Method for obtaining indirect time-of-flight depth map and corresponding sensor

Abstract

Embodiments of the present disclosure relate to a method for obtaining an indirect time-of-flight depth map and a corresponding sensor. In one embodiment, a method for obtaining a depth map by indirect time-of-flight in a photosensitive pixel circuit network divided into pixel circuit groups includes: performing at least one capture during which the pixels of the network are controlled by a demodulation signal; and introducing a phase shift into the demodulation signal with different values distributed in each pixel circuit group.

Description

[0001] Cross - Reference to Related Applications

[0002] This application claims the benefit of French Patent Application No. 2011151, filed on October 30, 2020, which is incorporated herein by reference. Technical Field

[0003] The present disclosure relates to integrated circuits, and more particularly to integrated sensors for obtaining depth maps by indirect time-of-flight (commonly denoted by the acronym "iToF"). Background Art

[0004] Conventionally, a time-of-flight measurement system measures the duration between the emission time of an optical signal (i.e., a light signal) and the time when the signal is received after reflection by an element present in the detection field (i.e., the field of view of the system in the illuminated area). The distance separating the detection field element and the system is reconstructed based on the measured duration, which is proportional to the speed of light.

[0005] An "iToF" indirect time-of-flight system emits an optical signal modulated at a modulation frequency and measures the phase difference of the received signal with respect to the modulation of the transmitted signal. The measurement of the phase difference can be obtained based on the amount of charge generated by the pixels of the receiver during an integration period controlled by a demodulation signal synchronized with the modulation of the optical signal.

[0006] Two samplings offset by 180° (i.e., two consecutive integrations controlled respectively by an in-phase demodulation signal and a demodulation signal offset by 180°) are sufficient to reconstruct the phase shift.

[0007] That is, techniques using more than two samplings, such as the technique known as "4-bin sampling" (using four samplings, where the differential components are offset by 90° with respect to the demodulation signal (i.e., the differential components are at 0° and 180° and the differential components are offset by 90° and 270°), are more commonly implemented because they have advantages, in particular due to the known and mastered signal offset 90° demodulation mechanism ("in-phase (I) / quadrature (Q)" in the field of radio frequency communication).

[0008] Increasing the modulation frequency improves the accuracy of the measured distance but also increases the measurement ambiguity (i.e., the periodicity of different distances results in the same phase difference), thus reducing the measurement range. That is, measurement techniques at several modulation frequencies allow limiting the problems related to ambiguity.

[0009] There is a need to propose systems for obtaining depth maps by indirect time-of-flight and at a high modulation frequency (e.g., greater than 100 MHz (megahertz) or even greater than 200 MHz), with a high resolution, such as approximately one million pixels.

[0010] However, during the integration phase, increasing the modulation frequency and resolution of a conventional system generates very high DC currents. For example, for a 1-megapixel sensor modulated at 200 MHz, it is close to 2 A (amperes), and the peak of the capacitive current also approaches 25 A in approximately one hundred picoseconds.

[0011] On the one hand, these current peaks generate very high "EMI" electromagnetic interference, and on the other hand, they cause the power supply voltage to drop and the ground voltage to rise, which specifically interferes with the switching of logic elements and degrades the synchronization of time-of-flight measurements.

[0012] Therefore, in a system that acquires depth maps with high resolution and high modulation frequency by indirect time-of-flight, it is necessary to propose solutions to overcome the problems caused by the above-mentioned current peaks. Summary of the Invention

[0013] According to an embodiment, a spatial distribution of phase shifts on the demodulation signal is proposed in the photosensitive pixel circuit matrix to disperse electromagnetic interference. The spatial distribution of the phase shift can be defined on a pre-established pixel circuit pattern, and the pattern can be one or more columns, one or more rows, or a rectangular sub-network of pixel circuits.

[0014] According to one embodiment, a method for acquiring a depth map by indirect time-of-flight in a photosensitive pixel circuit network divided into pixel circuit groups is proposed. The method includes: at least one capture, during which the pixel circuits of the network are controlled by a demodulation signal, and the method includes introducing phase shifts with different values into the demodulation signal and distributing them in each pixel circuit group.

[0015] For example, if the acquisition includes four captures, during which the demodulation signal used is provided with an offset of 180° or an offset of 90°, then the introduction of phase shifts with different values distributed in each pixel circuit group is performed on the demodulation signal, and the demodulation signal has a corresponding phase shift of 180° or 90° at each capture.

