CN105043539B - Method and device for operating a photodetector - Google Patents

Abstract

The invention relates to methods and devices for operating photodetectors. A method for operating a photodetector (100) is provided. The photodetector (100) has a plurality of individually activatable avalanche photodiodes (104) per image point (102). The avalanche photodiode (104) is configured in an activated state to provide an electrical pulse when receiving an amount of light. The avalanche photodiode (104) is light insensitive during the regeneration period after the electrical pulse. The method has the step of activating at least two of the avalanche photodiodes (104) of the image point (102). Here, the avalanche photodiodes (104) are activated offset from each other by a duration that is less than the regeneration period.

Description

Method and device for operating a photodetector

technical field

The invention relates to a method for operating a photodetector, a corresponding device and a corresponding computer program product.

Background technique

Photodiodes require a structurally determined amount of light in order to provide a minimum brightness value. Avalanche photodiodes can convert significantly smaller amounts of light into electrical signals.

DE 10 2009 029 376 A1 describes a photon detector with a paralyzable photon sensitive element and a distance measuring device with such a photon detector.

Contents of the invention

Against this background, a method for operating a photodetector, a device using the method and finally a corresponding computer program product are described according to the independent claims using the approach presented here. Advantageous refinements are given by the respective subclaims and the ensuing description.

Avalanche photodiodes (also known as single-photon avalanche diodes) trigger an electrical pulse when a minimum amount of light falls on the light-sensitive area of the avalanche photodiode. Light volumes are already achievable in the case of single photons. After an electrical pulse has been provided, the avalanche photodiode requires a fixed time until it is ready again to provide another electrical pulse in response to the incidence of a minimum amount of light. Light cannot be recorded during this time.

According to the approach presented here, the time of insensitivity or blindness is reduced by releasing an additional avalanche photodiode within which the minimum amount of light in the electrical pulse is imaged.

Introducing a photodetector having a plurality of individually activatable avalanche photodiodes per image point, wherein the avalanche photodiodes are configured in the activated state to provide an electrical pulse when receiving a light quantity, and the avalanche photodiodes after providing the electrical pulse Light insensitive during the regeneration cycle.

Furthermore, a method for operating a photodetector is presented, wherein the method has the following steps:

At least two of the avalanche photodiodes of the image point are activated, wherein the avalanche photodiodes are activated offset from each other by a period of time, wherein the period of time is less than the regeneration period.

A photodetector can be understood as an image sensor. A photodetector can have multiple image points or pixels. The image points can be arranged in a planar matrix. Avalanche photodiodes can be understood as single-photon photodiodes. The amount of light can depend on the implementation of the avalanche photodiode. The amount of light can be small so that a single photon hitting the avalanche photodiode can trigger a pulse. The regeneration period may be the length of time required for the avalanche photodiode to be ready to provide another pulse after providing a pulse. The duration may be the reciprocal of the sampling rate of the photodetector.

The avalanche photodiodes of the pixels can be activated in time sequence. After the last avalanche photodiode of a pixel has been activated, the first avalanche photodiode of a pixel can be reactivated. The avalanche photodiodes of the pixels can be activated sequentially. Continuous operation of the photodetectors is achieved by sequential activation.

The duration may be greater than the reciprocal of the number multiplied by the regeneration period in order to enable uninterrupted detection. The duration may depend on the number of avalanche photodiodes. The more avalanche photodiodes a pixel has, the shorter this time period can be set.

If the regeneration phase should last longer, this duration can be extended by a small duration, so that the first avalanche photodiode of a pixel and then all other avalanche photodiodes go through an idle phase as a safety reserve after their regeneration period. Continuous operation can thus be ensured.

The method may have the step of deactivating the avalanche photodiode. After the activation duration of the avalanche photodiode that started when the avalanche photodiode was activated has expired, a deactivation step can be carried out in order to avoid false detections. The activation duration may be how long the avalanche photodiode should remain activated. After the activation duration, a nominal regeneration period may be waited to elapse before resuming activation.

At least two avalanche photodiodes per image point can be activated simultaneously in order to quantify the amount of light detected. Here, one of the avalanche photodiodes can be triggered when a desired amount of light, for example one photon, is received. When the amount of light is twice as large, for example comprising two photons, two avalanche photodiodes can be triggered. The more avalanche photodiodes are activated simultaneously, the more gradations of intensity can be distinguished.

The method can start in response to a starting point in time. This duration can be variable. The starting time point may be, for example, the sending time point of the light pulse. Time can be measured from a starting point in time until one of the avalanche photodiodes of the pixel supplies an electrical pulse. From the elapsed time and the speed of light in the penetrating medium, the spatial path traveled by the light pulse can be determined. When the light pulse is reflected at the object, the distance to the object corresponds approximately to half a distance depending on the geometry of the photodetector.

