Lidar sensor and method of removing noise therefrom
By employing dual histogram processing and optical pulse control, the lidar sensor achieves high-sensitivity signal sensing and noise removal, solving the problem of increased noise and improving signal accuracy and precision.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- SOLIDVUE INC
- Filing Date
- 2022-04-01
- Publication Date
- 2026-05-29
AI Technical Summary
Existing lidar sensors suffer from increased noise when sensitivity is improved, making it difficult to simultaneously achieve high-sensitivity signal sensing and noise removal.
The system employs a dual histogram processing unit and an interference noise remover. The first histogram processing unit extracts data values greater than or equal to a first reference value from the histogram bin. The second histogram processing unit accumulates the data values and removes noise. Interference noise is also removed by controlling the light pulse emission delay time and time difference.
It achieves highly sensitive signal sensing and noise removal, offsetting the increase in noise caused by increasing the reverse voltage, and maintaining the accuracy and precision of the signal.
Smart Images
Figure CN115201781B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to a lidar sensor, and more particularly to a lidar sensor capable of removing background noise and a method for removing noise therefrom. Background Technology
[0002] In general, a lidar sensor is a device that emits laser pulses and receives reflected light from surrounding targets to accurately map the surrounding environment by measuring the distance, direction, material, features, and so on of the targets.
[0003] LiDAR sensors are used to make more precise observations of the physical properties of the atmosphere and to measure distances by taking advantage of the ability of lasers to generate pulse signals with high energy density and short period.
[0004] Furthermore, lidar sensors can be classified into time-of-flight (TOF) and phase-shift (PS) methods based on the laser signal modulation method.
[0005] The Time-of-Flight (TOF) method is a method for measuring distance by emitting a pulse signal from a laser and measuring the time it takes for the pulse signal to be reflected back from the object within a measurement range. The PS method is a method for calculating time and distance by emitting a laser beam that is continuously modulated to a specific frequency and measuring the amount of phase change of the signal reflected back from the object within a measurement range.
[0006] When lidar sensors are applied to external environments, high-sensitivity signal sensing and noise removal are crucial due to the presence of various types of background noise, such as sunlight and dark noise.
[0007] However, in lidar sensors, when the overvoltage of the reverse voltage of the single-photon avalanche diode (SPAD) is increased in order to increase sensitivity, there is a problem that the noise also increases while the sensitivity is improved.
[0008] Therefore, there is a need to develop a lidar sensor that can perform high-sensitivity signal sensing and noise removal as well as eliminate interference between lidar sensors. Summary of the Invention
[0009] 1. Technical issues
[0010] The technical objective of this invention is to provide a lidar sensor that can remove noise for the first time by extracting data values greater than or equal to a first reference value from the bins of a histogram using a first histogram processing unit, and remove noise a second time by extracting accumulated data values greater than or equal to a second reference value from the bins of a histogram containing data values extracted by a second histogram processing unit, thereby simultaneously performing high-sensitivity signal sensing and noise removal, and also provides a method for removing noise therefrom.
[0011] Furthermore, another technical objective of the present invention is to provide a lidar sensor that can remove interference noise between lidar sensors by controlling the light-emitting unit to change the light pulse emission delay time and light pulse time difference (TD) between the first beam and the second beam in response to histogram processing performed by the first histogram processing unit, and to provide a method for removing noise therefrom.
[0012] The technical problems to be solved by the present invention are not limited to the technical problems described above, and other technical problems not mentioned can also be clearly understood by those skilled in the art through the following disclosure.
[0013] 2. Solution to the problem
[0014] A lidar sensor according to one embodiment of the present invention includes: a light receiving unit configured to sense reflected light reflected from an object; a weight generation unit configured to generate weights based on the sensitivity of the reflected light from the light receiving unit; a first histogram processing unit configured to perform histogram processing on the sensed signal of the light receiving unit based on the generated weights and extract data values greater than or equal to a first reference value from the histogram bins, thereby removing noise for the first time; and a second histogram processing unit configured to perform histogram processing by accumulating the data values extracted by the first histogram processing unit and extract the accumulated data values greater than or equal to a second reference value from the histogram bins of the accumulated data values, thereby removing noise for the second time.
[0015] In an alternative implementation of the lidar sensor, the weight generation unit can divide the multiple single-photon avalanche diodes (SPADs) contained in each light-receiving pixel of the light-receiving unit into multiple subgroups, determine whether the number of SPADs in the subgroup that simultaneously senses reflected light is greater than or equal to a reference number, generate weights for the sensed signal based on the number of subgroups that are greater than or equal to the reference number, and output the generated weights to the first histogram processing unit.
[0016] In an alternative embodiment of the lidar sensor, when it is determined whether the number of SPADs that simultaneously sense reflected light in a subgroup is greater than or equal to a reference number, the weight generation unit can calculate the number of SPADs that simultaneously sense reflected light in the subgroup, compare the calculated number of SPADs with a preset reference number, classify the subgroups corresponding to the calculated number of SPADs, and calculate the number of classified subgroups when the calculated number of SPADs is greater than or equal to the reference number.
[0017] In an alternative implementation of the lidar sensor, the first histogram processing unit may check whether weights are generated during histogram processing of the sensed signal and may reflect the generated weights in the data values of the sensed signal, so as to perform histogram processing when generating weights.
