Method for determining the synchronization of a plurality of camera devices and sensor means during a safety test on a motor vehicle

Abstract

The invention relates to a method for determining the synchronization of a plurality of camera devices of a first measuring system and sensor means of a second measuring system during a safety test, in particular a crash test on a motor vehicle, wherein - Pulsed light with a defined pulse pattern is emitted by an externally synchronized lighting device before the safety test is carried out in a static state of the motor vehicle, which is detected by the camera devices of the first measuring system and by sensor means of the second measuring system, which are designed to measure brightness and / or electrical signals of the lighting device, wherein the defined pulse pattern is identified in the image sequences of the images detected by the camera devices by forming pixel sum signals and in the measurement signals of the sensor means, and it is checked whether the temporal position of peaks in the pixel sum signals and in the measurement signals corresponds to the temporal positions of a plurality of defined support points or not, and - during the execution of the safety test, pulsed light with a defined pulse pattern is emitted by the lighting device, wherein the intensity of the light is changed in a defined manner at a first time point, which in particular defines the time of impact of the motor vehicle on the collision barrier, and at least at a second time point, which in particular represents the end of the safety test, and it is checked by peak analysis whether the intensity changes were detected simultaneously by the camera devices in a defined static area of ​​the camera image and by the sensor means.

Description

[0001] The present invention relates to a method for determining a synchronicity of a plurality of camera devices of a first measuring system and sensor means of a second measuring system during a safety test, in particular a crash test, on a motor vehicle.

[0002] Safety tests, particularly crash tests conducted on motor vehicles, typically involve the use of multiple (optical and non-optical) sensor devices and camera devices that begin their measurements simultaneously at a manually or automatically triggered event time and record corresponding data, particularly measurement data from the sensor devices and image data from the camera devices, of the safety test for a predetermined measurement duration. In this context, two boundary conditions must be met: the temporal synchronization of all measuring transducers (i.e., all sensor devices and camera devices) used in the safety test, and the exact same triggering time for the measurements, in order to achieve comparability of the measurement data in terms of time relative to the event time.

[0003] DE 10 2019 115 091 B3 discloses a method for conducting a safety test with synchronization of the measurement technology, LED lighting, dummy technology, and cameras via a hall cycle. EP 3 487 267 A1 also shows cameras synchronized with the lighting, with the lighting reaching its greatest intensity at the moment of impact and then decreasing in intensity. US 2007 / 0 057 154 A1 describes a possibility for synchronizing sensors, which can also be used in crash tests, using temporal pulse sequences.

[0004] Safety tests typically utilize two separate, higher-level measurement systems. A first measurement system comprises all image-recording camera devices used in the safety test. A second measurement system primarily comprises non-optical sensors (e.g., accelerometers) and some optical sensors (e.g., optical distance sensors or brightness sensors) used in the safety test.

[0005] This separation of the two higher-level measurement systems allows them to operate independently of each other. The two higher-level measurement systems are unified by a synchronous signal generator, which sets a fixed clock frequency, specifically a 1000 Hz clock frequency, for both measurement systems.

[0006] Both systems use this measurement frequency (sometimes multiplied by a factor of 10). The event signal forms a trigger signal for both measurement systems and is triggered at a time t0 = 0 ms, for example, by the vehicle hitting a collision barrier or earlier by another external event. It is then distributed centrally to both measurement systems so that the sensors or camera devices begin recording data.

[0007] However, there is no evidence as to whether the two higher-level measuring systems or the sensors and camera devices used therein actually trigger at the same time and whether the synchronicity is actually maintained until the end of the measurement series (for example, until a time t1 = 300 ms).

[0008] A separate check of each sensor or camera device is very time-consuming and therefore usually not possible for the entire test setup. Deviations can be caused by technical errors in the sensor or camera devices themselves, an insufficient signal from the clock generator, or incorrectly configured parameters in a software program (e.g., an incorrect edge in the signal, etc.). Furthermore, the location of the sensor or camera devices (e.g., in the vehicle itself, in a test hall, or in a film pit) varies. This can lead to corresponding temporal deviations between the measurement signals or image signals captured by the sensor or camera devices.

