Laser radar and detection method thereof

Abstract

The invention provides a laser radar which comprises the components of an emitting device which comprises at least one light-emitting unit which is configured to emit a plurality of detection pulses for detecting an obstacle in one detection process; the detection device comprises at least one detection unit, each detection unit comprises a pixel array, and the pixel arrays can respond to echoes reflected by the detection pulses on the obstacles and generate detection data; the data processing device is coupled with the detection device and is configured to obtain detection data; the data processing device comprises a feedback processing module, and the feedback processing module is coupled with the transmitting device; in the primary detection process, the transmitting device firstly transmits a first group of detection pulses, and the first group of detection pulses comprise one or more detection pulses; the data processing device obtains a first group of detection data, wherein the first group of detection data is the detection data generated after the pixel array responds to echoes reflected by the first group of detection pulses on the obstacle; the feedback processing module is configured to determine a subsequent transmitting strategy of the transmitting device according to the first group of probe data. According to the laser radar, the subsequent transmitting strategy of the transmitting device is determined through a feedback mechanism, and the system power consumption of the laser radar can be effectively reduced.

Description

Technical Field

[0001] The present invention generally relates to the field of laser radar, and in particular to a laser radar and a detection method of a laser radar. Background Art

[0002] LiDAR is a radar system that emits laser beams to detect the position, speed and other characteristic quantities of a target. It is an advanced detection method that combines laser technology with photoelectric detection technology. LiDAR is widely used in autonomous driving, transportation communications, drones, intelligent robots, resource exploration and other fields due to its advantages such as high resolution, good concealment, strong anti-active interference ability, good low-altitude detection performance, small size and light weight.

[0003] Some LiDARs use detectors similar to Single Photon Avalanche Diodes (SPADs) for detection. SPADs are highly sensitive and can be triggered to respond by a single photon, but are easily affected by ambient light noise. On the other hand, SPADs have low photon detection efficiency (PDE) for the light bands commonly used for LiDAR detection, so the signal strength obtained by a single SPAD detection is weak, limiting the distance measurement capability of LiDARs. In a detection scan, SPADs may be triggered only a few times within the detection time window, so it is difficult to distinguish whether the detected light signal is an echo signal or ambient light noise, resulting in a low signal-to-noise ratio, which affects the accuracy of the detection results.

[0004] In order to improve the distance measurement performance of LiDAR and reduce the impact of noise, LiDAR can perform multiple repeated measurements during a detection process when detecting obstacles in the same field of view. One measurement is also called a scan, that is, a laser can emit light multiple times during its light-emitting time (i.e., emit multiple detection pulses), each light emission is called a scan, and each scan emits one detection pulse (the number of repeated measurements can reach 400-500 times, or even thousands of times, or more or less times). Figure 1a As shown in FIG. 1 , the laser radar performs 400 scans in one detection process. Figure 1b As shown in the figure, during each scan, the detection pulse emitted by the LiDAR is reflected by the obstacle and the echo generated by the ambient light noise, etc., which will trigger different numbers of SPADs at different times. Figure 1b The horizontal axis is time, and the vertical axis is the number of triggered SPADs. Figure 1bThe histogram in the upper right corner of the figure cannot effectively distinguish the echo signal from the noise signal. Therefore, the results of the second scan can be accumulated based on the results of the first scan, so that a higher number of echo signals can be accumulated at the time corresponding to the echo signal. This process is repeated until 400 scan results are accumulated (assuming that the total number of scans is 400) to obtain a histogram (refer to Figure 1b The histogram in the lower right corner of the figure) accumulates a strong peak at the time corresponding to the obstacle, thereby effectively distinguishing the echo signal from the noise signal, and measuring the distance of the obstacle at this time, and then obtaining a point in the lidar point cloud. This method of obtaining obstacle information based on the results of multiple scan accumulations can effectively improve the detection performance of lidars that use SPADs as detectors in some cases. However, due to the number of light emissions, the signal-to-noise ratio of weak echo signals is relatively low, especially for echo signals of long-distance, low-reflectivity targets, which have poor ability to trigger the detector. Therefore, the pulse peak obtained after a limited number of accumulations may still be low, and the signal-to-noise ratio is low.

[0005] In order to improve the signal-to-noise ratio, some laser radars use a convolution processing solution, that is, convolution processing is performed on the electrical signal output by each detection unit. This method can effectively improve the signal-to-noise ratio and is very beneficial for the actual distance measurement of the laser radar. However, in the existing solutions, most laser radars use the same size of convolution kernel and the same convolution step size during the convolution process, which leads to high computational consumption and weak adjustment ability. In addition, during a detection process, each laser of the laser radar emits the same number of times, and each corresponding detection unit also responds a corresponding number of times, which leads to high power consumption of the laser radar system.

[0006] Therefore, how to reduce computing consumption and power consumption of the LiDAR system while maintaining better detection performance (such as higher signal-to-noise ratio, ranging performance, accuracy of detection results, dynamic adjustment capabilities, etc.) is a need for continuous improvement in the LiDAR field.

[0007] The contents of the background technology section are merely the technologies known to the inventors and do not necessarily represent the prior art in the field. Summary of the invention

[0008] In response to one or more problems existing in the prior art, the present invention provides a laser radar, which, during a detection process, first emits a first group of detection pulses, performs feedback processing based on the detection data of the first group of detection pulses, and determines the subsequent emission strategy of the transmitting device, thereby achieving at least one of the detection performances including reducing computing consumption, reducing system power consumption, maintaining a high signal-to-noise ratio, ranging performance, accuracy of detection results, and dynamic adjustment capability.

[0009] The laser radar comprises:

[0010] A transmitting device, comprising at least one light emitting unit, wherein the light emitting unit is configured to emit a plurality of detection pulses in one detection process for detecting obstacles;

[0011] A detection device, comprising at least one detection unit, wherein each detection unit comprises a pixel array, wherein the pixel array can respond to an echo reflected by the detection pulse on an obstacle and generate detection data; and

[0012] a data processing device, coupled to the detection device and configured to acquire the detection data;

[0013] The data processing device comprises a feedback processing module, and the feedback processing module is coupled to the transmitting device;

[0014] Wherein, in one detection process, the transmitting device first transmits a first group of detection pulses, wherein the first group of detection pulses includes one or more detection pulses;

[0015] The data processing device acquires a first set of detection data, where the first set of detection data is detection data generated after the pixel array responds to the echo reflected by the first set of detection pulses on the obstacle;

[0016] The feedback processing module is configured to determine a subsequent transmission strategy of the transmitting device according to the first set of detection data.

[0017] According to one aspect of the present invention, the feedback processing module is configured to:

[0018] performing a first convolution process on the first set of detection data to obtain a first reconstructed pixel array; and

[0019] A subsequent emission strategy of the emission device is determined according to the first recombined pixel array.

