Signal processing method of light receiving element and device thereof

Abstract

The invention discloses a signal processing method of a light receiving element and a device thereof. The signal processing device accumulatively records the number of times of laser echo reception in N (N is a natural number of 2 or more) periods in a search memory, selects at least one period in which the number of times of reception reaches a candidate critical value as a candidate period, and divides the candidate period into M (M is a natural number of 2 or more) subdivided periods. And accumulatively recording the receiving times of the laser echoes in each subdivision time period into a candidate memory, and determining the receiving time point of the laser echoes on the basis of the receiving times of each subdivision time period accumulated in the candidate memory.

Description

Technical Field

[0001] Embodiments of the present invention relate to a signal processing method and apparatus for an optical receiving element, and more particularly, to a signal processing method and apparatus for a time-to-digital converter (TDC) of an optical receiving element for measuring time of flight (TOF) in a light detection and ranging (LiDAR) sensor, an image sensor, or the like. Background Art

[0002] Figure 1 FIG. 1 is an exemplary diagram of a conventional LiDAR. LiDAR 100 outputs a pulsed laser signal using a light emitting element such as a vertical cavity surface emitting laser (VCSEL) / edge emitting laser diode (EELD). The optical receiving element 120 senses the laser reflected from the object 130 (hereinafter referred to as laser echo). As an example, a single photon avalanche diode (SPAD) is the optical receiving element 120. LiDAR 100 measures the time (i.e., TOF) taken for the laser output from the light emitting element to be reflected back by the object 130 and determines the distance to the object 130.

[0003] In order to improve the resolution (angular resolution) of LiDAR 100, it is necessary to improve the pixel resolution of the optical receiving element 120 for sensing light. Figure 2 FIG. 2 is an exemplary diagram of the layout structure of an optical receiving element and a TDC of a LiDAR. Referring to Figure 2 , there are TDC elements 210 in each column of the pixel array 200. Therefore, if it is necessary to improve the pixel resolution, it is necessary not only to reduce the size of the pixels but also to reduce the size of the TDC 210 at the same time. Summary of the Invention

[0004]

Technical Problem

[0005] The technical problem to be solved by the embodiments of the present invention is to provide a signal processing method and apparatus that can improve the resolution of an optical receiving element by reducing the size of a histogram memory used in signal processing for measuring TOF and that can maintain performance even when the size is reduced.

[0006]

Technical Solution

[0007] To solve the above technical problem, a signal processing method for an optical receiving element according to an embodiment of the present invention includes the following steps: cumulatively recording the number of laser echo receptions in N (N is a natural number greater than or equal to 2) time periods into a search memory; selecting at least one time period in which the number of receptions reaches a candidate threshold value as a candidate time period; dividing the candidate time period into M (M is a natural number greater than or equal to 2) sub-time periods, and cumulatively recording the number of laser echo receptions in each sub-time period into a candidate memory; and determining the reception time point of the laser echo according to the number of receptions in each sub-time period cumulatively recorded in the candidate memory.

[0008] To solve the above technical problem, a signal processing device for an optical receiving element according to an embodiment of the present invention includes: a search memory for cumulatively recording the number of laser echo receptions in N (N is a natural number greater than or equal to 2) time periods; and a candidate memory for cumulatively recording the number of laser echo receptions in at least one time period corresponding to a predefined candidate threshold value, wherein the search memory and the candidate memory include a histogram memory, and the histogram memory includes a plurality of bins composed of a plurality of bits.

[0009]

Beneficial Effects

[0010] According to the embodiments of the present invention, the usage amount of the histogram memory for measuring TOF can be reduced. Since the usage amount of the histogram memory is small, a TDC can be implemented using a memory of a smaller size. In addition, the resolution of the optical receiving element can be improved by reducing the pixel and the size of the TDC of the optical receiving element. In addition, the size of the histogram memory can be reduced with almost no performance degradation. Description of the Drawings

[0011] Figure 1 is an example diagram of an existing LiDAR.

[0012] Figure 2 is an example diagram of the layout structure of the optical receiving element and the TDC of the LiDAR.

[0013] Figure 3 is an example diagram of the histogram memory used in TOF measurement.

[0014] Figure 4 and Figure 5 is using Figure 3 is an example diagram of the 1-step histogram method of the histogram memory.

[0015] Figure 6 is another example diagram of the histogram memory used in TOF measurement.

[0016] Figure 7 is using Figure 6An exemplary diagram of the 2-step histogram method for a histogram memory

[0017] Figure 8 It is a flowchart of an embodiment of a signal processing method for an optical receiving element capable of reducing the usage amount of a histogram memory according to an embodiment of the present invention.

[0018] Figure 9 It is an exemplary diagram of the memory structure of a signal processing device according to an embodiment of the present invention.

[0019] Figure 10 It is a flowchart of an example of a method for clearing a histogram memory in a candidate memory according to an embodiment of the present invention.

[0020] Figure 11 It is a diagram showing the time point for clearing judgment of a histogram memory in a candidate memory according to an embodiment of the present invention.

[0021] Figure 12 It is an exemplary diagram of a method for managing information for clearing a histogram memory according to an embodiment of the present invention.

[0022] Figure 13 It is an exemplary diagram of a method for periodically performing clearing judgment of a histogram memory according to an embodiment of the present invention.

[0023] Figure 14 It is another exemplary diagram of a method for clearing a histogram memory according to an embodiment of the present invention.

[0024] Figure 15 and Figure 16 It is an exemplary diagram of the scanning range of a retrieval memory according to an embodiment of the present invention.

[0025] Figure 17 and Figure 18 It is an exemplary diagram of a method for reducing the size of a retrieval memory according to an embodiment of the present invention.

[0026] Figure 19 It is an exemplary diagram of a method for reducing the size of a retrieval memory while increasing the sensing success rate according to an embodiment of the present invention.

[0027] Figure 20 It is an exemplary diagram of a method for loading a predetermined value for each pixel using a lookup table according to an embodiment of the present invention.

[0028] Figure 21 It is an exemplary diagram of a method for arranging a lookup table according to an embodiment of the present invention.