[0016] Therefore, due to the distribution of phase shifts on the demodulation signal in the network pixel circuit groups, the network pixel circuits are not all controlled by a single demodulation signal at the same time during capture, but the moments when the pixel circuits are controlled are distributed over time.

[0017] Therefore, the current peaks consumed in the network have lower intensities and are distributed over time, and the amplitude of the resulting electromagnetic interference decreases proportionally to the number of different phase shifts. In addition, the frequency of the electromagnetic interference is proportional to the number of different phase shifts.

[0018] According to one embodiment, the values of the phase shifts are discretely distributed within the period of the demodulation signal.

[0019] For example, "discretely distributed" means that the distribution in terms of number is voluntarily limited to a number that is much smaller than the maximum number of different phase shifts that can theoretically be introduced.

[0020] According to one embodiment, the network is arranged as columns and rows of pixel circuits, and the number of discrete values of the phase shift is between substantially one percent and approximately one tenth of the number of columns or rows. "Substantially" means, for example, "rounded up or down to the nearest integer".

[0021] In fact, providing a discrete number of phase shifts has implementation advantages, particularly with simpler and less costly means, and is also easier to control from the perspective of the calibration of the introduced phase shift (i.e., the accuracy of generating the phase shift).

[0022] In addition, the values of the phase shift can be evenly distributed or unevenly distributed within the period of the demodulated signal. However, an even distribution of the phase shift within the period of the demodulated signal allows for optimal minimization of electromagnetic interference.

[0023] And, regardless of the modulation frequency, for example, as opposed to any phase shift that causes uncontrolled spectral spreading, distributing the values of the phase shift relative to the period of the demodulated signal allows controlling the offset frequency of the resulting electromagnetic interference and performing reproducible spectral spreading. This is more advantageous in the case of acquisitions at several modulation frequencies.

[0024] According to one embodiment, the number of different values of the phase shift is selected such that the product of the frequency of the demodulated signal and said number lies outside the bandwidth of interest.

[0025] In fact, electromagnetic interference can be particularly problematic for certain frequencies of the operation of systems adjacent to the system for implementing depth map acquisition (e.g., within the bandwidth of interest that includes frequencies). In addition, certain materials of the system for implementing depth map acquisition can inherently attenuate frequencies beyond a given bandwidth, thus forming the bandwidth of interest in the context of this embodiment.

[0026] According to one embodiment, the network is arranged as columns and rows of pixel circuits, and the pixel circuit groups are segmented according to a periodic pattern on the columns and / or rows.

[0027] Selecting a periodic pattern in the plane of the pixel circuit network allows for distributing the current consumption in space, limiting the occurrence of constructive phenomena in electromagnetic interference generation, and making the system performance uniform.

[0028] According to one embodiment, the pixel circuit groups are segmented according to a periodic pattern of one or more columns such that each group includes several spatially discontinuous columns or several spatially discontinuous blocks of spatially continuous columns.

[0029] This embodiment is advantageous in terms of signal distribution when the architecture of the pixel circuit network provides a demodulation signal distribution by pixel circuit columns.

[0030] According to one embodiment, the method further includes: for each pixel circuit, calculating a time-of-flight phase difference between the demodulation signal and the optical signal received during the at least one capture, and calculating the time-of-flight phase difference includes, for each pixel circuit group, compensating for the phase shift introduced on the demodulation signal.

[0031] According to one embodiment, the compensation includes: for each pixel circuit group, adding the phase shift value introduced on the corresponding demodulation signal to the calculated time-of-flight phase difference modulo 360°.

[0032] In other words, after calculating the phase difference seen during the integration of the optical signal by the pixel circuit, during the reconstruction of the distance, the compensation is generally performed in a digital manner that can be easily parameterized in this regard.

[0033] According to one embodiment, calculating the time-of-flight phase difference includes: performing a trigonometric operation on the independent variables generated by at least two captures, during which the pixel circuit network is controlled by corresponding demodulation signals, and wherein the compensation includes, for each pixel circuit group, rotating the independent variable of the trigonometric operation by an angle equal to the phase shift value introduced on the corresponding demodulation signal.

[0034] In other words, the compensation is performed during the trigonometric calculation of the phase difference seen by the pixel circuit during the integration of the optical signal, which allows considering the compensation to directly provide the time-of-flight phase difference for distance reconstruction.