[0020] According to one aspect of the present invention, the feedback processing module is configured to:

[0021] Taking statistics on the data of the reconstructed pixels in the first reconstructed pixel array; and

[0022] According to the statistical results, the subsequent launch strategy of the launch device is determined.

[0023] According to one aspect of the present invention, the feedback processing module is configured to:

[0024] Counting whether there is an over-threshold pulse in the data of the recombined pixels in the first recombined pixel array and obtaining a pulse statistical result, wherein the over-threshold pulse is a pulse exceeding a first intensity threshold;

[0025] The subsequent transmission strategy of the transmitting device is determined according to the pulse statistical result.

[0026] According to one aspect of the present invention, the feedback processing module is configured to:

[0027] When the cross-threshold pulse exists in the data of the recombined pixels in the first recombined pixel array, further counting one or more of the following parameters: the intensity of each of the cross-threshold pulses, the obstacle distance corresponding to each of the cross-threshold pulses, the number of recombined pixels having the cross-threshold pulse, and the ratio of the recombined pixels having the cross-threshold pulse to all the recombined pixels in the first recombined pixel array;

[0028] According to the further statistical results, the subsequent launch strategy of the launch device is determined.

[0029] According to one aspect of the present invention, the feedback processing module is configured to perform ambient light compensation on data of at least part of the reconstructed pixels in the first reconstructed pixel array.

[0030] According to one aspect of the present invention, the transmission strategy includes one of the following:

[0031] continuing to transmit a second set of detection pulses, wherein the second set of detection pulses includes one or more detection pulses; and

[0032] Stop sending detection pulses and end this detection.

[0033] According to one aspect of the present invention, the continuing to transmit the second group of detection pulses comprises:

[0034] continuing to transmit the second set of detection pulses at a reduced transmission intensity; or

[0035] The second set of detection pulses continues to be emitted at a constant emission intensity.

[0036] According to one aspect of the present invention, the feedback processing module is configured to:

[0037] When an obstacle is detected based on the preset standard through the first group of detection data, the emission of detection pulses is stopped; otherwise, the emission of the second group of detection pulses continues.

[0038] According to one aspect of the present invention, the feedback processing module is configured to: feed back the transmission strategy to the transmission device, and the transmission device is configured to execute the transmission strategy.

[0039] According to one aspect of the present invention, during a detection process, when the transmitting device continues to transmit a second group of detection pulses, the data processing device is configured to: acquire a second group of detection data, wherein the second group of detection data is detection data generated after the pixel array responds to the echo reflected by the second group of detection pulses on the obstacle; the feedback processing module is configured to: determine the subsequent emission strategy of the transmitting device based on the second group of detection data.

[0040] According to one aspect of the present invention, the data processing device further includes: an output processing module, which is coupled to the feedback processing module. When the feedback processing module determines that the emission strategy is to stop emitting detection pulses, the output processing module is configured to: perform a second convolution processing on the detection data generated according to at least one group of detection pulses emitted by the emission device to obtain a second recombined pixel array, and generate a lidar point cloud based on the second recombined pixel array, wherein the detection data generated according to the at least one group of detection pulses includes the first group of detection data.

[0041] According to one aspect of the present invention, the second convolution processing includes: performing convolution processing on the detection data using multiple convolution kernels of different sizes to obtain multiple second recombined pixel arrays, and generating a lidar point cloud based on the multiple second recombined pixel arrays.

[0042] According to one aspect of the present invention, the output processing module is configured to perform ambient light compensation on at least a portion of the reorganized pixel data in the second reorganized pixel array.

[0043] According to one aspect of the present invention, the output processing module is configured to: determine a final recombined pixel array according to the status of the recombined pixel data in the plurality of second recombined pixel arrays, and generate a lidar point cloud according to the final recombined pixel array.

[0044] According to one aspect of the present invention, the output processing module is configured to determine the final recombined pixel array according to pulse threshold crossing states of at least two adjacent recombined pixels in the second recombined pixel array.

[0045] According to one aspect of the present invention, the output processing device is configured to determine the calculation method of the intermediate recombined pixel data according to the pulse threshold crossing state of at least two adjacent recombined pixel data in the second recombined pixel array.

[0046] According to one aspect of the present invention, the output processing device is configured to determine the calculation method of the intermediate recombined pixel data according to the pulse threshold crossing status of the at least two adjacent recombined pixel data and the distance and / or reflectivity of the corresponding obstacle.

[0047] According to one aspect of the present invention, the calculation method of the intermediate recombined pixel data includes one of the following: performing interpolation calculation on the intermediate recombined pixels, performing single-point calculation on the intermediate recombined pixels, and performing convolution calculation on the intermediate pixels.

[0048] The present invention also provides a detection method of a laser radar, wherein the laser radar comprises a transmitting device and a detecting device, wherein the transmitting device comprises at least one light-emitting unit, wherein the light-emitting unit transmits a plurality of detection pulses in one detection process, and the detecting device comprises at least one detection unit, wherein each detection unit comprises a pixel array; the detection method comprises: in one detection process,

[0049] S11: firstly transmitting a first group of detection pulses, wherein the first group of detection pulses includes one or more detection pulses;

[0050] S12: Responding to the echoes of the first group of detection pulses reflected on the obstacles through the pixel array of the detection unit and generating a first group of detection data; and

[0051] S13: Determine a subsequent transmission strategy according to the first set of detection data.

[0052] According to one aspect of the present invention, step S13 comprises:

[0053] performing a first convolution process on the first set of detection data to obtain a first reconstructed pixel array; and

[0054] A subsequent emission strategy of the emission device is determined according to the first recombined pixel array.

[0055] According to one aspect of the present invention, step S13 comprises:

[0056] Taking statistics on the data of the reconstructed pixels in the first reconstructed pixel array; and

[0057] According to the statistical results, the subsequent launch strategy of the launch device is determined.

[0058] According to one aspect of the present invention, the step of performing statistics on the data of the reconstructed pixels in the first reconstructed pixel array comprises:

[0059] Counting whether there is an over-threshold pulse in the data of the recombined pixels in the first recombined pixel array and obtaining a pulse statistical result, wherein the over-threshold pulse is a pulse exceeding a first intensity threshold;

[0060] The subsequent transmission strategy of the transmitting device is determined according to the pulse statistical result.

[0061] According to one aspect of the present invention, when the over-threshold pulse exists in the data of the reconstructed pixels in the first reconstructed pixel array, the step of performing statistics on the data of the reconstructed pixels in the first reconstructed pixel array further comprises:

[0062] Further counting one or more of the following parameters: the intensity of each of the over-threshold pulses, the obstacle distance corresponding to each of the over-threshold pulses, the number of reconstructed pixels having the over-threshold pulses, and the ratio of the reconstructed pixels having the over-threshold pulses to all the reconstructed pixels in the first reconstructed pixel array;

[0063] According to the further statistical results, the subsequent launch strategy of the launch device is determined.