The method can have the step of determining the starting point in time using the runtime signal. The runtime signal can be a time period that has been determined between emission of a light pulse and receipt of a light pulse in the case of reception of a temporally elapsed light pulse. That is, run-time signals can represent fundamental knowledge. The starting point in time of the method can thus be adapted if its light pulse is expected with low probability within this duration.

The method may have the step of setting a duration, wherein the duration is set smaller at a point in time closer to the start than at a point further away from the start in order to obtain a duration closer to the point in time than further away from the start. Greater temporal resolution at time points. A finer spatial resolution can be achieved by shorter durations closer to the starting time point, so detection can be performed with a coarser resolution in larger distances. With a finer resolution, the amount of light within a duration with a constant light intensity becomes smaller than with a coarser resolution. In other words, the sensitivity of the photodetector becomes greater with a coarser resolution.

The method can start in response to a starting point in time. More avalanche photodiodes may be activated simultaneously at a starting time point further away from the method than closer to the starting time point in order to obtain a greater sensitivity with increasing distance from the starting time point. At the same time, the more active avalanche photodiodes, the larger the possible impingement area for light becomes. Since the light pulse to be received is expected to have a lower intensity further away from the starting point in time, since it is attenuated by the medium, reliable reception can still be ensured by the greater sensitivity.

Furthermore, a device for operating a photodetector with a plurality of individually activatable avalanche photodiodes per pixel is presented, wherein in the activated state the avalanche photodiodes are designed to provide electrical pulse, and the avalanche photodiode is light-insensitive during the regeneration period following the electrical pulse, wherein the device has the following characteristics:

A device for activating at least two of the avalanche photodiodes of an image point, wherein the avalanche photodiodes are activated offset from each other by a period of time, wherein the period of time is less than the regeneration period.

The object on which the invention is based can also be solved quickly and efficiently by this embodiment variant in the form of the device of the invention.

A device can be understood as an electrical device which processes sensor signals and outputs control and/or data signals as a function of the sensor signals. The device can have an interface, which can be embodied in hardware and/or software. In the case of a hardware configuration, the interface can, for example, be part of a so-called system ASIC, which contains the various functions of the device. However, it is also possible for the interface to be a dedicated, integrated circuit or to consist at least partially of discrete components. In the case of a software configuration, the interface can be a software module which, for example, exists on a microcontroller in addition to other software modules.

Also advantageous is a computer program product or computer program with a program code which can be stored on a machine-readable carrier or storage medium, such as a semiconductor memory, hard disk memory or optical memory, and in particular when the program product or program When implemented on a computer or device, it can be used to execute, realize and/or control the steps of the method according to one of the above-described embodiments.

Description of drawings

The concept presented here is explained in more detail below by way of example with reference to the figures.

Figure 1 shows a diagram of a photodetector according to one embodiment of the present invention;

Figure 2 shows the time course of light detection in the case of a single avalanche photodiode;

3 shows a diagram of the detection times of an avalanche photodiode in the case of differently shaped light pulses;

FIG. 4 a shows a diagram of an image point with individually activatable avalanche photodiodes according to an exemplary embodiment of the invention;

Figure 4b shows a graphical representation of different light pulses detected by an avalanche photodiode activated offset by a duration according to one embodiment of the present invention;

FIG. 5 a shows a diagram of an image point with avalanche photodiodes that can be activated in groups according to an exemplary embodiment of the invention;

Figure 5b shows a graphical representation of different light pulses detected by groups of avalanche photodiodes activated offset by a duration according to one embodiment of the present invention;

FIG. 6 a shows a diagram of an image point with different numbers of avalanche photodiodes that can be activated according to an exemplary embodiment of the invention;

Figure 6b shows a graphical representation of different light pulses detected by different numbers of avalanche photodiode groups according to one embodiment of the present invention;

Figure 6c shows a graphical representation of different light pulses detected by different sized avalanche photodiode groups activated for different durations in accordance with one embodiment of the present invention;

Fig. 7a shows a diagram of an image point with avalanche photodiodes activated for different durations that can be shifted according to an embodiment of the present invention;

Figure 7b shows a graphical representation of different light pulses detected by groups of avalanche photodiodes activated for different durations in accordance with one embodiment of the present invention;

Figure 8 shows a graphical representation of light pulses detected by different sized avalanche photodiode groups at different sized intervals in accordance with one embodiment of the present invention;

Figure 9 shows a flow chart of a method for operating a photodetector according to one embodiment of the present invention; and

FIG. 10 shows a block diagram of a device for operating a photodetector according to an exemplary embodiment of the invention.

In the subsequent description of advantageous exemplary embodiments of the invention, identical or similar reference symbols are used for elements shown in different figures and having a similar effect, wherein a repeated description of these elements is omitted.

Detailed ways

FIG. 1 shows a diagram of a photodetector 100 according to an exemplary embodiment of the invention.