[0035] According to one embodiment, the method further includes transmitting an optical signal modulated by a modulation signal, and the demodulation signal is synchronized with the modulation signal.

[0036] According to another embodiment, an integrated circuit suitable for depth mapping by indirect time-of-flight is proposed. The integrated circuit includes: a photosensitive pixel circuit network divided into pixel circuit groups; control means configured to control the network pixels using a demodulation signal during at least one capture for acquiring a depth map, and the control means is configured to introduce phase shifts with different values distributed in each pixel circuit group in the demodulation signal.

[0037] According to one embodiment, the control means is configured to introduce phase shifts with values discretely distributed within the period of the demodulation signal.

[0038] According to one embodiment, the network is arranged in columns and rows of pixel circuits, and the control means is configured to introduce a number of discrete phase shift values between substantially one percent and substantially one-tenth of the number of columns or rows.

[0039] According to one embodiment, the control device is configured to introduce a number of different phase shift values, the number of different phase shift values being selected such that the product of the frequency of the demodulated signal and the number lies outside the bandwidth of interest.

[0040] According to one embodiment, the network is arranged as columns and rows of pixel circuits, and groups of pixel circuits are segmented according to a periodic pattern on the columns and / or rows.

[0041] According to one embodiment, groups of pixel circuits are segmented according to a periodic pattern of one or more columns such that each group includes a number of spatially discontinuous columns or a number of spatially discontinuous blocks of spatially contiguous columns.

[0042] According to one embodiment, the integrated circuit further includes a computing device configured to, for each pixel circuit, calculate the time-of-flight phase difference between the demodulated signal and the optical signal received during the at least one capture, and the computing device is configured to, for each group of pixel circuits, compensate for the phase shift introduced by the control device on the demodulated signal.

[0043] According to one embodiment, the computing device is configured to, for each group of pixel circuits, compensate for the phase shift introduced on the demodulated signal by adding the value of the corresponding phase shift to the modulo 360° of the calculated time-of-flight phase difference.

[0044] According to one embodiment, the control device is configured to control the pixel circuits of the network using respective demodulated signals during at least two captures of depth map acquisition, the computing device is configured to calculate the time-of-flight phase difference by performing a trigonometric operation on the independent variables generated by the at least two captures, and the computing device is configured to compensate for the phase shift introduced on the demodulated signal for each group of pixel circuits by pivoting at an angle equal to the corresponding phase shift value of the independent variable of the trigonometric operation.

[0045] According to one embodiment, the integrated circuit further includes a transmitting device configured to transmit an optical signal modulated by a modulation signal, and the control device is configured to generate the modulation signal and a demodulated signal synchronized with the modulation signal. Description of the Drawings

[0046] Figure 1 Illustrates an example of an integrated sensor of the indirect time-of-flight type configured to measure a depth map;

[0047] Figure 2 Shows from Figure 1 A given block Bk of pixel blocks B1 - B56 of an example;

[0048] Figure 3Schematically illustrates a first curve graph G1 and a second curve graph G2. The first curve graph G1 shows the peak current Ipk according to time T during capture in the network RES_PX, and the second curve graph G2 shows the amplitude EMI and frequency F of the electromagnetic interference generated by the peak current;

[0049] Figure 4 Shows a group of pixel circuits segmented according to a periodic pattern of several columns, such that each group includes several spatially discontinuous sets of spatially contiguous columns; and

[0050] Figure 5 Illustrates a method for compensating for the phase shift introduced on the demodulated signal. Detailed implementation

[0051] Figure 1 Illustrates an example of an integrated sensor CI of the "iToF" indirect time-of-flight type, which is intended to measure a depth map.

[0052] The integrated sensor CI includes an optical transmitter EM intended to emit an optical signal modulated at a modulation frequency. The optical transmitter EM includes, for example, an infrared laser diode and is modulated by a periodic modulation signal MOD, which is typically a step signal, and the modulation frequency is, for example, greater than 100 MHz.

[0053] The control device CMD is configured to control the illumination phase of the optical transmitter EM and in particular to generate the modulation signal MOD.

[0054] The control device CMD is also configured to control the integration phase (also referred to as capture) of the optical signal through a network of photosensitive pixel circuits RES_PX, the optical signal being from the reflection of the modulated optical signal emitted during illumination.