[0009] The object of the present invention is to provide a method for determining a synchronicity of a plurality of camera devices and sensor means during a safety test, in particular a crash test, on a motor vehicle, by means of which the synchronicity of these measuring sensors can be reliably determined in a simple manner.

[0010] This object is achieved by a method having the features of claim 1. The subclaims relate to advantageous developments of the invention.

[0011] In a method according to the invention for determining a synchronicity of a plurality of camera devices of a first measuring system and sensor means of a second measuring system in a safety test, in particular a crash test, on a motor vehicle, it is provided that - prior to conducting the safety test in a static state of the motor vehicle, pulsed light with a defined pulse pattern is emitted by an externally synchronized lighting device, which is detected by the camera devices of the first measuring system and by sensor means of the second measuring system, which are designed to measure brightnesses and / or electrical signals of the lighting device, wherein the defined pulse pattern is identified in the image sequences of the images captured by the camera devices by forming pixel sum signals and in the measurement signals of the sensor means, and it is checked whether the temporal position of peaks in the pixel sum signals and in the measurement signals of the sensor means corresponds to the temporal positions of a plurality of defined support points or not, and - during the performance of the safety test, pulsed light with a defined pulse pattern is emitted by the lighting device, wherein the intensity of the light is changed in a defined manner at a first point in time, which in particular defines the point in time of impact of the motor vehicle on the collision barrier, and at least at a second point in time, which in particular represents the end of the safety test, and peak analysis is used to check whether the intensity changes were recorded simultaneously by the camera devices in a defined static area of ​​the camera image and by the sensor means.

[0012] The method according to the invention enables a quick and simple verification of the synchronicity and temporally synchronous triggering of all camera devices of the first measuring system and the sensor devices of the second measuring system, which can detect at least one brightness value or a current (or another electrical measurement value). This is achieved by calculating the pixel sums before the actual start of the test and performing peak analyses in the pixel sums and in the measurement signals of the sensor devices of the second measuring system while the vehicle is in a static state (speed: v = 0 m / s). For this purpose, the pulsed illumination device of the test system is advantageously used as a synchronization means.The illumination device is pulsed by a central synchronization signal; in one embodiment, an additional, redundant timing device can optionally be used. The static state of the vehicle at the beginning of the activation of the illumination device before the actual start of the test enables, in particular, comparability of the brightness values ​​within an image stack. Furthermore, the exposure time of the camera devices is preferably set to the same pulse width of the light pulses generated by the illumination device. In principle, not all sensors used by the second measurement system can be tested, since the light changes, for example, do not affect acceleration sensors, etc.What can be checked, for example, is at least the synchronous triggering of the second measuring system itself as well as the evaluation of the support points with additional brightness sensors, which are usually not directly installed in the vehicle.

[0013] In one embodiment, it is proposed that, before conducting the safety test, the radiant power of the pulsed light emitted by the illumination device of the test facility be briefly changed at the support points. For example, the light pulses can be temporarily suspended completely so that no light is emitted. Furthermore, it is possible not to completely switch off the pulsed light for a short time, but to briefly reduce or increase the radiant power in a defined manner. These brief changes in the radiant power preferably have a duration selected such that these changes lead to the evaluable peaks in the pixel sum signals and enable high temporal resolution. This duration can preferably correspond to the pulse width of the pulsed light, since the exposure time of the camera devices is preferably set to this pulse width.

[0014] Preferably, a slightly larger pulse width of the light is set in the security test itself in order to prevent asynchronous or signal-induced exposure changes in the images during the actual security test.

[0015] In a further embodiment, it is proposed that an additional light signal be generated by the lighting device at time t0 = 0 ms (event time) and at time t1 (end of the safety test, for example, at time t1 = 300 ms) after the event triggered manually or automatically on the motor vehicle. This additional light signal can preferably have a duration corresponding to the pulse width of the pulsed light. In principle, such an additional light signal can also be generated and evaluated in the phase before the start of the actual safety test.

[0016] In an advantageous embodiment, additional defined sampling points can be used in a period between the start time (event time) and the end time of the safety test, and in particular for the measurement before the safety test, to check synchronicity. These sampling points are generated by the lighting device in the form of brighter or darker scenes. These sampling points can preferably also have a duration corresponding to the pulse width of the pulsed light.

