METHOD AND DEVICE FOR OPTICAL DISTANCE MEASUREMENT
Patent Information
- Application Number
- DE502019012495
- Authority / Receiving Office
- DE · DE
- Patent Type
- Patents
- Current Assignee / Owner
- Filing Date
- 2019-03-27
- Publication Date
- 2024-12-05
- Estimated Expiration
- 2039-03-27
AI Technical Summary
Existing LIDAR devices fail to distinguish between reflection from particle clouds and object reflections, leading to degraded transit time measurements and incorrect recognition of particle clouds as objects, especially in adverse weather conditions.
The method involves conducting multiple transit time measurements, generating histograms of received optical power versus transit time, and determining a correlation between the backscatter curve and sensitivity curve to differentiate between particle clouds and objects, thereby assigning a particle cloud feature to each measurement, which helps in determining the presence and confidence level of a particle cloud.
This approach effectively prevents the misclassification of particle clouds as objects, improving the accuracy of distance measurements and enabling better navigation in adverse weather conditions by providing a confidence measure for particle cloud detection and density assessment.
Description
Technical area
[0001] The invention relates to a method and a device for optical distance measurement. State of the art
[0002] LIDAR (Light Detection and Ranging) devices are well-known in the art. They comprise a transmitter unit for transmitting measurement pulses and a receiver unit for receiving reflected measurement pulses from objects within a measurement range. Based on the time-of-flight principle, the speed of light is used to determine the distance to the objects from which the measurement pulses were reflected.
[0003] The result of such time-of-flight measurements is a point cloud in which each point represents a single measurement, i.e., the reception of a transmitted and reflected measurement pulse. Overall, the quality of such a point cloud suffers greatly in difficult weather conditions, such as fog, dust, sea spray, etc. This is because the reflections from the particles and their corresponding points in the LIDAR point cloud are incorrectly detected as object reflections. Depending on the particle density, this can even lead to the complete failure of the device.
[0004] DE 10 2012 112987 B3 discloses an optoelectronic sensor and a method for detecting and determining the distance of objects.
[0005] US 2014 / 0324266 A1 describes an apparatus and method for determining weather conditions using vehicle-mounted sensors.
[0006] State-of-the-art distance measurement devices are not capable of detecting that they are located within a particle cloud. Consequently, the device incorrectly identifies a particle cloud as an object. This reduces the quality of the time-of-flight measurement. However, since the device does not detect this, the result of a time-of-flight measurement degraded by a particle cloud is treated with the same certainty—i.e., with the same degree of confidence—as a measurement under conditions without a particle cloud. An assessment of the quality of the resulting point cloud as a result of the time-of-flight measurement can only be made in downstream computing units that evaluate the supplied data over multiple time intervals. Description of the invention: task, solution, advantages
[0007] The present invention is based on the object of further developing a method and a device for optical distance measurement in such a way that the reflection of measuring pulses on a particle cloud can be distinguished from a reflection on an object.
[0008] The above-mentioned task is solved by a method for optical distance measurement, which includes the performance of several time-of-flight measurements.
[0009] A time-of-flight measurement comprises transmitting at least one measurement pulse, in particular exactly one or a plurality of measurement pulses, by means of a transmitting unit, as well as reflecting at least one transmitted measurement pulse, in particular of transmitted measurement pulses, from a particle cloud or from at least one object, and receiving at least one measurement pulse, in particular of reflected measurement pulses, by means of a receiving unit. The method comprises generating a point cloud based on the time-of-flight measurement. The point cloud is a result of the time-of-flight measurement.
[0010] Furthermore, the method can preferably determine the propagation time of each transmitted, reflected, and received measurement pulse. The method also includes, in particular, generating a histogram in which the received optical power (or an electronic signal corresponding to the measured optical power) is plotted against the determined propagation time. Furthermore, the determined propagation time of several measurement pulses can be averaged. A histogram with the averaged values can then also be generated.
[0011] Furthermore, each reflected measurement pulse received by the receiving unit results in a point in a point cloud, preferably in a 2D or 3D spatial coordinate system. According to the invention, at least one point in a point cloud is thus obtained as a result of a time-of-flight measurement. Furthermore, the determined time of flight of several received and reflected measurement pulses can be averaged, so that these together represent a point in a point cloud.
[0012] The method comprises generating a backscatter curve based on the time-of-flight measurement, wherein the method comprises evaluating the backscatter curve to detect at least one object and / or at least one particle cloud from which at least one measurement pulse was reflected. The method comprises providing a sensitivity curve for evaluating the backscatter curve, wherein the evaluation of the backscatter curve comprises determining a correlation between the sensitivity curve and the backscatter curve in order to determine, using the time-of-flight measurements, whether a particle cloud is located in a measurement area measured using the time-of-flight measurements and to assign a particle cloud feature to each time-of-flight measurement. The term "measured measurement area" refers in particular to the area through which the measurement pulse of the time-of-flight measurement passed.