[0018] In an alternative implementation of the lidar sensor, when checking whether weights are generated, the first histogram processing unit can check whether a sensing signal exists when no weights are generated, and can process the data value of the sensing signal to 1 through histogram if a sensing signal exists.
[0019] In an alternative implementation of the lidar sensor, when noise is removed for the first time, the first histogram processing unit checks whether the data value of each bin is greater than or equal to a first reference value. If the data value of a bin is greater than or equal to the first reference value, the difference between the data value of the bin and the first reference value can be calculated, and the calculated difference value can be extracted and output to the second histogram processing unit.
[0020] In an alternative implementation of the lidar sensor, when checking whether the data value of each bin is greater than or equal to a first reference value, the first histogram processing unit can treat the data value of the corresponding bin as noise, and remove the noise when the data value of the bin is less than the first reference value.
[0021] In an alternative implementation of the lidar sensor, when performing histogram processing, the second histogram processing unit can generate a bin corresponding to the first data value when it receives the first data value from the first histogram processing unit, generate or accumulate a bin corresponding to the second data value when it receives the second data value from the first histogram processing unit, and generate or accumulate a bin corresponding to the Nth data value when it receives the Nth data value from the first histogram processing unit.
[0022] In an alternative implementation of the lidar sensor, the second histogram processing unit may check whether the bin corresponding to the second data value is the same as the bin corresponding to the first data value when it receives the second data value from the first histogram processing unit. If the check result is the same, the second data value may be added to the first data value of the corresponding bin. If the check result is different, a new bin corresponding to the second data value may be generated.
[0023] In an alternative implementation of the lidar sensor, the second histogram processing unit may check whether the bin corresponding to the Nth data value is a previously generated bin when it receives the Nth data value from the first histogram processing unit. If the bin is a previously generated bin according to the check result, the Nth data value may be added to the data value of the previously generated bin. If the check result is different, a new bin corresponding to the Nth data value may be generated.
[0024] In an alternative implementation of the lidar sensor, during the second noise removal, the second histogram processing unit may check whether the Nth data value is the last data value when it receives the Nth data value from the first histogram processing unit, and may generate or accumulate the bin corresponding to the last data value when the Nth data value is the last data value based on the check result, and extract the accumulated data value that is greater than or equal to the second reference value from the bin of the histogram.
[0025] In an alternative implementation of the lidar sensor, when checking whether the Nth data value is the last data value, the second histogram processing unit may check whether the Nth data value is the last data value based on a preset number of histogram processing steps.
[0026] In an alternative implementation of the lidar sensor, during the second noise removal process, the second histogram processing unit can check whether the data value of each bin is greater than or equal to a second reference value, calculate the difference between the data value of the bin and the second reference value, and subsequently extract and output the calculated difference value when the data value of the bin is greater than or equal to the second reference value.
[0027] In an alternative implementation of the lidar sensor, when checking whether the data value of each bin is greater than or equal to the second reference value, the second histogram processing unit can treat the data value of the corresponding bin as noise, and remove the noise when the data value of the bin is less than the second reference value.
[0028] In an alternative implementation of the lidar sensor, the second reference value of the second histogram processing unit may be different from the first reference value of the first histogram processing unit.
[0029] In an alternative embodiment of the lidar sensor, the lidar sensor may further include: a light-emitting unit configured to emit a first beam and a second beam toward an object; and an interference noise remover configured to control the light-emitting unit to change the light pulse emission delay time and light pulse time difference (TD) between the first beam and the second beam in response to histogram processing performed by a first histogram processing unit.
[0030] In an alternative implementation of the lidar sensor, the interference noise remover can check whether the first histogram processing unit is performing first histogram processing, and can control the light-emitting unit to change the emission start time of the first beam and the second beam in different ways while the first histogram processing unit is performing first histogram processing, and keep the time difference (TD) of the light pulse between the first beam and the second beam constant.
[0031] In an alternative implementation of the lidar sensor, when checking whether the first histogram processing unit performs the first histogram processing, the interference noise remover can control the light-emitting unit to change the emission start time of the first beam and the second beam in a different way so that the current light pulse emission delay time is different from the previous light pulse emission delay time, and when the first histogram processing unit performs the Nth histogram processing instead of the first histogram processing, the current light pulse time difference (TD) is different from the previous light pulse time difference.
[0032] In an alternative implementation of the lidar sensor, the interference noise remover can modify the current optical pulse emission delay time corresponding to the Nth histogram processing and the previous optical pulse emission delay time corresponding to the (N-1)th histogram processing in different ways, and can also modify the current optical pulse time difference (TD) corresponding to the Nth histogram processing and the previous optical pulse time difference corresponding to the (N-1)th histogram processing in different ways.
[0033] In an alternative embodiment of the lidar sensor, the lidar sensor may further include a random number generation unit configured to generate two-dimensional random numbers, wherein the interference noise remover may modify the optical pulse emission delay time and optical pulse time difference (TD) between the first beam and the second beam based on the two-dimensional random numbers.
[0034] In an alternative embodiment of the lidar sensor, the random number generation unit may include first and second single-photon avalanche diodes (SPADs), a light shield configured to block the beam emitted to either the first or second SPAD, and a random number generator connected to the first and second SPADs to generate random numbers in response to an input pulse.