Photodetector 100 is shown schematically and here has, for example, four image points 102 or pixels 102 . Photodetector 100 may be referred to as a SPAD imager 2x2 and may record up to 100 events per count cycle. Pixels 102 are arranged in a rectangular grid consisting of rows and columns. Pixels 102 can be referred to as SPAD pixels 5×5 and can record up to 25 events per counting cycle. Each pixel 102 has a plurality of individually activatable avalanche photodiodes 104 per pixel 102 (only one avalanche photodiode 104 per pixel 102 is assigned a reference numeral for clarity). Avalanche photodiodes 104 may be referred to as SPAD microcells and each record one event per count cycle.

In this embodiment, the photodetector 100 has 25 avalanche photodiodes 104 per pixel. The avalanche photodiodes 104 are likewise arranged in a rectangular grid of rows and columns, with each row and column having five avalanche photodiodes 104 arranged side by side or one above the other. That is, pixel 102 is approximately square. In the activated state, the avalanche photodiodes 104 are configured to provide one electrical pulse per avalanche photodiode 104 receiving an amount of light. The amount of light may correspond to a single photon. Electrical pulses can be triggered by the amount of light. After the avalanche photodiode 104 has output an electrical pulse, it is light insensitive, ie blind, during the regeneration period. Photons incident during this time do not trigger the pulse. When the avalanche photodiode 104 is in a deactivated state, incident light does not trigger an electrical pulse even above the amount of light used to trigger the electrical pulse.

The SPAD receiver 100 can be implemented in different ways. In its simplest form, it consists of a single SPAD unit 104 . The SPAD unit can record or count the arrival of individual photons. However, due to the internal reset time of a few nanoseconds, such as 10 nanoseconds up to 50 nanoseconds, only limited count rates, such as 20 megacounts per second (MCounts/s) to 100 MCounts/s are possible, otherwise the SPAD Cell 104 is in saturation. Furthermore, individual SPAD units 104 do not record within the count whether a single photon or multiple photons have arrived.

According to one embodiment, a single avalanche photodiode 104 has no threshold for the amount of light used to trigger the electrical pulse. Every single photon can trigger activation. The background light thus in principle produces a certain disturbance count rate or activation per time unit. Furthermore, the activation can also be thermally induced, which can be referred to as the dark count rate. However, it is assumed that by means of effective light emitted by an effective light source, such as reflected laser pulses, a measurable increase in the total count rate or a significantly increased count rate is achieved and thus a distinction can be made between the interfering light count rate and/or the dark count rate or The noise component of the received power versus the effective count rate or signal component of the received power. In advantageous cases, the effective count rate is significantly greater than the interference light/dark count rate. In adverse situations, the relationship is reversed. However, the signal-to-noise ratio can be increased by the long measurement time with a large number of repetitions of the measurement and thus pulses can then also be detected.

To achieve higher sensitivity, multiple SPAD units 1104 may be operated side by side, eg in parallel in a matrix. This increases the effective area and records more photons. However, directional selectivity or angular resolution can also be achieved by means of a matrix arrangement, analogous to the imaging on multiple pixels in the case of a camera. A matrix arrangement of SPAD units 104 can be connected such that the number of arriving photons can be recorded. A SPAD imager 100 with high count rate and high spatial resolution can be achieved by the increased number of SPAD (macro) pixels 102 and microcells 104 within the superstructure 100 from multiple macro pixels 102 to the SPAD imager 100 , similar to the number of pixels in a camera. A high insensitivity to false counts due to ambient light can be achieved by the high count rate.

For example, the photodetector 100 can be used in connection with a driver assistance system of a vehicle. In driver assistance systems a large number of comfort functions such as automatic cruise control, parking guidance, lane keeping assistant, traffic sign recognition; and safety functions such as progressive safety control, lane departure warning have been installed on the market in recent years. Currently, highly automated driving functions are developed, in which responsibility is increasingly shifted from the driver to the vehicle. To achieve this, the most reliable sensor systems with an almost seamless and error-free detection of the vehicle environment are required. LiDAR sensors can meet this requirement.

Such systems are of particular interest in the case of different lidar technologies, in which light pulses are emitted in different spatial directions and the light reflected by the environment is detected by means of the most sensitive SPAD receivers (Single Photon Avalanche Diodes, Each photon) to receive.

FIG. 2 shows the time course of the detection of light in the case of a single avalanche photodiode. An avalanche photodiode is, for example, a component of a pixel, as shown in FIG. 1 . The flow is shown on the time curve. Here, time is depicted on one axis. Transverse to this axis provides the strength of the signal. If the avalanche photodiode is effectively turned on and an amount of light 200 falls on the avalanche photodiode, the avalanche photodiode triggers an electrical pulse 202 . In other words, the SPAD is ignited when a photon arrives. The intensity of the electrical pulse 202 is predetermined by the design of the avalanche photodiode. After the avalanche photodiode has output a pulse 202, during a regeneration period 204 or reset time 204, the avalanche photodiode becomes insensitive. The avalanche photodiode cannot provide an electrical signal during the regeneration period 204 even though the desired amount of light 200 falls on the avalanche photodiode. In other words, the SPAD does not record when a photon arrives. Even if a double amount of light 206 falls on a ready-to-receive and effectively switched-on avalanche photodiode, only electrical pulses 202 of a structurally determined size or intensity are provided. In other words, even if two photons arrive, the SPAD is simply ignited.