[0055] Each capture is performed simultaneously with the illumination, and the integration moment of the pixel is controlled by the demodulation signal DEMOD. The demodulation signal DEMOD is synchronized with the modulation signal MOD and allows the measurement of the phase difference between the received optical signal and the emitted optical signal to infer the distance between the reflecting object and the sensor CI therefrom.

[0056] The network of photosensitive pixel circuits RES_PX includes pixel circuits sensitive to the wavelength of the emitted optical signal and can have a density of approximately 1 MP (MegaPixel), for example 0.5 MP. The density of the network RES_PX is also referred to as "clarity".

[0057] The network of pixel circuits RES_PX is arranged to have columns of pixel circuits accessible via a column decoder DECY, and rows of pixel circuits accessible via a row decoder DECX.

[0058] The control device CMD is configured to control the pixel circuit network RES_PX using the demodulation signal DEMOD during capture (e.g., two or four captures) in the context of acquiring a depth map.

[0059] Furthermore, the computing device CAL is configured to calculate, based on the amount of charge generated by the network pixel circuit RES_PX controlled by the demodulation signal DEMOD, the phase difference Φ between the modulation of the transmitted optical signal and the modulation of the received optical signal during the at least one capture, also known as the time-of-flight phase difference.

[0060] The amount of charge generated by the pixel circuit can typically be provided to the computing device CAL via a voltage signal from an analog-to-digital converter incorporated in the column decoder DECY.

[0061] In addition to its function of cooperating with the row decoder DECX and decoding according to the architecture of the pixel circuit network RES_PX, the column decoder DECY can also be provided for typical processing such as subtraction of dark current.

[0062] Furthermore, the pixel circuit network RES_PX is divided into pixel circuit blocks, e.g., 56 blocks B1, B2, …, B56, each block including a number of columns, e.g., twelve columns Col1 - Col12( Figure 2 )

[0063] The division into pixel circuit blocks can be performed on the rows of the network RES_PX or on both rows and columns, i.e., on a rectangular portion of the pixel circuit network RES_PX.

[0064] Note that the division of the pixel circuit blocks is provided in the context of the demodulation signal distribution and can be purely virtual, i.e., the pixel circuit network RES_PX does not necessarily include structures specific to the defined divided blocks, e.g., as opposed to columns and rows specifically corresponding to conductive tracks contacting each pixel circuit of the network RES_PX.

[0065] The control device CMD is actually configured to introduce phase shifts Δ1 - Δ12( Figure 2 ) into the demodulation signal DEMOD with different values distributed in the respective pixel circuit groups.

[0066] From a hardware perspective, the control device CMD can be created by logic circuitry and includes a state machine to implement the control and scheduling functions of the sensor CI, and can further include means for generating signals, which are configured to specifically generate the modulation MOD and demodulation DEMOD signals and introduce the phase shifts Δ1 - Δ12.

[0067] In this regard, reference is made to Figure 2 ​

[0068] Figure 2 shows Figure 1 a given block Bk among the exemplary pixel blocks B1 - B56. In this example, block Bk includes twelve pixel circuit columns Col1 - Col12.

[0069] In this example, each photosensitive pixel circuit is of the "2 - tap" type, that is, capable of photogenerating charge simultaneously and significantly during two consecutive half - periods of the demodulation signal DEMOD.

[0070] The operation of the 2 - tap pixel is summarized by two inputs TAP1 and TAP2 that receive the demodulation signal DEMOD with a 180° offset, for example 0° on input TAP1 and 180° on input TAP2 in the first column Col1.

[0071] That is, there should be a distinction between the charge photogeneration of the 2 - tap mechanism and the second capture Capt2 ( Figure 4 ) controlled by the demodulation signal DEMOD with a 180° offset (specifically regarding Figure 4 ). In fact, the charge photogeneration of the 2 - tap mechanism can correspond to information segments collected in a "differential" manner during the control of a single demodulation signal DEMOD, while the second capture Capt2 ( Figure 4 ) is controlled using the "reference" demodulation signal DEMOD initially provided with a 180° offset relative to the modulation signal MOD.

[0072] The control device CMD is configured to generate a demodulation signal DEMOD referred to as "reference", which is fully synchronized with the modulation signal MOD that controls the optical signal emission. "Synchronization" means, for example, that the phase of the reference demodulation signal DEMOD is aligned with the phase of the modulation signal MOD (optionally offset by 180° or offset by 90° during consecutive capture acquisitions, regarding Figure 4 see below).