[0029] Figure 22 It is an exemplary diagram of the memory structure of each pixel according to an embodiment of the present invention.

[0030] Figure 23 It is another exemplary diagram of the memory structure of each pixel according to an embodiment of the present invention.

[0031] Figure 24 It is yet another exemplary diagram of the memory structure of each pixel according to an embodiment of the present invention.

[0032] Figure 25 It is an exemplary diagram of a method for reconstructing a candidate memory according to an embodiment of the present invention. Detailed implementation manners

[0033] Next, a signal processing method and device of a light receiving element according to an embodiment of the present invention will be described in detail with reference to the accompanying drawings.

[0034] The signal processing device of this embodiment can be used not only for LiDAR sensors or image sensors but also in various scenarios where TOF needs to be measured, and its application field is not limited to specific sensors. However, for the sake of convenience of description, hereinafter, the case where the signal processing device is the TDC of the light receiving element of a LiDAR sensor will be mainly described as an example.

[0035] Figure 3 It is an exemplary diagram of a histogram memory used in TOF measurement, Figure 4 and Figure 5 is Figure 3 an exemplary diagram of a 1-step histogram method using the histogram memory.

[0036] Referring to Figures 3 to 5 , in order to measure the time required for a signal (e.g., a laser pulse) output from a light emitting element to be reflected by an object and incident on the light receiving element, the signal processing device divides a predefined measurement range time (i.e., the dynamic range 410) into a plurality of time periods. Figure 3 and Figure 5 The example of

[0037] is a case where the dynamic range t is divided into 1012 (= 4096) time periods in order to implement 12-bit TOF. The dynamic range can be preset to various different values according to different embodiments.

[0038] For example, if a first pulse is output from the light-emitting element, the signal processing device records from which one of the 4096 time periods the first pulse is detected. If a second pulse is output from the light-emitting element, the signal processing device records from which one of the 4096 time periods the second pulse is detected. By the same method, the signal processing device sequentially records from which time period each of the multiple pulses continuously output from the light-emitting element is detected. In other words, as Figure 5 shown, after the LiDAR sensor continuously outputs multiple pulses, the time-of-flight (TOF) is determined by accumulating the number of signal detections in each time period.

[0039] Figure 3 The histogram memory in Figure 3 includes 4096 bits corresponding to each time period on the horizontal axis and 8 bits for accumulating and storing the number of times each laser echo is received on the vertical axis. Since the numerical range that can be represented by 8 bits is 0 to 255 (i.e., 2^8), the histogram memory can accumulate and store the number of times the laser echo is received up to 256 times in total. In the histogram memory, the column of bits corresponding to each time period (i.e., the vertical column of bits) is called a bin. The histogram memory of this embodiment, as Figure 5 includes a total of 4096 bins. Figure 3 The histogram memory in Figure 3 uses 32769 (= 8 * 4096) bits in total to accumulate and store the number of times the laser echo is received 256 times.

[0040] Figure 6 is another example diagram of the histogram memory used in TOF measurement. Figure 7 is a diagram showing Figure 6 an example of the 2-step histogram method using the histogram memory in Figure 6 .

[0041] Referring to Figure 6 and Figure 7 , the histogram memory includes 64 bits corresponding to 64 time periods on the horizontal axis and 8 bits on the vertical axis for accumulating and storing the number of times the laser echo is received 256 times. The signal processing device determines the reception time period of the laser echo by the 2-step histogram method 600.

[0042] First, the signal processing device divides the dynamic range 710 into 64 time periods and accumulatively stores the number of times the laser echo is received in each time period (720). For example, the signal processing device accumulatively stores the number of times the laser echo is received in each time period in each bin of the histogram memory.

[0043] The signal processing device determines the bin (i.e., time period) with the maximum cumulative count based on the cumulative reception count of the laser echo recorded in the histogram memory. For example, if the cumulative count of the sixth bin is the largest, the signal processing device determines the time period corresponding to the sixth bin as the candidate time period. Then, after dividing the candidate time period (e.g., the time period of the sixth bin) into 64 sub - time periods, the signal processing device assigns the sub - time periods to each bin of the histogram memory respectively. The signal processing device cumulatively stores the laser echo reception count in each bin of the histogram memory corresponding to the sub - time period (730).

[0044] As an embodiment, the signal processing device divides the dynamic range into 64 time periods and determines the reception time period 720 of the laser echo in the first round, and then divides the reception time period (i.e., candidate time period) determined in the first round into 64 time periods again and determines the reception time period of the laser echo in the second round (730). The first - round determination process 720 and the second - round determination process 730 are executed using the same histogram memory.

[0045] In this case, through the first - round determination process 720 and the second - round determination process 730, there is a time resolution of 64 * 64 = 4096, so it has the same time resolution as Figure 3 12 - bit - TOF. In addition, to determine the reception time period of the laser echo, a histogram memory of a total of 512 bits (= 64 * 8) is used. Compared with Figure 3 , the usage amount of the histogram memory can be reduced by approximately 98.43%. That is, if the 2 - step histogram method of Figure 4 is used instead of the 1 - step histogram method of Figure 3 , the usage amount of the histogram memory for measuring TOF can be significantly reduced. In addition, to achieve 12 - bit - TOF, not only the 2 - step histogram method of 6 bits / 6 bits, but also the resolutions of the first and second stages such as 7 bits / 5 bits can be variously deformed according to different embodiments.

[0046] Figure 7 The example of

[0047] Figure 4 and Figure 5In the 1-step histogram method 400, if the number of detections is increased, both the sensing success rate (SR) and precision will increase. However, there is a drawback of increased usage of the histogram memory. Figure 6 and Figure 7 In the 2-step histogram method 700, compared to the 1-step histogram method 400, the usage of the histogram memory is significantly reduced. However, in the 2-step histogram method 700, if the number of detections in the first-round determination process 720 is increased, only the sensing success rate increases, and if the number of detections in the second-round determination process 730 is increased, only the precision increases.