[0064] According to one aspect of the present invention, the step S13 comprises: performing ambient light compensation on at least part of the reorganized pixel data in the first reorganized pixel array.

[0065] According to one aspect of the present invention, the transmission strategy includes one of the following:

[0066] continuing to transmit a second set of detection pulses, wherein the second set of detection pulses includes one or more detection pulses; and

[0067] Stop sending detection pulses and end this detection.

[0068] According to one aspect of the present invention, the step of continuing to transmit the second group of detection pulses comprises:

[0069] continuing to transmit the second set of detection pulses at a reduced transmission intensity; or

[0070] The second set of detection pulses continues to be emitted at a constant emission intensity.

[0071] According to one aspect of the present invention, step S13 comprises:

[0072] When an obstacle is detected based on the preset standard through the first group of detection data, the emission of detection pulses is stopped; otherwise, the emission of the second group of detection pulses continues.

[0073] According to one aspect of the present invention, it also includes:

[0074] Feeding back the launch strategy to the launch device; and

[0075] The transmitting device is controlled to execute the fed-back transmitting strategy.

[0076] According to one aspect of the present invention, during a detection process, when the second group of detection pulses continues to be emitted, the detection method further includes:

[0077] Acquire a second set of detection data, where the second set of detection data is detection data generated after the pixel array responds to echoes of the second set of detection pulses reflected on obstacles; and

[0078] The subsequent transmission strategy is determined based on the second set of detection data.

[0079] According to one aspect of the present invention, when it is determined that the emission strategy is to stop emitting detection pulses, the detection method further includes: performing a second convolution processing on the detection data generated according to at least one group of detection pulses emitted by the transmitting device; obtaining a second recombinant pixel array, and generating a lidar point cloud based on the second recombinant pixel array, wherein the detection data generated according to the at least one group of detection pulses includes the first group of detection data.

[0080] According to one aspect of the present invention, the step of performing a second convolution processing on the detection data generated according to at least one group of detection pulses emitted by the transmitting device includes: performing convolution processing on the detection data using multiple convolution kernels of different sizes to obtain multiple second recombined pixel arrays, and generating a lidar point cloud based on the multiple second recombined pixel arrays.

[0081] According to one aspect of the present invention, the method further includes: performing ambient light compensation on at least a portion of the reorganized pixel data in the second reorganized pixel array.

[0082] According to one aspect of the present invention, it also includes: determining a final recombined pixel array according to the status of the recombined pixel data in the plurality of second recombined pixel arrays, and generating a lidar point cloud according to the final recombined pixel array.

[0083] According to one aspect of the present invention, the step of determining a final recombined pixel array based on the states of recombined pixel data in a plurality of second recombined pixel arrays comprises: determining the final recombined pixel array based on pulse threshold crossing states of at least two adjacent recombined pixels in the second recombined pixel arrays.

[0084] According to one aspect of the present invention, the step of determining the final recombined pixel array according to the states of the recombined pixel data in the plurality of second recombined pixel arrays further comprises: determining a calculation method for the intermediate recombined pixel data according to the pulse over-threshold states of at least two adjacent recombined pixel data in the second recombined pixel arrays.

[0085] According to one aspect of the present invention, the step of determining the final recombined pixel array based on the states of the recombined pixel data in the plurality of second recombined pixel arrays further comprises: determining a calculation method for the intermediate recombined pixel data based on the pulse over-threshold states of the at least two recombined pixel data and the distance and / or reflectivity of the obstacles to which they correspond.

[0086] According to one aspect of the present invention, the calculation method of the intermediate recombined pixel data includes one of the following: interpolation of the intermediate pixels, single-point calculation of the intermediate pixels, and convolution calculation of the intermediate pixels.

[0087] In summary, the scheme of the present invention has been introduced in detail. The laser radar of the present invention works based on a feedback mechanism. In a detection process, the transmitting device first transmits a first group of detection pulses, and the feedback processing module determines the subsequent emission strategy of the transmitting device based on the first group of detection data. The subsequent emission strategy may be to stop emitting light, or to continue to emit detection pulses with a reduced emission intensity, or to continue to emit detection pulses with an unchanged emission intensity. Thereby, the overall number of scans can be reduced, greatly saving the system power consumption of the laser radar. At the same time, each of the multiple light-emitting units can decide the total number of scans corresponding to the light-emitting unit according to its real-time detection results, thereby realizing adaptive adjustment of the number of emitted detection pulses, which can reduce the system power consumption while ensuring high detection accuracy within the entire field of view.

[0088] Furthermore, the feedback processing module performs the first convolution processing on the first group of detection data by adopting a larger step size and a smaller convolution kernel, so that the calculation amount of the first convolution processing is lower and the calculation speed is faster, which can reduce the system power consumption of the laser radar.

[0089] Furthermore, during the second convolution processing, the output processing module preferentially adopts a smaller convolution step size for the second convolution processing than the convolution step size in the first convolution processing. By adopting a coarse-first-then-fine approach, the amount of calculation can be reduced, the system power consumption of the lidar can be reduced, and ultimately a precise calculation result can be obtained.

[0090] Furthermore, the feedback processing module / output processing module can reduce the interference of ambient light by performing ambient light compensation on the data of at least part of the recombined pixels in the first / second recombined pixel array, thereby making the result more accurate.

[0091] In addition, for different detection targets, this solution can also adjust the total number of light emissions through a feedback mechanism. For close-range, high-reflectivity (strong echo signal) areas, fewer light emissions are used, while for long-range, low-reflectivity (weak echo signal) areas, more light emissions are used, thereby achieving adaptive adjustment of the number of light emissions based on real-time echo signals and reducing the system power consumption of the entire lidar.

[0092] In addition, the laser radar of the present invention can adjust the spatial resolution and signal-to-noise ratio. For close-range, high-reflectivity targets, it can achieve high spatial resolution; for long-range, weak-reflectivity targets, it can maintain a high signal-to-noise ratio and achieve longer detection capabilities. It can also alleviate the expansion problem of high-reflectivity objects and the problem of false targets, which is beneficial to improving the accuracy of the detection results and enabling the laser radar to dynamically maintain better detection performance throughout the entire field of view. BRIEF DESCRIPTION OF THE DRAWINGS

[0093] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0094] Figure 1a A schematic diagram showing a laser radar performing multiple measurements in one detection;

[0095] Figure 1b A schematic diagram showing a histogram obtained by accumulating multiple measurement results;

[0096] Figure 2 A schematic diagram of a laser radar according to an embodiment of the present invention is shown;

[0097] Figure 3 A schematic diagram of a detection device according to an embodiment of the present invention is shown;

[0098] Figure 4 A schematic diagram showing a detection device according to another preferred embodiment of the present invention is shown;

[0099] Figure 5 shows an enlarged view of a detection unit according to a preferred embodiment of the present invention;

[0100] Figure 6 A schematic diagram showing a feedback processing module performing a first convolution process on a first group of detection data according to a preferred embodiment of the present invention is shown;

[0101] Figure 7 A schematic diagram of a laser radar according to a preferred embodiment of the present invention is shown;

[0102] Figure 8 to Figure 10 A schematic diagram showing a second convolution process performed on a first group of detection data by an output processing module according to some preferred embodiments of the present invention;

[0103] Fig.11 A schematic diagram showing the positional relationship between adjacent recombined pixels and an intermediate recombined pixel; and

[0104] Fig.12 A flow chart of a laser radar detection method according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0105] In the following, only some exemplary embodiments are briefly described. As those skilled in the art will appreciate, the described embodiments may be modified in various ways without departing from the spirit or scope of the present invention. Therefore, the drawings and descriptions are considered to be exemplary and non-restrictive in nature.