[0013] Determining the correlation serves to determine whether a particle cloud is present in a measurement area measured using the time-of-flight measurements. In other words, it is determined whether a device for carrying out the method is located in and / or in front of a particle cloud. Advantageously, the method comprises determining, using each time-of-flight measurement, whether a particle cloud is present in the measurement area and assigning a particle cloud feature to the time-of-flight measurement.
[0014] The term "particle cloud feature" refers specifically to a particle cloud status and / or a particle cloud probability. The particle cloud feature is determined based on the determined correlation. A particle cloud status could indicate whether the correlation indicates that a particle cloud is present in the measurement area or not. The status could thus be presented as dual yes / no information. For example, the status could be "yes" if the correlation exceeds a previously set threshold and "no" if it is below it.
[0015] Furthermore, a particle cloud probability can be determined. This could be understood as a confidence measure that measurement pulses were reflected by a particle cloud. The method can thus determine a confidence measure between 0 and 100%, where 100% represents a very high correlation and thus it can be assumed that measurement pulses were reflected by a particle cloud, while 0% represents a negligible correlation and thus it can be assumed that no measurement pulses were reflected by a particle cloud. In particular, the confidence measure is determined based on the determined correlation.
[0016] In particular, the determination of the correlation serves to prevent a particle cloud from being mistakenly identified as an object.
[0017] The term "optical distance measurement" refers to the determination of distances using optical signals, in this case optical measurement pulses. The term "distance" refers to a distance. The distance traveled by the measurement pulse is the distance between the transmitting unit that emitted the measurement pulse and the reflecting object that reflected it, plus the distance between the reflecting object and the receiving unit that received the corresponding reflected measurement pulse. The reflected measurement pulses each represent backscattered signals of the emitted measurement pulses. The term "reflecting object" can be understood as an object and / or a particle cloud.
[0018] In particular, the method is used in driverless navigation of vehicles, especially automobiles. Above all, the method can be used for autonomous control of a vehicle. Furthermore, the method can be used to assist a driver of a vehicle. For this purpose, the distances to all reflective objects within a measurement range are determined.
[0019] A measuring pulse is in particular an optical, in particular electromagnetic, signal. The measuring pulse is advantageously a light pulse, i.e. a pulse with a wavelength that is not within the range visible to the human eye. Invisible infrared is preferably used for safety reasons. A measuring pulse preferably has a pulse width so that the measuring pulse can be understood as a time-limited portion of electromagnetic radiation. Since the measuring pulse is an electromagnetic signal and thus the speed of the measuring pulse is known, the distance traveled by the measuring pulse can be deduced from the propagation time of a measuring pulse using the speed of light.
[0020] The receiving unit can preferably comprise at least one detector, for example a photodetector, especially an optical detector. The receiving unit can comprise a plurality of receiving elements, preferably operating in linear or Geiger mode. The receiving elements operating in linear mode can be, in particular, avalanche photodiodes, PN photodiodes, PIN diodes, or photomultipliers, while the receiving elements operating in Geiger mode are preferably single-photon avalanche diodes (SPADs). In the latter mode, each photon is measured individually, and a report is made via the resulting histogram. The term "receiving element" refers primarily to pixels.
[0021] Furthermore, the transmitting unit can comprise a plurality of transmitting elements, each of which is formed, in particular, by a laser. In particular, the transmitting unit is to be understood as a "transmitting matrix," i.e., an array of transmitting elements, while the receiving unit is to be understood as a receiving matrix, i.e., an array of receiving elements. A matrix can be understood, in particular, as a three-dimensional, in particular plate-shaped, body, on one surface of which the corresponding elements, transmitting elements or receiving elements, are arranged.
[0022] The term "particle cloud" refers in particular to an accumulation of particles in the air. In particular, a particle cloud is an aerosol. Advantageously, these are solid and / or liquid, finely distributed suspended particles in the air. For example, they can be liquid, finely distributed liquid droplets in the air, such as fog or spray. Furthermore, they can be solid, finely distributed dust particles, such as dust or smog. Furthermore, a particle cloud can be understood as an exhaust cloud from the exhaust of a vehicle, which includes, in particular, oil and soot smoke. The dimensions, in particular the diameter, of the particles are typically less than 1 mm, in particular less than 100 µm, most preferably less than 50 µm.
[0023] In contrast to a particle cloud, an object within the meaning of the invention is not an aerosol. It is primarily a solid or liquid body.
