COMPUTER-IMPLEMENTED PROCESS, ELECTRONIC VEHICLE SYSTEM AND COMPUTER PROGRAM PRODUCT

Abstract

The application relates to a computer-implemented method for object detection in a point cloud (28) generated by means of an active optical sensor system (10), wherein points of the point cloud (28) have a respective intensity signal (30) of a respective received reflection, wherein the method comprises: • Transforming the respective intensity signal (30) so that the optical power of the respective intensity signal (30) can be determined from the transformed (32) and • Performing object detection using the transformed intensity signals (32). The application also concerns an electronic vehicle system and a computer program product.

Description

Technical field

[0001] The application relates to a computer-implemented method for object detection, an electronic vehicle system, and a computer program product. The electronic vehicle system includes an active optical sensor system, e.g., a lidar system, configured to generate a point cloud. background

[0002] Sensor systems for environmental perception can be based on wireless signals such as electromagnetic waves or sound. Accordingly, examples include sound-based ultrasonic sensors and electromagnetic wave-based sensors such as camera sensors, radar sensors, or lidar sensors.

[0003] Lidar technology (Lidar stands for Light Detection and Ranging) is an important sensor principle for environmental perception, based on electromagnetic waves and used in active optical sensor systems. A lidar system has an optical transmitter and an optical receiver. The transmitter can emit optical signals, i.e., light, which can be pulsed. In a lidar system, laser beams in the ultraviolet, visible, or infrared range can be used as the light source. The receiver reflects the transmitted light off a point, such as an object, within a detection area in the vicinity of the lidar system and receives it as received light. Using the transmitted light, the lidar system's evaluation unit can process the received light, for example, using a time-of-flight method.Time-of-Flight (TOF) measurements can be analyzed, and the spatial location and distance of the reflection points can be determined. Furthermore, it is possible to calculate the relative velocity. In this context, reflection or reflected light refers to any light that is thrown back, including light reflected by scattering or absorption-emission.

[0004] Light reflected from the surroundings can be detected in the receiving device by receiving sensors. Lidar system receiving sensors can have multiple receiving elements, called pixels, for opto-electrical conversion. The pixels can be configured to receive optical signals from different angles.

[0005] US2020 / 0072946 describes a lidar device comprising one or more optical elements configured to direct incident light in one or more directions. The lidar device further includes a receiver sensor with a plurality of pixels configured to output detection signals in response to the received light. Corrected image data is generated based on the detection signals and an expected optical point spread function of the lidar device. Overview

[0006] In a point cloud generated by an active optical sensor system, each point in the cloud exhibits a corresponding intensity signal from a received reflection. A computer-implemented method for object detection in such a point cloud exhibits: • Transforming the respective intensity signal so that the optical power of the respective intensity signal can be determined from the transformed signal and • Performing object detection using the transformed intensity signals.

[0007] The points in the point cloud can also be called reflection points, as they represent the locations in the vicinity of the active optical sensor system where the emitted optical signal was reflected. The respective intensity signal is a signal that depends on the intensity of the optical signal reflected at the point in the vicinity of the point cloud to which that intensity signal is assigned. The intensity is directly related to the energy of the received light. The intensity signal can, for example, correspond to the intensity of the received light or be directly derived from it. The point cloud generated by the active optical sensor system can exhibit specific solid angles, distances, and intensity signals of the received reflections for each reflection point.

[0008] The active optical sensor system can, in particular, include a lidar system. The active optical sensor system is configured to send and receive optical signals and to process them. The active optical sensor system comprises an optical transmitter, an optical receiver, and a processing unit. The receiver includes an optical receiver sensor, which can have multiple receiver pixels, each configured to convert an optical signal into an electrical signal. The receiver pixels can also be referred to as pixels. The receiver pixels of the optical receiver sensor contain light-sensitive components that capture light and convert it into an electrical signal, which can then be read and made available for further processing.The light-sensitive components can be photodiodes, charge-coupled devices (CCDs), CMOS sensors, photomultiplier tubes, SPADs or similar.

