METHOD AND SYSTEM FOR DETERMINING POSITIONS ON THE GROUND FROM AN AERIAL VEHICLE

Abstract

A method and system for determining positions on the ground from an aerial vehicle, including scanning the ground with a measuring beam to obtain a point cloud, wherein each point of the point cloud represents coordinates of a position on the ground in an aircraft-related coordinate system; segmenting a first subset of points within the point cloud, wherein the first subset of points shares a predefined first distortion class; segmenting a second subset of points within the point cloud, wherein the second subset of points shares a predefined second distortion class; geo-referencing the point cloud from the aircraft-related coordinate system to an Earth-related coordinate system; strip adjusting the geo-referenced first subset of points relative to the geo-referenced second subset of points for identifying a set of correction parameters; correcting the geo-referenced first subset of points by the set of correction parameters.

Claims

1. A method for determining positions on the ground (15) from an aerial vehicle (17), the method comprising: a) Scanning the ground (15) with a measuring beam (30) to obtain a point cloud (47), wherein each point (43) of the point cloud (47) represents coordinates of a position on the ground (15) in an aircraft-related coordinate system; b) Segmenting a first subset (44) of points (43) within the point cloud (47), wherein the first subset (44) of points (43) shares a predefined first distortion class (38); c) Segmenting a second subset (45) of points (43) within the point cloud (47), wherein the second subset (45) of points (43) shares a predefined second distortion class (36); d) Geo-referencing the point cloud (47) from the aircraft-related coordinate system to an Earth-related coordinate system; e) Strip adjusting the geo-referenced first subset (44) of points (43) relative to the geo-referenced second subset (45) of points (43) to identify a set of correction parameters; and f) Correcting the geo-referenced first subset (44) of points (43) by the set of correction parameters.

2. The method of claim 1, wherein there is a partial overlap between the positions on the ground (15) covered by the first subset (44) of points (43) and the positions on the ground (15) covered by the second subset (45) of points.

3. The method of claim 1, wherein a position of the UAV (17) is changed along a survey line (14) during the measurement.

4. The method of claim 1, wherein one or more of steps b) to f) are performed as post-processing steps after the UAV (17) has landed.

5. The method of claim 1, wherein more than two subsets (44, 45, 46) of points (43) are segmented from the point cloud (47).

6. The method of claim 1, wherein two or more subsets (44, 45, 46) of points (43) are corrected with sets of correction parameters.

7. The method of claim 1, wherein each point (43) of the point cloud (47) is tagged with a distortion class (35-42).

8. The method of claim 1, comprising the step of determining a correlation between a direction, a distance and/or a time stamp of the measuring beam (30) and a distortion class (35-42).

9. The method of claim 1, wherein the measuring beam (30) is deflected by a Risley prism (31, 32).

10. The method of claim 1, wherein the UAV (17) carries a measuring device (17) in form of a LiDAR device.

11. A system for determining positions on the ground (15) from an aerial vehicle (17), the system comprising a UAV (17) with a measuring device (18) for scanning the ground (15) with a measuring beam (30) to obtain a point cloud (47), so that each point (43) of the point cloud (47) represents coordinates of a position on the ground (15) in an aircraft-related coordinate system; the system further comprising a calculation module adapted for performing the following steps: a) Segmenting a first subset (44) of points (43) within the point cloud (47), wherein the first subset (44) of points (43) shares a predefined first distortion class (38); b) Segmenting a second subset (45) of points (43) within the point cloud (47), wherein the second subset (45) of points (43) shares a predefined second distortion class (36); c) Geo-referencing the point cloud (47) from the aircraft-related coordinate system to an Earth-related coordinate system; d) Strip adjusting the geo-referenced first subset (44) of points (43) relative to the geo-referenced second subset (45) of points (43) for identifying a set of correction parameters; e) Correcting the geo-referenced first subset (44) of points (43) by the set of correction parameters.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In the following, the invention is described in exemplary fashion on the basis of advantageous embodiments, with reference being made to the attached drawings. In detail:

[0046] FIG. 1: shows an embodiment of the inventive system;

[0047] FIG. 2: shows a schematic illustration of an inventive measuring system;

[0048] FIG. 3: shows a schematic illustration of distortion classes of the measuring system of FIG. 2;

[0049] FIG. 4: shows a point cloud obtained with the measuring system of FIG. 2;

[0050] FIG. 5: shows another embodiment of an inventive measuring system;

[0051] FIG. 6: shows the measuring system of FIG. 5 in a different state;

[0052] FIG. 7: a perspective view of the measuring system of FIG. 5;

[0053] FIG. 8: a scanning pattern of the measuring system of FIG. 7;

[0054] FIG. 9: another scanning pattern of the measuring system of FIG. 7;

[0055] FIG. 10: a point cloud obtained with the measuring system of FIG. 8.