[0017] In a preferred embodiment, it is possible to perform low-pass filtering of the pixel sum signal captured by the camera devices. This advantageously compensates for unwanted noise influences caused, for example, by minimal movements of the vehicle or bystanders that affect the overall light influence.

[0018] As already mentioned above, it is intended that, to check the synchronicity of the first measuring system with the second measuring system, current sensors for measuring the electrical power consumption of the lighting device and / or brightness sensors in the test facility are used, for example. From their measurement signals (if necessary after a likewise advantageous low-pass filtering), a comparison of the positions of the peaks is carried out with the pixel sum signals generated from the camera images. These sensor devices evaluate the time-dependent electrical power consumption of the lighting device and / or the brightness in the test hall, preferably at a frequency of at least 1000 Hz. In this way, the synchronicity of the data between the two measuring systems can be assessed before the test and also during the test by comparing the temporal positions of the peaks in the pixel sum signals.Here again, the recording time within a cycle is preferably set to be identical to the light pulse length.

[0019] In the safety test itself, pixel sum signal analysis is also possible, but only in non-moving parts of the image, the so-called "regions of interest" (ROI). There, brief bursts of illumination (such as at t = t0 = 0 ms and t = 300 ms) can be detected with much greater detail, even if the pixel sum of the overall image provides only limited information about the reference points, since many vehicle parts are flying around in the safety test and the vehicle is moving rapidly. By analyzing the pixel sums only in the ROIs, excessive illumination or darkening of the image is not necessary, whereas analyzing the overall image in a moving safety test would require very strong illumination to reliably detect the reference point.

[0020] The verification of the synchronicity of the camera devices of the first measuring system and the sensor devices of the second measuring system is thus carried out in two steps: The first step is performed before the start of the safety test (typically approximately 30 minutes to 1 hour before the start of the test). At this point, the vehicle is already in front of the collision barrier. For the motion analysis, reference images must be taken there anyway using the camera equipment of the first measurement system. The light is preferably pulsed with the same duration as the camera equipment of the first measurement system takes images (including peaks at the trigger event and, for example, at t = 300 ms, etc., and possibly several additional sampling points). This allows the pixel sum signals to be generated (because they are static). The trigger event is triggered manually by a person, preferably directly on the vehicle. The signals are then evaluated, the gradient calculated, and the position of the peaks assessed.The result is the output of the pixel sum signals for each of the camera devices of the first measurement system and for each of the sensors for brightness or current measurement of the second measurement system. The camera images themselves are not relevant. Therefore, the evaluation of the synchronicity of the measurement systems prior to the actual safety test is generally carried out in a static state (v = 0 m / s).

[0021] In a second step, during the actual safety test (t = -20 ms: start of measurements, t0 = 0 ms: impact time / event time, and t = 300 ms: end of the measurement series), the light emitted by the illumination device becomes slightly brighter or slightly dimmer at times t0 = 0 ms (impact time) and t = 300 ms (end of measurement). However, the light cannot be switched off completely, as the camera images are relevant at all times (even at t0 = 0 ms, event time). The pulse width of the pulsed light is preferably chosen to be wider (e.g., 500 ms) with a shorter exposure time for the camera devices of the first measurement system. The reason for this is that if the signals are shifted, the images become darker over time (until they are unexposed). Since the camera images are very relevant in the safety test, the risk of recording with the same pulse width is too high.Minor signal shifts can still be compensated for with wider light pulses. The actual (preliminary) check of the synchronism of the first measurement system with the second measurement system, or of the sensors of the respective first and second measurement systems with each other, is performed in step 1 anyway. In the experiment itself, only the peaks (especially at t0 = 0 ms and t = 300 ms) that are detectable even in the non-static state with a sufficiently large change in light intensity can be evaluated. This allows the position of the trigger event over time and the overall synchronism to be estimated for a period of 300 ms.

[0022] Overall, numerous advantages can be achieved using the method presented here.