[0106] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the accompanying drawings, which are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first" and "second" are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, "multiple" means two or more, unless otherwise clearly and specifically defined.

[0107] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection, an electrical connection, or can communicate with each other; it can be directly connected, or indirectly connected through an intermediate medium, it can be the internal connection of two elements or the interaction relationship between two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0108] In the present invention, unless otherwise clearly specified and limited, a first feature being "above" or "below" a second feature may include that the first and second features are in direct contact, or may include that the first and second features are not in direct contact but are in contact through another feature between them. Moreover, a first feature being "above", "above" and "above" a second feature includes that the first feature is directly above and obliquely above the second feature, or simply means that the first feature is higher in level than the second feature. A first feature being "below", "below" and "below" a second feature includes that the first feature is directly below and obliquely below the second feature, or simply means that the first feature is lower in level than the second feature.

[0109] The disclosure below provides many different embodiments or examples to realize different structures of the present invention. In order to simplify the disclosure of the present invention, the parts and settings of specific examples are described below. Of course, they are only examples, and the purpose is not to limit the present invention. In addition, the present invention can repeat reference numbers and / or reference letters in different examples, and this repetition is for the purpose of simplicity and clarity, which itself does not indicate the relationship between the various embodiments and / or settings discussed. In addition, the present invention provides various specific examples of processes and materials, but those of ordinary skill in the art can be aware of the application of other processes and / or the use of other materials.

[0110] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0111] The present invention provides a laser radar. Figure 2 FIG. 1 shows a schematic diagram of a laser radar 100 according to an embodiment of the present invention. Figure 2 As shown, the laser radar 100 includes a transmitting device 110, a detecting device 120 and a data processing device 130. The transmitting device 110 includes at least one light emitting unit 111 ( Figure 2 The exemplary embodiment shows a light emitting unit 111), the light emitting unit 111 is configured to emit multiple detection pulses L in one detection process for detecting obstacles OB; the detection device 120 includes at least one detection unit 121 ( Figure 2 One of the detection units 121 is shown as an example, wherein each detection unit 121 includes a pixel array ( Figure 2An 8×16 pixel array is shown as an example, the pixel array can respond to the echo L' reflected by the detection pulse L on the obstacle OB and generate detection data; the data processing device 130 is coupled to the detection device 120 and configured to obtain detection data; the data processing device 130 includes a feedback processing module 131, and the feedback processing module 131 is coupled to the transmitting device 110; wherein, in a detection process, the transmitting device 110 first transmits a first group of detection pulses, wherein the first group of detection pulses includes one or more detection pulses L; the data processing device 130 obtains a first group of detection data, wherein the first group of detection data is the detection data generated after the pixel array responds to the echo reflected by the first group of detection pulses on the obstacle OB; the feedback processing module 131 is configured to determine the subsequent emission strategy of the transmitting device 110 according to the first group of detection data. The laser radar of the present invention determines the subsequent emission strategy of the transmitting device through the feedback mechanism, thereby determining the number of detection pulses to be emitted according to the real-time detection results, which can effectively reduce the total number of detection pulses and reduce the system power consumption of the laser radar. For example, when the detection data generated by the echoes generated by the first group of detection pulses is sufficient to accurately calculate the obstacle information, you can choose to stop transmitting the remaining detection pulses in the detection process; and when the detection data generated by the echoes generated by the first group of detection pulses is insufficient to accurately calculate the obstacle information, you can choose to continue to transmit more detection pulses and generate detection data based on further detection results. The following is a detailed description.

[0112] Figure 3 A schematic diagram of a detection device 120 according to a preferred embodiment of the present invention is shown. Figure 3 As shown, the detection device 120 includes a plurality of detection units, for example, N detection units R1, R2, R3, ..., R N , where N is an integer greater than 1, and multiple detection units are arranged in a one-dimensional linear array along the vertical direction to cover the vertical field of view of the laser radar, and further realize scanning detection of the three-dimensional scene by mechanically rotating the laser radar's transceiver or by rotating the scanning mirror.

[0113] Figure 4 FIG. 1 shows a schematic diagram of a detection device 120 according to another preferred embodiment of the present invention. Figure 4 As shown, the detection device 120 includes a plurality of detection units R 11 , R 21 , …, R M1 , …, R 1N , R 2N , …, R MN , multiple detection units can be arranged in a two-dimensional array to form an N×M detection unit array, that is, an N-row and M-column detection unit array, where N and M are both integers greater than 1, and the present invention does not limit the size relationship between N and M.

[0114] Figure 5 FIG. 1 shows an enlarged view of a detection unit 121 according to a preferred embodiment of the present invention, as shown in FIG. Figure 5 As shown, the detection unit 121 includes a pixel array, each pixel array includes a plurality of pixels arranged in an array, such as 8×16=128 pixels exemplarily shown in the figure, each pixel may include one or more single photon avalanche diodes (SPAD), each single photon avalanche diode can be independently gated and addressed, and the data in one or more SPADs in each pixel can be accumulated to form the data of the pixel. It is understandable that the data of each SPAD can also be output and read independently.

[0115] The above embodiments have provided a detailed introduction to the detection device of the present invention. It should be noted that the present invention does not impose any restrictions on the number and arrangement of detection units in the detection device, the number and arrangement of pixels in each detection unit, and the number and arrangement of SPADs in each pixel. In practical applications, they can be flexibly arranged according to requirements.

[0116] According to a preferred embodiment of the present invention, the transmitting device 110 includes at least one light-emitting unit 111, and each light-emitting unit 111 corresponds to at least one detection unit 121. The echo generated on the obstacle by the detection beam emitted by each light-emitting unit 111 will be received and converted by the detection unit 121 corresponding thereto. On the one hand, the light-emitting units 111 can be paired with the detection units 121 one by one, that is, the number of light-emitting units 111 is equal to the number of detection units 121, and the echo generated on the obstacle by the detection beam emitted by each light-emitting unit 111 will be received and converted by the detection unit 121 paired therewith; on the other hand, the echo generated on the obstacle by the detection beams emitted by multiple light-emitting units 111 can also be received and converted by one detection unit 121; in addition, the detection beam emitted by one light-emitting unit 111 can also be subdivided into multiple beams, or gathered to multiple detection units 121 by means of light homogenization or the like.

