DE102020211306A1 - Cluster analysis of a point cloud - Google Patents

Abstract

The invention relates to a method for cluster analysis of a point cloud representing an environment of a time-of-flight sensor, comprising the following steps: Receiving point-cloud signals, which represent a point cloud comprising a number of points, which represents an area around the time-of-flight sensor detected by means of the time-of-flight sensor, with the number of points each having a pulse characteristic of echo pulses detected by the runtime sensor and scattered by one or more respective areas of objects in the vicinity of the runtime sensor, on the basis of which the multiple points were determined,for at least one point of the point cloud, determining spatial position information of the area corresponding to the at least one point based on the pulse characteristic of the echo pulse of the at least one point.The invention further relates to a device, a transit time sensor, a computer program, a machine-readable storage medium and a force f vehicle.

Description

The invention relates to a method for cluster analysis of a point cloud representing an environment of a transit time sensor, a device, a transit time sensor, a computer program, a machine-readable storage medium and a motor vehicle.

State of the art

The disclosure document US 2018/0299552 A1 discloses a lidar system.

The patent specification EP 1 358 508 B1 discloses a lidar system.

Disclosure of Invention

The object on which the invention is based is to be seen in the provision of a concept for efficient cluster analysis of a point cloud representing an environment of a time-of-flight sensor.

This object is solved by means of the respective subject matter of the independent claims. Advantageous configurations of the invention are the subject matter of the dependent subclaims.

• Empfangen von Punktwolkesignalen, welche eine mehrere Punkte umfassende Punktwolke repräsentieren, welche eine mittels des Laufzeitsensors erfasste Umgebung des Laufzeitsensors repräsentiert, wobei den mehreren Punkten jeweils eine Pulscharakteristik von mittels des Laufzeitsensors detektierten von einer oder mehreren jeweiligen Flächen von Objekten in der Umgebung des Laufzeitsensors gestreuten Echopulsen zugeordnet ist, basierend auf welchen die mehreren Punkte ermittelt wurden,
• für zumindest einen Punkt der Punktwolke Ermitteln einer Raumlageinformation der dem zumindest einen Punkt entsprechende Fläche basierend auf der Pulscharakteristik des Echopulses des zumindest einen Punktes.

According to a first aspect, a method for cluster analysis of a point cloud representing an environment of a transit time sensor, comprising the following steps:

• Receiving point cloud signals, which represent a point cloud comprising a number of points, which represents an environment of the time sensor detected by the time sensor, with the number of points each having a pulse characteristic of echo pulses detected by the time sensor and scattered by one or more respective surfaces of objects in the area around the time sensor is assigned, based on which the multiple points were determined,
• for at least one point of the point cloud, determining spatial position information of the area corresponding to the at least one point based on the pulse characteristic of the echo pulse of the at least one point.

According to a second aspect, a device is provided which is set up to carry out all the steps of the method according to the first aspect.

According to a third aspect, a transit time sensor is provided which comprises the device according to the second aspect.

According to a fourth aspect, a computer program is provided which comprises instructions which, when the computer program is executed by a computer, for example by the device according to the second aspect and/or by the transit time sensor according to the third aspect, cause this to carry out a method according to the first aspect to execute.

According to a fifth aspect, there is provided a machine-readable storage medium on which the computer program according to the fourth aspect is stored.

According to a sixth aspect, a motor vehicle is provided, comprising the device according to the second aspect or the runtime sensor according to the third aspect.

The invention is based on and includes the knowledge that the above object is achieved in that a pulse characteristic of the echo pulse is used to determine spatial position information of a surface of an object in the vicinity of the time-of-flight sensor, from which a pulse was reflected and thus became an echo pulse. It has been shown that the pulse characteristic of the echo pulse depends on the orientation of the surface relative to the time-of-flight sensor. For example, a surface that is oriented at an angle to the transit time sensor causes the echo pulse to broaden compared to a surface that is oriented frontally to the transit time sensor. Spatial position information of the surface can thus be determined efficiently in an advantageous manner.

Thus, in particular, the technical advantage is brought about that a concept for efficient cluster analysis of a point cloud representing an environment of a time-of-flight sensor is provided.

According to one embodiment, a pulse characteristic comprises a pulse width and/or a pulse shape.