[0048] In the 2-step histogram method 700, if one wants to increase the sensing success rate and precision, the number of detections in the first-round determination process 720 and the second-round determination process 730 needs to be increased. However, due to the limitation of the frame rate, the number of detections is restricted. Therefore, compared to the 1-step histogram method, the sensing success rate and precision structurally decrease. Thus, in the 2-step histogram method 700, a method capable of increasing the sensing success rate and precision is needed, and Figure 8 will be specifically described below.

[0049] Figure 8 is a flowchart of an embodiment of a signal processing method of an optical receiving element capable of reducing the usage of the histogram memory according to an embodiment of the present invention.

[0050] Referring to Figure 8 , the signal processing device cumulatively records the number of laser echo receptions in N (N is a natural number of 2 or more) time periods in the retrieval memory (S800). The retrieval memory for cumulatively storing the number of laser echo receptions in each time period may include at least one histogram memory. Figure 9 An example of a retrieval memory composed of one histogram memory will be described below. As another embodiment, in order to reduce the size of the retrieval memory, the scanning range of the histogram memory is made smaller than the dynamic range, and it will be described in Figure 16 . As still another embodiment, in order to increase the number of receptions cumulated in the retrieval memory for improving precision, a method of deleting the number of receptions cumulated in the histogram memory at regular intervals will be described in Figures 17 to 19 .

[0051] The signal processing device selects at least one time period in which the number of receptions reaches a predefined candidate threshold as a candidate time period (S810). As an embodiment, the candidate time period may be composed of only the reception time period reaching the candidate threshold, or as Figure 6The time period reaching the candidate threshold value and at least one time period in its vicinity may be included. However, for ease of explanation, hereinafter, it is assumed that the candidate time period includes only the time period reaching the candidate threshold value. The candidate threshold value may be set to various different values according to different embodiments. For example, if the number of bits in each bin of the retrieval memory is 2, the maximum number of received times that can be stored in each bin is 4 (= 2^2), and thus the candidate threshold value may be set to be less than or equal to 4.

[0052] The signal processing device divides the candidate time period into M (M is a natural number of 2 or more) sub - time periods, and cumulatively records the number of laser echo receptions in each sub - time period in the candidate memory (S820). As an embodiment, the number N of time periods of the retrieval memory and the number M of sub - time periods of the candidate memory may be the same or different. For example, N = M = 64 may be set.

[0053] As an embodiment, the candidate memory may include at least one histogram memory. For an example of a candidate memory including multiple histogram memories for multiple candidate time periods, it will be described in Figure 9 . As another embodiment, the candidate memory may clear (flush) and reuse the histogram memory of the candidate time period that does not reach a certain criterion, which will be described in Figures 11 to 14 .

[0054] The signal processing device determines the reception time point of the laser echo based on the number of receptions in each sub - time period cumulatively stored in the candidate memory (S830). For example, the sub - time period of the candidate time period with the largest cumulative number of receptions among the sub - time periods of multiple candidate time periods may be determined as the reception time point of the laser echo.

[0055] For the purpose of facilitating the understanding of this embodiment, although it is described in the order of the storage process (S800) of the retrieval memory, the candidate time period determination process (S810), and the candidate memory storage process (S820), the storage of the retrieval memory and the candidate time period determination processes (S800, S810) are always in a working state. That is, if the first candidate time period is determined, then the process of the subsequent second candidate time period is continued.

[0056] Figure 9 It is an example diagram of the memory structure of a signal processing device according to an embodiment of the present invention.

[0057] Referring to Figure 9 , the memory of the signal processing device for cumulatively storing the number of laser echo receptions in N time periods may generally include a retrieval memory 910 and a candidate memory 920. The retrieval memory 910 and the candidate memory 920 may be memories that can be logically distinguished or physically distinguished.

[0058] The retrieval memory 910 includes at least one histogram memory. This embodiment illustrates the case where the retrieval memory 910 includes one histogram memory. The histogram memory may include a plurality of bins, and each bin includes a plurality of bits. The signal processing device may divide the dynamic range into N time periods, and may respectively accumulate and store the number of received laser echoes in each time period into N bins of the histogram memory. For example, if the light-emitting element outputs continuous laser pulses 900, then some of the laser pulses 902, 904, 906 are reflected by the object and incident on the light-receiving element, and the signal processing device accumulates and stores the number of received laser echoes in each time period into the respective bins 912, 914, 916 of the retrieval memory. This embodiment is described based on the bins 912, 914, 916 corresponding to partial time periods. The signal processing device determines the candidate time periods based on the number of received echoes accumulated in each of the bins 912, 914, 916 in the retrieval memory 910.

[0059] The number of bits of the bins 912, 914, 916 that make up the histogram memory of the retrieval memory 910 can be set to various different values according to different embodiments. If the number of bits of the bins 912, 914, 916 is increased, the maximum value of the number of received echoes accumulated in each of the bins 912, 914, 916 of the retrieval memory 910 increases, thereby increasing the sensing success rate, but the size of the retrieval memory 910 also increases accordingly. Therefore, considering the size of the retrieval memory 910, the number of bits of each bin 912, 914, 916 of the histogram memory can be preset to various values according to different embodiments.

[0060] The candidate memory 920 includes at least one histogram memory 922, 924, 926. This embodiment illustrates the case of the candidate memory 920 including a plurality of histogram memories 922, 924, 926 in order to be able to determine the cumulative number of received echoes of multiple candidate time periods.

[0061] The histogram memories 922, 924, 926 in the candidate memory 920 include M bins respectively corresponding to M sub-time periods. The M sub-time periods are M intervals obtained by dividing the candidate time period. The number of bits of the bins of the histogram memories 922, 924, 926 in the candidate memory 920 can be preset to various different values according to different embodiments. If the number of bits of the bins is increased, the maximum value of the number of received echoes accumulated in the candidate memory 920 increases, and the accuracy can be improved, but the size of the candidate memory 920 also increases accordingly. Therefore, considering the size and accuracy of the candidate memory 920, the number of bits of the bins in the candidate memory 920 can be preset to various different values according to different embodiments.