[0073] The control device is configured to further introduce phase shifts Δ1 - Δ12 into the reference demodulation signal DEMOD, for example via a phase - locked loop PLL or a phase generator PGEN, to produce the offset demodulation signal DEMOD+Δi, where i = [1:12].

[0074] Each column Coli of block Bk, i = [1:12], receives the corresponding offset demodulation signal DEMOD+Δi, that is, having phase shifts Δ1 - Δ12 with different values distributed in each pixel circuit column Col1 - Col12.

[0075] In this advantageous example, the phase shift is introduced into the reference demodulation signal DEMOD with values that are discrete and uniformly distributed within the demodulation signal period, that is to say, for example, twelve offsets with a 30° interval from 0° to 330°.

[0076] This corresponds to an input TAP2 offset of 180°, six offsets from 180° to 330° at 30° intervals, and six offsets from 0° to 150° at 30° intervals.

[0077] Other discrete amounts of phase shift values can be provided, for example, sixty-four (64) offsets at 5.625° intervals.

[0078] From a more general perspective, the number of discrete values of the phase shift can be provided between one percent and one tenth of the total number of columns of the network RES_PX, for example, rounded down or up to the nearest integer. In Figure 1 and Figure 2 the example shown, this corresponds to a number between 6 and 68 different discrete values.

[0079] If there are groups of pixel circuits with various phase shift values formed with respect to the rows of the network RES_PX, the number of discrete values of the phase shift can be defined in the same way, with respect to the total number of rows in the network RES_PX.

[0080] Providing a "discrete" number of offsets is indeed advantageous to limit the complexity of phase shift generation. For example, the "discrete number" refers to a number that is deliberately limited to be much smaller than the maximum number of different phase shifts that can be introduced theoretically, that is, for example, the total number of columns in the pixel circuit network RES_PX, or more generally, the total number of pixel circuits in the network RES_PX.

[0081] In fact, the control device CMD for generating the phase shift can thus be easily created in a compact manner and further precisely calibrated on the expected value of the phase shift and according to the desired interval.

[0082] In fact, during the capture, this phase shift is applied column by column within the block Bk to all the blocks B1 - B56 of the pixel circuit network RES_PX of the sensor CI.

[0083] In other words, in all blocks Bk (k = [1:56]), the pixel circuits of each individual column Coli (i = [1:12]) are controlled during capture using the same offset demodulation signal DEMOD + Δi.

[0084] Thus, a phase shift Δi is introduced into the demodulation signal DEMOD with different values distributed in the corresponding pixel circuit groups, which in this example correspond to a number of spatially discontinuous columns (one column Coli in each block Bk). In other words, the pixel circuit groups are segmented according to the pattern of one column with the periodicity of one block.

[0085] That is to say, as Figure 4 shown in the example, the pixel circuit groups can be segmented according to the periodic pattern of several columns such that each group includes several spatially discontinuous sets of spatially continuous columns.

[0086] Therefore, during capture, the phase shift distribution on the demodulation signals in the respective pixel circuit groups of the network results in the control of the respective pixel circuit groups of the network, and these controls are triggered at moments distributed over time.

[0087] Alternatively, the pixel circuit groups can be segmented according to a pattern that does not have spatial periodicity, for example, according to a pseudo-random "mixing" distribution in the pixel circuit network RES_PX. In such a pseudo-random distribution of the respective pixel circuit groups of the network RES_PX, the phase shift distribution on the demodulation signals results in the same effect that the control triggers are distributed over time.

[0088] In fact, the demodulation signal DEMOD controls the switching of transistors and the charge transfer in the network pixels. The current (especially capacitive current) circulating in the pixel circuit network RES_PX generated by these acquisition mechanisms can be very high and generate electromagnetic interference.

[0089] However, the distribution of the switching moments over time allows reducing electromagnetic interference.

[0090] In this regard, reference is made to Figure 3 .

[0091] Figure 3 During capture, the generation of current peaks Ipk according to the time T in the network RES_PX is schematically illustrated in the first graph G1, and the amplitude EMI and frequency F of the electromagnetic interference generated by the current peaks are shown in the second graph G2.

[0092] In both graphs G1 and G2, the dashed curves illustrate the conventional case in which all pixel circuits of the network RES_PX are controlled by the reference demodulation signal DEMOD at the same moment.

