Point group data processing device, point group data processing method, and program for point group data processing

Abstract

To secure accuracy and reliability when point group data of a plurality of different systems are integrated.SOLUTION: A data processing device 200 includes an adjustment calculation part 203 for performing adjustment calculation of the first point group data, by using second point group data and third point group data having accuracies higher than that of the first point group data, wherein the second point group data and the third point group data have accuracies higher than that of the first point group data, the second point group data has accuracy higher than that of the third point group data, the result of the adjustment calculation based on the second point group data is reflected on correction of the first point group data by a relatively large first weight, and the result of the adjustment calculation based on the third point group data is reflected on correction of the first point group data by a relatively small second weight.SELECTED DRAWING: Figure 2

Description

Technical Field

[0001] The present invention relates to a technique for handling point cloud data.

Background Art

[0002] Techniques for obtaining point cloud data by three-dimensional photographic surveying or laser scanning are known (for example, see Patent Document 1).

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] There are a plurality of known point cloud data obtained by different systems, and it is convenient if newly obtained point cloud data can be integrated therewith. Here, if errors can be redistributed among the plurality of point cloud data, the accuracy and reliability of the entire integrated point cloud data can be improved.

[0005] Against such a background, an object of the present invention is to ensure accuracy and reliability when integrating point cloud data of a plurality of different systems.

Means for Solving the Problems

[0006] The present invention includes an adjustment calculation unit that performs adjustment calculation of the first point cloud data using second point cloud data and third point cloud data with higher accuracy than the first point cloud data. The second point cloud data and the third point cloud data are more accurate than the first point cloud data, and the second point cloud data is more accurate than the third point cloud data. The result of the adjustment calculation based on the second point cloud data is reflected in the correction of the first point cloud data with a relatively large first weight, and the result of the adjustment calculation based on the third point cloud data is reflected in the correction of the first point cloud data with a relatively small second weight. This is a point cloud data processing device.

[0007] In the present invention, the first point cloud data is acquired while moving, the first weight becomes smaller as the distance from the first point cloud data on the path of the movement increases, and the second weight becomes smaller as the distance from the second point cloud data on the path of the movement increases. In the present invention, the second point cloud data and the third point cloud data are associated with information regarding accuracy.

[0008] In the present invention, when the first point cloud data is point cloud data obtained by laser scanning, point cloud data obtained by laser scanning is selected as the second point cloud data and the third point cloud data. When the first point cloud data is point cloud data obtained based on a captured image, point cloud data obtained based on the captured image is selected as the second point cloud data and the third point cloud data. In the present invention, the first point cloud data, the second point cloud data, and the third point cloud data may include point cloud data based on BIM. In the present invention, the first point cloud data, the second point cloud data, and the third point cloud data may include point cloud data acquired using a total station.

[0009] The present invention performs adjustment calculations on the first point cloud data using second and third point cloud data with higher precision than the first point cloud data. The second and third point cloud data are more precise than the first point cloud data, and the second point cloud data is more precise than the third point cloud data. The results of the adjustment calculations based on the second point cloud data are reflected in the correction of the first point cloud data with a relatively large first weight, and the results of the adjustment calculations based on the third point cloud data are reflected in the correction of the first point cloud data with a relatively small second weight. This is a point cloud data processing method.

[0010] The present invention is a program that can be read and executed by a computer. The computer is made to perform adjustment calculations on the first point cloud data using second and third point cloud data with higher precision than the first point cloud data. The second and third point cloud data are more precise than the first point cloud data, and the second point cloud data is more precise than the third point cloud data. The results of the adjustment calculations based on the second point cloud data are reflected in the correction of the first point cloud data with a relatively large first weight, and the results of the adjustment calculations based on the third point cloud data are reflected in the correction of the first point cloud data with a relatively small second weight. This is a program for point cloud data processing.

Effects of the Invention

[0011] According to the present invention, the accuracy and reliability are ensured when integrating point cloud data of multiple different systems.