[0009] When highly reflective objects are present in the vicinity of the active optical sensor system, such as retroreflective objects like traffic signs, problems can occur with the dynamics of the detected optical signal. Pixels receiving optical signals of particularly high intensity become saturated. Saturation means that more light reaches the pixels than they are capable of converting into an electrical signal. Optical signals reflected by such highly reflective objects can exhibit crosstalk to other pixels of the receiving sensor. Crosstalk is an undesirable optical artifact. The received signal of the other pixels is distorted by the crosstalk, leading to unwanted effects in the point cloud that can impair object detection. Crosstalk is also known as blooming.The saturated pixels can also be referred to as the source pixels of the crosstalk.

[0010] In various embodiments, the active optical sensor system can be configured as a scanning active optical sensor system, such as a scanning lidar system or laser scanner. Scanning active optical sensor systems emit optical signals that move in a scanning direction. This scanning motion can be achieved by deflecting the optical signal transmitted by the optical transmitter using a deflection device. The optical signal can be pulsed. To deflect the light, the deflection device can, for example, include one or more rotating mirrors and / or MEMS mirrors. Other deflection options include phased optical arrays or liquid crystal-based systems. Thus, the scanning motion allows information to be acquired about points that have different solid angles relative to the active optical sensor system.

[0011] A point cloud can be understood as a set of points, each with corresponding coordinates in a coordinate system, particularly a three-dimensional one. Two-dimensional point clouds are also conceivable. In the case of a three-dimensional point cloud, the three-dimensional coordinates can be determined, for example, by the direction of incidence (i.e., the solid angle) of a light ray reflected at a reflection point and the corresponding travel time or radial distance measured for that point. The three-dimensional coordinate system of the point cloud can be a three-dimensional polar coordinate system. However, the information can also be specified in Cartesian coordinates for each point. A conversion can then be performed to determine the solid angle. In addition to the spatial information, namely, for example, theIn addition to three-dimensional coordinates, which include the distance to the point, and the intensity signal, the point cloud can also contain additional information or measurement data for the individual points, such as color values ​​and / or gray values ​​of the received light and / or time information.

[0012] Object detection in the point cloud can be performed by an algorithm. Objects detected in this way can be important for use by other algorithms, for example, those used for autonomous or semi-autonomous driving, since the free space of the self-driving vehicle depends on the detected objects.

[0013] Using the described method, the point cloud can be processed in such a way that the effects of crosstalk can be mitigated and object detection using the point cloud can be improved. The processing includes filtering that is suitable both for removing or at least reducing crosstalk and for detecting real objects in the point cloud that are obscured by crosstalk.

[0014] By appropriately transforming the respective intensity signal, the optical power of that signal can be determined from the transformed signal. The effect of saturation does not occur with respect to the optical power after this transformation. Therefore, the optical power can be used in subsequent processing steps to perform the described filtering and to extract further information from the point cloud. Thus, by transforming the intensity signal, more information is retained, which can then be used for subsequent object detection.

[0015] In one embodiment of the method, the optical power of the intensity signal is linearly mapped by the transformation. This means that the transformed values ​​are determined from the optical power of the intensity signal via a linear mapping. The linearity of the mapping avoids saturation and preserves more information about how much light is actually received by the receiving pixel. The dynamics of the received optical signal can thus be better represented.

[0016] In one embodiment of the method, the intensity signal is transformed using a surface under the intensity signal. This surface can be determined, in particular, by integration over time. The surface corresponds to a measure of the received optical energy during the time period over which the integration was performed.

[0017] In one embodiment of the method, the area under the intensity signal is determined over a signal length of the received reflection. When using pulsed light as the emitted optical signal, the transformation can be performed, for example, by integrating the received light pulse of the optical signal. Here, integration can be performed, for example, over the entire duration of the received light pulse.

[0018] In one embodiment of the method, the area under the intensity signal is determined with respect to a sliding window function. For this transformation, the intensity signal can, for example, be integrated over the duration of the sliding window.

[0019] In one embodiment of the method, the object detection further features: • Identifying at least one point in the point cloud that is the source point of crosstalk, • Estimating the optical power of at least one source point using the respective transformed intensity signal, • Determining values ​​of the expected optical crosstalk of at least one source point within the point cloud using a point spread function of the active optical sensor system, • Comparison of the expected optical crosstalk values ​​of at least one source point with the respective values ​​received by the active optical sensor system, • Detecting objects using the result of the comparison.