[0093] The solid curves illustrate the influence of the distribution of the control moments over time, as described above with respect to Figure 1 and Figure 2 obtained by introducing a phase shift into the demodulation signal DEMOD with different values Δ1 - Δ12 distributed in each pixel circuit group.

[0094] Under normal circumstances, a current peak with a high intensity Ip, for example up to 25 A (amperes), is generated periodically with a period of 1 / fmod of a reference demodulation signal DEMOD, where fmod is the frequency of the reference demodulation signal, for example greater than 100 MHz.

[0095] Therefore, electromagnetic interference is generated at a frequency fmod according to a high amplitude A.

[0096] When the pixel circuits are controlled by an offset demodulation signal DEMOD+Δi, where N different phase shifts are distributed among N groups of pixel circuits of the network RES_PX, the switching at the same moment is N times less, but the switching times of different moments of the pixel circuit groups are N times more.

[0097] Therefore, the current peak intensity Ip / N is basically equal to the intensity of Ip in the normal case divided by N, and relative to the normal case, the generation period of the current peak is also divided by N, 1 / (N*fmod).

[0098] Therefore, the electromagnetic interference is offset by a factor of N at a frequency of N*fmod and is generated according to a lower amplitude A / N.

[0099] Therefore, regarding Figure 1 and Figure 2 The described embodiments and implementations result in current peaks consumed in the network with a lower intensity and discrete distribution over time. The amplitude of the electromagnetic interference generated by the current peaks decreases proportionally to the number of different phase shifts, and the frequency of the electromagnetic interference is offset proportionally to the number of different phase shifts.

[0100] In one aspect, the number N of different values of the phase shift Δi introduced into the demodulation signal DEMOD can be selected such that the amplitude A / N of the electromagnetic interference is low enough not to interfere with the operation of the sensor CI or its nearby locations.

[0101] In another aspect, the number N of different values of the phase shift Δi introduced into the demodulation signal DEMOD can also be selected such that the frequency of the electromagnetic interference N*fmod is outside the bandwidth of interest.

[0102] The bandwidth of interest can, for example, correspond to the operating frequency band of the sensor CI, such that electromagnetic interference with a frequency outside this interval does not interfere with the operation of the sensor CI. In addition, structural elements of the sensor CI, such as a Faraday cage-type shield in the package housing, can absorb and attenuate electromagnetic waves outside the bandwidth of interest. In the second case, the electromagnetic interference whose amplitude has been divided by N can thus be additionally attenuated.

[0103] It should be noted that the representations of the curves G1 and G2 correspond to phase shift values introduced in a manner that is uniformly distributed over the period of the demodulated signal. This particularly results in the spectrum of the electromagnetic interference being located only at frequencies equal to the above product N*fmod.

[0104] That is, a non-uniform distribution of the phase shift over the period of the demodulated signal will result in a wider spectral distribution of the electromagnetic interference.

[0105] In this case, the non-uniform distribution and the number N of the phase shifts can be jointly selected such that the spectrum or a majority of the spectrum of the electromagnetic interference lies outside the bandwidth of interest.

[0106] Thus, with respect to Figure 1 and Figure 2 the described embodiments and implementations allow the elimination of the constraints related to electromagnetic interference, which enables an increase in the resolution of the pixel circuit network and the frequency of the modulation signal in a system for acquiring depth maps by indirect time-of-flight.

[0107] Recall that distance measurement by indirect time-of-flight is based on the time-of-flight phase difference of the received optical signal, evaluated based on the demodulated signal, and the embodiments and implementations illustrated below with reference to Figure 4 and Figure 5 allow for the compensation of the phase shift introduced on the demodulated signal for each pixel circuit group in the evaluation of the time-of-flight phase difference.

[0108] Figure 4 On the one hand, another example of dividing the photosensitive pixel circuit network into pixel circuit groups is illustrated, and on the other hand, a method for compensating the phase shift Δi introduced on the demodulated signal DEMOD is illustrated.

[0109] In this example, the pixel circuit network RES_PX is first divided into blocks Bk of columns as described above with respect to Figure 1 each block Bk including twelve columns Col1 - Col12.

[0110] In this example, the pixel circuit groups in each block Bk include two consecutive columns Col1 / 2, Col3 / 4, …, Col11 / 12.

[0111] Thus, the pixel circuit groups are divided according to a periodic pattern of two columns such that each group includes a number of spatially non-contiguous sets of two spatially consecutive columns.