When a stronger light pulse arrives, for example due to a laser pulse reflected at a nearby object, the counting takes place at the beginning of the pulse or at the rising edge of the pulse.

FIG. 3 shows a diagram of detection times 300 of an avalanche photodiode with differently shaped light pulses 302 , 304 , 306 . The detection times 300 are represented in a graph which provides the time on the abscissa and the signal strength on the ordinate. The light pulses 302, 304, 306 respectively start and end at minimum intensity. Light pulses 302 , 304 , 306 each have an intensity maximum 308 which occurs approximately after half the pulse duration. The first light pulse 302 has a short pulse duration and a small maximum intensity 308 . The second light pulse 304 has a small pulse duration and a large maximum intensity 308 . Third light pulse 306 has a greater pulse duration and a greater maximum intensity 308 , wherein third light pulse 306 has a lower maximum intensity 308 than second light pulse 304 . The detection time point 300 is shortly after the start of one of the light pulses 302 , 304 , 306 , respectively, when the intensity of the light exceeds the detection limit. In other words, the SPAD ignites at the beginning of the pulse. That is, the light pulses 302, 304, 306 are detected with a small time offset from the actual onset. The shape of the light pulses 302, 304, 306 is not detected because a single avalanche photodiode outputs its electrical pulse 202 when it records the amount of light and cannot provide further pulses during its regeneration period immediately afterwards. The electrical pulse 202 has a different time offset from the maximum value 308 depending on the shape of the light pulses 302 , 304 , 306 .

The approach presented here makes it possible, for example, not only to determine rising pulse edges 300 in lidar systems, but to detect and evaluate the overall pulse shape. The change in the shape of the received pulse relative to the shape of the transmitted pulse gives the extent and position to atmospheric disturbances, such as rain, fog, snow, spray, or reflective object surfaces, such as "soft" bushes, "hard" cars, " A hint of a long, sloping" wall.

FIG. 4 a shows a representation of an image point 102 with individually activatable avalanche photodiodes 104 according to an exemplary embodiment of the invention. Pixel 102 essentially corresponds to one of the pixels in FIG. 1 . Image point 102 here has 16 individually activatable avalanche photodiodes 104 .

In other words, Fig. 4a shows SPAD pixels 4x4. Up to 16 events per count cycle are spread equidistantly over 32 ns in a two nanosecond grid.

In one embodiment, the SPAD pixel 102 is composed of 16 SPAD units 104 whose individual reset times are, for example, 32 ns. By activating and deactivating the SPAD unit 104 stepwise at intervals of 2 ns (corresponding to a distance of 15 cm), the SPAD unit 104 is active within 2 ns, respectively. After traversing all SPAD units 104 , restart with the first SPAD unit 104 . This mode of operation can be referred to as rolling. Interval scanning or interval sampling is achieved in this manner. If a light pulse arrives, the pulse duration can be detected with this embodiment, but not yet the pulse height.

Fig. 4b shows a diagram of different light pulses 302, 304, 306 detected by avalanche photodiodes activated offset by a duration according to an embodiment of the present invention. Here, light pulses 302 , 304 , 306 are detected by a pixel with 16 avalanche photodiodes as in FIG. 4 a . In this embodiment, each avalanche photodiode has a regeneration period 204 of 32ns. The avalanche photodiodes are activated with a two-nanosecond offset from each other. Thus, the avalanche photodiode may output an electrical pulse 202 , pass through its regeneration period 204 or reset time 204 and be activated again directly thereafter. The duration can also be greater than the inverse of the number 16 multiplied by the regeneration period of 32 ns in order to enable uninterrupted detection. The avalanche photodiode is deactivated again after the expiry of a period of two nanoseconds, even though it has not provided electrical pulses, in order to avoid false detections. Thus, this duration corresponds to the activation duration per cycle of the avalanche photodiode. The avalanche photodiode then remains deactivated for a duration of inactivity corresponding to regeneration period 204 until the next activation. The remaining 15 avalanche photodiodes of the pixel are activated and deactivated offset such that one of the avalanche photodiodes of the pixel is offset by at least two nanoseconds in the active state and is therefore ready to receive. Once the light pulses 302, 304, 306 result in a significantly increased count rate, a two nanosecond time resolution results by shifting the avalanche photodiode by two nanoseconds upon activation. That is, the pulse length 400 of the light pulses 302, 304, 306 can be determined exactly to two nanoseconds. The avalanche photodiodes of the pixels are activated here in chronological order. In this case, the first avalanche photodiode of the pixel is reactivated after the last avalanche photodiode of the pixel has been activated.

