Device, method and system for detecting a vegetation object in the vicinity of a vehicle

Abstract

Device (12) for detecting a vegetation object (14) in the environment (16) of a vehicle (18), comprising: an input interface (26) for receiving sensor data from a radar sensor (20) with information about the vehicle's environment, wherein the sensor data includes scan points (32) with associated point velocities; a processing unit (28) for identifying an object in the vicinity of the vehicle based on a clustering of scan points according to the point velocities of the scan points; and an evaluation unit (30) for classifying the object based on an evaluation of the point velocities of the scan points assigned to the object, in order to distinguish between a vegetation object (14) and a living object (22), wherein the evaluation unit (30) is designed to calculate a frequency spectrum of the point velocities of the scan points assigned to the object (32) and to classify the object as a vegetation object (14) if the frequency spectrum does not include pronounced high points; to distinguish between a person, encompassing a pedestrian or a cyclist, and an animal based on the point velocities of the scan points (32) assigned to the object, if the object has been classified as a living object (22); and is trained to recognize typical movement patterns based on a comparison with predefined pattern data; and is trained to recognize a living object (22) that is in front of or behind a vegetation object (14).

Description

[0001] The present invention relates to a device for detecting a vegetation object in the vicinity of a vehicle and a corresponding method. The invention further relates to a system for detecting a vegetation object.

[0002] Modern vehicles (cars, vans, trucks, motorcycles, etc.) incorporate a variety of driver assistance systems that provide the driver with information and control individual vehicle functions semi- or fully automatically. Sensors detect the vehicle's surroundings and other road users. Based on the collected data, a model of the vehicle's environment can be generated. With the ongoing development of autonomous and semi-autonomous vehicles, the influence and scope of driver assistance systems are constantly expanding. The development of increasingly precise sensors makes it possible to perceive the environment and traffic and to control individual vehicle functions fully or partially without driver intervention. Driver assistance systems can thus contribute significantly to increased road safety and improved driving comfort.

[0003] A key requirement here is the detection and recognition of the vehicle's surroundings. Environmental sensors, such as radar, lidar, ultrasonic, and camera sensors, collect sensor data containing information about the environment. Based on this collected data, and potentially also considering data available within the vehicle, objects in the vehicle's vicinity can then be detected. Based on these detected objects, the behavior of an autonomous or semi-autonomous vehicle can, for example, be adapted to the current situation.

[0004] Radar sensors represent an important data source for environmental perception. A radar sensor emits a signal towards a target (object), which is reflected back. The reflection is received by the radar sensor, and its location can be reconstructed. When a dynamic target or object is detected (pedestrian, vehicle, animal, etc.), a frequency change occurs in the carrier frequency of the radar signal (Doppler effect). This frequency change corresponds to the movement of the target (Doppler frequency). Vibrations or rotations of the target can generate additional modulations. This effect is also known as the micro-Doppler effect. Based on the analysis of a micro-Doppler signature of a dynamic object, object recognition may be possible.In this context, for example, Garcia-Rubia and Kilic, “Analysis of Moving Human Micro-Doppler Signature in Forest Environments”, 2014, reveal an approach to detecting and investigating human movements in a forest environment based on a micro-Doppler signature.

[0005] EP 3 349 038 A1 discloses a method for classifying objects near a vehicle, comprising the steps of: outputting radar signals from the vehicle and radar detection from reflections of these signals from objects near the vehicle; based on one or more parameters of the detections, clustering the detections into one or more clusters representing objects near the vehicle; for a cluster, determining a measure of the dispersion of at least one detection parameter; determining a spreading function, wherein the spreading is a function of the dispersion of one or more detection parameters; determining a value of a density function based on the number of detections that lie within the spreading function; and classifying the object based on the density function value.

[0006] US Patent 9229102B1 discloses systems and methods for detecting targets on the opposite side of a wall. In some aspects, the techniques include providing a user with an indication that portions of the reflected radar signals have been blocked by radio frequency blocking material on the wall. In other aspects, the techniques involve identifying candidate targets as multipath echoes or motion-induced errors based on a correlation map.

[0007] US Patent 10,205,457 B1 discloses a target detection and imaging system comprising a radar unit and at least one ultra-low phase noise frequency synthesizer. The radar unit is configured to detect the presence and characteristics of one or more objects in various directions. The radar unit may include a transmitter for sending at least one radio signal and a receiver for receiving the at least one radio signal returned by the one or more objects. The system helps to clearly and promptly detect and categorize people on the road in order to provide timely corrective input to the autonomous vehicle.