[0024] The term "backscatter curve" is to be understood as a curve that can be used to determine the optical power received by the receiving unit in relation to the determined propagation time. In particular, the backscatter curve is the histogram generated from a propagation time measurement. The received optical power is preferably converted into an electronic signal, which is plotted in relation to the determined propagation time. To generate the backscatter curve, a corresponding electronic signal and the propagation time of at least one transmitted and received measurement pulse are therefore preferably determined based on the measured optical power. The at least one corresponding pair of values is plotted on a coordinate system with the electronic signal as the y-axis and the propagation time as the x-axis.
[0025] In other words, the backscatter curve is the raw signal of the time-of-flight measurement. The raw signal is the analog or digital representation of the reflected optical signal—in other words, the electronic signal into which the received reflected optical power is converted. The backscatter curve is the result of all transmitted, reflected, and received measurement pulses of the same time-of-flight measurement.
[0026] When using receiving elements operating in linear mode, a time-of-flight measurement primarily involves the transmission of a single measurement pulse. Time-of-flight information from a large number of photons is already taken into account during the reception of the individual transmitted and reflected measurement pulse, so that the generated histogram of a single measurement pulse represents the backscatter curve. This is because individual photons are reflected by closer particles in the particle cloud, while other photons are reflected by more distant particles.
[0027] When using receiver elements in Geiger mode, however, the time-of-flight measurement involves transmitting a large number of measurement pulses, with the histogram displaying the averaged time-of-flights. The histogram can also display the backscatter curve.
[0028] The term "sensitivity curve" of the receiving unit refers in particular to the sensitivity curve of the receiving unit, in particular of a device for implementing the method. The sensitivity curve describes, among other things, how sensitively the receiving unit or individual receiving elements of the receiving unit react to the distance from the receiving unit to a reflective object. In other words, the sensitivity curve describes the influence of the distance of the reflective object, from which a measuring pulse is reflected, on the optical power detected, assuming the same optical power of the measuring pulse. Furthermore, other factors can influence the sensitivity curve, such as the divergences between a transmitting and receiving unit of a device and / or the distance between the transmitting and receiving units.Other significant factors influencing the sensitivity curve include the optical and mechanical design of the receiving unit and / or the transmitting unit, as well as the electronic components used, such as the receiving elements and / or the transmitting elements, particularly diodes, and / or amplifiers. The sensitivity curve is therefore specific to the receiving unit, particularly the device or sensor.
[0029] The sensitivity curve represents an overall increase, in other words a "peak", in a coordinate system with the electronic signal into which the optical power has been converted on the y-axis and the time of flight on the x-axis, preferably with a rising edge and a falling edge.
[0030] The shape of the sensitivity curve results from the fact that with increasing distance, the expected optical power of the backscattered signal naturally decreases. However, due to the factors mentioned above, maximum power is only achieved at a certain distance from the receiving unit.
[0031] The method may include providing a single common sensitivity curve for the entire receiving unit and / or the entire device. Furthermore, the method may include providing a sensitivity curve for each receiving element of the receiving unit. The method may advantageously generate a backscatter curve for each individual receiving element. In other words, each time-of-flight measurement may be assigned to a receiving element. Based on the backscatter curve and the sensitivity curve, the correlation for each receiving element may be determined and optionally averaged.
[0032] Furthermore, only a portion of the receiving elements of the receiving unit, e.g., only one receiving element, can be used for particle cloud detection. In other words, the number of receiving elements used for particle cloud detection is smaller than the total number of receiving elements.
[0033] Then, using this receiving element, the backscatter curve and sensitivity curve are determined.
[0034] The evaluation involves determining a correlation between the sensitivity curve and the backscatter curve. This is advantageous because the sensitivity curve exhibits very similar characteristics to a section of a backscatter curve resulting from reflection from a particle cloud.
[0035] Put simply, this is because a particle cloud extends over a certain distance range from a device for carrying out the method, so that some measuring pulses are reflected by particles located closer to the cloud, while other measuring pulses are reflected by particles located further away in the cloud. Overall, reflections are typically received from all distances to a device for carrying out the method that include the particle cloud. Precisely this type of reflection is represented by the sensitivity curve, which maps how a receiving unit, preferably a device for carrying out the method, reacts to reflections from all different distances to a receiving unit, in particular to a device for carrying out the method.
[0036] The course of the sensitivity curve and the section of a backscatter curve attributable to reflection from a particle cloud are therefore very similar. In particular, the shape of the curves is similar. By determining a correlation between the sensitivity curve and the backscatter curve, it can be determined whether at least one section of the backscatter curve has a similar course, particularly its shape, to the sensitivity curve, and whether this section can therefore be attributed to reflection from a particle cloud.