DETAILED DESCRIPTION

[0056] In FIG. 1, a UAV 17 is operated remotely by a ground-based operator (not shown). The operator uses a remote control for controlling the UAV 17 via a radio frequency (RF) link. Mission planning information including flight control and operating instructions for the UAV 17 are either stored in a controller memory of the remote control 10 or a in controller within a body 19 of the UAV 17. The UAV 17 carries a measuring device in form of a LiDAR system 18. The LiDAR system 18 comprises a LASER scanner, an Inertial Measurement Unit (IMU), and a GNSS receiver. The LiDAR system 18 is powered by the UAV 17.

[0057] The UAV 17 is shown in flight above an area on the ground 15, wherein the UAV 17 follows a survey line 14. The LiDAR scanner 18 mounted on the UAV emits LASER beams 30 with varying directions so that the ground is scanned with a scanning pattern 16 that is enclosed with a circular envelope 20. The light of the LASER beams 30 is reflected back to the LiDAR system 18 by objects on the ground. From the direction of the LASER beams 30 and the time of flight of the LASER light, the position of the object relative to the UAV 17 can be determined. The scanning of the ground with the LASER beams 30 results in a point cloud 47, wherein each point 43 of the point cloud 47 represents a position on the ground, respectively, the position of an object on the ground.

[0058] The data obtained with the LiDAR system 18 provides position data relative to an aircraft-related coordinate system. By using additional information from the navigation system of the UAV 17, namely from the IMU and the GNSS receiver of the UAV 17, the position data of the point cloud 47 is geo-referenced as 3D points in an Earth-related coordinate system.

[0059] The LiDAR system 18 comprises a wedge-shaped prism 21 being rotatable about a central axis 22 of the LiDAR system 18. The light beam 30 being emitted along the central axis 22 is deflected by the wedge-shaped prism 21 so that the light beam 30 after the prism 21 encloses an angle 23 with the central axis 22. When the prism 21 is rotated, the light beam 30 scans a circular scanning pattern 16 on the ground 15.

[0060] In FIG. 2 the LiDAR LASER scanner 18 is formed by single rotating prism 21. The LASER beam 30 is deviated from the central axis 22 with an angle 23. When the prism 21 is rotating at a given angular velocity, the beam intersects the plane 15 and forms the circle 16. In the top view of the scanning pattern 16 shown in FIG. 3 the circular scanning path is divided in four sectors 25, 26, 27, 28. The error in the orientation or in the angle measurement of the prism 21 has the effect that the LiDAR system 18 provides roughly correct position data in two of the sectors, namely sectors 27, 28. In sector 25 the distance to the ground 15 is measured too short, in sector 26 the distance to the ground 15 is measured too long.

[0061] When the UAV 17 follows survey line 14 position data is provided from an area on the ground 15, wherein the same positions on the ground 15 are first scanned with the measuring beam 30 being in sector 26 and later are scanned with the measuring beam 30 being in sector 25 of the scanning pattern 16. When being geo-referenced the coordinates of one and the same position on the ground 15 that are measured with in sector 25 will not match with the coordinates that are measured in sector 26, but the coordinates for one and the same position on the ground lie in two different parallel planes. With the inventive method the mismatch between different sets of coordinates belonging to the same position on the ground is minimized.

[0062] In the example of FIG. 3 different distortion classes are correlated with each of the sectors 25, 26, 27, 28. The points within the point cloud having the same distortion class share the same level of systematic distortion, namely “too short” for sector 25, “too long” for sector 26 and “roughly correct” for sectors 27, 28. When following the scanning pattern 16, the measuring beam 30 has a plurality of different directions relative to the central axis 22 of the LiDAR system 18. Each of these directions is correlated with a distortion class. Each single point that is taken within the scanning pattern is tagged with a distortion class. The points of the point cloud sharing a first distortion class form a first subset of the point cloud, the points of the point cloud sharing a second distortion class form a second subset of the point cloud and so on.

[0063] The point cloud as measured provides coordinates in the aircraft-related coordinate system. By using additional information from the navigation system of the UAV 17 the point cloud can be geo-referenced which means that each set of coordinates for a position on the ground is transformed from the aircraft-related coordinate system to the Earth-related coordinate system.

[0064] Within the aircraft-related coordinate system the different subsets of points do not stand in an obvious relation relative to each other. After geo-referencing the deviation between two points from different subsets being related to the same position on the ground can be determined and a strip adjustment algorithm can be applied for minimizing the mismatch. The strip adjustment will result in a set of correction parameters that can be applied to equation (2) to compute each of the points of one subset of points for reducing the mismatch between this subset of points and another subset of points. According to the invention position data in a coordinate system (Earth-related coordinate system) that is different from the coordinate system of the LiDAR device 18 is thus used for correcting a systematic error within the LiDAR device 18.

[0065] In FIG. 4 a point cloud 47 obtained with the LiDAR system 18 of FIG. 2 is shown. The scanning pattern 16 is divided in five distortion classes 36-40. Based on the distortion classes 36-40 the points 43 of the point cloud 47 separated in five subsets of points 43. The points 43 in the first distortion class 36 form a first subset 44 of points 43 (marked with a star in FIG. 4). The points in the second distortion class 37 form a second subset 45 of points 43 (marked with a triangle). The points in the third distortion class 38 form a third subset 46 of points (marked with a dot). As the UAV 17 moves along survey line 14 each of the subsets 44, 45, 46 of points covers a virtual strip 48 on the ground 15. Based on the overlapping regions 49 between the different virtual strips the strip adjustment step can be performed.