FIG. 5 a shows a diagram of an image point 102 with avalanche photodiodes 104 that can be activated in groups according to an exemplary embodiment of the invention. Pixel 102 essentially corresponds to one of the pixels in FIG. 1 . In contrast, pixel 102 here has 16 individually activatable diode groups 500 . The diode groups 500 are composed of four avalanche photodiodes 104 , respectively. That is to say that at least two avalanche photodiodes 104 per image point 102 can be activated simultaneously in order to quantify the detected light quantity. In other words, Figure 5a shows a SPAD pixel 4x4x4, where four SPADs can be activated simultaneously.

In one embodiment, a SPAD pixel 102 is composed of 64 SPAD units 104, four of which are each activated at the same time. As a result, pulse heights can be detected in four height levels in addition to pulse durations.

A SPAD receiver 102 is presented using a special circuit technology with which the signal shape of the incoming light can be detected. In addition, the circuit technique presented here enables a large dynamic range.

In this case, the individual SPAD units 104 are activated or deactivated temporally in a targeted manner. In this way, some SPAD units 104 can record the rising edge of a pulse, for example, while other SPAD units 104 are activated at a slightly delayed point in time and likewise record the pulse or the pulse shape. Pulses can thus be sampled in time by multiple different delays. In this case, the number of SPAD units 104 assigned to the delay can be variably designed with respect to time, so that a higher dynamic range and/or finer sampling and thus measurement accuracy can be achieved.

Fig. 5b shows a diagram of different light pulses 302, 304, 306 detected by groups of avalanche photodiodes activated offset by a duration according to an embodiment of the present invention. This group corresponds to the diode group in Fig. 5a. The light pulses 302, 304, 306 correspond to the illustrations in FIGS. 3 and 4b. As in FIG. 4 b , the diode groups are activated for a duration offset by two nanoseconds from each other and deactivated again after two nanoseconds in order to enable a quasi-continuous detection of the light pulses 302 , 304 , 306 . Additionally, four intensity levels 502, 504, 506, 508 can be distinguished because either one avalanche photodiode provides an electrical pulse 202 within the two nanoseconds, two avalanche photodiodes together provide Two electrical pulses 202 , three avalanche photodiodes jointly provide three electrical pulses 202 within the two nanoseconds, or four avalanche photodiodes jointly provide four electrical pulses 202 within the two nanoseconds. Here, four pulses 202 represent the highest intensity level 508 ; three pulses 202 represent the second highest intensity level 506 ; two pulses 202 represent the third highest intensity level 504 and one pulse 202 represents the lowest intensity level 502 . No pulses represent no intensity level, since then no amount of light is received. In addition to the pulse length 400 , the pulse shape of the light pulse can be detected via the intensity levels 502 , 504 , 506 , 508 .

A large dynamic range can be covered by the scheme presented here. Up to 140 dB can be achieved here. In this way, extremely weakly reflecting objects with extremely few photons at long distances and extremely strongly reflective objects with extremely many photons in the near field can be detected simultaneously without measuring errors in the distance measurement Inaccuracy or the receiver is no longer capable of measuring due to saturation.

In the approach presented here, the temporal pulse position and thus the measured runtime corresponding to the interval value can be determined precisely because the overall pulse shape is detected.

The temporally delayed activation or deactivation of the SPAD units enables the sampling of the optical signals 302 , 304 , 306 by the targeted, temporal activation and deactivation of individual SPAD units in a complex of several SPAD units. .

FIG. 6 a shows a diagram of an image point 102 with a different number of avalanche photodiodes 104 that can be activated according to an exemplary embodiment of the invention. Pixel 102 essentially corresponds to the pixel in FIG. 5a. In contrast, the image point has 64 diode groups 500 each consisting of four avalanche photodiodes 104 . That is, a pixel 102 has 264 avalanche photodiodes 104 . The avalanche photodiodes 104 of the diode group 500 can be activated jointly. Diode groups 500 can be controlled individually. Here too, multiple diode groups 500 can be activated jointly and simultaneously. The diode groups 500 form an 8x8 matrix. In other words, Figure 6a shows SPAD pixels 4x4x(4-16), where four to 16 SPADs can be activated simultaneously.

In one embodiment, the SPAD pixel 102 consists of 256 SPAD units 104, of which initially four are each activated at the same time, but their number increases to 16 over time. The receiver thus becomes more sensitive to reflections that are further away, generally weaker. An increase in dynamic range ensues.