[0062] The signal processing device determines a candidate time period during which the accumulated number of receptions reaches the candidate threshold value by retrieving the memory 910. The time period of the bin 914 whose reception number first reaches the candidate threshold value is selected as the first candidate time period, and the signal processing device allocates the histogram memory 922 of the candidate memory 920 to the first candidate time period (i.e., the time period of the bin 914). Then, if a second candidate time period (i.e., the time period of the bin 916) is determined in the retrieval memory 910, the signal processing device allocates the histogram memory 924 of the candidate memory 920 to the second candidate time period. Using the same method, the multiple histogram memories 922, 924, 926 in the candidate memory 920 are sequentially allocated to each candidate time period (i.e., the respective time periods of the bins 912, 914, 916).

[0063] If the number of the histogram memories 922, 924, 926 in the candidate memory 920 is small, and multiple noises are first determined as multiple candidate time periods, there is a problem that the reception time point of the laser echo cannot be accurately detected. To improve the sensing success rate, the number of the histogram memories 922, 924, 926 in the candidate memory 920 can be increased, but in this case, there is a disadvantage that the size of the candidate memory 920 increases. As an example, a solution to this problem includes a method of clearing (flushing) the histogram memories 922, 924, 926 of the candidate time periods that do not meet certain conditions and reusing them after the histogram memories 922, 924, 926 in the candidate memory 920 are allocated to the candidate time periods. This will be described again in Figure 10 .

[0064] Figure 10 is a flowchart of an example of a method for clearing the histogram memory in the candidate memory according to an embodiment of the present invention. Figure 11 is a diagram showing the clearing judgment time point of the histogram memory in the candidate memory according to an embodiment of the present invention.

[0065] Refer to Figure 10 and Figure 11 , the signal processing device determines the number of laser echo receptions (S1000) during a predefined flush check period for each candidate time period. For example, the signal processing device can divide the candidate time period into M sub-time periods, and can cumulatively store the number of laser echo receptions in the M sub-time periods in each bin of the histogram memory. The signal processing device determines the total accumulated number of receptions (i.e., the total of the accumulated number of receptions in the M bins) during the period from the time point when the histogram memory is allocated to the candidate time period to the flush check period. As Figure 11 shown, the flush check period can be set based on the number of output times of the laser pulse 1100.

[0066] For a candidate time period during which the total number of receptions does not reach a predefined flush threshold during the flush check interval, the signal processing device flushes the histogram memory allocated to the candidate time period (S1010, S1020, S1030). For example, Figure 9 in the example of, if the total number of receptions accumulated in the histogram memory 922 of the first candidate time period during the flush check interval is less than the flush threshold, the values of the histogram memory 922 allocated to the first candidate time period are initialized to be allocated to other candidate time periods.

[0067] As an embodiment, the signal processing device may determine whether to perform repeated flushing at a certain period instead of determining whether to flush the candidate time period at once. A corresponding example will be shown in Figure 13 .

[0068] As another embodiment, for a candidate time period that satisfies certain conditions, the signal processing device may no longer perform the determination of whether to flush, and for multiple candidate time periods allocated with histogram memories, it does not repeatedly perform the determination of whether to flush each time. A corresponding example will be shown in Figure 14 .

[0069] Figure 12 is an example diagram of a method for managing information for flushing a histogram memory according to an embodiment of the present invention.

[0070] Referring to Figure 12 , the signal processing device stores and manages the total number of receptions 1200, 1210, 1220 and the number of laser emissions 1230, 1240, 1250 of each of the histogram memories 922, 924, 926 in the candidate memory. For example, starting from the time point when the first histogram memory 922 is allocated to the first candidate time period, the signal processing device accumulatively records the number of laser echo receptions 1200 recorded in the first histogram memory 922 (i.e., the sum of the receptions of each bin of the first histogram memory 922) and the number of laser emissions 1230 output by the light emitting element.

[0071] For example, if the number of laser emissions during the flush check interval is defined as K times, the signal processing device accumulatively records the number of laser pulses output by the light emitting element to the number of laser emissions 1230 starting from the allocation time point of the first histogram memory 922. If the number of laser emissions 1230 reaches K, it determines whether to flush the first histogram memory 922. This embodiment illustrates an example of defining the flush check interval based on the number of laser pulses output by the light emitting element, which is only an example, indicating that various values representing a certain time period can be used as the flush check interval.

[0072] Figure 13This is an example diagram of a method for periodically performing a clearing determination of a histogram memory according to an embodiment of the present invention.

[0073] Referring to Figure 13 , the signal processing device periodically determines whether to clear each histogram memory. For example, if the clearing check interval of the first histogram memory 922 has elapsed, the signal processing device uses the method described in Figure 10 to determine whether to clear the first histogram memory 922 (1300). If the first histogram memory 922 is not cleared at the first round of determination time point, the signal processing device determines whether to clear the first histogram memory 922 (1310) in the second round when the clearing check interval elapses again. Using the same method, the signal processing device can perform N - cycle determinations on whether to clear the first histogram memory (1320).

[0074] As an embodiment, the clearing threshold value for determining whether to clear can be the same value or different values in each cycle. For example, the first clearing threshold value for determining whether to clear in the clearing check interval of the first cycle and the second clearing threshold value for determining whether to clear in the clearing check interval of the second cycle can be different predetermined values.

[0075] If, by reducing the size of the clearing check interval, the cumulative number of received times accumulated in the histogram memory is periodically determined multiple times whether to clear, then noise can be quickly deleted with a low clearing threshold value, and the time for noise to occupy the histogram memory can be shortened.

[0076] Figure 14 This is another example diagram of a method for clearing a histogram memory according to an embodiment of the present invention.

[0077] Referring to Figure 14 , the signal processing device, as in Figure 13 , periodically determines whether to clear the histogram memory. The signal processing device uses an upper threshold value and lower threshold values 1400, 1410 when determining whether to clear.

[0078] For example, if the total number of received times accumulated in the first histogram memory 922 is greater than the upper threshold value, the signal processing device no longer performs the process of determining whether to clear the first histogram memory 922. That is, the candidate time period of the first histogram memory 922 is determined as a candidate and will not be discarded from the candidate memory.