[0117] Preferably, the light emitting unit 111 may be a vertical cavity surface emitting laser (VCSEL) or an edge emitting laser (EEL). When the emitting device 110 includes a plurality of light emitting units 111, the plurality of light emitting units 111 may be arranged as a one-dimensional linear array or a two-dimensional surface array according to the arrangement of the detection unit 121, which are all within the protection scope of the present invention.

[0118] According to one embodiment of the present invention, during the detection process of the laser radar, the detection beam L emitted by the light-emitting unit 111 can be deflected by a scanning device (such as a rotating mirror), or the transmitting device 110 can be driven by the rotor to poll and emit the detection beam L at a certain horizontal angle (such as 0.05°, etc.), and the corresponding pixel array of the detection unit 121 can respond to the echo L' reflected by the detection beam L on the obstacle OB and convert it into an electrical signal. By moving the field of view, the laser radar can complete the scanning of the entire scene of the surrounding environment, thereby realizing detection within a certain horizontal field of view. The data processing device 130 can obtain a three-dimensional point cloud of the surrounding environment by splicing the three-dimensional point cloud of each field of view area. However, the laser radar of the present invention is not limited to the mechanical rotating laser radar or scanning laser radar introduced above, and can also be a solid-state laser radar, that is, it does not require any mechanical rotating structure, and can also cover a certain spatial field of view through the surface array setting of the transmitting device and the detection device.

[0119] In some laser radars, each light-emitting unit executes the same emission strategy during multiple detection processes, that is, emits the same number of detection pulses for detecting obstacles, and the pixel array of the corresponding detection unit also responds a corresponding number of times, which results in high system power consumption and poor adaptive adjustment capability of the laser radar. The present invention makes improvements on this. During a detection process, the light-emitting unit 111 first emits a first group of detection pulses, and the pixel array of the corresponding detection unit 121 responds to the echo of the first group of detection pulses reflected on the obstacle and generates a first group of detection data. The data processing device 130 obtains the first group of detection data, and the feedback processing module 131 determines the subsequent emission strategy of the emission device 110 based on the first group of detection data. Specifically, in a detection process, assuming that the total number of times each light-emitting unit 111 emits light is Nsweep times (for example, 400 times), each light-emitting unit 111 first emits a first group of detection pulses, such as 1, 50, 100, etc., recorded as N1, where N1 < Nsweep, and the pixel array of the corresponding detection unit 121 responds to the echo reflected by the first group of detection pulses on the obstacle and generates a first group of detection data. The first group of detection data can be the histogram data obtained by accumulating the echo response signals of N1 scans. The data processing device 130 can collect the electrical signals output by each pixel in the pixel array in parallel (such as echo waveforms, or the triggering number and triggering time of SPAD), and obtain the first group of detection data after N1 accumulations; the feedback processing module 131 determines the subsequent emission strategy of the transmitting device 110 based on the first group of detection data.

[0120] According to a preferred embodiment of the present invention, the subsequent transmission strategy of the transmitting device 110 includes one of the following: continuing to transmit the second group of detection pulses, wherein the second group of detection pulses includes one or more detection pulses, which are recorded as N2, wherein N1+N2≤Nsweep; and stopping transmitting the detection pulses to end the current detection. The following describes how the feedback processing module 131 determines the subsequent transmission strategy of the transmitting device 110 based on the first group of detection data.

[0121] According to an embodiment of the present invention, the feedback processing module 131 is configured to determine the subsequent emission strategy of the transmitting device 110 based on the first set of detection data. For example, it can directly determine whether there is an over-threshold pulse based on the first set of detection data, or the signal-to-noise ratio is high enough to decide whether to continue to emit the second set of detection pulses. Specifically, the feedback processing module 131 can be configured to count whether there is an over-threshold pulse in the first set of detection data and obtain pulse statistics results, and determine the subsequent emission strategy of the transmitting device 110 based on the pulse statistics results, wherein the over-threshold pulse is a pulse that exceeds the first intensity threshold. According to a preferred embodiment of the present invention, when there is an over-threshold pulse in the first set of detection data, the feedback processing module 131 is further configured to further count one or more of the following parameters: the intensity of each over-threshold pulse; the obstacle distance corresponding to each over-threshold pulse (the obstacle distance calculated based on the over-threshold pulse); the number of over-threshold pulses; the proportion of over-threshold pulses in the first set of detection data; and is configured to determine the subsequent emission strategy of the transmitting device 110 based on the further statistical results. Additionally or alternatively, the feedback processing module 131 may determine the subsequent emission strategy of the transmitting device 110 based on the signal-to-noise ratio of the first group of detection data. According to a preferred embodiment of the present invention, the feedback processing module 131 is configured to perform a first convolution processing on the first group of detection data to obtain a first reorganized pixel array, and determine the subsequent emission strategy of the transmitting device 110 based on the first reorganized pixel array. When performing the first convolution processing on the first group of detection data, preferably, a smaller convolution kernel conv and a larger convolution step stride may be used. The convolution kernel may also be called a convolution window or an accumulation window, and the size of the convolution kernel is P×Q (in the form of a matrix of P rows and Q columns), which may cover P×Q pixels, wherein P and Q are integers greater than or equal to 1, and the two may be equal or unequal, and the minimum size of the convolution kernel is the size of a pixel, that is, conv=1×1.

[0122] Figure 6 FIG. 4 is a schematic diagram showing a first convolution process performed on a first set of detection data according to a preferred embodiment of the present invention. Figure 6As shown, in this embodiment, the size of the convolution kernel conv=1×1, the convolution step length stride=3, in the first convolution processing process, the convolution kernel conv=1×1 is gradually moved in a preset direction (for example, from left to right, from top to bottom) with a convolution step length stride=3 as a unit, and each time the convolution kernel covers all pixels (1 pixel) The output of the electrical signal (for example, an echo waveform) is accumulated once to obtain a reconstructed pixel, and the 8×16 pixel array of the entire detection unit is traversed to obtain a 3×6 reconstructed pixel array (the first reconstructed pixel array), where a dark grid represents a reconstructed pixel, refer to Figure 6 , the dotted box frames one of the recombined pixels and its corresponding single pixel (original pixel), Figure 6 The lower right part shows a signal waveform of a reconstructed pixel (original pixel). The feedback processing module 131 performs the first convolution processing on the first group of detection data by using a smaller convolution kernel conv and a larger convolution step length stride, and can obtain a first reconstructed pixel array based on part of the first group of detection data, thereby reducing the amount of calculation, improving the calculation speed, and saving system power consumption. It should be noted that although the present embodiment is introduced by taking the size of the convolution kernel conv=1×1 and the convolution step length stride=3 as an example, it does not constitute a limitation of the present invention. In the first convolution processing process, the size of the convolution kernel can be other sizes, and the convolution step length can also be other step lengths. Preferably, the convolution step length is greater than the side length of the convolution kernel. In the first convolution processing process, the feedback processing module 131 can obtain the first reconstructed pixel array based on part or all of the first group of detection data, and can quickly determine whether the current detection result can detect obstacles accurately enough, and the feedback processing module 131 adopts the convolution processing method to reduce the calculation consumption, save system power consumption, and improve the signal-to-noise ratio of the detection data.