According to one embodiment, the transit time sensor is set up to emit pulses for scanning the environment in different spatial directions and to determine a distance from the scattering, in particular reflective, surface or surfaces of an object by measuring the transit time of the echo pulses reflected back, i.e. to carry out a transit time measurement.

According to one embodiment, the transit time sensor is one of the following sensors: LiDAR sensor, radar sensor and ultrasonic sensor.

With a LiDAR sensor, the emitted pulse is a laser pulse, which can also be referred to as a LiDAR pulse. The echo pulse can be referred to as a LiDAR echo.

In the case of a radar sensor, the emitted pulse is a radar pulse. The echo pulse can be referred to as a radar echo.

With an ultrasonic sensor, the emitted pulse is an ultrasonic pulse. The echo pulse can be referred to as an ultrasonic echo.

According to one embodiment, the transit time sensor is a transit time sensor for a motor vehicle.

The spatial position information includes, for example, a spatial position, in particular an orientation of the surface, in particular a roll angle, in particular a yaw angle, in particular a pitch angle.

In one embodiment it is provided that a respective spatial position information of the corresponding areas is determined for several points of the point cloud, wherein depending on the respective spatial position information at least one area cluster is determined, to which at least some of the several points are assigned.

This brings about the technical advantage, for example, that the area cluster can be determined efficiently. In particular, those points of the plurality of points are assigned or combined to form a surface cluster which have the same respective spatial position information and/or which have the same respective spatial position information within a predetermined tolerance range.

In one embodiment, it is provided that the pulse characteristic includes a start time and an end time, with the end time of one point being compared with the start time of the other point for two directly adjacent points in the point cloud, with it being determined based on the comparison whether the two points correspond to two different faces separated by a gap or whether the two points correspond to a common continuous face. For example, the start time can be defined as the 50% point of a rising edge of the echo pulse and the end time as the 50% point of a falling edge of the echo pulse. The percentages refer to the maximum of the echo pulse (100%).

This brings about the technical advantage, for example, that it can be efficiently determined whether a gap or a distance is provided between the two surfaces or whether the two surfaces directly adjoin one another, ie form a common continuous surface.

In one embodiment it is provided that when two area clusters are determined, the two immediately adjacent points are defined in such a way that one point is assigned to one of the two area clusters and that the other point is assigned to the other of the two area clusters, wherein if the two points correspond to a common continuous surface, the two surface clusters are combined into one object cluster.

This brings about the technical advantage, for example, that surfaces of an object can be efficiently assigned to it.

One embodiment provides that the pulse characteristic includes a start time and an end time, with the end time of one point being compared with the start time of the other point for two directly adjacent points in the point cloud, with the spatial position information being determined based on the comparison.

This brings about the technical advantage, for example, that the position information can be determined efficiently. In particular, the orientation of the surface relative to the transit time sensor can be determined efficiently in this way.

In one embodiment it is provided that movement information of the object associated with the corresponding area is determined based on the determined spatial position information of the corresponding area.

This brings about the technical advantage, for example, that the movement information can be determined efficiently. The movement information includes, for example, a speed of the object and/or a direction of movement of the object.

One embodiment provides that a movement of the object is predicted based on the movement information determined.

This brings about the technical advantage, for example, that the movement of the object can be efficiently predicted.

In one embodiment it is provided that a determined surface cluster the determined Spatial position information is assigned from one of the points of the area cluster.

This brings about the technical advantage, for example, that the area cluster and thus also the associated object can be efficiently tracked based on the assigned spatial position information. Furthermore, this brings about the technical advantage that the area cluster can be efficiently distinguished from another area cluster.

Technical functionalities of the device according to the second aspect and/or of the transit time sensor according to the third aspect result analogously from corresponding technical functionalities of the method according to the first aspect and vice versa. Device features and/or features of the transit time sensor result in particular from corresponding method features and vice versa.

In one embodiment, the method according to the first aspect is a computer-implemented method.

In one embodiment, the method according to the first aspect is carried out by means of the device according to the second aspect and/or by means of the transit time sensor according to the third aspect.