Figure 6b shows a graphical representation of different light pulses 304, 302 detected by different numbers of avalanche photodiode groups according to one embodiment of the present invention. The light pulses 302 , 304 correspond to the light pulses in FIG. 3 . That is, the light pulse 302 has a small pulse length 400 and a small maximum intensity as in FIG. 4 . Likewise, light pulse 304 has a small pulse length 400 and a high maximum intensity as in FIG. 4 . The light pulses 302, 304 are shown in a graph which provides time on its abscissa. The intensity level or pulse height of the light is provided on the ordinate. Time starts at start point 600 . The starting point 600 here represents a point in time at which a light pulse is emitted. Light pulses 302, 304 represent reflected light of emitted light pulses. In this exemplary embodiment, light pulse 304 with a large maximum intensity is detected closer to starting point 600 , ie earlier than light pulse 302 with a small maximum intensity. The different maximum intensities represent the attenuation of emitted and reflected light when penetrating the medium through which the light is transmitted. The light pulse 304 is detected in four intensity levels 502 , 504 , 506 , 508 . In other words, four SPADs are activated, whereby the pixels are less sensitive and less resolved. Light pulses 302 are detected in 16 intensity levels. In other words, 16 SPADs are activated. As a result, light pulse 302 can be resolved more sensitively or better than light pulse 304 . In order to set the different sensitivities, four avalanche photodiodes are simultaneously activated per time step in the case of light pulse 304 and 16 avalanche photodiodes are simultaneously activated per time step in the case of light pulse 302 . That is, more avalanche photodiodes are simultaneously activated further away from the starting point 600 than closer to the starting point 600 in order to achieve greater sensitivity with increasing separation from the starting point 600 .

Fig. 6c shows a graphical representation of different light pulses 304, 302 detected by different sized avalanche photodiode groups activated for different time offsets in accordance with one embodiment of the present invention. This illustration substantially corresponds to the illustration in FIG. 6b. In addition, two known points in time 602 , 604 are used here. Time points 602, 604 are based on basic knowledge. This basic knowledge can be obtained, for example, in previous actions. That is, the starting point 600 is determined using the runtime signal. For example, one light pulse 302 , 304 each has been recorded shortly before two known points in time 602 , 604 . Time points 602, 604 mark the start of light pulses 302, 304, respectively. From the starting point 600 to the first point in time 602 , the pixels are moved with one diode group per time step, resulting in a resolution with four intensity levels. Starting from a first point in time 602 , pixels are operated with four diode groups per time step. This results in a resolution of 16 intensity levels, as in FIG. 6b. Additionally, starting from the first point in time 602, the time step is shortened from 2 ns to half a nanosecond. This results in a more precise time resolution. After the second light pulse 304 has been read in, until a second point in time 604 the pixels are again operated with one diode group per time step, that is to say with four intensity levels. Here, the time step size can again be increased to 2 ns per time step. Likewise, the time step size can be kept at half a nanosecond per time step. From the second point in time 604 onwards, the image sensor is operated again with four diode banks per time step, which in turn leads to 16 possible intensity levels. Here, the time step size is set to 2ns per time step. With an enlarged time step, the avalanche photodiode is effective for longer periods of time. This allows more light to fall on the avalanche photodiode per time step, similar to a video camera with a longer exposure time. That is, the image sensor is now more sensitive.

By targeting and temporally activating and deactivating individual SPAD units in a complex of multiple SPAD units, it is possible to variably design the SPADs activated/deactivated during the runtime of emitted and reflected light the number of units. The number of activated cells may increase with runtime corresponding to the distance of the object. Thus the sensitivity of the receiver to weaker signals of more distant objects can be increased.

In one exemplary embodiment, time points 602 , 604 are known, for example from previous measurements, which are to be detected in a targeted manner. Activation of the SPAD unit first takes place directly before this point in time 602 , 604 . Increase the number of simultaneously active SPAD units and/or decrease the sampling rate. In this way, the sensitivity and/or precision can be increased in a targeted manner by means of the existing basic knowledge.

FIG. 7 a shows a diagram of an image point 102 with an avalanche photodiode 104 activated for different durations that can be shifted according to an exemplary embodiment of the invention. The pixel corresponds to the pixel in FIG. 6 a and has 264 avalanche photodiodes 104 . In contrast, the diode array 500 here includes 16 avalanche photodiodes 104 for 16 intensity levels. That is, a pixel 102 has 16 diode groups 500 . In other words, Fig. 7a shows a SPAD pixel 4x4x16 with an adaptable temporal grid.

In one embodiment, a SPAD pixel 102 is composed of 256 SPAD units 104, 16 of which are each activated at the same time. Initially, the step-by-step activation and deactivation of the SPAD unit 104 is set at 0.5 ns intervals and therefore upsampled. As the time increases or the object is expected to move away, the sampling time is increased to 2ns. In this way, in addition to all the previous advantages, the distance accuracy is increased in the vicinity. Furthermore, the dynamic range is improved again, since fewer photons fall with smaller sampling steps.

Fig. 7b shows a graphical representation of different light pulses 304, 302 detected by groups of avalanche photodiodes activated for different time offsets in accordance with one embodiment of the present invention. The light pulses 302, 304 correspond to the light pulses in FIG. 6b. As in Figure 6c, the light pulses 304 are sampled at intervals of 0.5 ns. The light pulses 302 are sampled at intervals of 2 ns as in FIG. 6c. The interval here depends on the time interval of starting point 600 . The longer the elapsed time from the starting point 600, the larger the interval. The two light pulses 302, 304 are sampled with 16 intensity levels. In other words, 16 SPADs each are activated. A high time resolution results with an interval of 0.5 ns, a lower time resolution results with an interval of 2 ns. That is, the process presented herein begins in response to start time point 600 . Here, the duration is variable. The duration is set smaller closer to the starting time point 600 than further away from the starting time point 600, so that a larger time resolution.