[0037] The term "correlation" refers to the relationship between the backscatter curve and the sensitivity curve. The correlation is a measure of how similar at least one section of the backscatter curve and the sensitivity curve are, particularly their shape. This section of the backscatter curve could correspond to the reflection from a particle cloud.
[0038] The backscatter curve primarily comprises at least one peak, preferably several peaks. A peak is present in particular when it clearly contrasts with the noise of the backscatter curve, in other words, with a corresponding deviation of the optical power values at the same time of flight. A peak is present when the peak has a maximum optical power that preferably corresponds to at least twice the noise.
[0039] In particular, determining the correlation serves to determine whether at least a section of the backscatter curve exhibits a similarity to the sensitivity curve. The term "section" refers, in particular, to a part of the backscatter curve that includes an increase in the backscatter curve. In particular, the determined correlation can be used to determine whether a particle cloud was measured during the time-of-flight measurement—in other words, whether measurement pulses from the time-of-flight measurement were reflected by a particle cloud.
[0040] According to the invention, the method comprises detecting a particle cloud and preferably determining a confidence measure for the detection of the particle cloud. In particular, the confidence measure is determined based on the determined correlation. The confidence measure is a measure of the quality, or in other words, the credibility, of the detection of a particle cloud.
[0041] Preferably, the method includes outputting a message regarding the at least one particle cloud feature. An output could, for example, be as follows: "Particle cloud status 'yes' with a particle cloud probability of 75%."
[0042] The method comprises performing a plurality of time-of-flight measurements, wherein the particle cloud status of the time-of-flight measurement is preferably "Yes" or "No" in each case. The method comprises setting a first threshold, wherein a particle cloud is detected if the number of time-of-flight measurements with a particle cloud status of "Yes" exceeds the first threshold. The first threshold can be defined as "1," so that, for example, a particle cloud is detected even with a time-of-flight measurement with a particle cloud status of "Yes." Furthermore, the first threshold can be defined such that it amounts to 10%, most preferably 20%, of all time-of-flight measurements.
[0043] Furthermore, the method may comprise setting a second threshold for the particle cloud probability, wherein a particle cloud is detected when the number of runtime measurements with a particle cloud probability above the second threshold exceeds the first threshold.
[0044] The confidence measure is specifically expressed as a probability. The method thus determines a confidence measure between 0 and 100%, where 100% represents a very high correlation and the method is therefore very certain that measurement pulses were reflected by a particle cloud, while 0% represents a vanishingly low correlation and thus it can be assumed that no measurement pulse was reflected by a particle cloud.
[0045] Preferably, the method includes outputting a message regarding the detection of a particle cloud and the determined confidence level. An output could, for example, read: "A particle cloud was detected here with a confidence level of 75%."
[0046] In particular, the correlation is determined by convolving the backscatter curve with the sensitivity curve. Furthermore, determining the correlation may involve a Fourier transform, in particular a fast Fourier transform. Determining the correlation may further involve using an optimal filter adapted to the sensitivity curve. This optimal filter may be trained beforehand, with the method comprising training the optimal filter using a neural network. In particular, the optimal filter may be trained using an autoencoder. Furthermore, the optimal filter may be trained using deep learning.
[0047] To provide the sensitivity curve, the method may comprise simulating the sensitivity curve.
[0048] Alternatively or additionally, the sensitivity curve can be measured. For example, at least one object can be placed at different distances from a receiving unit, without the presence of a particle cloud, whereby the receiving unit receives the reflected optical power and converts it into a corresponding electronic signal. Overall, this results in the course of the electronic signal as a measure of the reflected optical power as a function of the distance or the propagation time, and thus the sensitivity curve. Compared to a simulation, a measurement, particularly end-of-line, has the advantage that it also includes the electrical behavior of the receiving unit, in particular the device for carrying out a method, in other words the sensor, as well as the sensor-specific mechanical and optical deviations.Furthermore, a simulation involves much more effort and could be incomplete.
[0049] A measurement of the sensitivity curve can be performed specifically for the type of device, in other words, the sensor, used to perform the method. The term "sensor type" primarily refers to the type of sensor, influenced by the model and the corresponding technology, for example, influenced by the detector type and / or the optics and / or the measurement principle.
[0050] In particular, the method considers a possible scaling of the backscatter curve depending on the density of a particle cloud when determining the correlation. In particular, the reflection from a particle cloud depends on its density. The denser the particle cloud, the higher the signal, and thus indirectly the optical power, in the backscatter curve. In other words, the backscatter curve can be scaled by a factor that takes the density of the particle cloud into account. The denser the particle cloud, the more the backscatter curve is compressed in the x-axis direction. This scaling can change the shape of the backscatter curve in relation to the sensitivity curve.Although the height of the increase in the backscatter curve, which is due to a reflection from a particle cloud, is therefore dependent on its density, the basic shape is nevertheless so similar to that of the sensitivity curve that even with particle clouds of different density, a correlation can be found, which indicates that the reflection is the reflection from a particle cloud.
