brightness (reflection intensity) at the measurement point.

The point cloud data A and the point cloud data B are point cloud data measured when traveling on the same road at different dates and times. Generally, even when traveling on the same road, the point cloud data measured by the 3D rider includes an error due to factors such as the performance limit of the sensor. That is, even if the point cloud data is obtained by measuring the same feature, the position coordinates of the measurement points measured at different dates and times cause a deviation on the world coordinate system. The data processing device 10 of this embodiment basically adjusts the positions of the point cloud data A and the point cloud data B by the following four steps.

(First step)
First, in the first step, the data processing device 10 converts each point cloud data into an ortho image which is a two-dimensional image. An orthophoto image is an image obtained by correcting an aerial photograph taken from the sky of a certain area by orthographic projection, and is a two-dimensional image of a certain area viewed from the sky. In this embodiment, the data processing device 10 converts the point cloud data A to generate the ortho image A, and converts the point cloud data B to generate the ortho image B. The conversion from the point cloud data to the ortho image can be performed by various known methods. Normally, by this conversion, one pixel of the ortho image contains a plurality of point cloud data.

(Second step)
Next, the data processing device 10 generates association information between the pixels in the ortho image A and the corresponding pixels in the ortho image B. This association information is information indicating the amount of deviation between the pixels in the ortho image A and the corresponding pixels in the ortho image B. This linking information is also hereinafter referred to as "pixel-to-pixel linking information". The inter-pixel linking information corresponds to the first information of the present invention.

(Third step)
Next, the data processing device 10 generates linking information between point cloud data (hereinafter, also referred to as “point cloud linking information”) based on the pixel-to-pixel linking information. The point cloud association information is information indicating the amount of deviation between a predetermined measurement point in the point cloud data A and the measurement point in the corresponding point cloud data B. The point cloud association information corresponds to the second information of the present invention.

(4th step)
Then, the data processing device 10 adjusts the positions of the measurement points included in the point cloud data A and the measurement points included in the point cloud data B based on the point cloud association information. For example, by moving the position of a certain measurement point in the point cloud data B by the amount of deviation indicated by the point cloud association information corresponding to the measurement point, the position of the measurement point matches the measurement point corresponding to the measurement point in the point cloud data A. Let me.

As described above, in this embodiment, the 3D point cloud data is once converted into an ortho image which is a two-dimensional image, the positional deviation between the images of the ortho image is detected, and this is expanded into the positional deviation between the 3D point clouds. .. This eliminates the need to perform complicated calculations on the 3D point cloud data to detect the positional deviation.

[Details of position adjustment processing]
Next, the above position adjustment process will be described in detail. FIG. 3 is a flowchart of the position adjustment process. This process is realized by the data processing device 10 executing a program prepared in advance.

First, the data processing device 10 acquires the point cloud data A and the point cloud data B from the point cloud DB 5A and the point cloud DB 5B (step S11). The point cloud data A and the point cloud data B are data measured by a 3D rider while traveling on the same road or the like at different dates and times. Now, it is assumed that the point cloud data of the road surface is measured by a 3D rider while traveling on the T-junction shown in FIG. 4 (A) in the direction of the arrow. A T-shaped road surface paint (hereinafter referred to as "Mark M") is provided on the road surface of the T-junction. Therefore, the point cloud data A and the point cloud data B each include the region of the mark M. Since the mark M is a white paint and has a high reflection intensity, the brightness of the measurement point corresponding to the mark M is high.

Next, the data processing device 10 generates an ortho image A from the point cloud data A and generates an ortho image B from the point cloud data B (step S12). Here, each pixel of the orthoimages A and B includes a plurality of point cloud data at the corresponding positions.

FIG. 4B shows the relationship between the pixels of the ortho image and the point cloud data. Note that FIG. 4B shows an ortho image of the region including the mark M. The ortho image is composed of a plurality of pixels P. Further, one pixel P of the ortho image includes a plurality of measurement points of the point cloud data. In the example of FIG. 4B, one pixel P of the ortho image includes 6 × 6 = 36 measurement points. Now, assuming that one pixel in the upper right of the ortho image of FIG. 4 (B) is a pixel Px, as shown in the enlarged view, 3 × 3 = 9 measurement points in the lower left of the image Px are road surface paints. It is a measurement point on the mark M and has high brightness. On the other hand, the measurement points other than the nine measurement points are the measurement points on the road surface and have lower brightness than the mark M.