FIG. 8 shows a graphical representation of light pulses 302 detected by different sized avalanche photodiode groups at different sized intervals in accordance with one embodiment of the present invention. Light pulse 302 corresponds to a short light pulse with a small amplitude in FIG. 3 . Light pulses 302 are sampled at 16 intensity levels.

By the targeted, temporal activation and deactivation of individual SPAD units in a complex of several SPAD units, the number of activated/deactivated SPAD units can be adapted highly dynamically to the received signal.

In this case, the control of the activation time and deactivation time as well as the number of activated and deactivated receivers can originate from the evaluation of previous sampled values of previous light signals. That is, a dynamic adaptation to the measurement situation is possible.

Signal shapes such as pulse signals can be detected by means of the approach presented here. The result is an increase in dynamic range and an increase in measurement accuracy.

In one embodiment, a highly dynamic adaptive adaptation of the activation time point and the deactivation time point of the SPAD unit is performed. All SPAD cells are active in the ground state in order to achieve the highest sensitivity. It is checked here whether more SPAD cells are ignited within a predetermined period of time than would be expected by the background light, ie whether a rising pulse edge 800 is possible. The background light is thus faded out and only a small number of SPAD units are in the reset phase. The number of SPAD cells fired at pulse edge 800 represents the pulse height at pulse edge 800 . From this it is determined how many SPAD cells that have not yet been fired should be deactivated as a reserve for the remaining pulses 302 . In addition, the time periods for the next time grid can be adapted and/or dead times can be inserted 802 .

The time grid can be shortened in order to be able to detect rising pulse heights with a smaller dynamic range. Within this next time grid, the number of activated and ignited SPAD units is again determined and the pulse height is derived from this. If the new pulse height is higher, activate eg a larger number of still deactivated SPAD units and/or shorten the effective time grid for the next cycle by inserting larger dead times 802 . If the pulse height is lower, reduce the number and/or reduce the dead time 802 . By controlling the activated number of SPAD units and adapting the dead time 802 by each time grid described here, it is possible to completely detect the pulse 302 with a limited number of SPAD units. Even in the most unfavorable case where the pulse length is greater than the reset time, the sum of initially ignited SPAD units and all activated SPAD units over the pulse duration 400 is no higher than the number of available SPAD units.

In other words, all SPADs are activated before the start of the pulse 800, none of which are triggered. When the number of triggered SPADs is greater than the limit: N1<<N0 SPAD is activated, and the recognition edge starts 800. The number of SPADs triggered after the edge start 800 increases: N3 > N2 > N1 , or the time grid is smaller or the dead time 802 is larger. After the pulse height the number drops: N4>N3, or the time grid is larger or the dead time 802 is smaller. After pulse 302 are activated as SPADs and none are triggered.

FIG. 9 shows a flowchart of a method 900 for operating a photodetector according to an exemplary embodiment of the invention. The photodetector has multiple individually activatable avalanche photodiodes per pixel. The avalanche photodiode is configured in an activated state to provide an electrical pulse when receiving an amount of light. Avalanche photodiodes are light insensitive during the regeneration period after the electrical pulse. Method 900 has an activation step 902 . At least two of the avalanche photodiodes of the image point are activated in an activation step 902 . The avalanche photodiodes are activated with a time offset from each other. This duration is less than the regeneration period.

The scheme presented here describes a dynamic single photon avalanche diode SPAD timing.

Time measurement can be performed by a fast counter, such as a TDC, time-to-digital converter, which is started when the laser beam is emitted and stopped when the pulse reaches the receiver. The counter reading then corresponds to the light travel time and thus to the distance. In this way, the point in time of the respective received pulse can be determined, usually at the rising edge.

In all exemplary embodiments shown here, the photon events can be counted starting from the rising pulse edge within a fixed or variable time grid corresponding to the distance grid. The histogram thus formed simulates the overall pulse shape. For technical realization, for example, a counter library can be used which, after the common start of the emission time points for the laser beams, stops step by step with a set time grid after the detection of a pulse edge and at each interval corresponding to the distance grid Events are counted and stored in steps. The length of the counter bank is significantly smaller than solutions for full time sampling and AD conversion. This reduces the data overhead, required chip area and costs for further processing. If the counter bank is designed rollingly, it is also possible to detect several received pulses of a transmitted pulse in succession and achieve multi-target capability. Histogram contents can be copied into histogram memory extremely quickly. The memory overhead of the histogram memory is much smaller than that for the total sampling. For example, one microsecond of sampling corresponding to 150m requires 1000 memory cells in the nanosecond grid. The counter bank and the histogram require 45 memory cells assuming a pulse duration of 15 ns and three pulses to be stored.