[0008] US Patent 5,867,257A discloses a battlefield system for detecting and analyzing vibrations corresponding to an invariant characteristic of a target of interest. The system includes a transmitter for generating a transmitting laser beam by amplifying a primary coherent laser signal. The target of interest can be an enemy soldier. The control and display module enables the transmitter and receiver to operate in an agile search mode, generating spectra that indicate the enemy soldier, and a signature classification mode, analyzing the spectra for the invariant characteristic of the enemy soldier, which the enemy soldier may be located in. Furthermore, a method for operating a battlefield detection system with a micro-Doppler ladar system and a signal processor is described.

[0009] One challenge lies in the radar-based detection of static objects (trees, shrubs, bushes, houses, walls, etc.). If a detected object turns out to be vegetation rather than a wall, for example, an autonomous or semi-autonomous vehicle might react differently in the event of an accident.

[0010] Based on this, the present invention addresses the problem of providing an improved approach for detecting objects in the environment of a vehicle. In particular, vegetation objects are to be detected and correctly classified. Reliable detection is to be enabled.

[0011] To solve this problem, the present invention relates in a first aspect to a device for detecting a vegetation object in the vicinity of a vehicle, comprising: an input interface for receiving sensor data from a radar sensor with information about the vehicle's environment, wherein the sensor data includes scan points with associated point velocities; a processing unit for identifying an object in the vehicle's environment based on a clustering of scan points according to the point velocities of the scan points; and an evaluation unit for classifying the object based on an evaluation of the point velocities of the scan points assigned to the object, in order to distinguish between a vegetation object and a living object.

[0012] Furthermore, one aspect of the invention relates to a system for detecting a vegetation object in the vicinity of a vehicle, comprising: a device as previously described; and a radar sensor to detect the vehicle's surroundings in order to comprehensively provide sensor data scan points with associated point velocities.

[0013] Further aspects of the invention relate to a method designed according to the device and a computer program product with program code for carrying out the steps of the method when the program code is executed on a computer, as well as a storage medium on which a computer program is stored which, when executed on a computer, effects an execution of the method described herein. Preferred embodiments of the invention are described in the dependent claims. It is understood that the features mentioned above and those to be explained below can be used not only in the combinations specified, but also in other combinations or individually, without departing from the scope of the present invention. In particular, the system and the method or the computer program product can be designed according to the embodiments described for the device in the dependent claims.

[0014] According to the invention, sensor data from a radar sensor is received and processed via an input interface. This sensor data includes information about the vehicle's surroundings. In particular, the sensor data includes individual scan points, preferably a plurality of scan points. Each scan point includes a point velocity associated with it. In particular, the data from a radar sensor already present on a vehicle can be used. A point velocity is understood to be, in particular, micro-Doppler information of a scan point, i.e., information that represents a movement or velocity of the object from which the radar signal is reflected. In particular, a micro-Doppler signature, i.e., a plurality of point velocities, is received.

[0015] The received sensor data is first processed to identify an object. This involves clustering based on point velocities. A multiple of scan points are identified as belonging to the same object if they exhibit similar movement. In the next step, the identified object is classified. Classification refers specifically to the identification of the object, i.e., its assignment to a pre-existing category. According to the invention, the classification is performed to distinguish between vegetation and living organisms.

[0016] A vegetation object is understood to be, in particular, a tree, bush, or thicket. A living object is understood to be, in particular, a human being or an animal. The invention is based on the fact that vegetation objects can move under certain circumstances, and this movement produces a characteristic micro-Doppler signature, even if the vegetation object is actually part of a static environment. The position of the vegetation does not change. According to the invention, it is exploited that there are weather conditions that cause vegetation objects to move. For example, the leaves of a tree move in the wind or due to rain.

[0017] Previous approaches have already used micro-Doppler signatures to detect pedestrians or living objects. The present invention relates to an extension of this to the detection of vegetation objects. Compared to the prior art, the approach according to the invention thus allows for purely radar-based identification of objects in the static environment. This enables improved knowledge of the surroundings. Based on the detected vegetation objects, functions of an autonomous or semi-autonomous vehicle, for example, can be implemented. No additional sensors are required, which saves costs.