[0112] Furthermore, Figure 4 the process of acquiring a depth map according to the 4-bin sampling technique is also illustrated, the process including four consecutive captures Capt1, Capt2, Capt3, Capt4, the principle of the 4-bin sampling technique being known to those skilled in the art.

[0113] In summary, the 4-bin sampling technique uses four captures, i.e., four integrations with corresponding irradiations, controlled by the components of the demodulated signal in the quadratic phase shift.

[0114] In other words, in each capture Capt1, Capt2, Capt3, Capt4, the reference demodulated signal is aligned with the modulation signal MOD for transmitting the optical signal at 0°, 180°, 90°, and 270° respectively.

[0115] The amount of charge photogenerated by each pixel during capture (commonly referred to as "bin") provides a piece of information corresponding to the superposition of the phase difference of the received optical signal relative to the phase of the demodulated signal.

[0116] For example, during the first capture Capt1, the pixel provides a piece of information related to the demodulated signal in phase Bin_0; during the second capture Capt2, the pixel provides a piece of information related to the demodulated signal Bin_180 offset by 180°; during the third capture Capt3, the pixel provides a piece of information related to the demodulated signal Bin_90 offset by 90°; during the fourth capture Capt4, the pixel provides information related to the demodulated signal Bin_270 offset by 270°.

[0117] Incidentally, the "2-tap" operation of the pixel allows each value of Bin to be collected twice in two corresponding captures offset by 180° (e.g., the information piece Bin_0 is collected by TAP1 of the first capture Capt1 and by TAP2 of the second capture Capt2). This is particularly advantageous in terms of ambient noise (usually "offset") and matching.

[0118] The reconstruction of the phase difference Φ (also known as the "time-of-flight phase difference") is performed by the computing device CAL, and for each pixel circuit, it includes a trigonometric operation TRIGO of the type "arctan(Q / I)", where the independent variable Q represents the phase difference offset by 90° obtained through the trigonometric relationship of the throughputs Bin_90 and Bin_270, and the independent variable I represents the phase difference obtained through the trigonometric relationship of the throughputs Bin_0 and Bin_180.

[0119] The phase shifts Δi respectively introduced on the demodulated signal DEMOD+Δi of each pixel circuit group correspond to the rotation of known angles in the independent variables Q and I.

[0120] Therefore, in order to compensate for the phase shift Δi introduced on the demodulated signal DEMOD+Δi, for each pixel circuit group, the trigonometric calculation can be adapted by pivoting the independent variables Q and I by an angle equal to the respectively introduced phase shift value Δi.

[0121] The distance can then be directly calculated by knowing the phase difference Φ compensated for each pixel circuit group.

[0122] Figure 5 Another method for compensating the phase shift Δi introduced on the demodulated signal DEMOD is illustrated.

[0123] In this method, the calculation device CAL is configured to obtain the value of the phase difference Φ for each pixel by means of a conventional trigonometric calculation, that is, independently of the group to which the pixel belongs and for that group, without taking into account the phase shift Δi introduced on the demodulated signal DEMOD+Δi.

[0124] The calculation device CAL is configured to compensate each phase shift Δi introduced on the demodulated signal based on the calculated value of the phase difference Φ. Since the calculation of the phase difference Φ is performed with respect to the offset signal DEMOD+Δi, the calculation device CAL adds the corresponding value of the offset Δi to the value of the phase difference Φ.

[0125] If the calculated phase difference Φ is between 0° and 360°-Δi, the actual phase difference Φout is equal to Φ+Δi; if the calculated phase difference Φ is between 360°-Δi and 360°, the actual phase difference Φout is equal to Φ-360°+Δi. This is equivalent to adding the value of the phase shift Δi to the calculated time-of-flight phase difference Φ and taking the modulo 360°.

[0126] Furthermore, the present invention is not limited to these embodiments and implementations and includes all its alternatives. For example, introducing a phase shift into the demodulated signal with different values distributed in the pixel circuit group can be applied to mechanisms other than the described mechanism, in particular the 4-bin sampling technique or the 2-tap mechanism of the pixels, which are given only as advantageous examples and are not restrictive.

Claims

1. A method for obtaining a depth map by indirect time-of-flight in a photosensitive pixel circuit network divided into pixel circuit groups, the method comprising: Performing at least one capture during which the pixel circuits of the network are controlled by a demodulation signal; Introducing a phase shift into the demodulation signal with different values distributed among each pixel circuit group; And For each pixel circuit, calculating a time-of-flight phase difference between the demodulation signal and the optical signal received during the at least one capture, wherein calculating the time-of-flight phase difference includes compensating for the value of the phase shift introduced on the demodulation signal for each pixel circuit group.