[0020] Determining the source points, which are associated with retroreflective objects, for example, can be done in various ways. For instance, the intensity signal can be analyzed to identify patterns or anomalies that indicate crosstalk. The analysis can also be performed in the frequency domain and optionally, for example, using filters.

[0021] In particular, for the identified source points, the received optical power is then estimated using the transform.

[0022] Determining the expected optical crosstalk values ​​of at least one source point within the point cloud is done using a point spread function of the active optical sensor system. For this purpose, the expected value of the optical crosstalk for points in the point cloud originating from the source point is determined. This expected value is calculated using the point spread function of the active optical sensor system. The point spread function can be determined, for example, during a calibration phase of the active optical sensor system.

[0023] In embodiments, the values ​​of the expected optical crosstalk of the at least one source point are determined by convolution of the transformed intensity signal of the source point with the point spread function of the active optical sensor system. This yields the expected value of the intensity signal of the expected optical crosstalk of the source point to other points in the point cloud. The expected values ​​of the optical crosstalk of the at least one source point to the other points depend on the optical power.

[0024] The expected optical crosstalk values ​​of at least one source point are then compared with the respective values ​​of the point cloud generated by the active optical sensor system. The expected values ​​were determined, for example, in the previous step. The point cloud values ​​are stored, for example, in memory.

[0025] In embodiments, the value of the expected intensity signal of the optical crosstalk from at least one source point to other points in the point cloud is compared with the respective values ​​of the intensity signal of the other points present in the point cloud generated by the active optical sensor system. The values ​​of the points stored in the point cloud depend on the optical power.

[0026] Object detection is then performed using the result of the comparison. For example, large deviations from the expected value may indicate a real object located within the crosstalk area. Small deviations from the expected value may indicate crosstalk without obscuring a real object near at least one source point. In particular, an object can be detected if the comparison reveals a difference above a predefined threshold.

[0027] Following this, filtering can be performed to reduce crosstalk in the point cloud or to remove crosstalk from the point cloud.

[0028] In one embodiment of the method, the active optical sensor system comprises receiving pixels configured to generate an intensity signal for each received reflection as an electrical receiving signal. The electrical receiving signal corresponds to the optical intensity signal. The transform of the intensity signal also maps intensities in the saturation range of at least one receiving pixel to the optical power of the intensity signal in such a way that the optical power of the intensity signal in the saturation range can be determined from the transform. The transform, therefore, does not saturate but rather maps to the optical power, which is preferably one-to-one or approximate.

[0029] In embodiments of the method, at least one source point is assigned to a respective receiving pixel in the saturation region. This means that the pixel of a source point receives so much light that it reaches saturation.

[0030] In embodiments, at least one receiving pixel of the optical receiving sensor includes at least one SPAD. A SPAD (Single Photon Avalanche Diode) is a highly sensitive light sensor capable of detecting single photons. This type of detector utilizes the avalanche breakdown mechanism in a semiconductor material to generate a measurable electrical charge in response to the reception of a single photon. This enables SPADs to detect very small amounts of light. Due to their high sensitivity, SPADs are also susceptible to crosstalk. Therefore, the described method is particularly advantageous for receiving pixels with SPADs. Alternatively, the receiving pixel could also include a photodiode, a charge-coupled device (CCD), CMOS sensors, photomultiplier tubes, or similar devices instead of the SPAD.

[0031] In a method for at least partially automated vehicle control, a point cloud is generated using an active optical sensor system in the vehicle. This point cloud contains an intensity signal corresponding to a received reflection for each point in the vehicle's environment. The received reflection is the reflection of the optical signal emitted by the active optical sensor system at a point in the environment. Optionally, the point cloud also includes the respective solid angles and distances for the reflection points.

[0032] The described computer-implemented object detection method is carried out by means of at least one processing unit in the vehicle, and the vehicle is guided at least partially automatically depending on the result of the object detection. The processing unit can be designed separately from the sensor system and connected to it via a data line, or it can be integrated with the sensor system into a common unit.

[0033] An electronic vehicle system comprises a processing unit and an active optical sensor system. The active optical sensor system is configured to generate a point cloud, where the point cloud contains a corresponding intensity signal for each point in the vehicle's environment, representing a received reflection. Optionally, the point cloud also includes the corresponding solid angles and distances for the reflection points. The processing unit is configured to transform the respective intensity signals so that the optical power of each intensity signal can be determined from the transformed signal. Furthermore, the processing unit is configured to perform object detection using the transformed intensity signals.