[0023] The method makes it possible to detect a deviation from the start of measurement of the measuring systems used during the safety test at time t0 = 0 ms (event time) and at at least one other sampling point within a measurement duration typical for safety tests (in particular, approximately 300 ms) by detecting asynchronous recordings from the camera equipment. This is achieved by determining the position of the peaks in the pixel sum signals at the sampling points, at least at the beginning and end of the safety test. In particular, the evaluation of the measurements before the actual safety test in the static state (v = 0 m / s) is also used to detect possible influences and setting errors before the actual safety test.

[0024] In addition, the method allows an evaluation of the peak positions of all camera devices among each other, so that all camera devices of the first measurement system used in the safety test can be checked for correct settings.

[0025] By evaluating the gradient in the preferably filtered pixel sum signals of the individual sensor devices and camera devices, it is also possible to determine whether the exposure duration of the camera devices matches the duration of the light pulses emitted by the illumination device and whether these are synchronized to this extent. Should a shift occur here, even with the same exposure duration of the camera devices and the emitted light pulse width, a gradient > or < 0 would be detectable in the filtered pixel sum signals.

[0026] A further advantage is that, in particular with the aid of optical sensor means, in particular brightness sensor means, and electrical sensor means of the second measuring system, it can be checked whether the second measuring system and the first measuring system operate synchronously with one another.

[0027] A further advantage arises from the fact that storing images with people in them is prohibited due to data protection regulations. The analysis described above allows a person in the measurement setup to verify the external trigger (manual activation) before the safety test, while simultaneously evaluating the synchronization of the sensors and camera devices used and the absolute position of the measurement start time at time t0 in compliance with data protection regulations for forwarding the measurement results to third parties.

[0028] Further features and advantages of the present invention will become clear from the following description of a preferred embodiment with reference to the accompanying drawings. Fig. 1 a graphical representation illustrating the basic timing of a safety test and the associated measurement duration, Fig. 2 a graphical representation illustrating an asynchronous recording of light pulses by means of a camera device with an equally long exposure time of the camera device, Fig. 3 a graphical representation of filtered pixel sum signals from camera devices of a first measuring system and a sensor means of a second measuring system.

[0029] With reference to Fig. Figure 1 illustrates the basic chronological sequence of a safety test, particularly a crash test, on a motor vehicle. Such safety tests typically employ two separate, higher-level measurement systems. A first measurement system comprises all image-recording camera devices used in the safety test. A second measurement system, in contrast, primarily comprises non-optical sensor devices and some optical sensor devices used in the safety test.

[0030] The recordings of the camera devices of the first measuring system and the sensor means of the second measuring system during the crash test begin at a time t = -20 ms before the actual impact of the motor vehicle on a collision barrier, which occurs at a time t = 0 ms (time of impact), and then usually last for about 300 ms, so that a typical measurement and evaluation period when carrying out the actual safety test ranges from t = -20 ms to t = 300 ms and is evaluated accordingly.

[0031] At time t = -20 ms, which defines the starting point of the safety test and can be triggered either automatically or manually, the image-recording camera devices of the first measuring system and the sensors of the second measuring system ideally begin their measurements simultaneously, so that appropriate synchronization should be ensured between all measuring sensors (i.e., all sensors and camera devices of both measuring systems). The corresponding measurement data are recorded up to a time t = t1 (for example, up to a time t1 = 300 ms), which represents the end of the safety test, and processed and evaluated by corresponding evaluation devices of the two measuring systems or, alternatively, by a central evaluation device jointly for both measuring systems.

[0032] When carrying out the safety tests, the synchronicity of the two measuring systems with each other and, in particular, the synchronicity of the camera devices of the first measuring system are particularly important for calculating the trajectories of different vehicle points.

[0033] In the following, a procedure is explained in more detail which makes it possible to check these synchronicities in a simple and reliable way.

[0034] A first aspect of the method is that, before the actual safety test begins and measurements are started by all measuring sensors, a lighting device in the test facility emits pulsed light with a pulse pattern defined in terms of intensity and frequency, which is then recorded by the camera devices of the first measuring system. The generation of the pulsed light with the defined pulse pattern begins well before the start of the actual safety test (typically approximately 1 hour to 30 minutes before the actual start of the test). This pulsed light is recorded by the relevant camera devices and evaluated by the evaluation device. Before the actual start of the safety test, among other things, the pixel sums of the individual images from the relevant camera devices are determined in the static state of the vehicle (v = 0 m / s).For this purpose, the pulsed illumination system of the test facility is used as a synchronization means—pulsed by a central synchronization signal, possibly with an additional redundant timing system. At specific points in time, which form support points, the emission of the pulsed light is briefly suspended. Alternatively, the radiant power of the pulsed light emitted by the illumination system can be deliberately reduced or increased at these times. In a further embodiment, it is also possible to generate an additional light signal at these times.