[0018] In an advantageous embodiment, the evaluation unit for classifying the object is configured based on the number of scan points assigned to the object. Preferably, the evaluation unit for classifying the object as a vegetation object is configured when the number of scan points assigned to the object exceeds a predefined threshold. Classification is based on the size of the object, which is reflected in the number of scan points assigned to the object. A large object comprises a higher number of scan points than a smaller object. Typically, a vegetation object is larger than a living object. Therefore, vegetation objects and living objects can be distinguished based on size. This achieves a simple and efficiently computable evaluation with high differentiation accuracy.

[0019] In a preferred embodiment, the processing unit for identifying the object is designed to evaluate the standard deviation of the point velocities of the scan points assigned to the object. The processing unit exploits the fact that leaves or parts of a vegetation object move with similar point velocities in the wind or rain. This is not typically the case for pedestrians, cyclists, or other living objects. Therefore, if the standard deviation of the point velocities remains below a certain threshold, there is an increased probability that the detected object is a vegetation object. This improves the discrimination accuracy.

[0020] In one embodiment, the evaluation unit is configured to classify the object as a vegetation object if the point velocities of the scan points assigned to the object are below a predefined threshold. Typical movements caused by wind are slow compared to the movements of living objects. Therefore, a distinction can be made based on velocity.

[0021] In one embodiment, the evaluation unit is configured to calculate a signal-to-noise ratio (SNR) for the point velocities of the scan points assigned to the object. A living object, especially a pedestrian, causes higher SNR values ​​than a vegetation object, for example, due to the movement of its arms and legs. Based on this, a ratio between the expected SNR of a living object and a vegetation object can be defined. This results in improved differentiation.

[0022] Preferably, the evaluation unit is designed to classify the object as a vegetation object when the SNR is below a predefined SNR threshold. Preferably, this predefined SNR threshold is based on typical human movements. Comparing the movement with this threshold allows for efficient data analysis.

[0023] According to the invention, the evaluation unit is designed to calculate a frequency spectrum of the point velocities of the scan points assigned to the object. The evaluation unit is designed to classify the object as a vegetation object if the frequency spectrum does not include any pronounced peaks. The distinction based on the frequency spectrum of the point velocities is based on the fact that the leaves of a vegetation object typically move at different speeds. Therefore, different frequencies result in the point velocities of the scan points. This results in a frequency spectrum that does not exhibit any pronounced peaks. In contrast, the frequency spectrum of a pedestrian would typically exhibit one or more pronounced peaks.

[0024] According to the invention, the evaluation unit is designed to distinguish between a human, including a pedestrian or cyclist, and an animal based on the point velocities of the scan points assigned to the object, provided the object has been classified as a living being. The evaluation unit can be configured to recognize typical movement patterns based on a comparison with predefined pattern data. Based on this, it is then possible not only to distinguish between a living being and a vegetation object, but also to further identify a living being. In particular, different living beings can be distinguished based on typical movement patterns. A movement pattern can, for example, be characterized by a preferred frequency. This enables further improved object detection in the environment.

[0025] According to the invention, the evaluation unit is designed to detect a living object located in front of or behind vegetation. Based on an analysis of the frequency spectrum or other approaches, a mixed object can also be detected, i.e., a living object partially obscured by vegetation or vice versa. This detection allows for a further improvement in the assessment of a vehicle's surroundings.

[0026] In one configuration, the evaluation unit is designed to perform a freespace boundary classification. The recognition accuracy can be further improved.

[0027] In a preferred embodiment, the processing unit and / or the evaluation unit is configured to apply a time-based sliding window method. A sliding window method is understood to be a method in which scan points are aggregated and analyzed over a predefined period. This can further improve the accuracy of object identification and classification.

[0028] The vehicle's surroundings include, in particular, the visible area around the vehicle as measured by a radar sensor mounted on the vehicle. A radar sensor can also comprise multiple sensors, enabling, for example, a 360° all-around view and thus capturing a complete image of the surroundings. The sensor data from a radar sensor includes, in particular, distance, point velocity (corresponding to micro-Doppler information), elevation angle, and azimuth angle for various detections. A scan point is understood to be a single point, i.e., a single detection. Typically, a large number of scan points are generated during a radar sensor measurement cycle. A measurement cycle is understood to be a single traverse of the visible spectrum. The sensor data acquired in a measurement cycle can be referred to as a radar target list.Clustering refers to the grouping of scan points. Different scan points are identified as belonging together or to the same object. A predefined threshold can be specified based on previously conducted test series. Alternatively, a predefined threshold can be adjusted during operation.