[0035] Meanwhile, a method for removing noise from a lidar sensor comprising a light receiving unit, a weight generation unit, a histogram processing unit, and an interference noise removal unit according to an embodiment of the present invention includes: sensing reflected light reflected from an object by the light receiving unit; generating weights by the weight generation unit based on the reflective light sensitivity of the light receiving unit; performing histogram processing on the sensed signal of the light receiving unit based on the generated weights by the histogram processing unit and extracting data values greater than or equal to a first reference value from the histogram bins, thereby removing noise for the first time; and performing histogram processing by the histogram processing unit accumulating the extracted data values and extracting accumulated data values greater than or equal to a second reference value from the histogram bins of the accumulated data values, thereby removing noise for the second time.
[0036] 3. Beneficial effects
[0037] The effects of the lidar sensor according to the present invention and the method for removing noise therefrom will be described below.
[0038] In this invention, noise can be removed for the first time by extracting data values greater than or equal to a first reference value from the bin of the histogram using a first histogram processing unit, and for the second time by extracting accumulated data values greater than or equal to a second reference value from the bin of the histogram containing the data values extracted by the second histogram processing unit, thereby simultaneously performing highly sensitive signal sensing and noise removal.
[0039] Furthermore, in this invention, by means of histogram processing performed by the first histogram processing unit, the light emission unit can be controlled to change the light pulse emission delay time and light pulse time difference (TD) between the first beam and the second beam, thereby removing interference noise between lidar sensors.
[0040] Furthermore, in this invention, even when combined with a double histogram by increasing Vex, it has the effect of counteracting the interference caused by the increase in dark count rate due to the increase in Vex, and the SPAD sensitivity (photon detection probability) increases due to the increase in Vex.
[0041] In general, the reason for not intentionally increasing Vex is that DCR increases rapidly. However, in this invention, since a double histogram can be used to remove the increased Vex, noise can be removed even when Vex is intentionally increased.
[0042] Further applicability of the invention will become apparent from the following detailed description. However, it should be understood that various modifications and variations within the spirit and scope of the invention will be readily apparent to those skilled in the art, and that the detailed description and specific embodiments (e.g., preferred embodiments of the invention) are given by way of example only. Attached Figure Description
[0043] Figure 1 This includes block diagrams and circuit diagrams for describing the lidar sensor according to the present invention.
[0044] Figure 2 This is a view used to illustrate the noise removal effect of the lidar sensor according to the present invention.
[0045] Figures 3A to 3C This is a view used to describe the process of removing interference noise according to the present invention.
[0046] Figure 4 This is a set of views used to describe the outdoor sensitivity of the lidar sensor according to the present invention.
[0047] Figure 5 This is a view used to describe the indoor sensitivity of the lidar sensor according to the present invention. Detailed Implementation
[0048] The invention will be described in more detail below with reference to the accompanying drawings.
[0049] The suffixes “module” and “unit” used for components in the following description are given only for the convenience of writing the manual, and “module” and “unit” are used interchangeably.
[0050] Furthermore, embodiments of the present invention will be described in detail with reference to the accompanying drawings and the descriptions therein, but the present invention is not limited or constrained by these embodiments.
[0051] As far as possible, widely used and common terms have been selected for use in this specification while taking into account the functionality of the invention. However, these terms may vary according to the invention, the habits of those skilled in the art, or the emergence of new technologies. Furthermore, in certain cases, terms may be arbitrarily chosen by the applicant, and in such cases, their meanings will be described in the description of the invention. Therefore, it should be understood that the terms used in this specification should be interpreted based on their actual meaning and the content of the entire specification, and not merely their names.
[0052] Figure 1 This includes block diagrams and circuit diagrams for describing the lidar sensor according to the present invention.
[0053] like Figure 1As shown, the lidar sensor of the present invention includes: a light receiving unit 100 that senses reflected light reflected from an object; a weight generation unit 200 configured to generate weights based on the sensitivity of the reflected light of the light receiving unit 100; and a histogram processing unit 300 that performs histogram processing on the sensed signal of the light receiving unit 100 based on the generated weights and extracts data values greater than or equal to a first reference value to remove noise.
[0054] Here, the light receiving unit 100 can be configured as a pixel array including a plurality of light receiving pixels 110.
[0055] In addition, each light receiving pixel 110 may include a plurality of single-photon avalanche diodes (SPADs) 120 that convert the sensed reflected light into electrical signals.
[0056] For example, the light receiving unit 100 may be configured as a pixel array, wherein 30 light receiving pixels 110 are set as a 30×1 array, but the present invention is not limited thereto.
[0057] Here, each light receiving pixel 110 can be configured as a SPAD array with 16 single-photon avalanche diodes (SPADs) 120 arranged in a 4×4 array, but the invention is not limited thereto.
[0058] Next, the weight generation unit 200 may include: a determiner 210, which divides the plurality of single-photon avalanche diodes (SPADs) 120 contained in each light receiving pixel 110 of the light receiving unit 100 into a plurality of subgroups 130, and determines whether the number of SPADs 120 in the subgroups 130 that simultaneously sense reflected light is greater than or equal to a reference number; a weight generator 220, which generates weights for the sensed signal based on the number of subgroups 130 that are greater than or equal to the reference number; and an output device 230, which outputs the generated weights to the histogram processing unit 300.