[0035] Through these intensity changes, which preferably have a duration corresponding to the pulse width of the pulsed light, corresponding sampling points are obtained at these times. The defined pulse pattern in the image sequences of the images captured by the camera devices is identified by forming pixel sum signals. This check is carried out to determine whether the temporal position of peaks in the pixel sum signals corresponds to the temporal positions of the sampling points. The static state of the vehicle enables comparability of the brightness values ​​within a stack of images. In addition, the exposure time of the camera devices is set to the same pulse width of the lighting device. By forming the pixel sum for each individual image, an evaluation of the gradient and the peaks of the pixel sum signals at the sampling points is possible.

[0036] Furthermore, the lighting device can be used to generate additional bright or dark scenes, preferably with a duration corresponding to the pulse width of the pulsed light. This provides additional sampling points at defined times, which can also be used to evaluate synchronicity before conducting the actual safety test.

[0037] This makes it possible to check whether all camera devices of the first measuring system are working synchronously with each other.

[0038] During the actual safety test, the radiant power of the pulsed light emitted by the illumination device is reduced or increased at time t = t0 = 0 ms (event time / impact time) and at time t = t1 = 300 ms (end of the test). In another alternative embodiment, it is also possible to generate an additional light signal at these times. These intensity changes, which preferably have a duration corresponding to the pulse width of the pulsed light, produce corresponding additional sampling points—an additional first sampling point and a further second sampling point—at times t0 and t1. The defined pulse pattern in the image sequences of the images captured by the camera devices is identified by generating pixel sum signals.In this case, it is checked whether the temporal position of peaks in the pixel sum signals corresponds to the temporal positions of the support points or not, whereby at least the first support point, which represents the time of the impact of the motor vehicle, and the second support point, which represents the time of the end of the safety test, are used as support points.

[0039] In the time interval between the two times t0 and t1, additional bright or dark scenes can be generated using the lighting device, preferably with a duration corresponding to the pulse width of the pulsed light. This provides additional sampling points at defined times within the time interval between the event time t0 and the end time t1 of the safety test, which can also be used to evaluate synchronicity.

[0040] Fig. Figure 1 shows the basic measurement sequence using the illumination device as a synchronization means (at least two sampling points, in this case at t0 = 0 ms and at t1 = 300 ms, recording frequency 1000 Hz). For easier signal filtering, the camera devices begin recording at t = -20 ms before the trigger event (total light output 1a). The maximum recording time for each individual image recorded by the camera devices is 1 ms at a recording frequency of 1000 Hz. At times t = 0 ms and t = 300 ms, the respective total light output 1b, 1c exhibits deviations resulting from the defined intensity changes at these times and which can be used to determine synchronicity.

[0041] Fig. Figure 2 illustrates an asynchronous recording of the light pulses generated by the illumination device using a camera device. Below, Fig. 2 to detect the light pulses 100a, 100b, 100c of the illumination device with a constant pulse duration of 1 ms and a defined amplitude due to the timing. In Fig. Figure 2 above shows the image 200a, 200b, 200c of these light pulses, recorded by one of the camera devices of the first measurement system. The camera device's exposure times are identical, but the times between the pulses are different. The image 200a, 200b, 200c of the camera device is shifted by a few microseconds and thus time-delayed, resulting in an undesirable asynchrony.