[0051] In a further step, the method may include detecting a density of the particle cloud and preferably determining a confidence measure for the density detection. In particular, the method may include taking into account the height of a peak. Preferably, the evaluation includes an estimation of the maximum height of the increase in the backscatter curve that was assigned to the reflection from a particle cloud, from which the density of the particle cloud can be deduced. The confidence measure is a measure of the quality, in other words, the credibility, of the detected density.
[0052] The confidence measure for density detection is specifically expressed as a probability. The method thus determines a confidence measure between 0 and 100%, where 100% means that the evaluation clearly indicates the detected density, and the method is therefore very certain that the particle cloud has exactly this density. 0% means that the evaluation was inconclusive, and it is therefore very uncertain whether the particle cloud actually exhibits the detected density.
[0053] Preferably, the method includes outputting a message regarding the density detection and the determined confidence level. An output could, for example, be as follows: "The density of the particle cloud comprises X particles per volume with a confidence level of 80%."
[0054] A message can be addressed primarily to a driver of a vehicle, where the vehicle is being navigated using the present method and / or the driver is being assisted using the present method. The output serves, in particular, to inform the driver that the vehicle is in a particle cloud and / or is moving toward a particle cloud.
[0055] In particular, the method can comprise performing at least one further measurement to measure measurement data from at least one further sensor modality. The term "sensor modality" is in particular the German term for the English term "sensor modality." In other words, the sensor modality is a sensor category. Sensors thus differ in particular by their sensor modality or sensor category. In particular, the sensor modality determines the measurement method used to generate the measurement data. The sensor modality is preferably lidar, radar, image, or ultrasound. This means that the corresponding sensor of the modality is preferably a lidar sensor, a radar sensor, an image sensor, in particular a camera, or an ultrasound sensor. The method preferably comprises measuring radar measurement data, image data, and / or ultrasound data.
[0056] Preferably, the method may comprise determining a confidence measure for the detection of a particle cloud for the measurement data of each sensor modality, wherein the method further comprises comparing the determined confidence measures. In particular, an overall confidence measure is determined, which takes into account the confidence measures of the individual sensor modalities. This overall confidence measure may be output along with a message.
[0057] In particular, the corresponding time-of-flight measurement can be assigned to each point in a point cloud. In a further step, the method according to the invention comprises determining a confidence measure for the time-of-flight measurement, in particular for the corresponding points in a point cloud, taking into account the detection of a particle cloud.
[0058] In particular, the confidence measure can be understood as the quality of a time-of-flight measurement, i.e., the quality or probability that its results are degraded, particularly due to reflection from a particle cloud. The higher the confidence measure, the less likely it is that the measurement is degraded. Overall, the present method thus enables an assessment of the quality of the time-of-flight measurement.
[0059] Once a particle cloud is detected, the confidence measure for the time-of-flight measurement that is degraded due to the particle cloud is adjusted. In particular, the confidence measure is adjusted compared to a time-of-flight measurement in which no particle cloud was detected. In particular, the confidence measure is determined depending on the detection of a particle cloud and / or the detection of its density and / or the confidence measure for the detection of the particle cloud and / or the confidence measure for the detection of the density.
[0060] Furthermore, the method may comprise creating a 3D depth map that describes the distances and / or directions at which a particle cloud and / or the detected objects are arranged relative to a device for carrying out the method.
[0061] Since the method comprises a plurality of runtime measurements, the steps described above can preferably be carried out for each of the measurements.
[0062] According to the invention, the confidence measure for the travel time measurement is assigned to the corresponding points of a point cloud. The method according to the invention comprises generating a point cloud that includes points from different travel time measurements. By assigning the confidence measures, the point cloud can contain points with different confidence measures.
[0063] The method particularly comprises object recognition, wherein the method comprises distinguishing between objects and particle clouds. In particular, during object recognition, elevations in the backscatter curve are observed and classified. In order to assess whether an elevation in the backscatter curve should actually be assigned to an object, information about the similarity, i.e. the correlation between the backscatter curve and the sensitivity curve, can be relevant. If there is sufficient correlation between an elevation and the sensitivity curve, it can be assumed that the elevation does not result from reflection from an object, but from reflections from a particle cloud. The method primarily comprises classifying elevations as particle clouds or objects, wherein the method further preferably comprises determining a confidence measure for each classification.In particular, the confidence measure is determined based on the determined correlation. The confidence measure is a measure of the quality, or in other words, the credibility, of the classification.