[0029] The invention is described and explained in more detail below with reference to some selected embodiments in conjunction with the accompanying drawings. These show: Fig. 1 a schematic representation of a vehicle with a system according to the invention; Fig. 2 a schematic representation of a device for detecting a vegetation object according to the present invention; Fig. 3 a representation of the inventive approach for detecting vegetation objects; Fig. 4 an exemplary schematic representation of two frequency spectra for distinguishing between a living object and a vegetation object; Fig. 5 a schematic representation of standard deviations of point velocities; Fig. 6. Another schematic representation of standard deviations; Fig. 7 a schematic representation of a further embodiment of a system according to the invention; and Fig. 8 a schematic representation of a method according to the invention.

[0030] The Fig. Figure 1 shows a schematic representation of a system 10 according to the invention. The system 10 comprises a device 12 for detecting a vegetation object 14 in the environment 16 of a vehicle 18. The system 10 further comprises a radar sensor 20 with which the environment 16 of the vehicle 18 is detected. The environment 16 includes the various objects that can be detected by the radar sensor 20 (trees, houses, traffic signs, other road users, pedestrians, cyclists, animals, etc.). These various objects in the environment 16 reflect the radar waves. For each individual scan point, its position, distance to the radar sensor, and, in particular, a point velocity are recorded. The point velocity corresponds to micro-Doppler information and represents a movement of the target point relative to the radar sensor. Even small movements of an object detected by the radar sensor 20 cause a frequency change in the reflected signal.

[0031] According to the invention, individual objects in the vicinity 16 of the vehicle 18 are identified based on the received point velocities. For each object, it is then determined whether the object is a vegetation object 14 (in particular a tree, bush, shrub, etc.) or a living object 22 (in particular a pedestrian, cyclist, animal, etc.). In the illustrated embodiment, the system 10 is integrated into a vehicle 18 and provides, for example, a driver assistance system with information about the surroundings. It is understood that other arrangements are conceivable in other embodiments. In particular, the device 12 can also be arranged partially or completely outside the vehicle, for example as a cloud service or as a mobile device with a corresponding app.

[0032] The invention exploits the fact that vegetation objects 14 move under certain weather conditions and are thus distinguishable from other static objects, such as a house 24. In windy weather, for example, similar movement vectors of different scan points belonging to a vegetation object are detected within shorter time windows. In particular, movements radial to the sensor are detected in both positive and negative directions.

[0033] In the Fig. Figure 2 schematically illustrates a device 12 according to the invention. The device 12 comprises an input interface 26, a processing unit 28, and an evaluation unit 30. The device 12 according to the invention can, for example, be integrated into a vehicle control unit or implemented as a separate module. It is possible that the device according to the invention is implemented partially or completely in software and / or hardware. The various units can preferably be configured as a processor, a processor module, or software for a processor.

[0034] Input interface 26 receives sensor data from a radar sensor. Input interface 26 can be implemented as a hardware connector, for example. It is also possible for input interface 26 to be a software interface for data reception. The received sensor data includes a point velocity (which can also be called radial velocity) that describes the movement of the detected target. It is understood that the sensor data can include further information. In this respect, the received sensor data includes information about the vehicle's surroundings, insofar as objects in the vehicle's vicinity can be detected based on the sensor data. A radar sensor can, for example, perform four to five hundred or several thousand measurements per single measurement cycle. Typically, several measurement cycles are performed per second.For example, a measurement frequency can be between 16 Hz and 18 Hz. It goes without saying that the sensor data can also be accumulated over a period of time.

[0035] In processing unit 28, the received sensor data is evaluated to identify individual objects in the vehicle's vicinity. To identify these objects, the scan points are clustered based on their point velocities. For example, accurate clustering can be achieved based on the standard deviation of the measured point velocities of a vegetation object. Additional information can also be used for clustering. For instance, gating can be performed around a detected tree trunk.