Brief Description of the Drawings

[0012]

Figure 1

Figure 2

Figure 3

Modes for Carrying Out the Invention

[0013] 1. First Embodiment (Overview) FIG. 1 shows an overview of this embodiment. In FIG. 1, a worker 100 is shown walking along path C while carrying a laser scanner system 110 and performing a laser scan of the surroundings. It is also possible to mount the laser scanner system 110 on a moving body (such as a vehicle, a mobile robot, an unmanned aerial vehicle (UAV), etc.) and perform a laser scan while moving it.

[0014] Here, based on GNSS (Global Navigation Satellite System) data and IMU (inertial measurement unit) data, data on the position and orientation of the laser scanner 111 equipped in the laser scanner system 110 during the scan is acquired. However, since GNSS data contains errors (especially in the case of single positioning) and there are also problems with shaking, the data on the position and orientation of the laser scanner 111 during the scan contains errors.

[0015] Here, it is assumed that there are a plurality of regions along path C or near it where point cloud data in the absolute coordinate system has already been obtained. The absolute coordinate system is the coordinate system used in maps and GNSS, for example, a coordinate system in which coordinates are specified by latitude, longitude, and altitude.

[0016] In the example of FIG. 1, it is assumed that point cloud data 1 has already been obtained in region 1 and point cloud data 2 has already been obtained in region 2. Also, it is assumed that the point cloud data acquired by the laser scanner 111 overlaps with point cloud data 1 and point cloud data 2.

[0017] Here, the point cloud data 1 obtained in region 1 is point cloud data obtained by a stationary laser scanning device for surveying. When acquiring the point cloud data 1, the stationary laser scanning device for surveying used is in a state where its position and orientation are accurately measured. For this reason, since it is a laser scanner device for surveying, the point cloud data 1 has relatively high position accuracy.

[0018] The point cloud data 2 is point cloud data obtained by SfM using the captured images of the camera mounted on the UAV. This is point cloud data of the object to be captured obtained by continuously capturing images with the mounted camera while flying the UAV, based on the principle of three-dimensional photogrammetry. In the acquisition of the point cloud data 2, a plurality of reference points with known positions are arranged on the ground, and absolute coordinate values are given to the obtained point cloud data. However, due to factors such as the position accuracy of the reference points, the number of reference points, and the attitude stability of the flying UAV, the accuracy of the point cloud data 2 is worse compared to the point cloud data 1. For example, when the accuracy of the point cloud data 1 is 5 mm or less, the accuracy of the point cloud data 2 is a value such as 2 cm or less.

[0019] Note that even though the point cloud data 2 is relatively low in accuracy, it is more accurate than the point cloud data obtained by the laser scanning system 110.

[0020] In this embodiment, adjustment calculations are performed using the point cloud data 1 and the point cloud data 2 to improve the accuracy and reliability of the point cloud data obtained by the laser scanning system 110. At this time, the adjustment calculations are performed in consideration of the accuracy of the point cloud data 1 and the point cloud data 2.

[0021] That is, in the adjustment calculations for improving the accuracy and reliability of the point cloud data obtained by the laser scanning system 110, the contribution (weight) of the relatively high-accuracy point cloud data 1 to the adjustment calculations is made relatively high, and the contribution (weight) of the relatively low-accuracy point cloud data 2 to the adjustment calculations is made relatively low.

[0022] If the degree of the above contributions in the point cloud data 1 and the point cloud data 2 is the same, the relatively low-accuracy data of the point cloud data 2 will have an impact, and the high accuracy of the point cloud data 1 cannot be utilized. By making the degree of contribution of the point cloud data 2 to the adjustment calculations relatively low, this problem is alleviated, and the accuracy of the point cloud data obtained by the laser scanning system 110 can be improved.