[0123] The following describes in detail a preferred embodiment of how the feedback processing module 131 determines the subsequent emission strategy of the emission device 110 according to the first reorganized pixel array.

[0124] According to a preferred embodiment of the present invention, the feedback processing module 131 is configured to perform statistics on the data of the recombined pixels in the first recombined pixel array, and determine the subsequent emission strategy of the emission device 110 according to the statistical results. Specifically, the feedback processing module 131 is configured to perform statistics on whether there are over-threshold pulses in the data of the recombined pixels in the first recombined pixel array and obtain pulse statistical results, and determine the subsequent emission strategy of the emission device 110 according to the pulse statistical results, wherein the over-threshold pulse is a pulse exceeding the first intensity threshold.

[0125] According to a preferred embodiment of the present invention, when the over-threshold pulse exists in the data of the recombined pixels in the first recombined pixel array, the feedback processing module 131 is further configured to further count one or more of the following parameters: the intensity of each over-threshold pulse; the obstacle distance corresponding to each over-threshold pulse (the obstacle distance calculated based on the over-threshold pulse); the number of recombined pixels with over-threshold pulses; the ratio of the recombined pixels with over-threshold pulses to all the recombined pixels in the first recombined pixel array; and is configured to determine the subsequent emission strategy of the transmitting device 110 based on the results of the further statistics.

[0126] Specifically, the feedback processing module 131 is configured to determine whether the further statistical result meets the preset standard, wherein the preset standard can be one or more of the following: the intensity of the over-threshold pulse is greater than or equal to the second intensity threshold, wherein the second intensity threshold is greater than or equal to the aforementioned first intensity threshold; the obstacle distance corresponding to the over-threshold pulse is less than the distance threshold; the number of recombined pixels with over-threshold pulses reaches the number threshold; the ratio of the recombined pixels with over-threshold pulses to all the recombined pixels in the first recombined pixel array reaches the ratio threshold. For example, when the intensity of the over-threshold pulse reaches the second intensity threshold, and the ratio of the recombined pixels with over-threshold pulses to all the recombined pixels in the first recombined pixel array reaches the ratio threshold (for example, 90%), or when the number of recombined pixels with over-threshold pulses reaches the number threshold, and the obstacle distance corresponding to the over-threshold pulse is less than the distance threshold, it means that the further statistical result meets the preset standard. It should be noted that the preset standard is not limited to the above-mentioned situation, and may also include other preset standards, which are all within the protection scope of the present invention.

[0127] When the further statistical results meet the preset standard, it means that the signal strength of the data of the recombined pixels in the first recombined pixel array is strong and the signal-to-noise ratio is high. The feedback processing module 131 can detect obstacles and accurately calculate obstacle information through the first set of detection data based on the preset standard. The feedback processing module 131 is configured to determine that the subsequent emission strategy of the transmitting device 110 is to stop emitting detection pulses and end this detection when it is determined that an obstacle is detected through the first set of detection data based on the preset standard. The feedback processing module 131 feeds back the emission strategy to the transmitting device 110, and the transmitting device 110 is configured to execute the emission strategy.

[0128] On the contrary, when the further statistical results do not meet the preset standard, it means that the signal strength of the data of the recombined pixels in the first group of pixel arrays is weak, the signal-to-noise ratio is low, and the feedback processing module 131 cannot detect obstacles accurately enough through the first group of detection data based on the preset standard. The feedback processing module 131 is configured to determine that the subsequent emission strategy of the transmitting device 110 is to continue to emit the second group of detection pulses when it is determined that the obstacle cannot be detected accurately enough through the first group of detection data based on the preset standard. Similarly, the feedback processing module 131 feeds back the emission strategy to the transmitting device 110, and the transmitting device 110 is configured to execute the emission strategy.

[0129] According to another preferred embodiment of the present invention, when there is no over-threshold pulse in the data of the recombined pixels in the first recombined pixel array, it means that the signal strength of the data of the recombined pixels in the first recombined pixel array is weak, the signal-to-noise ratio is low, and the feedback processing module 131 cannot detect the obstacle accurately enough through the first set of detection data. At this time, the feedback processing module 131 determines that the subsequent emission strategy of the transmitting device 110 is: continue to emit the second set of detection pulses. Similarly, the feedback processing module 131 feeds back the emission strategy to the transmitting device 110, and the transmitting device 110 is configured to execute the emission strategy.

[0130] The above embodiment introduces the specific process of the feedback processing module 131 determining the subsequent emission strategy of the transmitting device 110 according to the first group of detection data. The feedback processing module 131 is configured to count whether there is an over-threshold pulse in the data of the recombined pixels in the first recombined pixel array. When there is an over-threshold pulse in the data of the recombined pixels in the first recombined pixel array, it further counts whether the over-threshold pulse meets the preset standard. If it meets the preset standard, it means that according to the first group of detection data, it is sufficient to detect the obstacle. The feedback processing module 131 determines that the subsequent emission strategy of the transmitting device 110 is: stop emitting detection pulses and end this detection. At this time, the transmitting device 110 does not need to continue to emit detection pulses, and it is not necessary to emit all the Nsweep detection pulses that the light-emitting unit 111 can emit, thereby reducing the actual number of emissions of the light-emitting unit 111, and reducing the actual number of responses of the pixel array of the corresponding detection unit 121, thereby greatly saving the system power consumption of the laser radar. On the contrary, if the preset standard is not met, or when there is no over-threshold pulse in the data of the recombined pixel in the first recombined pixel array, it means that the first set of detection data is not enough to accurately detect the obstacle, and the feedback processing module 131 determines that the subsequent emission strategy of the transmitting device 110 is: continue to emit the second set of detection pulses. At this time, the transmitting device 110 continues to emit the second set of detection pulses for detection. Based on the feedback mechanism, the dynamic adjustment capability of the laser radar has been greatly improved.