[0059] Here, when it is determined whether the number of SPADs 120 that simultaneously sense reflected light is greater than or equal to a reference number, the weight generation unit 200 can calculate the number of SPADs 120 that simultaneously sense reflected light in the subgroup, compare the calculated number of SPADs 120 with a preset reference number, classify the subgroups 130 corresponding to the calculated number of SPADs 120, and calculate the number of classified subgroups 130 when the calculated number of SPADs 120 is greater than or equal to the reference number.
[0060] Furthermore, when determining whether the number of SPADs 120 that simultaneously sense reflected light is greater than or equal to a reference number, the weight generation unit 200 may set the reference number to approximately 50% of the total number of SPADs 120 in each subgroup 130, but the present invention is not limited thereto.
[0061] For example, the weight generation unit 200 may set the reference number to 2 when the total number of SPADs 120 in each subgroup 130 is 4, but the present invention is not limited thereto.
[0062] Furthermore, when generating weights, the weight generation unit 200 can generate the weights of the sensing signals in such a way that the weight of the sensing signal corresponding to the larger quantity of the subgroup 130 that is determined to be greater than or equal to the reference quantity is higher than the weight of the sensing signal corresponding to the smaller quantity of the subgroup 130 that is determined to be greater than or equal to the reference quantity.
[0063] For example, when the number of subgroups 130 in each light receiving pixel 110 is 4, the weight generation unit 200 can generate a weight of 2 for the sensing signal corresponding to subgroup 130 when the number of subgroups 130 is greater than or equal to the reference number, generate a weight of 4 for the sensing signal corresponding to subgroup 130 when the number of subgroups 130 is greater than or equal to the reference number, generate a weight of 6 for the sensing signal corresponding to subgroup 130 when the number of subgroups 130 is greater than or equal to the reference number, and generate a weight of 8 for the sensing signal corresponding to subgroup 130 when the number of subgroups 130 is greater than or equal to the reference number. However, the present invention is not limited to this.
[0064] Next, the histogram processing unit 300 may include a first histogram processing unit 310 and a second histogram processing unit 320.
[0065] Here, the first histogram processing unit 310 can be configured as pre-historical diagram (PRH) circuit logic, such as... Figure 1 As shown.
[0066] The first histogram processing unit 310 performs histogram processing of the sensing signal of the light receiving unit 100 based on the generated weights, and is configured to extract data values greater than or equal to the first reference value from the histogram bins, thereby removing noise for the first time.
[0067] Here, when performing histogram processing of the sensing signal, the first histogram processing unit 310 can check whether weights have been generated, and if weights have been generated, the first histogram processing unit 310 can perform histogram processing by reflecting the generated weights in the data value of the sensing signal.
[0068] For example, when checking whether weights have been generated, the first histogram processing unit 310 can check whether a sensing signal exists if no weights have been generated, and can process the data value of the sensing signal to 1 through histogram if a sensing signal exists.
[0069] Furthermore, when checking for the presence of a sensing signal, the first histogram processing unit 310 can process the data value of the sensing signal to 0 using histogram when there is no sensing signal.
[0070] In addition, when noise is removed for the first time, the first histogram processing unit 310 can check whether the data value of each bin is greater than or equal to the first reference value. When the data value of a bin is greater than or equal to the first reference value, the difference between the data value of the bin and the first reference value can be calculated, and the calculated difference value can be extracted and output to the second histogram processing unit 320.
[0071] In addition, when checking whether the data value of each bin is greater than or equal to the first reference value, the first histogram processing unit 310 can treat the data value of the corresponding bin as noise, and remove the noise when the data value of the bin is less than the first reference value.
[0072] Next, the second histogram processing unit 320 can be configured as post-histogram (POH) circuit logic, such as... Figure 1 As shown.
[0073] The second histogram processing unit 320 can perform histogram processing by accumulating the data values extracted by the first histogram processing unit 310, and can extract data values greater than or equal to the second reference value from the histogram of accumulated data values, thereby removing noise for the second time.
[0074] Here, when performing histogram processing, the second histogram processing unit 320 can generate a bin corresponding to the first data value when it receives the first data value from the first histogram processing unit 310, generate or accumulate a bin corresponding to the second data value when it receives the second data value from the first histogram processing unit 310, and generate or accumulate a bin corresponding to the Nth data value when it receives the Nth data value from the first histogram processing unit 310.
[0075] For example, when the second histogram processing unit 320 receives the second data value from the first histogram processing unit 310, it can check whether the bin corresponding to the second data value is the same as the bin corresponding to the first data value. If the check result is the same, the second data value can be added to the first data value of the corresponding bin. If the check result is different, a new bin corresponding to the second data value can be generated.
[0076] Furthermore, when the second histogram processing unit 320 receives the Nth data value from the first histogram processing unit 310, it can check whether the bin corresponding to the Nth data value is a previously generated bin. If the bin is a previously generated bin according to the check result, the Nth data value can be added to the data value of the previously generated bin. If the check result is different, a new bin corresponding to the Nth data value can be generated.
[0077] In addition, when removing noise for the second time, the second histogram processing unit 320 can check whether the Nth data value is the last data value when it receives the Nth data value from the first histogram processing unit, and can generate or accumulate the bin corresponding to the last data value when the Nth data value is the last data value according to the check result, and then extract the accumulated data value that is greater than or equal to the second reference value from the bin of the histogram.