[0034] In one embodiment, the vehicle system is configured to steer the vehicle at least partially automatically, depending on the result of object detection. The detected objects are relevant for this at least partially automatic vehicle control because they represent, for example, an obstacle and / or another road user.

[0035] A computer program product contains instructions which, when executed by a computing unit, cause the computing unit to carry out the described computer-implemented procedure, or, when executed by the described electronic vehicle system, cause the electronic vehicle system to carry out the described procedure for at least partially automatic driving of the vehicle. List of characters

[0036] The following section provides further explanation and description of exemplary implementations of this application with reference to the figures. They show Fig. 1 schematically a vehicle with an active optical sensor system and a computing unit, Fig. 2 optical intensity signals, Fig. 3 + Fig. 4 transformed intensity signals, Fig. 5 a transformed intensity signal and a signal of optical power, Fig. 6 schematically a folding using a point spreading function, Fig. 7 a method for object detection.

[0037] The same reference symbols are used in the figures for identical or similar elements. Representations in the figures may not be to scale. Character description

[0038] Fig. Figure 1 schematically shows a vehicle 20, for example a passenger car. The vehicle 20 has an electronic vehicle system which includes an active optical sensor system 10, e.g. a lidar system, and a processing unit 24. The optical sensor system 10 is located in a front area of ​​the vehicle 20 and the environment 22 it detects is located in front of the vehicle 20 in the direction of travel.

[0039] The active optical sensor system 10 comprises an optical transmitter 12, an optical receiver 14, an optical deflector 16, and an evaluation unit 18. The evaluation unit 18 can include a processor, an FPGA, or similar device for processing data.

[0040] The optical transmitter 12 emits an optical signal L. It has a light source for emitting, for example, laser light. The optical receiver 14 receives the optical signal L reflected in the surroundings 22.

[0041] Optionally, the optical transmitter 12 can transmit the optical signal L in pulses. The pulsed optical signal L has short periods during which the optical signal L is transmitted. This can be referred to as a pulse. Between the pulses, the optical transmitter 12 does not transmit any optical signal L. The reflections of the pulses in the environment 22 are then received by the optical receiver 14.

[0042] The optical deflection device 16 is configured to deflect the optical signal L transmitted by the optical transmitter 12 into the environment 22 and to deflect the optical signal L reflected from the environment 22 to the optical receiver 14. The deflection device 16 can be controlled such that the optical signal L performs a scanning motion 26 across the environment 22. For example, the deflection device 16 can include a rotating mirror device that performs a rotational movement to deflect the optical signal L such that the scanning motion 26 is carried out by the optical signal L. During the rotational movement, the angular position of the deflection device 16 is changed.

[0043] The optical receiving device 14 comprises an optoelectronic receiving sensor, also called a detector. The receiving sensor can, for example, have point-shaped sensors, so-called pixels or receiving pixels, which can be arranged in rows or areas. A pixel can, for example, have one or more avalanche photodiodes (APDs) or one or more single-photon avalanche diodes (SPADs). The optoelectronic detector can receive light, in particular the optical signal L, and convert it into electrical receiving signals. The electrical signals can be processed by the evaluation device 18.

[0044] SPADs can also be triggered by very small amounts of incident light, especially single incident photons. The intensity signal, particularly its amplitude, can depend, for example, on the number of triggered SPADs within a trigger cycle (i.e., during a received light pulse) for pixels containing SPADs. Reflections from highly reflective surfaces in the environment result in a large amount of reflected light, and the number of triggered SPADs can quickly become limited in terms of amplitude, thus failing to adequately represent the measurement dynamics. The described dynamics at highly reflective surfaces, especially retroreflective objects, can lead to signal crosstalk to other pixels. This crosstalk can also be referred to as blooming.

[0045] The problems of insufficient dynamics and crosstalk can be addressed, particularly for receiving sensors with SPADs, by the method described below.