In one embodiment, the method according to the first aspect includes carrying out a runtime measurement using the runtime sensor. According to one embodiment, carrying out the transit time measurement includes sending out one or more pulses. According to one embodiment, carrying out the transit time measurement includes detecting one or more echo pulses scattered, in particular reflected, from one or more surfaces of one or more objects in the vicinity of the transit time sensor.

In one embodiment it is provided that the point cloud is provided with the determined area cluster or with the determined object cluster.

A surface within the meaning of the description is a boundary surface of an object, in particular a three-dimensional object. Faces of an object thus delimit the object. The object is therefore delimited by the surfaces, which can also be called side surfaces. The faces of an object form a surface of the object.

The abbreviation "at least one" stands for "one or more".

The abbreviation "bzw." stands for "respectively". The term “or” stands for “respective”, which stands for “and/or” in particular.

• 1 ein Ablaufdiagramm eines Verfahrens zum Clusteranalysieren einer eine Umgebung eines Laufzeitsensors repräsentierende Punktwolke,
• 2 eine Vorrichtung,
• 3 einen Laufzeitsensor,
• 4 ein maschinenlesbares Speichermedium,
• 5 eine Fläche in einer Umgebung eines LiDAR-Sensors,
• 6 zwei Flächen in einer Umgebung eines LiDAR-Sensors,
• 7 - 9 jeweils ein Objekt in einer Umgebung eines LiDAR-Sensors in unterschiedlichen Ausrichtungen zum LiDAR-Sensor,
• 10 eine durchgehende Fläche in einer Umgebung eines LiDAR-Sensors,
• 11 zwei nicht miteinander verbundene Flächen in einer Umgebung eines LiDAR-Sensors und
• 12 ein Kraftfahrzeug.

Embodiments of the invention are shown in the drawings and explained in more detail in the following description. Show it:

• 1 a flow chart of a method for cluster analysis of a point cloud representing an environment of a transit time sensor,
• 2 a device
• 3 a transit time sensor,
• 4 a machine-readable storage medium,
• 5 an area in an environment of a LiDAR sensor,
• 6 two surfaces in an environment of a LiDAR sensor,
• 7 - 9 one object each in an environment of a LiDAR sensor in different orientations to the LiDAR sensor,
• 10 a continuous surface in an environment of a LiDAR sensor,
• 11 two unconnected surfaces in an environment of a LiDAR sensor and
• 12 a motor vehicle.

The same reference symbols can be used below for the same features.

• Empfangen 101 von Punktwolkesignalen, welche eine mehrere Punkte umfassende Punktwolke repräsentieren, welche eine mittels des Laufzeitsensors erfasste Umgebung des Laufzeitsensors repräsentiert, wobei den mehreren Punkten jeweils eine Pulscharakteristik von mittels des Laufzeitsensors detektierten von einer oder mehreren jeweiligen Flächen von Objekten in der Umgebung des Laufzeitsensors gestreuten Echopulsen zugeordnet ist, basierend auf welchen die mehreren Punkte ermittelt wurden,
• für zumindest einen Punkt der Punktwolke Ermitteln 103 einer Raumlageinformation der dem zumindest einen Punkt entsprechende Fläche basierend auf der Pulscharakteristik des Echopulses des zumindest einen Punktes.

1 shows a flowchart of a method for cluster analysis of a point cloud representing an environment of a time-of-flight sensor, comprising the following steps:

• Receiving 101 point cloud signals, which represent a point cloud comprising a number of points, which represents an environment of the time sensor detected by means of the time sensor, with the number of points each having a pulse characteristic of one or more respective surfaces of objects in the area around the time sensor, which are detected by the time sensor and scattered is associated with echo pulses, based on which the multiple points were determined,
• for at least one point of the point cloud determining 103 spatial position information of the area corresponding to the at least one point based on the pulse characteristic of the echo pulse of the at least one point.

2 shows a device 201 which is set up to carry out all steps of the method according to the first aspect.

3 shows a transit time sensor 301 comprising the device 201 according to the second aspect.

In one embodiment, the time-of-flight sensor is one of the following sensors: LiDAR sensor, radar sensor, and ultrasonic sensor.

4 shows a machine-readable storage medium 401 on which a computer program 403 is stored.

The computer program 403 includes instructions which, when the computer program 403 is executed by a computer, for example by the device 201 and/or by the transit time sensor 301, cause this to carry out a method according to the first aspect.