[0131] During a detection process, when the transmitting device 110 needs to continue to transmit the second group of detection pulses, the transmitting device 110 can continue to transmit the second group of detection pulses with a reduced emission intensity. Alternatively, the transmitting device 110 can also continue to transmit the second group of detection pulses with a constant emission intensity. According to a preferred embodiment of the present invention, when there is an over-threshold pulse in the data of the recombined pixel in the first recombined pixel array, but the over-threshold pulse does not meet the preset standard, the transmitting device 110 can continue to transmit the second group of detection pulses with a reduced emission intensity for supplementary detection to avoid erroneous judgment of obstacles, and reducing the light intensity can also save power consumption. According to another preferred embodiment of the present invention, when the distance of the obstacle is determined to be relatively far according to the data of the recombined pixel in the first recombined pixel array, but the number and / or intensity of the over-threshold pulses are high, there may be obstacles with high reflectivity, and the transmitting device 110 can continue to transmit the second group of detection pulses with a reduced emission intensity for supplementary detection to reduce the expansion of the high reflectivity object and achieve more accurate detection. It should be noted that the present invention does not limit the specific implementation method of reducing the emission intensity, for example, it can be achieved by reducing the current or reducing the voltage. According to another preferred embodiment of the present invention, when there is no over-threshold pulse in the data of the reconstructed pixels in the first reconstructed pixel array, the emission device 110 may continue to emit the second group of detection pulses at a constant emission intensity.

[0132] In a detection process, when the transmitting device 110 continues to transmit the second group of detection pulses, the data processing device 130 is configured to obtain a second group of detection data, which is the detection data generated after the pixel array responds to the echo reflected by the second group of detection pulses on the obstacle; the feedback processing module 131 is configured to determine the subsequent emission strategy of the transmitting device 110 according to the second group of detection data. Specifically, the second group of detection pulses includes one or more detection pulses, which are recorded as N2. The data processing device 130 can collect and accumulate the signals generated after the pixel array responds to the echo reflected by the N2 detection pulses on the obstacle to generate the second group of detection data, and can accumulate the second group of detection data with the first group of detection data to obtain accumulated data, and determine the subsequent emission strategy of the transmitting device 110 according to the accumulated data in the manner of determining the subsequent emission strategy of the transmitting device 110 according to the first group of detection data, and repeat this cycle until it is determined that the subsequent emission strategy of the transmitting device 110 is to stop transmitting detection pulses, or the total number of transmitted detection pulses reaches Nsweep, and then the detection is terminated.

[0133] According to a preferred embodiment of the present invention, during one detection process, the feedback operation of the feedback processing module 131 can be performed multiple times. For example, each time the transmitting device transmits a group (or a certain number) of detection pulses, the feedback processing module 131 determines the emission strategy of the transmitting device 110 based on the detection data of the group of detection pulses and performs a feedback operation. The feedback processing module 131 can determine the subsequent emission strategy of the transmitting device 110 multiple times by performing multiple feedback operations, and the transmitting device 110 can execute the emission strategy multiple times, which can not only effectively reduce the system power consumption of the laser radar and improve the dynamic adjustment capability of the laser radar, but also is conducive to the balance between the laser radar's signal-to-noise ratio, accuracy, distance measurement capability, system power consumption and other detection performances, and is conducive to improving the laser radar's adaptive capability.

[0134] In order to more accurately determine the emission strategy of the emission device 110, according to a preferred embodiment of the present invention, the feedback processing module 131 is configured to perform ambient light compensation on the data of at least part of the recombined pixels in the first recombined pixel array. Specifically, the feedback processing module 131 can first perform pulse detection on the data of the first recombined pixel, detect whether there is a pulse exceeding the first intensity threshold in the data of the first recombined pixel, and if there is a pulse exceeding the first intensity threshold, perform ambient light compensation on the pulse data exceeding the first intensity threshold. Alternatively, the feedback processing module 131 can also first perform ambient light compensation on the pulse data of the first recombined pixel, and then detect whether there is a pulse exceeding the first threshold in the compensated data. The present invention does not limit the order of over-threshold pulse detection and ambient light compensation. In a specific implementation, the ambient light compensation data can be determined by collecting baseline data by the feedback processing module 131, or it can also be determined by other ambient light collection devices (not shown in the figure). After the ambient light compensation data is determined, the feedback processing module 131 compensates the ambient light compensation data to the pulse peak value exceeding the first intensity threshold, or it can also compensate the ambient light compensation data to the first intensity threshold. The feedback processing module 131 can reduce the influence of ambient light on the data of the recombined pixels in the first recombined pixel array by performing ambient light compensation on the data of at least part of the recombined pixels in the first recombined pixel array, so that the data of the recombined pixels in the first recombined pixel array is more accurate. Therefore, the feedback processing module 131 can also perform statistics on the data of the recombined pixels in the first recombined pixel array after ambient light compensation to obtain more accurate statistical results, so as to more accurately determine the subsequent emission strategy of the transmitting device 110, taking into account the reduction of power consumption and the improvement of accuracy of the laser radar system, which is conducive to further improving the detection performance of the laser radar.

[0135] The above-mentioned embodiments introduce how the feedback processing module 131 determines the subsequent emission strategy of the transmitting device based on the first set of detection data, that is, performing a first convolution processing on the first set of detection data to obtain a first recombined pixel array, and determining the subsequent emission strategy of the transmitting device based on the pulse statistics of the recombined pixel data in the first set of recombined pixel array. If the obstacle is detected accurately enough based on the current scanning results, the detection pulse can be stopped and the detection is ended. This achieves a preliminary judgment based on a small number of scanning results, thereby reducing the total number of light emissions of the transmitting device and the total number of responses of the detection device, greatly saving the system power consumption of the lidar. On the contrary, if the current scanning result is not accurate enough to detect the obstacle, the second group of detection pulses will continue to be emitted, and feedback will be determined based on the accumulated result of the second group of detection data and the first group of detection data. If the obstacle is detected accurately enough based on the accumulated result, the emission of detection pulses can be stopped to end the detection. If the accumulated result is still not accurate enough to detect the obstacle, the third group of detection pulses will continue to be emitted, and feedback will be determined based on the accumulated result of the third group of detection data and the first two groups of detection data. If the obstacle is detected accurately enough based on the accumulated result, the emission of detection pulses can be stopped to end the detection. If the accumulated result is still not accurate enough to detect the obstacle, the fourth group of detection pulses will continue to be emitted, and so on. After each group of detection pulses is emitted, This group of detection data is accumulated with the previous groups of detection data, and feedback judgment is performed based on the accumulated result of this group of detection data and the previous groups of detection data. This cycle is repeated for multiple feedback judgments until the obstacle is detected accurately enough based on the accumulated result, or the total number of transmitted detection pulses reaches Nsweep. The feedback processing module 131 can then determine that the subsequent transmission strategy of the transmitting device 110 is to stop transmitting detection pulses and end this detection, thereby reducing the overall number of scans and greatly saving the system power consumption of the laser radar. Each light-emitting unit can decide the total number of scans corresponding to the light-emitting unit based on its real-time detection results, thereby achieving adaptive adjustment of the number of transmitted detection pulses. While reducing system power consumption, it can also ensure high detection accuracy throughout the field of view.