[0079] If the total number of received times accumulated in the first histogram memory 922 is less than the lower threshold value, the signal processing device clears the first histogram memory 922. That is, the candidate time period of the first histogram memory 922 is deleted from the candidates for determining the received time period of the laser echo. If the total number of received times of the first histogram memory 922 is less than the upper threshold value and less than the lower threshold value, the signal processing device will re-determine whether to clear the first histogram memory 922 in the next cycle. As an embodiment, the signal processing device may only use the upper threshold value to determine whether to clear at the end of the determination clearing cycle (1420).

[0080] With the upper threshold value and the lower threshold value, the determination of whether to clear the candidate time period exceeding the upper threshold value is no longer performed, so the calculation amount can be reduced. That is, according to this embodiment, for high-intensity signals, there is no need to repeatedly perform clearing checks. The candidate time periods with small received times determined as noise can be quickly cleared, thereby improving the turnover rate of the candidate memory, and further improving the overall logic efficiency.

[0081] Figure 15 and Figure 16 is an example diagram of the scanning range of the retrieval memory according to an embodiment of the present invention.

[0082] Referring to Figure 15 , in order to measure the TOF of one point, the signal processing device scans a plurality of laser pulses 1500 output from the light emitting element with a dynamic range (i.e., a plurality of laser pulses within the burst time), and accumulatively records the laser echo reception times in the retrieval memory 1530. That is, the scanning ranges 1510, 1520 of the retrieval memory 1530 are consistent with the dynamic range. For example, in order to detect the candidate time period, the signal processing device may divide the dynamic ranges 1510, 1520 into 64 time periods, and accumulatively record the accumulated times of each time period in 64 bins of the retrieval memory 1530.

[0083] Referring to Figure 16, the signal processing device divides the dynamic range into at least two scanning ranges 1610 and 1620, and accumulatively records in the retrieval memory 1630 the number of received laser echoes scanned in the scanning ranges 1610 and 1620. For example, when the dynamic range is divided into two scanning ranges 1610 and 1620, the signal processing device scans the number of received laser echoes of the laser pulses output in the first half of the burst time among the laser pulses 1600 output by the light emitting element within the first scanning range 1610, and accumulatively stores it in the retrieval memory 1630. Then, after resetting the retrieval memory 1630, it scans the number of received laser echoes of the laser pulses output in the second half of the burst time with the second scanning range 1620, and accumulatively stores it in the retrieval memory 1630.

[0084] Figure 16 The scanning ranges 1610 and 1620 of the retrieval memory 1630 are Figure 15 Half of the scanning ranges 1500 and 1510 of the retrieval memory 1530, so Figure 16 Can be set to have the same resolution as Figure 15 And the number of bins is Figure 15 Half. That is, Figure 16 In the embodiment of Figure 15 The size of the retrieval memory 1530 can be reduced to

[0085] In this embodiment, although an example of dividing the dynamic range into two scanning ranges 1610 and 1620 of the same size is illustrated, this is only an example, and the number of scanning ranges 1610 and 1620 for distinguishing the dynamic range and the size of each scanning range 1610 and 1620 can be variously deformed according to the embodiment. For example, the intensity of the laser echo returned from a short distance is greater than the intensity of the laser echo returned from a long distance, and the scanning range 1610 in the first half can be set to be smaller than the scanning range 1620 in the second half. According to the size of the scanning range 1620, the number of bins and the number of bits of the histogram memory are dynamically variable. As another embodiment, the number of laser pulses scanned in the scanning range 1610 in the first half and the number of laser pulses scanned in the scanning range 1620 in the second half can be the same or different. As another embodiment, the laser pulses output in the first half are scanned with the scanning range 1620 in the second half, and the laser pulses output in the second half are scanned with the scanning range 1610 in the first half.

[0086] Figure 17 And Figure 18 Is an example diagram of a method for reducing the size of the retrieval memory according to an embodiment of the present invention.

[0087] Refer to Figure 17 And Figure 18, the size of the search memory can be reduced by decreasing the number of bits allocated to the histogram memory bins that make up the search memory. Figure 18 The left diagram of Figure 17 is a process of determining a histogram of the number of receptions based on the left search memory of Figure 18 The right diagram of Figure 17 is a process of determining a histogram of the number of receptions based on the right search memory of Figure 18 As can be seen from the left curve graphs 1800, 1802, 1804, 1806 of Figure 18 even if the number of detections increases, each bin does not become a saturation state. As can be seen from the right curve graphs 1810, 1812, 1814, 1816 of

[0088] Figure 19 is an example diagram of a method for reducing the size of the search memory and improving the sensing success rate according to an embodiment of the present invention.

[0089] Referring to Figure 19 , the signal processing device can improve the sensing success rate by subtracting a predefined value (hereinafter referred to as the subtracted value) from the accumulated number of receptions in the search memory at each fixed period (hereinafter referred to as the subtraction period). For example, after accumulating the number of laser echo receptions in the histogram memory within a certain period of time 1900, at each subtraction period, the subtracted value (for example, 3) (1910, 1920) can be subtracted from the number of receptions 1912, 1922 in each bin. When the accumulated number of receptions in each bin is less than the subtracted value, the signal processing device can initialize the bin to 0. The signal processing device repeatedly executes the process of subtracting the subtracted value at each subtraction period, and then determines the reception time period of the laser echo based on the final accumulated number of receptions (1930).

[0090] According to various purposes, the candidate threshold value, the subtraction period, and the subtracted value can be adjusted to various values. For example, by reducing the candidate threshold value, the candidate time period can be found quickly. However, when the noise is relatively large, there is a defect that a larger number of histogram memories are required in the candidate memory. The candidate threshold value can be reduced in an environment with less noise.