[0023] Note that in the region far from the point cloud data 1 and close to the point cloud data 2, the point cloud data 2 becomes effective. Therefore, in order to improve the accuracy and reliability of all the point cloud data obtained by the laser scanning system 110, the adjustment calculation using the point cloud data 2 is effective. By adjusting the above weight distribution, this effectiveness can be utilized.

[0024] (Laser Scanning System) As shown in FIG. 1, the laser scanning system 110 includes a laser scanner 111, a GNSS position measuring device 112, an IMU 113, a data storage unit 114, a communication device 115, a user interface 116, and a camera 117.

[0025] The laser scanner 111 includes a rotating part that rotates vertically, and ranging light is emitted in pulses from this rotating part. By performing this pulsed emission while moving, laser scanning of the surroundings is performed. The form of the laser scanning is not limited to this example, and forms such as electronically scanning, performing laser scanning by reciprocating the emitting part, and performing laser scanning by a rotating mirror or a reciprocating mirror can also be adopted.

[0026] The GNSS position measuring device 112 measures the position of the laser scanning system 110 using GNSS (Global Navigation Satellite System). Here, relative positioning (DGPS method or RTK-GPS method) is preferable if possible, but due to cost constraints, etc., single positioning may also be used. In the case of single positioning, a certain error is included in the measurement value.

[0027] The IMU 113 is an inertial measurement device that measures changes in the attitude of the laser scanning system 110 and the applied acceleration. The positional and attitudinal relationships of the laser scanner 111, the GNSS position measuring device 112, and the IMU 113 are known.

[0028] The data storage unit 114 stores data, programs, and laser scan data necessary for the operation of the laser scan system 110. The communication device 115 performs communication with the outside. The communication is performed by wire or wirelessly. The user interface 116 includes an operation panel for operating the laser scan system 110 and a display for displaying various types of information. A form in which a PC, tablet, or smartphone is used as the operation panel and display is also possible. In this case, communication is performed between the PC, tablet, or smartphone and the user interface 116 via the communication device 115.

[0029] The camera 117 captures an image of the object to be laser scanned. The positional and postural relationships of the camera 117, the GNSS positioning device 112, and the IMU 113 are known.

[0030] (Data processing device) FIG. 2 is a block diagram of the data processing device 200. The data processing device 200 performs data processing on the point cloud data (laser scan data) obtained by the laser scanner system 110.

[0031] The data processing device 200 is a computer including a CPU, a storage device, various arithmetic units, and various interface devices, and is realized by a computer program for realizing the illustrated functional units being executed by the above CPU and arithmetic units. Forms of realizing the data processing device 200 include a form realized using a general-purpose computer (for example, a personal computer), a form realized by dedicated hardware, and a form realized using a data processing server.

[0032] The data processing device 200 includes a measurement data reception unit 201, a point cloud data generation unit 202, an adjustment calculation unit 203, a storage unit 204, and a communication device 205.

[0033] The measurement data reception unit 201 receives the laser scan data acquired by the laser scan system 110, the already obtained point cloud data 1 and point cloud data 2. The laser scan data acquired by the laser scan system 110 is data on the position (position of the optical origin), orientation, direction from the laser scanner 111, and distance at the time of measurement of each point. When using captured images, the measurement data reception unit 201 receives the image data of a large number of captured images that serve as the basis for SfM.

[0034] The point cloud data generation unit 202 generates point cloud data based on the laser scan data received by the measurement data reception unit 201.

[0035] The laser scan data acquired by the laser scan system 110 is obtained by performing a laser scan while moving the laser scanner 111, and the position of the optical origin of the laser scanner is different for each point obtained during the movement.

[0036] Here, the time of the laser scanner 111 when a certain specific scan point is obtained is measured, and the position and orientation of the laser scanner 111 in the absolute coordinate system at that time are obtained from the GNSS data and IMU data. Therefore, the position of each scan point in the absolute coordinate system can be calculated. By performing this calculation for each point of the laser scan point cloud, the position of the laser scan point cloud acquired by the laser scanner system 110 while moving in the absolute coordinate system is obtained. As a result, point cloud data based on the laser scan point cloud acquired by the laser scanner system 110 while moving is obtained. This process is performed by the point cloud data generation unit 202. Note that since the GNSS data contains errors, there are errors in the positions of the points of the calculated point cloud data.