[0046] The evaluation unit 18 is configured to control the transmission of the optical signal L as a function of the angular position of the deflection device 16. The evaluation unit 18 is further configured to evaluate the transmitted and received optical signal L. The evaluation data generated in this way can be transmitted from the evaluation unit 18 to the processing unit 24 of the electronic vehicle system of the vehicle 20.

[0047] A point cloud 28 can be generated from the evaluation data produced by the analysis. Information about each point in the point cloud 28 is provided, depending on the evaluation data. In particular, the points in the point cloud 28 contain information about the intensity of the received optical signal L, as well as the spatial location and distance of the reflection points in the environment. The reflection points are those points where the reflection of the optical signal L emitted and received by the sensor system 10 occurred.

[0048] The point cloud 28 can be used, for example, to detect objects in the environment 22, to determine the distance to such objects, and / or to perform further evaluations. The point cloud 28 can be generated, for example, in the optical sensor system 10 from the evaluation data, or the point cloud 28 can be generated, for example, in the processing unit 24 from the raw data of the optical sensor system 10.

[0049] The computing unit 24 can, for example, be configured as the central vehicle computer of the vehicle 20, in which data from several sensor systems of the vehicle 20 can be received, evaluated and / or further processed. The computing unit 24 can, for example, be used to implement autonomous or semi-autonomous driving functions.

[0050] The optical sensor system 10 can, for example, be mounted or integrated at the front of the vehicle 20. Optical sensor systems 10 are also possible for other parts of the vehicle 20, e.g., for surround-view functions, such as on the sides and / or rear of the vehicle 20. It is also possible to arrange optical sensor systems 10, such as lidar systems, and / or other sensor systems such as radar, ultrasound, etc., on the vehicle 20, particularly in corner areas of the vehicle 20.

[0051] The optical sensor system 10 can be used to detect stationary or moving objects in the environment 22. Such objects can include things such as vehicles, people, animals, plants, obstacles, road surface irregularities, especially potholes or stones, road boundaries, traffic signs, open areas, especially parking lots, precipitation, or the like.

[0052] Fig. Figure 2 shows example intensity signals from 30 different pixels. The intensity signals represent the number of triggered SPADs over time as their amplitude. The x-axis is the time axis. The time axis is divided into time intervals, or bins, during which the triggered SPADs are counted and plotted in the y-direction.

[0053] The SPADs of the receiver sensor are triggered by the incident optical signal L, which can also be described as a triggering process. Examples are shown in... Fig. 2. The triggered SPADs per pixel are plotted as amplitude over a so-called trigger cycle, i.e., over the duration of a received light pulse of the optical signal L. Two pulses are received as reflections of the emitted optical signal L.

[0054] In Fig. Figure 2 shows that in the left received pulse, some of the pixels are saturating, meaning that too much light is being received for the amplitude to be represented by the number of triggered SPADs. Therefore, the measurement dynamics of the respective pixels in saturation are not sufficient. In the left pulse of Fig. 2. This refers to the reflection from a highly reflective object, which drives some pixels into saturation.

[0055] The right received impulse from Fig. 2 is not in saturation. It could be a reflection from a normal object located near the highly reflective reflection point.

[0056] When processing point cloud 28, it is advantageous to distinguish the artifact crosstalk from the real object. This application takes advantage of the fact that the dynamics of the intensity signal can be mapped as a function of the area under the intensity signal. For example, the integrated height of the intensity signal 30 over its full signal length or its pulse length can be used as a function of the area.

[0057] For example, a sliding, integrating window (FIR filter, finite impulse response filter) can be used to map the effect as a function of the area. This window can then correspond, for example, to the expected total length of a pulse. These dimensions must be related to their relative optical power according to their length.

[0058] Fig. Figure 3 shows the transformed 32 of the intensity signals 30 of Fig. 2. The intensity signals 30 of Fig. 2 were transformed with an FIR filter of height 1 and with 20 time bins, i.e. a width of 20 bins, i.e. integrated over the time window of 20 time units.

[0059] Fig. Figure 4 shows the relationship between relative optical power and the maximum amplitude of intensity signals 30 without transformation (white dots) and those transformed with a 10-bin FIR filter (black dots) and with a 20-bin FIR filter (transformed 32). The transformation with the 20-bin FIR filter maps the optical power 34 up to the maximum amplitude in such a way that the amplitude can be reconstructed from the optical power 34.