[0091] As another example, in an environment with a lot of noise, increasing the deletion value helps to remove noise, but in an environment where the signal strength (i.e., the received strength of the laser echo) is small, there is a drawback that the signal and the noise are removed together. Therefore, after calculating the minimum value of the signal (i.e., the laser echo) reflected at the maximum distance to be measured, according to the noise level at the action time point, the candidate threshold value can be adjusted as small as possible, and the deletion period and the deletion value can be adjusted to appropriate values.

[0092] As other embodiments, when using Figure 16 in the partial scan ranges 1610 and 1620, the sizes of the signals scanned in the respective scan ranges 1610 and 1620 can be different from each other. Therefore, the candidate threshold value, the deletion period, and the deletion value for each of the scan ranges 1610 and 1620 can be set separately.

[0093] Figure 20 is an example diagram of a method for loading set values according to pixels using a lookup table according to an embodiment of the present invention.

[0094] Referring to Figure 20 , the signal processing device can store and manage the set values for the optimal operation of each pixel in a lookup table. The lookup table stores and manages various set values mapped to the light intensity. For example, in an environment where the external light (i.e., the background light) is strong, even if the signal returned from a distance cannot be determined, the size of the candidate threshold value can be increased, the deletion period can be shortened, and the deletion value can be increased to completely remove the noise. As yet another example, in an environment where the external light is weak, by reducing the candidate threshold value, a candidate time period can be quickly found, and by adopting a method similar to the 1-step histogram method, the maximum value of the cumulative number of candidate time periods in the candidate memory can be increased, thereby improving the accuracy.

[0095] To this end, the signal processing device first measures the background light intensity 2000 of each pixel and loads the set value (2010) corresponding to the background light intensity from the lookup table. Moreover, the signal processing device measures the TOF (2020) and reads it out (2030). The set value loaded from the lookup table may include Figure 7 and Figure 8 the candidate threshold value for selecting a candidate time period as shown, Figures 10 to 15 the clearing threshold value for determining whether to clear as shown, Figure 16 the partial scan range as shown, Figure 17 the deletion period and the deletion value as shown, etc. The background light intensity of each pixel is determined by measuring the external light with a light receiving element, and there is no need for the light emitting element to output a laser.

[0096] Figure 21 is an example diagram of an arrangement method of a lookup table according to an embodiment of the present invention.

[0097] Reference Figure 21 As shown in Figure 21 , the signal processing device may include a plurality of look-up tables 2170 and 2180. The look-up tables 2170 and 2180 store set values mapped to various light intensities. When the light receiving element processes the laser echo in row units, one look-up table 2170 or 2180 may be provided for each of the plurality of columns 2100, 2110, 2120, and 2130. This embodiment presents a case where there are two look-up tables 2170 and 2180.

[0098] After the signal processing device first determines the light intensity 2160 of each pixel (i.e., the intensity of the background light), it refers to the look-up tables 2170 and 2180 based on the light intensity 2160 and loads the set values of each pixel. For example, after the signal processing device loads the set value 2140 of the search memory and the set value 2150 of the candidate memory from the look-up tables 2170 and 2180, it performs Figures 7 to 19 the method on this pixel.

[0099] This embodiment shows two look-up tables 2170 and 2180, but this is only an example. The look-up tables 2170 and 2180 may be one or at least three. Alternatively, each of the columns 2100, 2110, 2120, and 2130 may have a look-up table 2170 or 2180. Although an increase in the number of look-up tables 2170 and 2180 can shorten the time required for indexing and loading each of the columns 2100, 2110, 2120, and 2130, it will increase the area of the look-up tables 2170 and 2180. Therefore, according to different embodiments, an appropriate number of look-up tables 2170 and 2180 can be implemented in consideration of factors such as loading time and area.

[0100] Figure 22 FIG. is an exemplary diagram of the memory structure of each pixel according to an embodiment of the present invention.

[0101] Reference Figure 22 As shown in Figure 22 , each of the pixels 2200, 2202, and 2204 has a search memory 2210, 2212, and 2214 and a candidate memory 2220, 2222, and 2224. Each pixel may have a signal processing device. That is, the signal processing device of each pixel may use the search memory 2210, 2212, and 2214 and the candidate memory 2220, 2222, and 2224 to determine the reception time period of the laser echo.

[0102] In terms of performance, the more histogram memories in the candidate memory, the better. However, the more there are, the drawback is that the area of the candidate memory increases. Except for the case where noise accumulates in the candidate memory, there only needs to be one histogram memory laser per pixel. However, due to the existence of noise, multiple histogram memories are required in the candidate memory. To reduce the size of the candidate memory and at the same time reduce the performance degradation caused by noise, the candidate memory can be shared. This will be described in Figure 23 as follows.

[0103] For ease of explanation, this embodiment illustrates three pixels 2200, 2202, 2204. However, this is only an example, and the number of pixels constituting the sensor array can be different according to different embodiments. Just for ease of explanation, in the following embodiments including this embodiment, the explanation will be based on three pixels 2200, 2202, 2204.

[0104] Figure 23 It is another example diagram of the memory structure of each pixel according to an embodiment of the present invention.

[0105] Referring to Figure 23 , the signal processing device uses a candidate memory 2340 obtained by integrating the histogram memories of multiple pixels 2300, 2302, 2304. The number of candidate time periods detected by each pixel 2300, 2302, 2304 will be different. For example, one candidate time period is retrieved in the first pixel 2300, while five candidate time periods are retrieved in the second pixel. If the number of histogram memories in the candidate memories 2220, 2222, 2224 of each pixel is 3, then there will be a surplus of histogram memories in the candidate memory 2200 of the first pixel 2300, and there will not be enough histogram memories in the candidate memory of the second pixel 2302.

[0106] Thus, in this embodiment, the histogram memories of multiple pixels 2300, 2302, 2304 are integrated and used. Since the candidate memory is shared between adjacent pixels, the number of histogram memories in the candidate memory of each pixel can be reduced, and thus the area of the signal processing device can be reduced. Just to distinguish which pixel the candidate time period recorded in the candidate memory belongs to, address information 2352 is required.