[0078] Here, when checking whether the Nth data value is the last data value, the second histogram processing unit 320 can check whether the Nth data value is the last data value based on a preset number of histogram processing steps.
[0079] Furthermore, during the second noise removal process, the second histogram processing unit 320 can check whether the data value of each bin is greater than or equal to the second reference value, and can calculate the difference between the data value of the bin and the second reference value. Subsequently, when the data value of the bin is greater than or equal to the second reference value, the calculated difference value can be extracted and output.
[0080] Here, when checking whether the data value of each bin is greater than or equal to the second reference value, the second histogram processing unit 320 can treat the corresponding bin's data value as noise and remove the noise when the bin's data value is less than the second reference value.
[0081] Furthermore, the second reference value of the second histogram processing unit 320 may be different from the first reference value of the first histogram processing unit 310.
[0082] For example, the second reference value of the second histogram processing unit 320 may be greater than the first reference value of the first histogram processing unit 310.
[0083] Furthermore, the first reference value of the first histogram processing unit 310 and the second reference value of the second histogram processing unit 320 can be set based on the noise value and the signal sensitivity of the sensor, but the present invention is not limited thereto.
[0084] Additionally, the lidar sensor of the present invention may also include: a dual-emitting unit 400 that emits a first beam and a second beam toward a target; and an interference noise remover 500 that controls the dual-emitting unit 400 to change the optical pulse emission delay time and optical pulse time difference (TD) between the first beam and the second beam in response to histogram processing performed by the histogram processing unit 300.
[0085] Here, in this invention, the first beam and the second beam can be emitted by a single light-emitting unit instead of dual light-emitting units 400.
[0086] For example, when the optical pulse emission delay time between the first beam and the second beam changes, the interference noise remover 500 can change the optical pulse emission delay time between the first beam and the second beam by adjusting the emission start time of the first beam and the second beam to be different from each other.
[0087] Furthermore, when the optical pulse time difference (TD) between the first beam and the second beam changes, the interference noise remover 500 can change the optical pulse time difference between the first beam and the second beam by adjusting the optical paths of the first beam and the second beam to be different from each other.
[0088] Next, the interference noise remover 500 can check whether the first histogram processing unit 310 performs the first histogram processing, and can subsequently change the emission start time of the first beam and the second beam in different ways when the first histogram processing unit 310 performs the first histogram processing, and can control the dual light-emitting unit 400 to keep the light pulse time difference (TD) between the first beam and the second beam constant.
[0089] Here, when checking whether the first histogram processing unit 310 performs the first histogram processing, the interference noise remover 500 can control the dual light-emitting unit 400 to change the emission start time of the first beam and the second beam in a different way so that the current light pulse emission delay time is different from the previous light pulse emission delay time, and when the first histogram processing unit 310 performs the Nth histogram processing instead of the first histogram processing, it makes the current light pulse time difference (TD) different from the previous light pulse time difference.
[0090] For example, the interference noise remover 500 can change the current optical pulse emission delay time corresponding to the Nth histogram processing and the previous optical pulse emission delay time corresponding to the (N-1)th histogram processing in different ways, and can also change the current optical pulse time difference (TD) corresponding to the Nth histogram processing and the previous optical pulse time difference corresponding to the (N-1)th histogram processing in different ways.
[0091] Furthermore, the interference noise remover 500 can control the dual light-emitting unit 400 to keep the time difference (TD) between the first beam and the second beam constant while performing the Nth histogram processing.
[0092] In addition, the lidar sensor of the present invention may also include a random number generation unit 600 for generating two-dimensional random numbers.
[0093] Here, the interference noise remover 500 can change the optical pulse emission delay time and optical pulse time difference (TD) between the first beam and the second beam based on two-dimensional random numbers.
[0094] For example, the random number generation unit 600 may include first and second single-photon avalanche diodes (SPADs), a light shield that blocks the light beam emitted to either the first or second SPAD, and a random number generator connected to the first and second SPADs to generate random numbers in response to an input pulse, but the present invention is not limited thereto.
[0095] Here, the random number generation unit 600 can generate a 5-digit two-dimensional random number, but the present invention is not limited thereto.
[0096] Furthermore, when the light pulse emission delay time and light pulse time difference (TD) between the first beam and the second beam change, the interference noise remover 500 can control the random number generation unit 600 to generate two-dimensional random numbers.
[0097] As described above, in this invention, noise can be removed for the first time by extracting data values greater than or equal to a first reference value from the histogram bins only by the first histogram processing unit 310, and for the second time by extracting accumulated data values greater than or equal to a second reference value from the histogram bins containing the data values extracted by the second histogram processing unit, thereby simultaneously performing highly sensitive signal sensing and noise removal.
[0098] Furthermore, in this invention, interference noise between lidar sensors can be removed by controlling the dual-emission unit 400 to change the optical pulse emission delay time and optical pulse time difference (TD) between the first beam and the second beam in response to histogram processing performed by the first histogram processing unit 310.
[0099] Figure 2 This is a view used to illustrate the noise removal effect of the lidar sensor according to the present invention.
[0100] like Figure 2 As shown, the left-hand graph illustrates the traditional histogram method, while the upper-left graph shows the case where Vex (reverse voltage) is low and there is no sunlight. Due to the lack of sunlight, the histogram data representing noise (error) is small, but because Vex is low, the histogram data representing signal sensitivity does not exceed VTH (threshold voltage).