5 shows a transit time sensor 501. There is an object 503 in the vicinity of the transit time sensor 501. The object 503 is aligned obliquely with respect to the transit time sensor 501.

Exemplary propagation directions of pulses, which are transmitted by means of the propagation time sensor 501, are represented symbolically by several lines with the reference number 505.

A point cloud 510 , which represents the surroundings of the transit time sensor 501 , is determined or generated by corresponding transit time measurements using the transit time sensor 501 .

A first point 507 and a second point 509, which are assigned to the object 503, are drawn as an example for this point cloud 510.

A radial distance between the first point 507 and the second point 509 is identified by a double arrow with the reference number 511 .

6 shows a situation that is essentially analogous to that in 5 facts shown.

As a difference, a first object 601 and a second object 603 are provided here in the vicinity of the transit time sensor 501 . Both objects 601, 603 are aligned frontally to the travel time sensor 501.

The corresponding point cloud is in 6 denoted by reference numeral 604.

For example, the point cloud 604 includes a first point 605 and a second point 607.

The first point 605 is associated with the first object 601 . The second point 607 is associated with the second object 603 .

A radial distance between the first point 605 and the second point 607 is marked with a double arrow with the reference number 609 .

The radial distance 511 between the first point 507 and the second point 509 according to FIG 5 is equal to the radial distance 609 between the first point 605 and the second point 607, respectively 6 .

Known cluster analysis methods provide for comparing the determined radial distance with a threshold value, with the known cluster analysis methods providing in particular for specifying that the corresponding points belong to the same object if the corresponding radial distance is less than or less than or equal to the threshold value.

But there in the 5 and 6 the corresponding radial distance between the corresponding points is the same, the known cluster analysis methods would not be able to distinguish that in 5 the two points 507, 509 belong to a common object 503, whereas the two points 605, 607 belong to different objects 601, 603.

According to the concept described here, however, such a distinction is made possible in an advantageous manner, which will be explained in more detail below.

7 shows an object 700 comprising a surface 701, which is aligned frontally to a time-of-flight sensor 501.

The surface 701 is divided into four surface sections: a first surface section 703, a second surface section 705, a third surface section 707 and a fourth surface section 709.

A first point 711 is drawn in the first surface section 703 . A second point 713 is drawn in the second surface section 705 . A third point 715 is drawn in the third surface section 707 . A fourth point 717 is drawn in the fourth surface section 709 .

On the one hand, the four points represent a location at which a pulse emitted by the transit time sensor 501 strikes and is reflected by it back in the direction of the transit time sensor 501. These four points can therefore correspond in particular to points in a point cloud.

The four points 711, 713, 715, 717 have a spatial extension, which is intended to symbolize that the transmitted pulses also have a spatial extension.

The four points 711, 713, 715, 717 are each directly adjacent points. Directly adjacent points denote points between which there is no further point.

In 7 a time axis 719 is also drawn in, on which the respective echo pulses reflected by the four points 711, 713, 715, 717 are drawn.

In this respect, a first echo pulse 721, which is reflected by the first point 711, is drawn on the time axis 719. A second echo pulse 723 , which is reflected by the second point 713 , is plotted on the time axis 719 . A third echo pulse 725, which is reflected by the third point 715, is plotted on the time axis 719. A fourth echo pulse 727, which is reflected by the fourth point 717, is plotted on the time axis 719.

Since the surface 701 is oriented frontally to the transit time sensor 501, a yaw angle and a pitch angle and a roll angle of the surface 701 relative to the transit time sensor 501 are each 0°.

The effect of this is that the four echo pulses 721, 723, 725, 727 have the same pulse characteristics, in particular the same pulse shape, as the pulses emitted by the travel time sensor 501 in the direction of the surface 701.

The four echo pulses 721 , 723 , 725 , 727 thus essentially overlap completely and have a common start time 729 and a common end time 731 .

The concept described here analyzes a respective pulse characteristic of the four echo pulses 721, 723, 725, 727, which in the present example according to 7 has as a result that the pulse characteristics of the four echo pulses 721, 723, 725, 727 correspond to the emitted pulses.

From this information it is determined that the pitch angle, yaw angle and roll angle of surface 701 must all be 0°. In this respect, it is concluded from this that the surface 701 is aligned frontally with respect to the transit time sensor 501 .