[0064] The confidence measure is specifically expressed as a probability. The method thus determines a confidence measure between 0 and 100%, where 100% indicates that the correlation clearly indicates that this is the respective classified reflection object, while 0% indicates that the correlation is ambiguous, so it is uncertain whether the increase is due to reflection from a particle cloud or an object.
[0065] Preferably, the method includes outputting a message regarding the classification and the determined confidence level. An output could, for example, read: "Here, an object was detected at distance x with a confidence level of 60%."
[0066] In a further aspect, the present invention comprises a device for distance measurement, wherein the device comprises a receiving unit and a transmitting unit for performing multiple time-of-flight measurements. The transmitting unit serves to transmit at least one measuring pulse, in particular a plurality of measuring pulses, while the receiving unit serves to receive at least one reflected measuring pulse, in particular reflected measuring pulses that were previously transmitted and reflected by a particle cloud or by at least one object. The device is designed to carry out a method described above.
[0067] Furthermore, the device comprises, in particular, an evaluation unit designed to evaluate the backscatter curve for object detection, in particular of at least one object and / or a particle cloud from which measurement pulses were reflected. Furthermore, the evaluation unit is designed to determine a correlation between the sensitivity curve and the backscatter curve. Furthermore, the evaluation unit can be designed to detect objects and / or recognize a particle cloud and / or recognize objects and / or classify the density of a detected particle cloud and / or output feedback and / or determine a confidence measure and / or differentiate between particle clouds and objects. The evaluation unit can comprise digital signal processors, e.g., ASICs and / or FPGAs and / or microprocessors, etc.
[0068] The device is primarily a LIDAR device, in particular a scanning LIDAR device or a flash LIDAR device. In particular, the device comprises a LIDAR sensor. Furthermore, the device may comprise a radar sensor, an image sensor, in particular a camera, or an ultrasonic sensor. Furthermore, the device may be a driver assistance system.
[0069] In a further aspect, the invention comprises a computer program product comprising a computer-readable storage medium on which a program is stored which enables a computer, after being loaded into the memory of the computer, to carry out a method described above, optionally in interaction with a device described above.
[0070] Furthermore, the invention relates to a computer-readable storage medium on which a program is stored which enables a computer, after it has been loaded into the memory of the computer, to carry out a method described above, optionally in interaction with a device described above. Short description of the drawings
[0071] They show schematically: Figure 1 shows a process diagram of a method according to the invention; Figure 2 shows a sensitivity curve; Figure 3 shows an inventive device for carrying out a method according to the invention; Figure 4 shows an inventive device with a particle cloud and an object; Figure 5 shows an inventive device with a particle cloud and an object; Figure 6 shows a backscattering curve together with a sensitivity curve; and Figure 7 shows another backscattering curve together with a sensitivity curve. Preferred embodiments of the invention
[0072] Figure 1shows a process diagram of a method (100) according to the invention for distance measurement. The method comprises carrying out (101) at least one transit time measurement, wherein a transit time measurement comprises transmitting (102) at least one measuring pulse by means of a transmitting unit (12), reflecting (103) at least one transmitted measuring pulse, and receiving (104) at least one reflected measuring pulse by means of a receiving unit (11). The method can further comprise carrying out multiple transit time measurements.
[0073] The method further comprises generating (105) a backscatter curve (20) based on the time-of-flight measurement, as well as providing (106) a sensitivity curve (21). The sensitivity curve (21) can be simulated (107) and / or measured (108).
[0074] The method (100) further comprises the evaluation (109) of the backscatter curve (20) for object detection (113). The evaluation (109) comprises the determination (110) of a correlation between the sensitivity curve (21) and the backscatter curve (20). The correlation can be determined by means of convolution (111) and / or by using (112) an optimal filter. The determination (110) of the correlation serves to determine, by means of the at least one time-of-flight measurement, whether a particle cloud (27) is located in a measurement area measured by means of the time-of-flight measurement and to assign a particle cloud feature (120) to the time-of-flight measurement. In particular, the method comprises such an assignment (120).
[0075] Furthermore, the evaluation (109) can include object detection (113). The method (100) can include detecting (114) a particle cloud (27). For this purpose, the method can primarily comprise setting a first threshold, wherein a particle cloud is detected (114) if the number of runtime measurements with a particle cloud status of "Yes" exceeds the first threshold. Alternatively or additionally, the method can comprise setting a second threshold for the particle cloud probability, wherein a particle cloud is detected (114) if the number of runtime measurements with a particle cloud probability above the second threshold exceeds the first threshold.