[0101] Furthermore, in the middle left graph, due to the increase in Vex, it can be seen that although some data values of the histogram representing signal sensitivity exceed VTH, the data values of the histogram representing noise also increase, thus producing an error.
[0102] Additionally, in the lower left figure, it can be seen that the data values representing noise in the histogram also increase as Vex increases, and multiple errors are introduced due to sunlight.
[0103] Conversely, in the figure on the right, according to the histogram method of the present invention, high-sensitivity signal sensing and noise removal can be performed simultaneously through two-step histogram processing by means of a first histogram processing unit 310 configured as pre-histogram (PRH) circuit logic and a second histogram processing unit 320 configured as post-histogram (POH) circuit logic.
[0104] The first histogram processing unit of the front histogram (PRH) can extract data values greater than or equal to the first reference value from the histogram's representation of the sensed signal based on the generated weights, thereby removing noise for the first time.
[0105] Furthermore, when performing histogram processing on the sensed signal, the first histogram processing unit can check whether weights have been generated, and if weights have been generated, the first histogram processing unit 310 can perform histogram processing by reflecting the generated weights in the data values of the sensed signal.
[0106] As described above, the first histogram processing unit can check whether the data value of each bin is greater than or equal to the first reference value. When the data value of a bin is greater than or equal to the first reference value, it can calculate the difference between the data value of the bin and the first reference value, and extract the calculated difference value and output the difference value to the second histogram processing unit 320.
[0107] In addition, the first histogram processing unit can treat the data value of the corresponding bin as noise and remove the noise when the data value of the bin is less than the first reference value.
[0108] Next, the second histogram processing unit of the post-historical diagram (POH) can perform histogram processing by accumulating the data values extracted by the first histogram processing unit 310, and can extract the accumulated data values that are greater than or equal to the second reference value from the histogram of accumulated data values, thereby removing noise for the second time.
[0109] Here, the second histogram processing unit can generate a bin corresponding to the first data value when it receives the first data value from the first histogram processing unit, generate or accumulate a bin corresponding to the second data value when it receives the second data value from the first histogram processing unit, and generate or accumulate a bin corresponding to the Nth data value when it receives the Nth data value from the first histogram processing unit.
[0110] Furthermore, the second histogram processing unit can check whether the Nth data value is the last data value when it receives the Nth data value from the first histogram processing unit, and can generate or accumulate the bin corresponding to the last data value when the Nth data value is the last data value according to the check result, and can subsequently extract the accumulated data value that is greater than or equal to the second reference value from the bin of the histogram.
[0111] Here, the second histogram processing unit can check whether the data value of each bin is greater than or equal to the second reference value, and can calculate the difference between the data value of the bin and the second reference value. It can then extract and output the calculated difference value when the data value of the bin is greater than or equal to the second reference value.
[0112] In addition, the second histogram processing unit can treat the data value of the corresponding bin as noise and remove the noise when the data value of the bin is less than the second reference value.
[0113] Figures 3A to 3C This is a view used to describe the process of removing interference noise according to the present invention.
[0114] like Figures 3A to 3C As shown, the lidar sensor of the present invention can control the dual-emission unit to change the optical pulse emission delay time (PP) and optical pulse time difference (TD) between the first beam and the second beam, thereby responding to the histogram processing performed by the histogram processing unit to remove interference noise through the interference noise remover.
[0115] The interference noise remover can check whether the first histogram processing unit of the front histogram (PRH) has performed the first histogram processing, and can subsequently change the emission start time (PP) between the first beam and the second beam in different ways, and can control the dual emission unit to keep the light pulse time difference (TD) between the first beam and the second beam constant while the first histogram processing unit performs the first histogram processing.
[0116] Here, the interference noise remover can control the dual light-emitting unit to change the emission start time (PP) between the first beam and the second beam in a different way so that the current light pulse emission delay time is different from the previous light pulse emission delay time, and when the first histogram processing unit performs the Nth histogram processing instead of the first histogram processing, it makes the current light pulse time difference (TD) different from the previous light pulse time difference.
[0117] For example, the interference noise remover can change the current optical pulse emission delay time corresponding to the Nth histogram processing and the previous optical pulse emission delay time corresponding to the (N-1)th histogram processing in different ways, and can also change the current optical pulse time difference (TD) corresponding to the Nth histogram processing and the previous optical pulse time difference corresponding to the (N-1)th histogram processing in different ways.
[0118] In addition, the interference noise remover can control the dual-emission unit to keep the time difference (TD) between the first and second beams constant while performing the Nth histogram processing.
[0119] Therefore, as Figure 3B As shown, in the front histogram (PRH) processing, when the time difference of the light pulse (TD) of the sensed signal is the same as the time difference of the light pulse (IFTD) of the interference noise, the interference noise between the lidar sensors is not removed, and the sensed signal and the interference noise can be transmitted together to the back histogram (POH) processing.
[0120] However, while performing the Nth histogram processing, since the time difference of light pulses (TD) of the induced signal and the time difference of light pulses (IFTD) of the interference noise are different, the interference noise will be removed even when the IF signal, which is the interference noise, is strong, and only the induced signal can be transmitted to the post-histogram (POH) processing.