[0060] Fig. Figure 5 shows a transformed intensity signal 32 on the left and a signal 34 of the optical power on the right. The power signal 34 can be calculated from the transformed signal 32.

[0061] Within the point cloud 28, highly reflective points can be identified whose associated pixels become saturated upon reception and which can generate strong crosstalk as source points. This identification can be estimated, for example, using the magnitude of the intensity signal 30, i.e., its maximum amplitude. Additionally or alternatively, other methods can also be used to identify highly reflective points, i.e., potential source points. For example, ghost points, which are intrinsically generated by the high intensities in the evaluation unit 18, e.g., in a readout chip of the receiving sensor in the evaluation unit 18, can also be used to identify the source points.

[0062] Fig. Figure 6 schematically shows the convolution of an optical signal L reflected from a highly reflective point, a source point, with a point spread function 36 of the optical system of the active optical sensor system 10. The optical signal L from the source point is shown on the left. The point spread function 36 is shown on the right. The convolution of the two functions is shown below. The point spread function 36 can depend on, for example, environmental conditions such as rain, fog, etc., and can be selected accordingly.

[0063] The folding is performed along the readout direction of the receiving sensor. The folding results in an expected mathematical description of the crosstalk to other points in point cloud 28 and is shown in the lower part of Fig. Figure 6 is shown. This can now be compared in the next step with the values ​​determined by the active optical sensor system 10 for the other points. The values ​​determined by the active optical sensor system 10 are shown in Fig. Figure 6 is also shown. The range in which expected values ​​and measured values ​​differ is shown in Fig. 6 circled below.

[0064] The comparison between the expected crosstalk and the measured point cloud 28 can be used to distinguish false positives (apparent signals) and true positive points (real objects) in the crosstalk and to filter accordingly if necessary.

[0065] For comparison, the transformed value 32 or a signal dependent on it, such as the optical power 34, is used. For this purpose, the transformed value 32 and / or the signal dependent on it, such as the optical power 34, can be stored in the point cloud 28.

[0066] Optionally, it is also possible to calculate the values ​​back to the raw point cloud and perform the comparison on this basis, which, however, requires more computation.

[0067] Elevated values ​​relative to the expected crosstalk suggest real objects, while values ​​close to the expected value suggest simple crosstalk.

[0068] Optionally, a threshold can be specified, which can be considered a confidence interval. The measured difference is then compared to the threshold. If the difference is above the threshold, a real object is present. If the difference is below the threshold, it is crosstalk.

[0069] In the Fig. In the comparison shown below, a less reflective object is placed next to a highly reflective object.

[0070] The next step is filtering. This allows, for example, the filtering out of false positive signals caused by crosstalk. Points in the 3D point cloud 28 that originate from real objects at that point are retained. Filtering can be achieved, for example, by calculating a simple difference in the optical signal strength. The retained points in the point cloud can optionally be converted to a desired scale suitable for further processing.

[0071] In Fig. Figure 7 is a schematic representation of an embodiment of the described method for object detection.

[0072] The method can be advantageously applied, for example, to receiving sensors that have pixels with SPADs. The method can also be applied, for example, to a lidar system that operates as a laser scanner. The laser scanner can be configured, for example, to rasterize vertical laser lines along the horizontal direction. Crosstalk can be filtered out during the post-processing of the point cloud 28. This post-processing can take place, for example, in the evaluation unit 18 and / or in the processing unit 24.

[0073] The method makes it possible to filter the points in such a way that both points that only arise from crosstalk are filtered out, and real objects in the point cloud 28 that are superimposed by crosstalk can be detected.

[0074] The procedure indicates: 700: Intensity signals 30 of the points in point cloud 28 are converted into a measure that exhibits a dynamic range allowing even high optical power levels to be represented. This conversion can also be referred to as a transformation. The converted values ​​are based, for example, on a linear mapping of the optical power of the intensity signals.