A substantial overlapping of the respective echo pulses means in particular an overlapping within the scope of measurement accuracy and/or within the scope of homogeneity of a surface structure of surface 701.

8th shows surface 701 with a yaw angle greater than 0° relative to runtime sensor 501 and with a pitch angle of 0° and with a roll angle of 0° relative to runtime sensor 501.

In this case, this oblique orientation of the surface 701 relative to the transit time sensor 501 leads to a pulse broadening of the corresponding echo pulses 711, 713, 715, 717.

Furthermore, a pulse shape of the four echo pulses 721, 723, 725, 727 is rectangular.

An analysis of the respective pulse characteristics of the four echo pulses 721, 723, 725, 727 determines this, so that it can be determined based on this that the yaw angle is greater than 0° and the pitch angle and the roll angle are 0° relative to transit time sensor 501.

The second echo pulse 723 and the fourth echo pulse 727 have a common first start time 801 and a common first end time 803 . The first echo pulse 721 and the third echo pulse 725 have a common second start time 805 and a common second end time 807 .

The second echo pulse 723 and the fourth echo pulse 727 essentially completely overlap. The first echo pulse 721 and the third echo pulse 725 essentially completely overlap.

At the in 8th The orientation of the surface 701 shown relative to the transit time sensor 501 results in the corresponding orientation in that the first end time 803 corresponds to the second start time 805 .

9 Figure 7 shows face 701 in another orientation. In this case, a yaw angle is greater than 0°, an inclination angle is greater than 0° and a roll angle is equal to 0° in relation to transit time sensor 501.

The corresponding echo pulses 721, 723, 725 and 727 are analogous to those 7 and 8th plotted on the timeline 719.

In this case, the second echo pulse 723 and the third echo pulse 725 essentially completely overlap. A respective pulse shape of the three echo pulses 721, 723, 725 and 727 is triangular.

10 shows a common continuous surface 1001. The surface 1001 is provided with a first point 1003 and a second point 1005, which, analogous to the four points 711, 713, 715 and 717, are intended to identify a respective location at which a time-of-flight sensor 501 sent respective pulse strikes and from there is scattered back to the travel time sensor 501, in particular reflected.

The two points 1005 and 1003 are immediately adjacent to each other.

An echo pulse scattered from the second point 1005 is identified by reference number 1007 and is referred to below as the fifth echo pulse.

An echo pulse scattered from the first point 1003 is identified by reference numeral 1009 and is referred to below as the sixth echo pulse.

These two echo pulses 1007, 1009 are plotted on a time axis 719.

The fifth echo pulse 1007 has a third start time 1011 and has a third end time 1013 . The sixth echo pulse 1009 has a fourth start time 1015 and a fourth end time 1017 . The third end time 1013 corresponds to the fourth start time 1015.

Due to the continuous surface 1001 and due to the fact that the two points 1005 and 1003 are directly adjacent to one another, a respective course or a respective pulse shape of the two echo pulses 1007, 1009 is analogous to that in 8th pulse shape shown is rectangular.

From this information it can be advantageously concluded that the surface 1001 is a continuous surface.

In contrast, the 11 two surfaces separated by a gap: a first surface 1101 and a second surface 1103. The gap between the two surfaces is indicated by reference numeral 1109.

A first point 1105 is assigned to the first surface 1101 . A second point 1107 is assigned to the second area 1103 .

The echo pulse corresponding to the second point 1107 is provided with the reference number 1111 and is referred to below as the seventh echo pulse. The seventh echo pulse has a fifth start time 1115 and a fifth end time 1117 .

An eighth echo pulse corresponding to the second point 1105 is denoted by reference numeral 1113 and has a sixth start time 1119 and a sixth end time 1121 .

The two echo pulses 1111 and 1113 are plotted on a time axis 719.

Due to the fact that the two surfaces 1101 and 1103 are spaced apart from one another by a gap 1109 and that the two points 1105 and 1107 are immediately adjacent to one another, the gap 1109 is mapped as the time interval between the two echo pulses 1111 and 1113.

This means that the fifth end time 1117 is not equal to the sixth start time 1119 and is smaller than this.