[0037] Incidentally, it is also possible to obtain point cloud data from the image captured by the camera 117. In this case, selection of stereo images from a plurality of images with different viewpoints and partial overlap, extraction of feature points in the stereo images, identification of the correspondence of feature points between the stereo images, and calculation of the positions of the feature points in the absolute coordinate system by the forward intersection method are performed in the point cloud data generation unit 202. In the process of obtaining point cloud data from the captured images, measurement of the camera position by the surveying device and / or calculation of the position and orientation of the camera using the reference point target are performed. The former technique is described in, for example, Japanese Patent Application Laid-Open No. 2019-45425 and Japanese Patent Application Laid-Open No. 2019-138842.

[0038] The adjustment calculation unit 203 performs adjustment calculations to improve the accuracy and reliability of the point cloud data (hereinafter referred to as target point cloud data) obtained by the laser scanning system 110.

[0039] The principle of the adjustment calculation will be described below. The adjustment calculation is basically based on the same principle as the bundle adjustment calculation used in three-dimensional photogrammetry.

[0040] First, known point cloud data to be used for the adjustment calculation related to the target point cloud data is selected. This known point cloud data is point cloud data that overlaps at least partially with the target point cloud data. The position of each point of this known point cloud data is described using the absolute coordinate system of each point.

[0041] After selecting the known point cloud data, matching (identification of the correspondence) between the target point cloud data and the known point cloud data is performed. In this process, approximate matching using the position information of the target point cloud data and the known point cloud data is performed, and further detailed matching of points using template matching or the like is performed.

[0042] When matching the target point cloud data with the known point cloud data, adjustment calculations are performed using each point of the known point cloud that has been matched with the target point cloud data. Here, the target point cloud data has relatively large errors, and the known point cloud data has relatively small errors. Therefore, for a specific point in the known point cloud data, there is a deviation between the already obtained coordinates and the coordinates with the laser scanner 111 as the origin (viewpoint). The position of the laser scanner 111 is adjusted so that this deviation at each point becomes smaller.

[0043] For example, assume that as a result of the above matching process, point P1 of the target point cloud data and point p1 of point cloud data 1 are obtained as corresponding points. Also, let the position on the path C of the laser scanner 111 when point P1 is obtained be i. Here, since the target point cloud data has relatively large errors, the position data (coordinate values) of P1 in the absolute coordinate system and the position data (coordinate values) of point p2 do not coincide and there is a deviation. Therefore, the direction line starting from the laser scanner 111 located at position i on path C in Figure 1 and heading towards point P1 does not pass through point p1 but passes through a distant position (if there is no error, it would pass through point p1). Therefore, the parameters of the position and orientation of the laser scanner 111 at position i on path C are adjusted so that this deviation is corrected. This process is performed for all positions i (i = 1, 2, 3, 4, ···) on path C for each point for which the above matching is achieved, and conditions are explored such that the total deviation becomes minimum and the calculation results converge. This process is performed in the adjustment calculation.

[0044] In this adjustment calculation, each parameter is slightly displaced, and the calculation is performed so that the calculation results converge within a predetermined range. As a result of this adjustment calculation, the position and orientation of the laser scanner 111 on path C regarding the above corresponding points are corrected. Here, the interval between positions i on path C is, for example, 0.1 m to 2 m. Narrowing this interval allows for finer correction of the position and orientation of the laser scanner 111, but the processing burden increases.

[0045] In this way, the position of the movement trajectory C of the laser scanner 111 and the posture of the laser scanner 111 on this movement trajectory are corrected based on the known point cloud data (point cloud data 1 in the above case). Then, by using the corrected position and posture of the laser scanner to correct (recalculate) the target point cloud data, the accuracy and reliability of the target point cloud data are improved. This is the basic principle of SfM in this embodiment.