[0036] In evaluation unit 30, the previously identified objects are classified. This includes evaluating the point velocities of the scan points assigned to each object. The classification distinguishes between vegetation and living objects. In other words, it determines whether the identified (moving) object is vegetation or living. After initial motion detection, the classification can be based, for example, on the fact that vegetation typically occupies a larger area than living objects. A cluster of measurement points from a tree is usually larger than a cluster of measurement points from a pedestrian. If the object's cluster size is comparable to that of a pedestrian or an animal, it will not be recognized as vegetation.

[0037] In the Fig. Figure 3 schematically illustrates the inventive approach of recognizing a vegetation object based on its movements. Wind or rain sets the vegetation object 14 in motion. Individual scan points 32, which are assigned to the vegetation object 14, then exhibit a characteristic movement that differs from the movement of a living object.

[0038] In the Fig. Figure 4 is an example evaluation of a frequency spectrum for distinguishing between a vegetation object (top figure in Fig. 4) and a living object (bottom image in Fig. 4) shown. Typical SNR values ​​for living objects, especially pedestrians and vegetation, can be used for differentiation. The reflective characteristics of a pedestrian exhibit higher SNR values. Based on this, a so-called PVR (Pedestrian-to-Vegetation Ratio) can be defined. Classification accuracy is improved when this PVR value is high.

[0039] During the Fig. The distinction shown in section 4, based on a representation in the frequency domain (which can be determined, for example, using a Fast Fourier transform), is used to evaluate typical movements. Micro-Doppler measurements of a vegetation object usually suggest a characteristic of white noise ( Fig. 4 above). This results in a similar amplitude for different frequencies of the spectrum. There are no pronounced peaks for individual frequencies. In contrast, the frequency domain representation of a micro-Doppler measurement of a pedestrian shows a high amplitude at certain frequencies ( Fig. 4 below). This is because human movement is often regular or periodic, for example through regular arm swinging, etc. Based on the representation in the frequency domain, an accurate distinction can be made between a vegetation object and a living object.

[0040] In the Fig. 5 and Fig. Figure 6 schematically illustrates another approach to distinguishing between a vegetation object and a living object in windy weather. The measured velocities v are plotted on the vertical axis; the horizontal axis t corresponds to a time axis. A distinction is made between the temporal progression of measurement points 34 of a pedestrian as a living object (circles) and measurement points 36 of a vegetation object (crosses). As shown, the standard deviation 38 for the measurement points 36 of the vegetation object is usually smaller than for the measurement points 34 of the living object. If the standard deviation 38 is larger (cf. Fig. 6), the distinction based on this criterion becomes more difficult or impossible.

[0041] It goes without saying that various classification criteria can be combined to arrive at a common value. For example, a probability value for the presence of a vegetation object can be calculated from a combination of several criteria. If this probability value then exceeds a predetermined threshold, a vegetation object is detected. Furthermore, it is possible to differentiate between different living objects. For instance, the movements of a cyclist, a pedestrian, or even a dog exhibit different characteristics, which can be used as a basis for differentiation by evaluating the measured point velocities. For example, the differentiation can be based on an evaluation in the frequency domain. It is also possible to detect living objects located in front of or behind a vegetation object.In essence, an object is identified that encompasses both living and vegetation elements. Such a mixed object can also be referred to as a diffuse target. For example, a so-called freespace boundary classification can be used for its classification.

[0042] In the Fig. Figure 7 schematically illustrates another embodiment of a system 10 according to the invention. The radar sensor 20 is located inside the vehicle 18. The sensor data from the radar sensor 20 are transmitted to the device 12 via a communication unit 40 and evaluated within the device. The device 12 can transmit a classification result back to the vehicle 18. For example, in the illustrated embodiment, the device 12 can be designed as a mobile device, with communication functioning via Bluetooth or another short-range, low-energy communication standard. It is also possible for the device 12 to be offered as an internet service, in which case communication can take place, for example, via the mobile network.

[0043] Finally, in the Fig.Figure 8 schematically illustrates a method according to the invention. The method comprises the steps of receiving S10 sensor data, identifying S12 an object, and classifying S14 the object. The method can, for example, be implemented as software that runs on a vehicle control unit.

[0044] The invention has been comprehensively described and explained with reference to the drawings and the description. The description and explanation are to be understood as examples and not as limiting. The invention is not limited to the disclosed embodiments. Other embodiments or variations will become apparent to a person skilled in the art when using the present invention and upon a detailed analysis of the drawings, the disclosure, and the subsequent claims.