Now, when the ortho image A and the ortho image B are obtained, the data processing device 10 extracts the pixels of the feature points in the ortho images A and B (hereinafter, also referred to as “feature point pixels”) (step S13). ). Here, the feature point is typically a corner of the road surface paint or the like, and is a part on the road surface whose shape, color, or the like is different from that of the periphery. The feature points are used to identify pixels that correspond to each other in the ortho image A and the ortho image B. In this example, the data processing device 10 uses the upper right corner of the mark M as a feature point, and the pixel Px including the upper right corner of the mark M as a feature point pixel. That is, the data processing device 10 extracts the pixel Px including the upper right corner of the mark M in the ortho images A and B. As can be understood from FIG. 4B, not all the pixels constituting the ortho images A and B include the feature points, and some pixels among the pixels constituting the ortho images A and B. Will be extracted as feature point pixels.

Next, the data processing device 10 determines representative points in the feature point pixels extracted from each of the ortho images A and B (step S14). In this example, the data processing device 10 determines a representative point in the feature point pixel Px. FIG. 5 shows an example of representative points. The representative point is a point representing the pixel among a plurality of measurement points included in the feature point pixel, and can be determined by any of the following methods.

In the first method, a measurement point having a brightness closest to the brightness value of the feature point pixel is set as a representative point. In the example of FIG. 5, since the feature point pixel Px includes a part of the mark M having high reflection intensity, the brightness of the pixel itself is high. Therefore, the measurement point D having the brightness closest to the brightness of the pixel Px is set as a representative point. In the example of FIG. 5, the measurement point R1 having the brightness closest to the brightness of the pixel Px is determined as the representative point.

In the second method, the measurement point closest to the position of the center of gravity of the point cloud included in the feature point pixel is set as the representative point. In the example of FIG. 5, the measurement point R2 closest to the position of the center of gravity of the point cloud included in the feature point pixel Rx is determined as the representative point. In the third method, a region having a brightness higher than a predetermined value is specified in the feature point pixel, and the measurement point closest to the position of the center of gravity of the region is set as a representative point. In the example of FIG. 5, the pixel Px has a region having high brightness corresponding to the mark M, and the measurement point R3 closest to the position of the center of gravity is determined as the representative point.

Since the reflection intensity of the mark M, which is a road surface paint, varies depending on the road surface condition such as the weather, it is preferable to normalize and compare the brightness values as necessary when determining the brightness of the pixels. When the representative point is determined by any of the above methods, the data processing device calculates the amount of deviation between the representative points in the feature point pixels extracted from each of the ortho images A and B, and the inter-pixel linking information. Is generated (step S15). FIG. 6 shows a method of generating inter-pixel association information. Now, as shown in FIG. 6A, it is assumed that the feature point pixel Pa of the ortho image A includes a plurality of measurement points Da, and the representative point Ra is determined from them. Further, it is assumed that the feature point pixel Pb of the ortho image B includes a plurality of measurement points Db, and the representative point Rb is determined from them. It is assumed that the feature point pixels Pa and Pb are pixels corresponding to the same position.

In this case, as shown in FIG. 6B, the data processing device 10 calculates the amount of deviation (Δx, Δy, Δz) between the representative points Ra and Rb, and uses this as the inter-pixel linking information. That is, (Δx, Δy, Δz) is obtained as the inter-pixel association information between the feature point pixel Pa of the ortho image A and the feature point pixel Pb of the ortho image B. Here, Δz is a component in the vertical direction of the paper surface in FIG. 6B, that is, a component in the height direction. Specifically, Δz is the amount of deviation between the height component of the feature point pixel Pa and the height component of the feature point pixel Pb. In this way, the data processing device 10 determines representative points for all the feature point pixels extracted in the ortho images A and B, and generates inter-pixel association information. In the example of FIG. 6, the inter-pixel associative information shows a positional deviation having only a translation component, but in reality, due to distortion of the 3D point cloud data or the like, the positional deviation including the rotation component or Positional shifts that accompany enlargement / reduction occur. Therefore, the inter-pixel linking information may include misalignment information having such rotation and enlargement / reduction components.

When the inter-pixel association information is generated in this way, the data processing device 10 generates the inter-pixel association information based on the inter-pixel association information (step S16). For example, when the inter-pixel association information (Δx, Δy, Δz) shown in FIG. 6B is obtained, the data processing device 10 uses the point cloud included in the feature point pixel Pa of the ortho image A and the ortho image. Let (Δx, Δy, Δz) be the point cloud association information with the point cloud included in the feature point pixel Pb of B. As a result, the amount of misalignment at the point cloud level included in the ortho image can be obtained. For pixels other than the feature point pixels, the point cloud association information for the feature point pixels existing in the vicinity thereof may be used. Further, when there are a plurality of feature point pixels in the vicinity, the linking information between the plurality of point groups may be distributed and applied based on the positional relationship and the distance between the plurality of feature point pixels.