The approach presented here can be used in 2D and 3D lidar sensors for environment detection for driver assistance systems. Also, use in other products is possible. For example, the approach presented here can be used in service robotics, in gesture recognition in the interior of a motor vehicle, in spatial measurement, in workspace monitoring, in construction space monitoring.

FIG. 10 shows a block diagram of a device 1000 for operating a photodetector according to an exemplary embodiment of the invention. In this case, the photodetector corresponds essentially to the photodetector shown in FIG. 1 . The arrangement 1000 has a device 1002 for activating at least two of the avalanche photodiodes of an image point of a photodetector. The device is configured to activate the avalanche photodiodes offset from each other for a duration of time. In this case, the duration is shorter than the regeneration period.

The exemplary embodiments described and shown in the figures are chosen as examples only. Different exemplary embodiments can be combined with each other completely or with regard to individual features. An exemplary embodiment can also be supplemented by features of other exemplary embodiments.

Furthermore, method steps presented here can be carried out repeatedly and in a different sequence than that described.

If an embodiment includes an "and/or" association between a first feature and a second feature, this can be read as such that the embodiment has the first feature and the second feature according to one embodiment and also has the second feature according to another embodiment. Or only have the first feature or only have the second feature.

Claims (12)

1. A method (900) for operating a photodetector (100) having a plurality of individually activatable avalanche photodiodes (104) per image point (102), wherein the avalanche The photodiode (104) is configured in an activated state to provide an electrical pulse (202) when receiving an amount of light (200), and the avalanche photodiode (104) after providing the electrical pulse (202) during a regeneration period ( 204), wherein the method (900) has the steps of: activating (902) at least two of the avalanche photodiodes (104) of the image point (102), wherein the avalanche photodiodes (104) are activated offset from each other by a duration, wherein the duration is less than the regeneration period (204 ). 2. The method (900) according to claim 1, wherein the avalanche photodiodes (104) of the pixels (102) are activated in time sequence, wherein the last avalanche photodiodes (104) of the pixels (102) are activated After the photodiode (104), the first avalanche photodiode (104) of said image point (102) is reactivated. 3. The method (900) according to one of the preceding claims, wherein in the activation step (902) the duration is multiplied by the regeneration period (204) by the reciprocal of the quantity, in order to enable uninterrupted detection. 4. The method (900) according to any one of the preceding claims, having the step of deactivating an avalanche photodiode (104), wherein said avalanche photodiode starts when activating (902) said avalanche photodiode (104) After the activation duration of (104) has expired, a deactivation step is carried out in order to avoid false detections. 5. The method (900) according to one of the preceding claims, wherein at least two avalanche photodiodes (104) per image point (102) are simultaneously activated in the activation step (902) in order to quantify the detected light quantity ( 200). 6. The method (900) according to one of the preceding claims, wherein the method (900) starts in response to a starting time point (600), wherein in the activation step (902) the duration is variable of. 7. The method (900) as claimed in claim 6, comprising the step of determining the starting point in time (600) using a runtime signal. 8. The method (900) according to claim 6 or 7, having the step of setting said duration, wherein closer to said starting time point (600) than further away from said starting time point ( 600) to set the duration smaller in order to obtain a greater temporal resolution closer to the starting time point (600) than further away from the starting time point (600). 9. The method (900) according to any one of the preceding claims, wherein the method (900) starts in response to a starting point in time (600), wherein in an activation step (902) further away from the method At the starting time point (600) of (900), more avalanche photodiodes (104) are activated simultaneously than closer to the starting time point (600), so that as the distance from the starting time point (600 ) for greater sensitivity. 10. A device (1000) for operating a photodetector (100), wherein said photodetector (100) has a plurality of individually activatable avalanche photodiodes (104) per image point (102), wherein the avalanche The photodiode (104) is configured in an activated state to provide an electrical pulse (202) when receiving an amount of light (200), and the avalanche photodiode (104) after providing the electrical pulse (202) during a regeneration period ( 204), wherein the device (1000) has the following characteristics: A device (1002) for activating at least two of the avalanche photodiodes (104) of an image point (102), wherein the avalanche photodiodes (104) are activated offset from each other by a duration, wherein the duration is less than the Regeneration cycle (204). 11. A photodetector (100) with a plurality of individually activatable avalanche photodiodes (104) per image point (102), wherein the avalanche photodiodes (104) are configured in an activated state for when the received light amount ( 200) provides an electric pulse (202), and the avalanche photodiode (104) is light-insensitive in the regeneration period (204) after providing the electric pulse (202), wherein the avalanche of the image point (102) At least two of the photodiodes (104) can be activated offset from each other by a duration, wherein the duration is less than the regeneration period (204). 12. A machine-readable storage medium having stored thereon a computer program set up to carry out all the steps of the method according to claim 1.