[0121] Furthermore, in the back histogram (POH), even when the induced signal and interference noise are transmitted together from the front histogram (PRH), the interference noise can be removed because only the induced signal accumulates more than the interference noise, thus revealing the correct TOF value.
[0122] In addition, such as Figure 3C As shown, the lidar sensor of the present invention may further include a random number generation unit for generating two-dimensional random numbers.
[0123] That is, the interference noise remover of the present invention can change the optical pulse emission start time (PP) and optical pulse time difference (TD) between the first beam and the second beam based on two-dimensional random numbers.
[0124] For example, the random number generation section can be configured to pass through the circuitry of a noisy SPAD, which blocks the light beam and intentionally forms defects in a portion of the p-type substrate to generate noise.
[0125] Two noisy SPADs are connected to a phase-locked discriminator (PLA), and the random number code can be configured based on whether the noise is generated in SPAD A or SPAD B first, with a value of 1 appearing.
[0126] For example, a value of 0 is displayed when noise is first generated in SPAD A, and a value of 1 is output when noise is first generated in SPAD B. Figure 3A This process is repeated 5 times to generate a 5-digit two-dimensional random number.
[0127] As described above, the random number generation unit can generate a 5-digit two-dimensional random number, but the present invention is not limited thereto.
[0128] Figure 4 This is a view used to describe the outdoor sensitivity of the lidar sensor according to the present invention, and Figure 5 This is a view used to describe the indoor sensitivity of the lidar sensor according to the present invention.
[0129] like Figure 4 and Figure 5 As shown, the present invention measures the sensing signal for a distance of approximately 4m-32m in an external environment with sunlight of approximately 105klx, and Figure 4 The left-hand view in the image is a graph showing a non-linear relationship based on the measured distance. Figure 4 The right-hand view in the image is a graph showing the accuracy of approximately 1000 depth measurements for each point, and... Figure 5 It is a graph showing the nonlinearity of the measured distance in the internal environment.
[0130] As can be seen from the above, the lidar sensor of the present invention can sense highly sensitive signals and has increased accuracy even in noisy environments with sunlight of about 105 klx.
[0131] As described above, in this invention, noise can be removed for the first time by extracting data values greater than or equal to a first reference value from the histogram bins using a first histogram processing unit, and for the second time by extracting accumulated data values greater than or equal to a second reference value from the histogram bins containing the data values extracted by the second histogram processing unit, thereby simultaneously performing highly sensitive signal sensing and noise removal.
[0132] Furthermore, in this invention, histogram processing performed by the first histogram processing unit removes interference noise between lidar sensors by controlling the light-emitting unit to change the light pulse emission delay time and light pulse time difference (TD) between the first beam and the second beam.
[0133] The method for removing noise from lidar sensors according to the present invention will be described below.
[0134] First, in this invention, reflected light reflected from an object can be sensed.
[0135] Subsequently, in this invention, weights can be generated based on the sensitivity of the reflected light from the light receiver.
[0136] Here, in this invention, the multiple single-photon avalanche diodes (SPADs) contained in each light-receiving pixel of the light-receiving unit can be divided into multiple subgroups. It can be determined whether the number of SPADs in the subgroup that simultaneously senses reflected light is greater than or equal to a reference number, and the weight of the sensing signal can be generated based on the number of subgroups that are greater than or equal to the reference number.
[0137] Next, in this invention, histogram processing of the sensing signal of the optical receiver can be performed based on the generated weights, and data values greater than or equal to the first reference value can be extracted from the histogram to remove noise for the first time.
[0138] Here, in this invention, it is possible to check whether weights have been generated, and when weights have been generated, histogram processing is performed by reflecting the generated weights in the data values of the sensed signal.
[0139] For example, in this invention, it is possible to check whether the data value of each bin is greater than or equal to a first reference value. When the data value of a bin is greater than or equal to the first reference value, the difference between the data value of the bin and the first reference value can be calculated. Furthermore, the data value of the corresponding bin can be treated as noise, and the noise can be removed when the data value of the bin is less than the first reference value.
[0140] Subsequently, in this invention, the extracted data values can be accumulated to perform histogram processing, and the accumulated data values that are greater than or equal to the second reference value can be extracted from the histogram of accumulated data values, thereby removing noise for the second time.
[0141] Here, in this invention, a warehouse corresponding to the first data value can be generated when the first data value is received, a warehouse corresponding to the second data value can be generated when the second data value is received, and a warehouse corresponding to the Nth data value can be generated or accumulated when the Nth data value is received.
[0142] Furthermore, in this invention, when the Nth data value is received, it can be checked whether the Nth data value is the last data value, and when the Nth data value is the last data value according to the check result, a bin corresponding to the last data value can be generated or accumulated, and then the accumulated data value that is greater than or equal to the second reference value can be extracted from the bin of the histogram.
[0143] Here, in this invention, it is possible to check whether the data value of each bin is greater than or equal to the second reference value, calculate the difference between the data value of the bin and the second reference value, and then extract and output the calculated difference value when the data value of the bin is greater than or equal to the second reference value. Furthermore, the data value of the corresponding bin can be treated as noise, and the noise can be removed when the data value of the bin is less than the second reference value.
[0144] Furthermore, in this invention, interference noise between lidar sensors can be removed by changing the optical pulse emission delay time and optical pulse time difference (TD) between the first beam and the second beam in response to histogram processing.