Then, the data processing device 10 adjusts the position between the point cloud data A and the point cloud data B based on the generated point cloud association information (step S17). For example, in the example of FIG. 6, the position of each measurement point constituting the point cloud data B is shifted based on the deviation amount indicated by the obtained point cloud association information to obtain the deviation of the point cloud data A and B. Eliminate. Further, when the point cloud linking information is a position shift including a rotation component or a position shift accompanied by enlargement / reduction, the data processing device 10 uses the point cloud data based on the linking information including them. However, corrections such as rotation and enlargement / reduction are applied to eliminate the misalignment. The point cloud data A and B in which the misalignment is eliminated can be used for generating map information or the like as 3D point cloud data obtained at the same point.

[Modification example]
In the above embodiment, the position between the point cloud data is adjusted by using a road paint or the like provided on the road surface. Therefore, as preprocessing, only the point cloud data corresponding to the road surface is extracted from the 3D point cloud data measured by the 3D rider, and the above processing is applied to the point cloud data corresponding to the road surface. Is preferable. For example, the above position adjustment process is applied to the remaining point cloud data obtained by removing the point cloud of a three-dimensional object such as a sign in advance from the 3D point cloud data obtained by the 3D rider. As a result, the amount of point cloud data to be processed is reduced, and efficient processing becomes possible.

In the above embodiment, the data processing device 10 uses the point cloud data A and the point cloud data B measured at different dates and times to eliminate the deviation between the point cloud data A and B. Not limited to. Even if the point cloud data A and the point cloud data B are measured by different vehicles on the same date and time, the above data processing can be executed. The 3D rider mounted on each of a plurality of vehicles may have different sensor performance, which causes an error in the measurement data. However, the processing of the present invention is not related to the performance of the 3D rider. However, it is possible to eliminate the deviation of the point cloud data.

5A, 5B Point cloud database 10 Data processing device D Point cloud data measurement points M mark P, Pa, Pb Ortho image pixels R1 to R3, Ra, Rb Representative points

Claims (13)