[0076] After detecting (114) a particle cloud, the method (100) can determine (114a) a confidence measure for the detection. Furthermore, the method (100) can comprise the detection (115) of a density of the particle cloud and the determination (115a) of a confidence measure for the detection of the density. The method (100) can comprise the detection (116) of objects, wherein a distinction can be made (118) between particle clouds (27) and objects (26). Furthermore, the method can comprise outputting (117) a message regarding the detection of a particle cloud and / or its density. Furthermore, the evaluation (109) of the backscatter curve (20) can comprise the creation (119) of a 3D depth map that indicates the distances at which a detected particle cloud and / or detected objects (26) are located from a device (10) for carrying out the method (100).
[0077] Figure 2shows a sensitivity curve (21). The sensitivity curve is a representation of an electronic signal (30), into which the received optical power has been converted and which is thus considered an indirect measure of the optical power, over the propagation time (31). Figure 2 shows, on the one hand, how a sensitivity curve (21) was measured by individual measurements (22). The individual measurements (22) are based on measurements in which an object is placed at different distances and thus different travel times (31) from a receiving unit (11). For each distance or each travel time (31), the optical power is measured after reflection from the object and converted into an electronic signal (30) which is plotted against the travel time. The sensitivity curve (21) is thus obtained by interpolation. Furthermore, the sensitivity curve (21) can be simulated. In this case, too, the same curve results.
[0078] Figure 3 shows a device (10) according to the invention for carrying out a method (100) according to the invention. The device (10) according to the invention comprises a receiving unit (11) and a transmitting unit (12). The transmitting unit (12) serves to transmit measuring pulses, while the receiving unit (11) serves to receive the reflected measuring pulses. Overall, the receiving unit (11) and the transmitting unit (12) serve to carry out time-of-flight measurements. The evaluation unit (13) serves to evaluate the backscatter curve as a result of the time-of-flight measurement, specifically with the aid of the sensitivity curve (21). Furthermore, the device (10) has a control unit (14) which is designed to control the receiving unit (11), the transmitting unit (12) and / or the evaluation unit (13) accordingly.
[0079] Figure 4shows a device (10) according to the invention, in front of which a particle cloud (27) is located in the immediate vicinity and an object (26) is located at a distance from the device (10) and the particle cloud (27).
[0080] Figure 5 shows a device (10) according to the invention, in front of which a particle cloud (27) is located in the immediate vicinity. An object (26) is located in the particle cloud.
[0081] The corresponding backscatter curve (20) obtained after performing a runtime measurement based on the Figure 4 is in Figure 6 The backscatter curve (20) is an indirect representation of the optical power versus the propagation time (31), in which an electronic signal (30), into which the received optical power is converted, is plotted against the propagation time. Several peaks (23) can be seen in the backscatter curve (20). Such peaks (23) are typically associated with an object.
[0082] In Figure 6The sensitivity curve (21) is also shown in dotted form. In the course of determining a correlation, it is found that at least the section of the backscatter curve (20) that includes the left elevation (23) correlates strongly with the sensitivity curve (21). This is also clearly shown in Figure 6 The similarity in the shape of the backscatter curve (20) and the sensitivity curve (21) is clearly evident. Based on the high correlation, the left-hand elevation (23) is classified as the result of a reflection from a particle cloud (27). It is therefore an elevation (25) of the backscatter curve (20) due to a reflection from a particle cloud (27).
[0083] With regard to the right elevation (23), no correlation, or at least no sufficient correlation, can be determined, so that the elevation can be recognized with sufficient probability as an elevation (24) due to reflection from an object. In this way, particle clouds (27) can be clearly distinguished from objects (26).