[0075] Object detection is then performed in 702-710: 702: Identifying source points. Source points are those points that are associated with highly reflective, e.g., retroreflective, objects. 704: Determination of the optical power of the source points. 706: Convolution of the optical signals L of the source points with the point spreading function 36 of the optical receiver system of the active optical sensor system 10. From this, the effect of the source points on other points of the point cloud 28 is obtained. 708: Comparison of the result from 706 with the optical power signal 34 of the other points, which was determined from the transformation to 700. The deviation of the result from 706 with the signal determined as a function of 700 is classified and / or assessed in relation to a predefined threshold. The following assessments are possible: a. High deviations indicate a real object that lies within the crosstalk range. b. Low deviations indicate the crosstalk artifact 710: Based on the assessment from 708, points of the point cloud 28 can be filtered to remove crosstalk and obtain real objects in the point cloud 28. QUOTES INCLUDED IN THE DESCRIPTION

[0000] This list of documents cited by the applicant was automatically generated and is included solely for the reader's convenience. The list is not part of the German patent or utility model application. The DPMA accepts no liability for any errors or omissions. Cited patent literature

[0000] US 2020 / 0072946

[0005]

Claims

[1] Computer-implemented method for object detection in a point cloud (28) generated by means of an active optical sensor system (10), wherein points of the point cloud (28) exhibit a respective intensity signal (30) of a respective received reflection, wherein the method comprises: Transforming the respective intensity signal (30) such that the optical power of the respective intensity signal (30) can be determined from the transformed (32) and Performing object detection using the transformed intensity signals (32). [2] Method according to claim 1, wherein the optical power (34) of the intensity signal (30) is linearly mapped by the transformation (32). [3] Method according to claim 1 or 2, wherein the transformation of the intensity signal (30) is carried out using an area under the intensity signal (30). [4] Method according to claim 3, wherein the area under the intensity signal (30) is determined over a signal length of the received reflection. [5] Method according to claim 3, wherein the area under the intensity signal (30) is determined with respect to a sliding window function. [6] Method according to any of the preceding claims, wherein the object detection further comprises: Determining at least one point of the point cloud (28) which is the source point of crosstalk, Estimating the optical power (34) of at least one source point using the respective transformed intensity signal (32), Determining values ​​of the expected optical crosstalk of at least one source point within the point cloud (28) using a point spreading function (36) of the active optical sensor system (10), Comparison of the values ​​of the expected optical crosstalk of at least one source point with the respective values ​​received by the active optical sensor system (10), Detecting objects using the result of the comparison. [7] Method according to claim 6, wherein the values ​​used in the comparison depend on the respective optical power (34). [8] Method according to claim 6 or 7, wherein an object is detected when a difference is found in the comparison which is above a predefinable threshold. [9] Method according to one of the preceding claims, wherein the active optical sensor system (10) has receiving pixels which are configured to generate an intensity signal (30) for each received reflection and wherein the transform (32) also maps intensities in the saturation range of at least one receiving pixel to the optical power (34) of the intensity signal (30) in such a way that the optical power (34) of the intensity signal (30) in the saturation range can be determined from the transform (32). [10] Method according to claim 9, wherein the at least one source point is assigned to a respective receiving pixel in the saturation area. [11] Method for at least partially automatic driving of a vehicle (20), wherein a point cloud (28) is generated by means of an active optical sensor system (10) of the vehicle (20), wherein the point cloud (28) has a respective intensity signal (30) of a respective received reflection for points in an environment (22) of the vehicle (20), wherein a computer-implemented object detection method according to one of the preceding claims is carried out by means of at least one computing unit (24) of the vehicle (20), and the vehicle (20) is guided at least partially automatically depending on a result of object detection. [12] Electronic vehicle system comprising a computing unit (24) and an active optical sensor system (10) configured to generate a point cloud (28), wherein the point cloud (28) has a respective intensity signal (30) of a respective received reflection for points in an environment of the vehicle (20), wherein the computing unit (24) is configured to transform the respective intensity signals (30) in such a way that the optical power (34) of the respective intensity signal (30) can be determined from the transform (32) and object detection can be carried out using the transformed intensity signals (32). [13] Vehicle system according to claim 12, wherein the vehicle system is configured to guide the vehicle (20) at least partially automatically depending on a result of object detection. [14] Computer program product with commands that when executed by a computing unit (24), cause the computing unit (24) to carry out a computer-implemented method according to one of claims 1 to 10, or when performed by an electronic vehicle system according to claim 13, cause the electronic vehicle system to perform a method according to claim 11.

Images