[0042] To verify the synchronization of the first measuring system, to which the camera devices are assigned, with the second measuring system, to which the remaining sensor means of the test setup are assigned (primarily without visual sensor means), current sensors for measuring the electrical power consumption of the lighting device and / or brightness sensors positioned in the test facility are preferably used to measure and compare the position of the peaks. These sensor means of the second measuring system measure the time-dependent power consumption of the lighting device and / or the brightness in the test hall in parallel with the measurements of the camera devices of the first measuring system, preferably at a frequency of at least 1000 Hz. Here, too, the recording time within a cycle is set equal to the light pulse length.The measurement signals obtained in this way can also be used to perform peak analyses for each of the sensors used. This allows the synchronization of the data between the individual systems to be assessed both before and during the test.

[0043] Preferably, the measurement signals from the sensors used in the second measurement system are additionally subjected to low-pass filtering. This low-pass filtering advantageously reduces unwanted noise caused, for example, by minimal vehicle movements or bystanders.

[0044] In Fig. 3 shows corresponding measurement signals 300 of a sensor means and pixel sum signals 400, 500 as examples. The top of the figure shows Fig. 3 shows a low-pass filtered measurement signal 300 from a brightness sensor of the second measuring system, and in the center, a low-pass filtered pixel sum signal 400 from a first camera device of the first measuring system. Both are synchronous with each other and correctly capture the sampling points at times t = t0 and t = t1. The corresponding peaks of the two signals 300, 400 are located at the correct time positions. It can also be seen that the gradients of the filtered signals 300, 400 are 0 between the two sampling points.

[0045] In comparison, Fig. 3 below shows the filtered pixel sum signal 500 of a second camera device of the first measuring system. This pixel sum signal 500 is not temporally synchronized with the two remaining signals 300, 400 of the brightness sensor means of the second measuring system and the first camera device of the first measuring system, respectively. The times of the two sampling points, which should have been recorded at times t = t0 and t = t1, are incorrect here, and the slope of the filtered pixel sum signal 500 between the two sampling points is < 0.

[0046] The evaluation of the peak positions of the pixel sum signals of all camera devices relative to each other makes it possible to check the correct settings of the camera devices used during the safety test. By evaluating the gradient of the individual pixel sums, it can be checked whether the exposure time of the respective camera device matches the exposure time of the lighting device and is thus synchronized. If a shift occurs here, a gradient > 0 or < 0 in the filtered pixel sums of the individual images - as in Fig. 3 below - can be seen. The correspondence of the signals 300, 400 of the brightness sensor means of the second measuring system and the first camera device of the first measuring system in Fig. 3 illustrates their synchronicity.

Claims

[1] Method for determining a synchronicity of a plurality of camera devices of a first measuring system and sensor means of a second measuring system in a safety test, in particular a crash test on a motor vehicle, wherein - prior to carrying out the safety test in a static state of the motor vehicle, pulsed light with a defined pulse pattern is emitted by an externally synchronized lighting device, which is detected by the camera devices of the first measuring system and by sensor means of the second measuring system, which are designed to measure brightnesses and / or electrical signals of the lighting device, wherein the defined pulse pattern is identified in the image sequences of the images captured by the camera devices by forming pixel sum signals and in the measurement signals of the sensor means, and it is checked whether the temporal position of peaks in the pixel sum signals and in the measurement signals corresponds to the temporal positions of a plurality of defined support points, and - during the performance of the safety test, pulsed light with a defined pulse pattern is emitted by the lighting device, wherein the intensity of the light is changed in a defined manner at a first point in time, which in particular defines the point in time of impact of the motor vehicle on the collision barrier, and at least at a second point in time, which in particular represents the end of the safety test, and it is checked by peak analysis whether the intensity changes were recorded simultaneously by the camera devices in a defined static area of ​​the camera image and by the sensor means. [2] Method according to claim 1, characterized by that before carrying out the safety test, the radiant power of the pulsed light emitted by the lighting device is briefly changed at the support points. [3] Method according to one of claims 1 or 2, characterized bythat, before carrying out the safety test, additional defined support points are used, which are generated in the form of brighter or darker scenes by the lighting device. [4] Method according to one of claims 1 to 3 characterized by that a low-pass filtering of the pixel sum signals of the camera devices of the first measuring system is carried out. [5] Method according to one of claims 1 to 4, characterized by that a low-pass filtering of the measuring signals of the sensor means of the second measuring system is carried out. [6] Method according to one of claims 1 to 5, characterized by that the slope of the pixel sum signals is evaluated.

Images