A time interval between the fifth end time 1117 and the sixth start time 1119 is therefore provided.

Appropriate analysis of the pulse characteristics of the two echo pulses 1111 and 1113 and analysis of the corresponding start and end times can advantageously be used to determine that the two points 1105 and 1107 are each assigned to areas 1101 and 1103 separated by a gap 1109.

12 shows a motor vehicle 1201.

The motor vehicle 1201 includes the runtime sensor 301 of 3 .

In summary, the concept described here provides a point clustering method based on the shape of the echo of the sampled points. The time-of-flight sensor described here, in particular a LiDAR sensor, not only detects the arrival time of the echoes from the object surface, but also the exact pulse shape of the echoes. The pulse shape advantageously allows the two-dimensional dihedral angle (yaw and pitch) of the detected surface to be identified. In addition, according to one embodiment, the pulse shape, the echo start time and the echo end time of neighboring points are analyzed to determine the exact orientation of the surface. Advantageously, this allows to cluster all points belonging to an area and to identify whether different detected areas are connected to each other or not. The time-of-flight sensor, especially LiDAR sensor, can therefore provide either separate clusters for each face of an object or one cluster for each connected object.

• - Extraktion weiterer Informationen aus den Echos (Pulsformanalyse).
• - Direkte Messung des Winkels und der Orientierung der Objektfläche an jedem Abtastpunkt.
• - Hochleistungs-Clustering-Verfahren mit reduzierter Über- und/oder Unter-Clustering.
• - Geringerer Berechnungsaufwand, da weniger Cluster verfolgt werden müssen.
• - Die Extraktion der Orientierung eines Objekts erlaubt es in vorteilhafter Weise, zukünftige Bewegungen vorherzusagen.

The advantages of the concept are in particular:

• - Extraction of further information from the echoes (pulse shape analysis).
• - Direct measurement of the angle and orientation of the object surface at each scanning point.
• - High performance clustering method with reduced over- and/or under-clustering.
• - Reduced computational effort as fewer clusters need to be tracked.
• - The extraction of the orientation of an object allows, in an advantageous way, to predict future movements.

According to one embodiment, a transit time sensor, in particular a LiDAR sensor, is provided, which analyzes the pulse shape of echo pulses (echoes) that are sent back from object surfaces. According to one embodiment, the transit time sensor has a transmission unit which emits, in particular short, pulses, in particular light pulses, in particular laser pulses. The pulses, in particular light pulses, preferably do not allow any spatial gaps between the various scanning points in the far field. If the pulses, in particular light pulses, are reflected from a surface of an object perpendicular to the transit time sensor, in particular LiDAR sensor, as for example in 7 shown, a short pulse with a pulse shape that is similar to the emitted pulse is detected or recorded by a detector unit of the transit time sensor. If the pulses are scattered, in particular reflected, by an object surface that is oriented at a certain angle (pitch angle or yaw angle greater than 0°) in one dimension relative to the time-of-flight sensor, as in 8th shown, the echo is stretched and a rectangular echo shape is recognized by the detector unit (more precisely, a convolution of a rectangular function with the shape of the emitted pulse, in particular a laser pulse, is measured or recognized).

If the pulses are reflected from a surface that is at a certain angle in two dimensions relative to the time-of-flight sensor, as in 9 shown (pitch angle and yaw angle greater than 0°), the echo is stretched and a triangular echo shape is recognized by the detector unit (more precisely, a convolution of a triangular function with the shape of the emitted pulse, in particular a laser pulse, is recognized). Depending on the exact yaw and pitch, the signal shape can also range between a triangular and rectangular shape.

According to one embodiment, the time-of-flight sensor analyzes the echo shape in order to decide which of the 7 until 9 scenarios shown is present and extracts the two-dimensional angle at which the surface is oriented relative to the time-of-flight sensor.

This can be done, for example, by using different customized filters for the different cases. In particular, the angle of the surface can be determined from the echo shape (pulse shape of the echo pulse) alone, but not the orientation (there are up to 8 possible orientations). For example, from a rectangular echo shape, the angle of the area can be extracted from the width of the rectangle, but in particular it cannot be determined whether this angle is yaw or pitch and in which direction the area corresponding to this yaw or pitch is.