[0046] In the above case, point cloud data 1 is selected as the known point cloud data. However, as the distance from point cloud data 1 increases, the improvement effect of the accuracy of the position and posture of the laser scanner 111 by the adjustment calculation on path C decreases. Therefore, in addition to point cloud data 1, adjustment calculations using point cloud data 2 are further performed to improve the accuracy of the target point cloud data as a whole.

[0047] In this example, as shown in FIG. 1, there are point cloud data 1 and point cloud data 2 as the known point cloud data. Therefore, adjustment calculations using both point cloud data 1 and point cloud data 2 are performed to improve the accuracy of the target point cloud data in as wide a range as possible.

[0048] Here, point cloud data 1 is relatively high-precision and point cloud data 2 is relatively low-precision. Weighting is performed corresponding to this difference in accuracy, and the result of the adjustment calculation is reflected in the movement trajectory of the laser scanner 111 and the posture on this movement trajectory according to this weighting. Thereby, the advantage of performing adjustment calculations in as wide a range as possible by using point cloud data 2 can be obtained, and the adverse effect on the result of the adjustment calculation of relatively low-precision point cloud data 2 can be reduced.

[0049] For example, taking m and n as weighting coefficients, the weighting of point cloud data 1 is set to m = 0.8, and the weighting of point cloud data 2 is set to n = 0.2. m = 0.8 reflects relatively high precision, and m = 0.2 reflects relatively low precision. Note that m + n = 1.

[0050] In this case, the contribution of the result of the adjustment calculation in point cloud data 1 is multiplied by a coefficient of 0.8, and in point cloud data 2, it is multiplied by a coefficient of 0.2. By doing so, the contribution of the result of the adjustment calculation based on the relatively less accurate point cloud data 2 is reduced, and the high accuracy of point cloud data 1 is utilized.

[0051] Hereinafter, a specific example will be described. For example, let the weighting coefficient of point cloud data 1 be m and the weighting coefficient of point cloud data 2 be n. Note that m + n = 1. Here, let the correction amounts of the position of the laser scanner 111 at position i on path C by the adjustment calculation based on point cloud data 1 be ΔX1, ΔY1, ΔZ1, and the correction amounts of the position by the adjustment calculation based on point cloud data 2 be ΔX2, ΔY2, ΔZ2. Let the correction amounts of the movement trajectory of the laser scanner 111 (based on the adjustment calculation using both point cloud data 1 and point cloud data 2) be Δx, Δy, and Δz. In this case, the correction amount of each component is Δx = mΔX1 + nΔX2, Δy = mΔY1 + nΔY2, Δz = mΔZ1 + nΔZ2.

[0052] Note that the attitude of the laser scanner is also corrected in a similar manner. The above processing is performed in the adjustment calculation unit 203.

[0053] Here, the case where two known point cloud data, point cloud data 1 and point cloud data 2, are used has been described, but this concept can be extended to the case where there are three or more known point cloud data.

[0054] The storage unit 204 stores the data and operation programs necessary for the operation of the data processing device 200, and the data obtained as a result of various processes in the data processing device 200. For example, in the case of FIG. 1, point cloud data 1 and point cloud data 2 are stored in the storage unit 204. A form in which point cloud data 1 and point cloud data 2 are stored in a storage device installed in another location and read from there is also acceptable. The communication device 205 communicates with external devices. The communication is performed wirelessly or by wire.

[0055] (An example of processing) FIG. 3 is a flowchart showing an example of the processing performed in the data processing apparatus 200. The processing in FIG. 3 is stored in the storage unit 204 or an appropriate storage medium and executed by the CPU of the data processing apparatus 200. Prior to the processing, the operator 100 walks along the path C while carrying the laser scanner system 110, and performs laser scanning during this movement to obtain point cloud data around the path C. This point cloud data becomes the object of adjustment calculation. It is assumed that the point cloud data 1 and the point cloud data 2 have been acquired in advance and their accuracies have also been obtained (or estimated).