[0145] For example, in this invention, the emission start times of the first beam and the second beam can be changed in different ways during the first histogram processing, and the time difference (TD) between the first beam and the second beam can be kept constant.
[0146] Furthermore, in this invention, the emission start time between the first beam and the second beam can be changed so that the current optical pulse emission delay time is different from the previous optical pulse emission delay time, and the current optical pulse time difference (TD) can be changed to be different from the previous optical pulse time difference when performing the Nth histogram processing instead of the first histogram processing.
[0147] That is, in this invention, the current optical pulse emission delay time corresponding to the Nth histogram processing and the previous optical pulse emission delay time corresponding to the (N-1)th histogram processing can be changed in different ways, and the current optical pulse time difference (TD) corresponding to the Nth histogram processing and the previous optical pulse time difference corresponding to the (N-1)th histogram processing can be changed in different ways.
[0148] Here, in this invention, a two-dimensional random number can be generated, and the optical pulse emission delay time and optical pulse time difference (TD) between the first beam and the second beam can be changed based on the two-dimensional random number.
[0149] The features, structures, effects, etc., described above in this invention are included in at least one embodiment of the invention and should not be limited to only one embodiment. Furthermore, those skilled in the art can combine or modify these features, structures, effects, etc., for other embodiments. It should be understood that anything related to these combinations and modifications is included within the scope of this invention.
[0150] Furthermore, while embodiments have been primarily described above, they are merely examples and do not limit the invention. Each component shown in these embodiments can be implemented through modifications. Moreover, it should be understood that differences associated with these modifications and applications are included within the scope of the invention as defined in the appended claims.
[0151] Explanation of reference numerals in the attached figures
[0152] 100: Optical receiver 200: Weighting unit
[0153] 300: Histogram processing unit; 400: Light emission unit
[0154] 500: Interference and noise removal unit; 600: Random number generation unit
Claims
1. A lidar sensor, comprising: A light receiver is configured to sense reflected light reflected from an object; A weight generation unit is configured to generate weights based on the sensitivity of the reflected light from the light receiving unit; The first histogram processing unit is configured to perform histogram processing on the sensing signal of the light receiver based on the generated weights, and extract data values greater than or equal to the first reference value from the histogram bins, thereby removing noise for the first time. as well as The second histogram processing unit is configured to perform histogram processing by accumulating the data values extracted by the first histogram processing unit, and to extract the accumulated data values that are greater than or equal to the second reference value from the bins of the histogram of accumulated data values, thereby removing noise for the second time. Specifically, during the first noise removal process, the first histogram processing unit checks whether the data value of each bin is greater than or equal to the first reference value. If the data value of a bin is greater than or equal to the first reference value, the unit calculates the difference between the data value of the bin and the first reference value, extracts the calculated difference value, and outputs the difference value to the second histogram processing unit.
2. A lidar sensor, comprising: A light receiver is configured to sense reflected light reflected from an object; A weight generation unit is configured to generate weights based on the sensitivity of the reflected light from the light receiving unit; The first histogram processing unit is configured to perform histogram processing on the sensing signal of the light receiver based on the generated weights, and extract data values greater than or equal to the first reference value from the histogram bins, thereby removing noise for the first time. as well as The second histogram processing unit is configured to perform histogram processing by accumulating the data values extracted by the first histogram processing unit, and to extract the accumulated data values that are greater than or equal to the second reference value from the bins of the histogram of accumulated data values, thereby removing noise for the second time. During the second noise removal process, the second histogram processing unit checks whether the data value of each bin is greater than or equal to the second reference value, calculates the difference between the data value of the bin and the second reference value, and then extracts and outputs the calculated difference value when the data value of the bin is greater than or equal to the second reference value.
3. The lidar sensor according to claim 2, wherein, When checking whether the data value of each bin is greater than or equal to the second reference value, the second histogram processing unit treats the corresponding bin's data value as noise, and removes the noise when the bin's data value is less than the second reference value.
4. The lidar sensor according to claim 2, further comprising: A light-emitting unit configured to emit a first beam and a second beam toward an object; as well as An interference noise remover is configured to control the light-emitting unit to change the light pulse emission delay time and light pulse time difference between the first beam and the second beam in response to histogram processing performed by the first histogram processing unit.
5. The lidar sensor according to claim 4, wherein, The interference noise remover checks whether the first histogram processing unit performs the first histogram processing, and when the first histogram processing unit performs the first histogram processing, it controls the light-emitting unit to change the emission start time of the first beam and the second beam in different ways, and keeps the light pulse time difference between the first beam and the second beam constant.
6. The lidar sensor according to claim 4, wherein, When checking whether the first histogram processing unit performs the first histogram processing, the interference noise remover controls the light-emitting unit to change the emission start time of the first beam and the second beam in a different way so that the current light pulse emission delay time is different from the previous light pulse emission delay time, and when the first histogram processing unit performs the Nth histogram processing instead of the first histogram processing, the current light pulse time difference is different from the previous light pulse time difference.
7. The lidar sensor according to claim 4, further comprising a random number generation unit configured to generate two-dimensional random numbers. The interference noise remover uses the two-dimensional random number to change the optical pulse emission delay time and optical pulse time difference between the first beam and the second beam.