[0066] In the embodiment of FIG. 5 the LiDAR system 18 has an optical system comprising a first optical element 31 in the form of a wedge-shaped prism and a second optical element 32 in the form of a wedge-shaped prism.

[0067] The two optical elements 31, 32 are coaxially aligned and rotatable about the central axis 22 of the LiDAR system 18. Such arrangement of optical elements 31, 32 is called a Risley prim. In FIG. 5 both prisms have the same angular position so that the thick portions of both prisms 31, 32 as well as the thin portions of both prisms 31, 32 are aligned. In this state of the Risley prism 31, 32 the LASER beam 30 has a maximum deflection. If both prisms 31, 32 are rotated synchronously the light beam 30 describes a circular scanning pattern 16 on the ground 15. The circular scanning pattern 16 in the aligned state of the prisms 31, 32 corresponds to the envelope 20 of the LiDAR system 18.

[0068] In FIG. 6 the angular positions of the two prisms 31, 32 are separated by 180° so that the thin portion of the first prism 31 is aligned with the thick portion of the second prism 32 and vice versa. In this counter-aligned state of the prisms 31, 32 the light beam 30 undergoes a lateral shift and propagates in a direction parallel to the central axis 22 towards the ground 15. When the two prisms 31, 32 are rotated synchronously the light beam 30 describes a circular scanning pattern 16 on the ground 15. The circular scanning pattern 16 encloses a blind area 33 around the central axis 22, which is not accessible with the LiDAR system 18. The region between the envelope 20 and the blind area 33 corresponds to the scanning region 34 of the LiDAR system 18.

[0069] If the Risley prisms 31, 32 are rotated at different angular velocities a more complex scanning pattern 16 is generated between the envelope 20 and the blind area 33. One example of such scanning pattern 16 is shown in FIG. 7. The exact form of the scanning pattern 16 depends on the ratio between the angular velocity of the first Risley prism 31 and the second Risley prism 32. The scanning pattern will be periodical if the ratio is rational, the scanning pattern will be non-periodical if the ratio is irrational.

[0070] The ratio of the angular velocities can be adjusted so that the scanning region 34 of the LiDAR system 18 is densely covered with the scanning pattern 16. This enables the LiDAR system 18 to provide a substantially complete picture of the scanned region 34 without altering the position of the UAV 17 having the LiDAR system 18 on board.

[0071] For obtaining position data from regions on the ground outside the scanning region 34 the position of the UAV 17 is changed. The two prisms 31, 32 can be rotated continuously during the measurement. The angular velocity of both prisms 31, 32 can be constant.

[0072] For any Risley prism 31, 32 a distortion map can be obtained, wherein the distortion map provides a level of distortion for each direction of the measuring beam 30 within the scanning region 34. If the level of distortion is defined on a relative scale the distortion can for example range from −1% to +1%. This means each direction of the measuring beam 30 within the scanning region 34 is correlated to a distortion level having a value between −1% and +1%. Based on the distortion level the beam directions can be classified in distortion classes. For example, four distortion classes can be defined, wherein the first distortion class includes beam directions having a level of distortion between −1% and −0.5%, wherein the second distortion class includes levels of distortion between −0.5% and 0%, wherein the third distortion class includes levels of distortion between 0% and 0.5% and wherein the fourth distortion class includes levels of distortion between 0.5% and 1%. Based on the distortion map each direction of the measuring beam 30 is correlated with a distortion class.

[0073] In FIG. 8 a part of a scanning pattern 16 is shown, wherein different sections of the scanning pattern 16 are correlated with different distortion classes 35-42. Each set of coordinates from the ground 15 that is obtained with a measuring beam 30 being within one of the sections of the scanning pattern 16 is tagged with a distortion tag identifying the level of distortion according to distortion classes 35-42.

[0074] Another scanning pattern 16 having a simple periodic form is shown in FIG. 9. The scanning pattern 16 is classified in six distortion classes 35-40. When the UAV 17 follows survey line 14 the scanning pattern 16 follows a path as indicated in FIG. 10. The points 43 sharing a common distortion class are marked with a common symbol (e.g., star, triangle). All the points having a common symbol form a subset of points of the point cloud obtained with the UAV 17 on survey line 14. In FIG. 10 three different subsets 44, 45, 46 are marked with reference numerals. The sum of all the subsets corresponds to the point cloud 47 obtained during the measurement along survey line 14.

[0075] Along survey line 14 each subset 44, 45, 46 of points forms a virtual strip over the scanned region of the ground 15. The virtual strip of subset 46 of points is marked with reference 48 in FIG. 10. By performing a strip adjustment in the Earth-related coordinate system based on the overlapping regions between the subsets of points a set of correction parameters is identified for each of the subsets of points. The correction parameters are applied to adjust the subsets relatively to each other within the Earth-related coordinate system using equation (2) so that the systematic distortion of the LiDAR system 18 is corrected.