[0056] When the processing starts, the point cloud data around the path C (hereinafter referred to as target point cloud data) obtained by the laser scanner system 110, the point cloud data 1, and the point cloud data 2 are acquired (step S101). Next, weight setting based on the accuracies of the point cloud data 1 and the point cloud data 2 is performed (step S102). Next, adjustment calculation regarding the target point cloud data based on the point cloud data 1 and the point cloud data 2 is performed (step S103). At this time, adjustment calculation reflecting the weighting set in step S102 is performed.

[0057] (Advantages) According to the present embodiment, known point cloud data with relatively low accuracy is relatively less contributed to the target point cloud data to be adjusted by adjustment calculation, whereby the error in the entire integrated point cloud data is optimally distributed. Therefore, the accuracy and reliability when integrating point cloud data of a plurality of different systems are ensured.

[0058] 2. Second Embodiment When the path C is long and the walking time of the operator 100 who walks while carrying the laser scanner system 110 becomes long, it is conceivable that the trajectory of the path C and the error in the posture during movement change halfway. This is because, during the movement of the operator 100, changes in the position of the navigation satellite and switching of the navigation satellite occur, and the state of the error in GNSS positioning changes. Also, because the change in the error of the IMU occurs with the passage of time.

[0059] In this case, it may not be appropriate to apply the result of the adjustment calculation using the known point cloud in a specific area to the entire path C. Therefore, the weighting regarding the result of the adjustment calculation based on a specific known point cloud is changed according to the moving time of the laser scanner system 110 (laser scanner 111) along the path C from the said known point cloud.

[0060] For example, in FIG. 1, consider a situation where the worker 100 walks from the area of the point cloud data 1 towards the area of the point cloud data 2, and during this time, the laser scanner 111 performs a laser scan of the surroundings.

[0061] In this case, let the weighting coefficient of the point cloud data 1 in FIG. 1 be mi, and the weighting coefficient of the point cloud data 2 be ni. Here, i is the elapsed time on the path C. The starting point of the elapsed time is the time at the edge portion on the side of the point cloud data 2 of the point cloud data 1 on the moving path C. The elapsed time is measured on the path from this starting point position towards the point cloud data 2. Note that mi + ni = 1. Also, assume that the weighting according to the accuracy described in the first embodiment is set at the intermediate time position between the point cloud data 1 and the point cloud data. The intermediate time position is calculated starting from the outer edge portion of the point cloud data. Also, assume that the time position i is between the point cloud data 1 and the point cloud data 2 on the path C.

[0062] Here, mi is a function that decreases as the distance from the outer edge of the point cloud data 1 on the path C increases. ni is a function that decreases as the distance from the outer edge of the point cloud data 2 increases.

[0063] Here, let the correction amounts of the position of the laser scanner 111 at the time position i of the path C by the adjustment calculation based on the point cloud data 1 be ΔX1, ΔY1, and ΔZ1. Also, let the correction amounts of the position of the laser scanner 111 at the time position i of the path C by the adjustment calculation based on the point cloud data 2 be ΔX2, ΔY2, and ΔZ2. Further, let the correction amounts of the position of the movement locus of the laser scanner 111 at the time position i on the final path C be Δxi, Δyi, and Δzi. In this case, Δxi = miΔX1 + niΔX2, Δyi = miΔY1 + niΔY2, and Δzi = miΔZ1 + niΔZ2. Similar processing is also performed regarding the posture.

[0064] 3. Third Embodiment As a method for evaluating the accuracy of known point cloud data, it is also possible to consider the time when the point cloud data was acquired. In this case, the accuracy of the new point cloud data is evaluated higher.