[0084] The backscatter curve (20) obtained after performing a runtime measurement based on the Figure 5 shows Figure 7 Because the object (26) is located within the particle cloud (27), the elevation (23) attributable to the object is positioned on top of the elevation resulting from the particle cloud. The signals have been added together at this point, so that even if an object (26) is located within a particle cloud (27), it can be clearly detected and distinguished from the particle cloud (27). List of reference symbols
[0085] 100Methods for measuring distance 101Performing at least one time-of-flight measurement 102Emitting at least one measurement pulse by means of a transmitting unit 103Reflection of at least one transmitted measurement pulse by at least one object 104Receiving at least one reflected measurement pulse by means of a receiving unit 105Generating a backscatter curve based on the time-of-flight measurement 106Providing a sensitivity curve 107Simulating a sensitivity curve 108Measuring a sensitivity curve 109Evaluating the backscatter curve 110Determining a correlation between the sensitivity curve and the backscatter curve 111Convolution 112Using a matched filter 120Assigning a particle cloud feature to the time-of-flight measurement 113Object detection 114Detecting a particle cloud 114aDetermining a confidence measure for detecting the particle cloud 115Detecting the density of the particle cloud 115aDetermining a confidence measure for detecting the density 116Detecting objects 117Outputting a message 118Distinguishing between particle clouds and objects 119Creating a 3D depth map 10Device 11Receiving unit 12Transmitting unit 13Evaluation unit 14Control unit 20Backscatter curve 21Sensitivity curve 22Individual measurements 23Increase 24Increase due to reflection from an object 25Increase due to reflection from a particle cloud 26Object 27Particle cloud 30electronic signal 31runtime
Claims
1. Method (100) for optical distance measurement, wherein the method (100) comprises performing (101) several runtime measurements, wherein a runtime measurement comprises transmitting (102) at least one measuring pulse by means of a transmitting unit (12), reflecting (103) at least one transmitted measuring pulse and receiving (104) at least one reflected measuring pulse by means of a receiving unit (11), wherein at least one point of each point cloud is obtained as the result of the runtime measurements, wherein the method (100) comprises generating (105) a backscatter curve (20) based on the runtime measurement, wherein the method (100) comprises evaluating (109) the backscatter curve (20) for object detection (113), wherein the method (100) comprises providing (106) a sensitivity curve (21) for evaluating (109) the backscatter curve (20), wherein evaluating (109) comprises determining (110) a correlation between the sensitivity curve (21) and the backscatter curve (20) in order to determine by means of the runtime measurements whether a particle cloud (27) is arranged in a measuring range measured by means of the runtime measurements and to assign (120) a particle cloud feature to each of the runtime measurements, wherein the method (100) comprises detecting (114) a particle cloud (27), wherein the method comprises determining a confidence measure for each runtime measurement taking into account the detection of the particle cloud (27), wherein the confidence measure for the runtime measurements is each assigned to the corresponding points of the point cloud of the corresponding runtime measurement, wherein the method comprises generating a point cloud comprising points from different runtime measurements.
2. The method (100) according to claim 1, wherein the method (100) comprises determining (114a) a confidence measure for detecting (114) the particle cloud (27).
3. The method (100) according to claim 1 or 2, wherein determining (110) the correlation comprises considering a possible scaling of the backscatter curve (20) depending on a density of a particle cloud (27).
4. The method (100) according to one of the preceding claims, wherein for providing (106) the sensitivity curve (21), the method (100) comprises simulating (107) the sensitivity curve (21).
5. The method (100) according to one of the preceding claims, wherein for providing (106) the sensitivity curve (21), the method (100) comprises measuring (108) the sensitivity curve (21).
6. The method (100) according to claim 5, wherein the method (100) comprises using a device (10) for optical distance measurement, wherein the method comprises measuring (108) the sensitivity curve (21) specific to the device (10).
7. The method according to claim 1, wherein the particle cloud status of the runtime measurements is "Yes" or "No", respectively, wherein the method comprises setting a first threshold, wherein a particle cloud is detected (114) when the number of runtime measurements with a particle cloud status of "Yes" exceeds the first threshold.
8. The method according to one of claims 1 or 6, wherein the method comprises setting a first threshold, wherein the method comprises setting a second threshold for the particle cloud probability, wherein a particle cloud is detected (114) when the number of runtime measurements with a particle cloud probability above the second threshold exceeds the first threshold.
9. The method (100) according to one of claims 1, 7 or 8, wherein the method (100) comprises detecting (115) a density of the detected particle cloud (27) and preferably determining (115a) a confidence measure for the detection (115) of the density.
10. The method (100) according to one of claims 1 or 7 to 9, wherein the method (100) comprises outputting (117) a message relating to the detected particle cloud (27) and preferably the confidence measure for detecting the particle cloud (27) and / or outputting a message relating to the density of the detected particle cloud (27) and preferably the confidence measure for detecting the density.
11. The method (100) according to one of the preceding claims, wherein the method (100) comprises object detection (113), wherein the method comprises distinguishing (118) between objects (26) and particle clouds (27).
12. The method (100) according to one of the preceding claims, wherein the receiving unit (11) comprises a plurality of receiving elements, wherein the method (100) comprises providing (106) a sensitivity curve for each receiving element.
13. Device (10) for optical distance measurement, wherein the device comprises a receiving unit (11) and a transmitting unit (12) for performing (101) several runtime measurements, wherein the device (10) is configured to carry out a method according to one of claims 1 to 12.
14. A computer program product comprising a computer-readable storage medium on which is stored a program enabling a computer, after having been loaded into the memory of the computer, to perform a method (100) according to one of claims 1 to 12, optionally in cooperation with a device (10) according to claim 13.
15. A computer-readable storage medium on which a program is stored which, after being loaded into the computer's memory, enables a computer to perform a method according to one of claims 1 to 12, optionally in cooperation with a device (10) according to claim 13.