2. The method according to claim 1, wherein the different values of the phase shift are discretely distributed within a period of the demodulation signal.

3. The method according to claim 2, wherein the network is arranged as columns and rows of pixel circuits, and wherein the number of discrete values of the phase shift is between substantially one percent and substantially one tenth of the number of columns or rows.

4. The method according to claim 2, wherein the number of discrete values of the phase shift is selected such that the product of the frequency of the demodulation signal and the number (N*fmod) lies outside the bandwidth of interest.

5. The method according to claim 1, wherein the network is arranged as columns and rows of pixel circuits, and wherein the pixel circuit groups are divided according to a periodic pattern on the columns and / or rows.

6. The method according to claim 5, wherein the pixel circuit groups are divided according to the periodic pattern of one or more columns such that each group includes a number of spatially discontinuous columns or a number of spatially discontinuous sets of spatially continuous columns.

7. The method according to claim 1, wherein the compensation includes, for each pixel circuit group, adding the value of the phase shift introduced on the corresponding demodulation signal to the calculated time-of-flight phase difference modulo 360°.

8. The method according to claim 1, wherein calculating the time-of-flight phase difference comprises: Performing a trigonometric operation on the in-phase (I) and quadrature (Q) components generated by at least two captures, during which the pixel circuit network is controlled by corresponding demodulation signals, and wherein the compensation includes: for each pixel circuit group, rotating the in-phase (I) and quadrature (Q) components of the trigonometric operation by an angle equal to the value of the phase shift introduced on the corresponding demodulation signal.

9. The method according to claim 1, further comprising transmitting an optical signal modulated by a modulation signal, the demodulation signal being synchronized with the modulation signal.

10. A sensor comprising: A photosensitive pixel circuit network divided into pixel circuit groups; And A controller configured to: During at least one capture for obtaining a depth map, control the pixel circuits of the network using a demodulation signal, and Introduce a phase shift into the demodulation signal with different values distributed among each pixel circuit group; And A calculator configured to: For each pixel circuit, calculate a time-of-flight phase difference between the demodulation signal and the optical signal received during the at least one capture, and For each pixel circuit group, compensate for the value of the phase shift introduced by the controller on the demodulation signal.

11. The sensor according to claim 10, wherein the controller is configured to introduce the phase shift with values discretely distributed within the period of the demodulation signal.

12. The sensor according to claim 11, wherein the network is arranged as columns and rows of pixel circuits, and wherein the controller is configured to introduce a plurality of discrete phase shift values between substantially one percent and substantially ten percent of the number of the columns or the rows.

13. The sensor according to claim 11, wherein the controller is configured to introduce a plurality of different phase shift values, and the number of the plurality of different phase shift values is selected such that the product (N*fmod) of the frequency of the demodulation signal and the number is outside the bandwidth of interest.

14. The sensor according to claim 10, wherein the network is arranged as columns and rows of pixel circuits, and wherein the pixel circuit groups are segmented according to a periodic pattern on the columns and / or the rows.

15. The sensor according to claim 14, wherein the pixel circuit groups are segmented according to a periodic pattern of one or more columns such that each group includes a number of spatially discontinuous columns or a number of spatially discontinuous sets of spatially continuous columns.

16. The sensor according to claim 10, wherein the calculator is configured to add, for each pixel circuit group, the value of the corresponding phase shift to the calculated time-of-flight phase difference modulo 360°.

17. The sensor according to claim 10, wherein the controller is configured to control the pixel circuits of the network with corresponding demodulation signals during at least two captures of the acquisition of the depth map, and wherein the calculator is configured to: calculate the time-of-flight phase difference by performing a trigonometric operation on the in-phase (I) and quadrature (Q) generated by the at least two captures, and for each pixel circuit group, compensate for the phase shift introduced on the demodulation signal by pivoting the in-phase (I) and the quadrature (Q) of the trigonometric operation by an angle equal to the value of the corresponding phase shift.

18. The sensor according to claim 10, further comprising a transmitter configured to transmit an optical signal modulated by a modulation signal, wherein the controller is configured to generate the modulation signal, and the demodulation signal is synchronized with the modulation signal.

Images