In particular, the echo pulse shape analysis determines the two-dimensional angle of the surface at each sample point, but not the orientation. However, orientation can be determined by comparing the start and end times of the echo pulse with the start and end times of the nearest neighbor echo pulses. For example, it's in 8th not possible to recognize based on an echo pulse whether the surface is tilted to the right or to the left in relation to the plane of the paper. But by comparing the start and end time of the respective echo pulses of the two neighboring points, it can be determined that the end time of the right sampling point corresponds to the start time of the left sampling point. Therefore, the orientation of the inclined surface can be detected.

In addition, based on the matching end and start times, it can be determined that the two scan points belong to the same continuous area (see also explanations for 10 ). If they came from different faces with a gap between faces, as in 11 exemplified, there would be a time gap between the two echo pulses, as in 11 shown.

In one embodiment it is provided that all points originating from a coherent area are combined to form a point cluster, which can be referred to as an area cluster.

In one embodiment, an object cluster is formed or determined, which includes all area clusters that are connected to one another. For example, a truck trailer would have several faces that are oriented at 90° to each other and connected together such that these faces form an object cluster. In contrast, a spherical object would only have one face.

According to one embodiment, the surface clusters are assigned a parameter which indicates the orientation of the surface. In particular, this parameter can be used to simplify the assignment and tracking of the individual area clusters via subsequent run-time measurements. In addition, this precise area detection enables precise measurements of the movements of the object. By tracking the surface orientation of the rear of a motor vehicle, it is possible, for example, to recognize when the motor vehicle initiates a lane change.

QUOTES INCLUDED IN DESCRIPTION

This list of documents cited by the applicant was generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Patent Literature Cited

• US 2018/0299552 A1 [0002]
• EP 1358508 B1 [0003]

Claims (10)

Method for cluster analysis of a point cloud representing an environment of a time-of-flight sensor (501), comprising the following steps: receiving (101) point-cloud signals, which represent a point cloud comprising a plurality of points (711, 713, 715, 709), which represent a point cloud measured by the time-of-flight sensor (501 ) detected environment of the transit time sensor (501), wherein the multiple points (711, 713, 715, 709) each have a pulse characteristic of one or more respective surfaces (701) of objects (700) detected by the transit time sensor (501) in the is assigned to scattered echo pulses in the vicinity of the travel time sensor (501), on the basis of which the multiple points (711, 713, 715, 709) were determined, for at least one point of the point cloud determining (103) spatial position information of the area (701) corresponding to the at least one point based on the pulse characteristic of the echo pulse of the at least one point. procedure after claim 1 , wherein for a plurality of points (711, 713, 715, 709) of the point cloud, a respective spatial position information of the corresponding areas (701) is determined, wherein depending on the respective spatial position information at least one area cluster is determined, to which at least some of the several points (711, 713 , 715, 709). procedure after claim 1 or 2 , where the pulse characteristic comprises a start time (729, 801, 805) and an end time (731, 803, 807), where for two directly adjacent points of the point cloud the end time (731, 803, 807) of one point with the start time (729 , 801, 805) of the other point is compared, based on the comparison determining whether the two points correspond to two different areas (1101, 1103) separated by a gap (1109) or whether the two points belong to a common continuous area (1001 ) correspond. Procedure according to claims 2 and 3 , wherein when two area clusters are determined, the two immediately adjacent points are defined in such a way that one point is assigned to one of the two area clusters and that the other point is assigned to the other of the two area clusters, wherein if the two points of a common continuous Correspond to area that two area clusters are combined into one object cluster. Method according to one of the preceding claims, in which the pulse characteristic comprises a start time (729, 801, 805) and an end time (731, 803, 807), with the end time of one point having the start time (729, 801, 805) of the other point is compared, the position information being determined based on the comparison. Method according to one of the preceding claims, wherein movement information of the object (700) associated with the corresponding area (701) is determined based on the determined spatial position information of the corresponding area (701). Device (201) set up to carry out all steps of the method according to one of the preceding claims. Transit time sensor (301, comprising the device (201) after claim 7 . Computer program (403), comprising instructions which, when the computer program (403) is executed by a computer, cause it to carry out a method according to one of Claims 1 until 6 to execute. Machine-readable storage medium (401) on which the computer program (403) after claim 9 is saved.

Images