[0107] More specifically, as described below, the signal processing device performs a search process for candidate time periods through the candidate selection unit 2320 and by using the search memories 2310, 2312, and 2314 of the respective pixels 2300, 2302, and 2304. That is, the process of determining the candidate time periods for the respective pixels can be performed independently. If the candidate selection unit 2320 retrieves a candidate time period in any one of the multiple pixels 2300, 2302, and 2304, the allocation unit 2330 allocates the histogram memory 2354 in the integrated candidate memory 2340 to the candidate time period. At this time, in order to distinguish in which pixel the candidate time period is retrieved, the address information 2352 is stored together. Since the search memories 2310, 2312, and 2314 and the pixels 2300, 2302, and 2304 are in a one-to-one correspondence, the address information of the search memory (for example, the start position address of the search memory, etc.) can be stored in the address information 2352.

[0108] Figure 24 It is another example diagram of the memory structure of each pixel according to an embodiment of the present invention.

[0109] Referring to Figure 24 , the signal processing device integrates and uses the search memories of the multiple pixels 2400, 2402, and 2404 and the candidate memories of the multiple pixels. For example, the search memory of each pixel can be Figure 16 the histogram memory (hereinafter, partial scan memory) shown for scanning a part of the dynamic range. Figure 16 Although it has the benefit of being able to reduce the size of the search memory, it does not accumulate all the laser echo reception counts within the burst time of one cycle. Instead, it only scans a part of the dynamic range in the first half of the burst time, accumulates the laser echo reception counts, and then resets the search memory. Then, it scans the remaining part of the dynamic range in the second half of the burst time and accumulates the laser echo reception counts. Therefore, the maximum value of the laser echo reception counts accumulated in each scan range becomes smaller, resulting in a situation where data for improving accuracy cannot be sufficiently ensured.

[0110] As a method for solving this problem, multiple pixels 2400, 2402, 2404 are combined and a retrieval memory 2420 that integrates partial scan memories of the respective pixels 2400, 2402, 2404 is used. As shown in this embodiment, when there are three pixels 2400, 2402, 2404, the partial scan memory of each pixel may be a memory that scans a scan range after equally dividing a dynamic range into three parts. If the scan ranges of the three pixels 2400, 2402, 2404 are integrated and used, all dynamic ranges can be scanned at once. The pixel integration unit 2410 integrates the three pixels 2400, 2402, 2404 and determines whether a laser echo is received, and accumulatively records the number of receptions in each time period in the integrated retrieval memory 2420. The configurations of the candidate selection unit 2430, the allocation unit 2440, and the integrated candidate memory 2450 including the address information 2452 and the histogram memory 2454, etc. are the same as those Figure 23 shown, and additional description thereof is omitted here. Depending on different usage environments, the signal processing device may be used only after integrating the candidate memory as shown in Figure 23 , or may be used after integrating the retrieval memory and the candidate memory as shown in Figure 24 .

[0111] Figure 25 is an exemplary diagram of a method for reconfiguring a candidate memory according to an embodiment of the present invention.

[0112] Referring to Figure 25 , the signal processing device can dynamically adjust the number of bins of the candidate memory and the number of bits of each bin. For example, the signal processing device may configure the histogram memory in the candidate memory as a first structure 2510, and the first structure includes 16 bins of 2 bits. The interval of the candidate time period (i.e., the resolution) is greater than the pulse width 2500 of the laser echo, so it is impossible to accurately confirm at which position in the candidate time period the laser echo exists. Therefore, the entire candidate time period is subdivided into sub-time periods, and the reception time point of the laser echo is first determined by the histogram memory of the first structure 2510 including a plurality of bins corresponding to each sub-time period.

[0113] If there is a bin in the histogram memory of the first structure 2510 whose cumulative count reaches a predefined first reconstruction threshold value (e.g., 22 when it is a 2-bit bin), the signal processing device transforms the histogram memory of the first structure 2510 into a second structure 2520, which includes 8 bins of 4 bits. For example, the signal processing device takes the bin that reaches the first reconstruction threshold value (e.g., the 6th bin) in the first structure 2510 as the center, retains a certain number of bins on both sides, deletes the remaining bins, and then assigns the bits of the deleted bins to the retained bins. That is, although the total number of bits in the histogram memory is the same, as the first structure 2510 is transformed into the second structure 2520, the maximum value of the received counts that can be cumulatively recorded in each bin becomes larger.

[0114] If there is a bin (e.g., the sixth bin) in the second structure 2520 whose cumulative received count reaches a second reconstruction threshold value (e.g., 24 when it is a 4-bit bin), the signal processing device takes the bin that reaches the second reconstruction threshold value as the center in the histogram memory of the second structure 2520, retains a certain number of surrounding bins, deletes the remaining bins, and then expands the number of bits of the retained bins, thereby transforming it into a third structure 2530. By this process, the cumulative received count of the laser echo existing in the candidate time period is increased, and the laser echo in the candidate time period can be accurately distinguished as noise or an actual signal.

[0115] Although this embodiment illustrates an example of the transformation process in which the first structure 2510 is transformed into the third structure 2530 through two transformations, this is only an example, and the number of times of the transformation process can be variously changed according to different embodiments. In addition, the number of bins retained centered on the bin that reaches the reconstruction threshold value in the histogram memories of the respective structures 2510, 2520, and 2530, etc., can be variously changed according to different embodiments.

[0116] The present invention can be implemented as computer-readable program code in a computer-readable recording medium. The computer-readable recording medium includes any type of recording device for storing data readable by a computer system. Examples of such computer-readable recording media include ROM, RAM, CD-ROM, magnetic tapes, floppy disks, and optical data storage devices, etc. In addition, the computer-readable recording medium can be distributed in a network-connected computer system and store and execute the computer-readable code in a distributed manner.

[0117] The present invention has been described above centered on the preferred embodiments. Those of ordinary skill in the technical field to which the present invention pertains can understand that within the scope not departing from the essential features of the present invention, the present invention can be implemented in other modified forms. Therefore, the disclosed embodiments should be construed as being merely descriptive and not for the purpose of limitation. The scope of the present invention is not defined by the detailed description of the present invention, but by the claims, and all differences within the equivalent scope should be construed as being included within the scope of the present invention.