An acquisition unit that acquires the first and second 3D point cloud data,
A conversion unit that converts the first and second three-dimensional point cloud data into first and second two-dimensional images, and
The first pixel in the first two-dimensional image and the second pixel in the second two-dimensional image corresponding to the same feature point are extracted, and in each of the first pixel and the second pixel. A first generation unit that generates first information indicating the positional relationship of representative points, and
A second generation unit that generates second information indicating the positional relationship between the first three-dimensional point cloud data and the second three-dimensional point cloud data based on the first information.
Equipped with a,
The first generation unit uses a point group having a brightness closest to the brightness of the first pixel among the point groups included in the first pixel as a representative point in the first pixel, and the point group included in the second pixel. A data processing device in which a point having a brightness closest to the brightness of the second pixel is a representative point in the second pixel. An acquisition unit that acquires the first and second 3D point cloud data,
A conversion unit that converts the first and second three-dimensional point cloud data into first and second two-dimensional images, and
The first pixel in the first two-dimensional image and the second pixel in the second two-dimensional image corresponding to the same feature point are extracted, and in each of the first pixel and the second pixel. A first generation unit that generates first information indicating the positional relationship of representative points, and
A second generation unit that generates second information indicating the positional relationship between the first three-dimensional point cloud data and the second three-dimensional point cloud data based on the first information.
With
In the first generation unit, the point closest to the center of gravity position in the point cloud included in the first pixel is set as the representative point in the first pixel, and the point group included in the second pixel is closest to the center of gravity position. A data processing device in which a point is a representative point in the second pixel. An acquisition unit that acquires the first and second 3D point cloud data,
A conversion unit that converts the first and second three-dimensional point cloud data into first and second two-dimensional images, and
The first pixel in the first two-dimensional image and the second pixel in the second two-dimensional image corresponding to the same feature point are extracted, and in each of the first pixel and the second pixel. A first generation unit that generates first information indicating the positional relationship of representative points, and
A second generation unit that generates second information indicating the positional relationship between the first three-dimensional point cloud data and the second three-dimensional point cloud data based on the first information.
With
In the first generation unit, among the point cloud included in the first pixel, the point closest to the position of the center of gravity of the feature region included in the first pixel is set as a representative point in the first pixel, and the second pixel A data processing device in which, among the included point clouds, the point closest to the position of the center of gravity of the feature region included in the second pixel is a representative point in the second pixel. The three-dimensional point cloud data is data detected on the road surface by an external sensor, and is
The data processing device according to claim 3 , wherein the feature area is a paint area on a road surface. The data processing apparatus according to claim 4 , wherein the feature points correspond to the corners of the paint. Based on the second information, according to claim 1 to 5, further comprising a position adjusting section for adjusting the positional deviation between the first 3 and the dimensional point group data and the second three-dimensional point group data The data processing apparatus according to any one item. A data processing method executed by a data processing device.
The acquisition process for acquiring the first and second 3D point cloud data, and
A conversion step of converting the first and second three-dimensional point cloud data into first and second two-dimensional images, and
The first pixel in the first two-dimensional image and the second pixel in the second two-dimensional image corresponding to the same feature point are extracted, and in each of the first pixel and the second pixel. The first generation step of generating the first information indicating the positional relationship of the representative points, and
A second generation step of generating second information indicating the positional relationship between the first three-dimensional point cloud data and the second three-dimensional point cloud data based on the first information.
Equipped with a,
In the first generation step, a point group having a brightness closest to the brightness of the first pixel among the point groups included in the first pixel is set as a representative point in the first pixel, and a point group included in the second pixel. A data processing method in which a point having a brightness closest to the brightness of the second pixel is a representative point in the second pixel. A data processing method executed by a data processing device.
The acquisition process for acquiring the first and second 3D point cloud data, and
A conversion step of converting the first and second three-dimensional point cloud data into first and second two-dimensional images, and
The first pixel in the first two-dimensional image and the second pixel in the second two-dimensional image corresponding to the same feature point are extracted, and in each of the first pixel and the second pixel. The first generation step of generating the first information indicating the positional relationship of the representative points, and
A second generation step of generating second information indicating the positional relationship between the first three-dimensional point cloud data and the second three-dimensional point cloud data based on the first information.
With
In the first generation step, the point closest to the center of gravity position of the point cloud included in the first pixel is set as the representative point in the first pixel, and the point cloud included in the second pixel is closest to the center of gravity position. A data processing method in which a point is a representative point in the second pixel. A data processing method executed by a data processing device.
The acquisition process for acquiring the first and second 3D point cloud data, and
A conversion step of converting the first and second three-dimensional point cloud data into first and second two-dimensional images, and
The first pixel in the first two-dimensional image and the second pixel in the second two-dimensional image corresponding to the same feature point are extracted, and in each of the first pixel and the second pixel. The first generation step of generating the first information indicating the positional relationship of the representative points, and
A second generation step of generating second information indicating the positional relationship between the first three-dimensional point cloud data and the second three-dimensional point cloud data based on the first information.
With
In the first generation step, among the point cloud included in the first pixel, the point closest to the position of the center of gravity of the feature region included in the first pixel is set as a representative point in the first pixel, and the second pixel is set. A data processing method in which, among the included point clouds, the point closest to the position of the center of gravity of the feature region included in the second pixel is set as a representative point in the second pixel. A program executed by a data processing device equipped with a computer.
Acquisition unit that acquires the first and second 3D point cloud data,
A conversion unit that converts the first and second three-dimensional point cloud data into first and second two-dimensional images.
The first pixel in the first two-dimensional image and the second pixel in the second two-dimensional image corresponding to the same feature point are extracted, and in each of the first pixel and the second pixel. A first generation unit that generates first information indicating the positional relationship of representative points,
A second generation unit that generates second information indicating the positional relationship between the first three-dimensional point cloud data and the second three-dimensional point cloud data based on the first information.
It makes the computer function as,
The first generation unit uses a point group having a brightness closest to the brightness of the first pixel among the point groups included in the first pixel as a representative point in the first pixel, and the point group included in the second pixel. A program in which a point having a brightness closest to the brightness of the second pixel is a representative point in the second pixel. A program executed by a data processing device equipped with a computer.
Acquisition unit that acquires the first and second 3D point cloud data,
A conversion unit that converts the first and second three-dimensional point cloud data into first and second two-dimensional images.
The first pixel in the first two-dimensional image and the second pixel in the second two-dimensional image corresponding to the same feature point are extracted, and in each of the first pixel and the second pixel. A first generation unit that generates first information indicating the positional relationship of representative points,
A second generation unit that generates second information indicating the positional relationship between the first three-dimensional point cloud data and the second three-dimensional point cloud data based on the first information.
To make the computer function as
The first generation unit uses the point closest to the center of gravity position of the point cloud included in the first pixel as a representative point in the first pixel, and is closest to the center of gravity position of the point cloud included in the second pixel. A program in which a point is a representative point in the second pixel. A program executed by a data processing device equipped with a computer.
Acquisition unit that acquires the first and second 3D point cloud data,
A conversion unit that converts the first and second three-dimensional point cloud data into first and second two-dimensional images.
The first pixel in the first two-dimensional image and the second pixel in the second two-dimensional image corresponding to the same feature point are extracted, and in each of the first pixel and the second pixel. A first generation unit that generates first information indicating the positional relationship of representative points,
A second generation unit that generates second information indicating the positional relationship between the first three-dimensional point cloud data and the second three-dimensional point cloud data based on the first information.
To make the computer function as
In the first generation unit, among the point cloud included in the first pixel, the point closest to the position of the center of gravity of the feature region included in the first pixel is set as a representative point in the first pixel, and the second pixel A program in which, among the included point clouds, the point closest to the position of the center of gravity of the feature region included in the second pixel is a representative point in the second pixel. A storage medium that stores the program according to any one of claims 10 to 12.

Images