[0045] In the patent claims, the words "comprise" and "with" do not preclude the presence of further elements or steps. The undefined article "a" or "an" does not preclude the presence of multiple elements. A single element or unit can perform the functions of several of the units mentioned in the patent claims. An element, unit, interface, device, and system can be implemented partially or completely in hardware and / or software. The mere mention of some measures in several different dependent patent claims is not to be understood as precluding the advantageous use of a combination of these measures. A computer program can be stored / distributed on a non-volatile data carrier, for example, on optical storage media or on a solid-state drive (SSD).A computer program can be distributed together with hardware and / or as part of hardware, for example via the internet or via wired or wireless communication systems. Reference punctuation in the patent claims is not to be understood as limiting. Reference sign 10 System 12 Device 14 vegetation object 16 Environment 18 vehicles 20 radar sensor 22 living object 24 26 Input interface 28 processing units 30 evaluation units 32 scan points 34 Measuring point pedestrian 36 measuring point vegetation object 38 standard deviation 40 communication units

Claims

[1] Device (12) for detecting a vegetation object (14) in the environment (16) of a vehicle (18), comprising: an input interface (26) for receiving sensor data from a radar sensor (20) with information about the vehicle's environment, wherein the sensor data includes scan points (32) with associated point velocities; a processing unit (28) for identifying an object in the vicinity of the vehicle based on a clustering of scan points according to the point velocities of the scan points; and an evaluation unit (30) for classifying the object based on an evaluation of the point velocities of the scan points assigned to the object, in order to distinguish between a vegetation object (14) and a living object (22), wherein the evaluation unit (30) is designed to calculate a frequency spectrum of the point velocities of the scan points assigned to the object (32) and to classify the object as a vegetation object (14) if the frequency spectrum does not include pronounced high points; to distinguish between a person, encompassing a pedestrian or a cyclist, and an animal based on the point velocities of the scan points (32) assigned to the object, if the object has been classified as a living object (22); and is trained to recognize typical movement patterns based on a comparison with predefined pattern data; and is trained to recognize a living object (22) that is in front of or behind a vegetation object (14). [2] Device (12) according to claim 1, wherein the evaluation unit (30) for classifying the object based on a number the scan points (32) assigned to the object are formed; and the evaluation unit is preferably designed to classify the object as a vegetation object (14) when the number of scan points assigned to the object exceeds a predefined threshold. [3] Device (12) according to one of the preceding claims, wherein the processing unit (28) is configured to identify the object based on an evaluation of a standard deviation (38) of the point velocities of the scan points (32) assigned to the object. [4] Device (12) according to one of the preceding claims, wherein the evaluation unit (30) is designed to classify the object as a vegetation object (14) when the point velocities of the scan points (32) assigned to the object are below a predefined threshold. [5] Device (12) according to one of the preceding claims, wherein the evaluation unit (30) is configured to calculate a signal-to-noise ratio, SNR, for the point velocities of the scan points (32) assigned to the object. [6] Device (12) according to one of the preceding claims, wherein the evaluation unit (30) is configured to classify the object as a vegetation object (14) when the SNR is below a predefined SNR threshold; and The predefined SNR threshold is preferably based on typical human movements. [7] Device (12) according to one of the preceding claims, wherein the evaluation unit (30) is configured to perform a freespace boundary classification. [8] Device (12) according to one of the preceding claims, wherein the processing unit (28) and / or the evaluation unit (30) is configured to apply a time-based sliding window method. [9] System for detecting a vegetation object (14) in an environment (16) of a vehicle (18), comprising: a device (12) according to any one of the preceding claims; and a radar sensor (20) for detecting the vehicle's surroundings in order to provide sensor data comprehensive scan points (32) with associated point velocities. [10] Method for detecting a vegetation object (14) in an environment (16) of a vehicle (18), comprising the steps: Receiving (S10) sensor data from a radar sensor (20) containing information about the vehicle's environment, wherein the sensor data includes scan points (32) with associated point velocities; Identifying (S12) an object in the vicinity of the vehicle based on a clustering of scan points according to the point velocities of the scan points; and Classifying (S14) the object based on an evaluation of the point velocities of the scan points assigned to the object, in order to distinguish between a vegetation object (14) and a living object (22). [11] Computer program product comprising program code for performing the steps of the method according to claim 10 when the program code is executed on a computer.

Images