Claims

1. A signal processing method for an optical receiving element, comprising the following steps: Accumulatively record the number of laser echo receptions in N (N is a natural number greater than or equal to 2) time periods into a retrieval memory; Select at least one time period in which the number of receptions reaches a candidate threshold value as a candidate time period; Divide the candidate time period into M (M is a natural number greater than or equal to 2) sub - time periods, and accumulatively record the number of laser echo receptions in each sub - time period into a candidate memory; And Based on the number of receptions in each sub - time period accumulatively stored in the candidate memory, determine the reception time point of the laser echo.

2. The signal processing method for an optical receiving element according to claim 1, wherein The candidate memory includes a plurality of histogram memories for recording the number of laser echo receptions in the candidate time period, and the step of accumulatively recording into the candidate memory includes the following steps: According to the order in which the number of receptions reaches the threshold value, respectively allocate histogram memories to the candidate time period; and Accumulatively record the number of laser echo receptions in multiple sub - time periods within the candidate time period into the allocated histogram memory.

3. The signal processing method for an optical receiving element according to claim 1, wherein Further include the following steps: Clear the histogram memories allocated to candidate time periods in which the number of laser echo receptions does not reach a clearing threshold value during a predefined clearing check interval.

4. The signal processing method for an optical receiving element according to claim 3, wherein Further include the following steps: Allocate the cleared histogram memories to new candidate time periods determined by the retrieval memory.

5. The signal processing method for an optical receiving element according to claim 3, wherein The clearing check interval is defined based on the number of output times of laser pulses.

6. The signal processing method for an optical receiving element according to claim 3, wherein Further include the following steps: Repeatedly execute the clearing step at a certain period.

7. The signal processing method for an optical receiving element according to claim 6, wherein The step of repeatedly executing the clearing step at a certain period includes the following steps: For a candidate time period in which the number of laser echo receptions exceeds an upper threshold value, end the step of repeated execution; and If the number of laser echo receptions is less than a lower threshold value, clear the histogram memory allocated to the candidate time period.

8. The signal processing method for an optical receiving element according to claim 1, wherein The step of accumulatively recording into the retrieval memory includes the following steps: Divide the dynamic range for measuring TOF into multiple scan ranges; and For each of the multiple scan ranges, respectively execute the process of recording the number of laser echo receptions in the scan range into the retrieval memory.

9. The signal processing method for an optical receiving element according to claim 8, wherein The number of laser echoes scanned in at least two of the multiple scan ranges is different from each other.

10. The signal processing method for an optical receiving element according to claim 1, wherein The step of accumulatively recording into the retrieval memory further includes the following steps: Accumulatively store the number of laser echo receptions in N (N is a natural number) time periods into a histogram memory including N bins; For the accumulative number of receptions in the time periods stored in the N bins, perform deletion using a predefined deletion value at a predefined deletion period.

11. The signal processing method for an optical receiving element according to claim 1, wherein Include a look - up table that stores candidate threshold values respectively mapped to multiple different light intensities, and use the look - up table to determine the candidate threshold value corresponding to the intensity of the background light measured by the light receiving element.

12. The signal processing method of the optical receiving element according to claim 1, characterized in that, Each pixel has the retrieval memory, and multiple pixels share the candidate memory.

13. The signal processing method of the optical receiving element according to claim 12, characterized in that, The step of accumulatively recording into the candidate memory includes the following steps: Store the pixel address information where the candidate time period is found and the number of laser echo receptions into the candidate memory.

14. The signal processing method of the optical receiving element according to claim 1, characterized in that, The retrieval memory is formed by integrating partial scan memories existing in a plurality of pixels, and the step of cumulatively recording into the retrieval memory includes the step of determining the number of laser echo receptions by combining the plurality of pixels.

15. The signal processing method of the optical receiving element according to claim 1, characterized in that, The step of cumulatively recording into the candidate memory includes the following steps: Storing the number of laser echo receptions into a histogram memory, the histogram memory including a plurality of bins each composed of a plurality of bits; and In the histogram memory, a certain number of bins are reserved centered around the bins of sub-time periods in which the number of receptions reaches a predefined reconstruction threshold value, and after deleting the remaining bins, the bits of the deleted bins are assigned to the reserved bins and the histogram memory is reconstructed.

16. A signal processing device for an optical receiving element, comprising: A retrieval memory for cumulatively recording the number of laser echo receptions in N (N is a natural number of 2 or more) time periods; And A candidate memory for cumulatively recording the number of laser echo receptions in at least one time period corresponding to a predefined candidate threshold value of the reception number, The retrieval memory and the candidate memory include a histogram memory, the histogram memory including a plurality of bins each composed of a plurality of bits.

17. The signal processing device for an optical receiving element according to claim 16, characterized in that, The candidate memory includes a plurality of histogram memories for being assigned to a plurality of candidate time periods.

18. The signal processing device for an optical receiving element according to claim 16, characterized in that, The retrieval memory is a histogram memory that cumulatively stores the number of laser echo receptions in a scan range smaller than the dynamic range used to measure the TOF.

19. The signal processing device for an optical receiving element according to claim 16, characterized in that, Further includes: A look-up table for storing an optimal candidate threshold value mapped to the light intensity, The candidate threshold value is determined based on the intensity of the background light of the light receiving element and with reference to the look-up table.

20. The signal processing device for an optical receiving element according to claim 19, characterized in that, The look-up table is provided in the pixel array in units of a plurality of rows or a plurality of columns.

21. The signal processing device for an optical receiving element according to claim 16, characterized in that, A plurality of pixels share the candidate memory.

22. The signal processing device for an optical receiving element according to claim 16, characterized in that, The retrieval memory is formed by integrating partial scan memories existing in a plurality of pixels, and the retrieval memory records the number of laser echo receptions determined by combining the plurality of pixels.

23. The signal processing device for an optical receiving element according to claim 16, characterized in that, In the histogram memory of the candidate memory, the number of bins and the number of bits of each bin are dynamically variable.

Images