[0065] 4. Fourth Embodiment Assume that there is first known point cloud data and second known point cloud data in an overlapping area as the known point cloud data. In this case, the one with higher accuracy is adopted.

[0066] 5. Fifth Embodiment In FIG. 1, assume that the point cloud data 1 is obtained in a narrow range and the point cloud data 2 is obtained in a wide range including the area of the point cloud data 1. This is because the point cloud data obtained using a UAV is obtained in a wide range, while the point cloud data obtained using a stationary laser scanner has a limited range. In this case, in the overlapping area of the two, the point cloud data 1 is adopted, and in other parts, the point cloud data 2 is adopted.

[0067] 6. Sixth Embodiment As a means of obtaining point cloud data, a camera can also be used. For example, in the laser scanning system 110 of FIG. 1, a monocular camera or a stereo camera is used instead of the laser scanner. By taking pictures while moving, pictures are obtained from a number of different viewpoints, and point cloud data of the object to be photographed can be obtained based on the principle of SfM. It is also possible to use both a laser scanner and a camera to obtain the point cloud data of the measurement object. The laser scanning system 110 may be equipped with other surveying devices.

[0068] 7. Seventh Embodiment Regarding the accuracy of known point cloud data (for example, point cloud data 1 and point cloud data 2 in FIG. 1), it may be ranked in advance as accuracy level 1 (high accuracy), accuracy level 2 (slightly high accuracy), accuracy level 3 (medium accuracy), accuracy level 4 (slightly low accuracy), accuracy level 5 (low accuracy), etc., and the weighting coefficient may be determined based on that information.

[0069] For example, the above ranks are determined based on the method of obtaining known point clouds and the model of the measuring device. For example, when obtaining point cloud data using a UAV, the acquisition of point cloud data is performed using equipment and software developed for that purpose. In this case, the accuracy may be guaranteed by the manufacturer providing the technology. This is the same for laser scanning devices. In such a case, the products to be used are registered in advance, and the rank regarding the above accuracy is obtained based on the registration information.

[0070] 8. Eighth Embodiment There is also a method of selecting known point cloud data according to the form of the target point cloud data. Generally speaking, there are two forms of point cloud data. The first form is point cloud data obtained by laser scanning. The second form is point cloud data obtained by three-dimensional photogrammetry.

[0071] The point cloud data of the first form is obtained as point cloud data in which points are regularly arranged. The point cloud data of the second form is obtained as point cloud data in which points are unevenly distributed (localized and existing). The second point cloud data has the feature points extracted from the image as points. Since the feature points are extracted from edges, protrusions, color boundary parts, etc., the points are partially unevenly distributed (for example, it becomes a point cloud in which points are distributed along the edge of the shape of the object).

[0072] In the matching (specification of the correspondence relationship) of two different point cloud data, it is easier to match and the matching error is smaller between the point cloud data of the first form or between the point cloud data of the second form. Therefore, when the target point cloud data is of the first form, the one of the first form is preferentially selected as the known point cloud data, and when the target point cloud data is of the second form, the one of the second form is preferentially selected as the known point cloud data. This suppresses the reduction in accuracy due to matching errors and incorrect matching.

[0073] 9. Others When the present invention is used, loop closure processing is not necessarily required. Of course, its use is not excluded. Loop closure processing is a technique for suppressing the occurrence of errors by making the start position and the goal position of the moving surveying device the same point and using this point as a constraint point in adjustment calculation.

[0074] The present invention can also be used for BIM (Building Information Modeling) data. When there are a plurality of BIM data, there may be a difference in accuracy level. The present invention can be used for the integration of the plurality of BIM data. In this case, the point cloud data of the object described in the BIM data is generated based on the BIM data. The point cloud data based on the BIM data is obtained by creating data in which the object described in the BIM data is drawn as a set of points. Then, the present invention is used to integrate a plurality of point cloud data based on each BIM data. At this time, weighting based on the accuracy of each BIM data is set for each target point cloud data.

[0075] The present invention can also be used for integrating point cloud data obtained by a laser scanner and point cloud data obtained from BIM. That is, among a plurality of point cloud data to be integrated, there may be point cloud data based on BIM data.

[0076] As high-precision point cloud data, point cloud data obtained using a surveying device capable of highly accurate point surveying such as a total station may be adopted. For example, point cloud data of local parts such as the corner parts of a structure can be obtained by a total station, and this point cloud data may be used as high-precision point cloud data (that is, point cloud data with a large weight).

[0077] For example, assume there is a wide-area first point cloud data. Also assume that there is one or more partial point cloud data with relatively high precision (the high-precision partial point cloud data obtained by the above total station) in this first point cloud data. In this case, by integrating the first point cloud data and the one or more partial point cloud data, the accuracy of the first point cloud data can be improved.

[0078] The method for obtaining the target point cloud to be subject to adjustment calculation is not limited to the exemplified method (the method of obtaining while moving). The target point cloud may be obtained using a laser scanner installed at a specific position.

Industrial Applicability

[0079] The present invention can be used in technologies dealing with point cloud data.

Explanation of Signs

[0080] 100... Worker, 110... Laser scanning system, 111... Laser scanner, 117... Camera.

Claims

1. A point cloud data processing device comprising an adjustment calculation unit that performs adjustment calculation of the first point cloud data using second point cloud data and third point cloud data with higher accuracy than the first point cloud data, wherein the second point cloud data and the third point cloud data are more accurate than the first point cloud data, the second point cloud data is more accurate than the third point cloud data, the result of the adjustment calculation based on the second point cloud data is reflected in the correction of the first point cloud data with a relatively large first weight, and the result of the adjustment calculation based on the third point cloud data is reflected in the correction of the first point cloud data with a relatively small second weight.

2. wherein the first point cloud data is acquired while moving, the first weight decreases as the distance from the first point cloud data on the path of the movement increases, and the second weight decreases as the distance from the second point cloud data on the path of the movement increases, according to the point cloud data processing device according to claim 1.

3. The point cloud data processing device according to claim 1, wherein the second point cloud data and the third point cloud data are associated with information regarding accuracy.

4. When the first point cloud data is point cloud data obtained by laser scanning, point cloud data obtained by laser scanning is selected as the second point cloud data and the third point cloud data, and when the first point cloud data is point cloud data obtained based on a photographed image, point cloud data obtained based on the photographed image is selected as the second point cloud data and the third point cloud data, according to the point cloud data processing device according to claim 1.

5. The point cloud data processing device according to claim 1, wherein the first point cloud data, the second point cloud data, and the third point cloud data include point cloud data based on BIM.

6. The point cloud data processing device according to claim 1, wherein the first point cloud data, the second point cloud data, and the third point cloud data include point cloud data acquired using a total station.

7. Performing adjustment calculation of the first point cloud data using second point cloud data and third point cloud data with higher accuracy than the first point cloud data, wherein the second point cloud data and the third point cloud data are more accurate than the first point cloud data, and the second point cloud data is more accurate than the third point cloud data, The result of the adjustment calculation based on the second point cloud data is reflected in the correction of the first point cloud data with a relatively large first weight, and the result of the adjustment calculation based on the third point cloud data is reflected in the correction of the first point cloud data with a relatively small second weight. A point cloud data processing method. **Claim 8** A program that can be read and executed by a computer, causing the computer to perform an adjustment calculation of the first point cloud data using second point cloud data and third point cloud data with higher accuracy than the first point cloud data, wherein the second point cloud data and the third point cloud data have higher accuracy than the first point cloud data, the second point cloud data has higher accuracy than the third point cloud data, the result of the adjustment calculation based on the second point cloud data is reflected in the correction of the first point cloud data with a relatively large first weight, and the result of the adjustment calculation based on the third point cloud data is reflected in the correction of the first point cloud data with a relatively small second weight. A program for point cloud data processing.

Images