[0136] Figure 7 A schematic diagram of a laser radar 200 according to a preferred embodiment of the present invention is shown. Figure 7As shown, the laser radar 200 includes a transmitting device 110, a detection device 120, a data processing device 130 and a storage device 140. The storage device 140 is coupled to the data processing device 130 and is used to store detection data. According to a preferred embodiment of the present invention, the data processing device 130 is also configured to store the detection data in the storage device 140 and read the detection data in the storage device 140, wherein the storage device 140 includes a plurality of storage units (not shown in the figure), each storage unit corresponds to the detection data of different pixels at different times, and the data processing device 130 can accumulate the electrical signal output by each pixel according to the triggering time and store it in the corresponding storage unit. The detection data output by the detection device 120, the first recombined pixel array generated according to the first group of detection data, and the second recombined pixel array (described in detail later) can all be stored in the storage device 140.

[0137] Reference Figure 7 According to a preferred embodiment of the present invention, the data processing device 130 further includes an output processing module 132, and the output processing module 132 can be directly or indirectly coupled and communicated with the feedback processing module 131. The output processing module 132 and the feedback processing module 131 can be integrated in the same chip, such as a DSP chip, or can be composed of two independent chips (not shown in the figure), which respectively perform different functions.

[0138] According to a preferred embodiment of the present invention, when the feedback processing module 131 determines that the emission strategy is to stop emitting detection pulses, the output processing module 132 is configured to: perform a second convolution process on the detection data generated according to at least one group of detection pulses emitted by the transmitting device 110 to obtain a second reorganized pixel array, and generate a lidar point cloud based on the second reorganized pixel array, wherein the detection data generated according to the at least one group of detection pulses includes the first group of detection data, and if other groups of detection pulses are continuously emitted after the first group of detection pulses, the detection data generated according to the at least one group of detection pulses also includes the detection data generated by the other groups of detection pulses. Specifically, when the feedback processing module 131 determines that the emission strategy is to stop emitting detection pulses based on the first reorganized pixel array, the output processing module 132 is configured to perform a second convolution process on the first group of detection data to obtain a second reorganized pixel array, and generate a lidar point cloud based on the second reorganized pixel array. On the contrary, when the feedback processing module 131 determines that the emission strategy is to continue to emit the second group of detection pulses according to the first reorganized pixel array, and determines that the emission strategy is to stop emitting detection pulses according to the accumulated results of the second group of detection data and the first group of detection data, the output processing module 132 is configured to perform a second convolution process on the accumulated results of the second group of detection data and the first group of detection data to obtain a second reorganized pixel array, and generate a laser radar point cloud according to the second reorganized pixel array. When the feedback processing module 131 determines that the emission strategy is to continue to emit the third group of detection pulses according to the accumulated results of the second group of detection data and the first group of detection data, and determines that the emission strategy is to stop emitting detection pulses according to the accumulated results of the third group of detection data and the first two groups of detection data, the output processing module 132 is configured to perform a second convolution process on the accumulated results of the third group of detection data and the first two groups of detection data to obtain a second reorganized pixel array, and generate a laser radar point cloud according to the second reorganized pixel array, and so on.

[0139] The following is a detailed description of the case where the feedback processing module 131 determines that the emission strategy is to stop emitting detection pulses according to the first reorganized pixel array, and the output processing module 132 is configured to perform a second convolution process on the first group of detection data to obtain a second reorganized pixel array, and generate a laser radar point cloud according to the second reorganized pixel array. For the case where the transmitting device 110 emits multiple groups of detection pulses, the processing method is similar.

[0140] According to one embodiment of the present invention, when the feedback processing module 131 determines that the emission strategy is to stop emitting detection pulses based on the first reorganized pixel array, the output processing module 132 is configured to perform a second convolution process on the first group of detection data, and the second convolution process includes using a convolution kernel of a certain size and a convolution step size to perform a single second convolution process on the first group of detection data to obtain a second reorganized pixel array, and a laser radar point cloud is generated based on the second reorganized pixel array. Specifically, in this embodiment, the detection unit includes an 8×16 pixel array, the size of the convolution kernel conv=3×3, and the convolution step size stride=1. 2 columns and 2 rows of pixels can be expanded on the right and bottom sides of the detection unit, respectively, and the data processing device 130 performs a second convolution process based on the expanded pixel array; it can be understood that it is also possible not to expand, and only perform a second convolution process based on the pixel array in the detection unit. During the second convolution processing, the convolution kernel conv=3×3 moves step by step in a preset direction (e.g., from left to right, from top to bottom) with a convolution step size of stride=1. Each time it moves, the electrical signals (e.g., echo waveforms) output by all pixels (9 pixels) covered by the convolution kernel are accumulated to obtain a recombined pixel. By traversing the 8×16 pixel array of the entire detection unit, a second recombined pixel array can be obtained. The output processing module 132 generates and outputs a laser radar point cloud based on the obtained second recombined pixel array. It is easy for those skilled in the art to understand that the output processing module 132 can also use a single convolution kernel of other sizes and a convolution step size to perform a second convolution processing on the first set of detection data. For example, the size of the convolution kernel can be 2×2, 4×4, or 5×5, etc., and the convolution step size can also be 2 or other values, which are all within the protection scope of the present invention. It should be understood that if the second convolution processing is to use the same convolution kernel size and convolution step size as the first convolution processing, for example, the convolution kernel conv=1×1 and the convolution step size stride=3, in this case, the output processing module 132 may not need to perform the second convolution processing operation, and directly perform the second convolution processing based on the aforementioned Figure 6 The first reorganized pixel array of 3×6 in the embodiment generates and outputs a laser radar point cloud. Preferably, the second convolution process uses a convolution kernel size and a convolution step size different from those of the first convolution process. In the first convolution process, for example, the convolution kernel size conv=1×1 and the convolution step size stride=3 (see Figure 6 ); in the second convolution processing, for example, the size of the convolution kernel conv=3×3, the convolution step length stride=1, by adopting a smaller convolution step length than the first convolution processing for the second convolution processing, a higher spatial resolution can be obtained, and by adopting a larger convolution kernel than the first convolution processing for the second convolution processing, the signal-to-noise ratio can be improved. In the specific implementation, it can be flexibly set according to needs.

[0141] According to a preferred embodiment of the present invention, when the feedback processing module 131 determines that the emission strategy is to stop emitting detection pulses according to the first recombined pixel array, the output processing module 132 is configured to perform a second convolution process on the first set of detection data, and the second convolution process also includes performing convolution processes on the first set of detection data using a plurality of convolution kernels of different sizes, respectively, to obtain a plurality of second recombined pixel arrays, and generating a laser radar point cloud according to the plurality of second recombined pixel arrays. Wherein, the plurality of different sizes are at least two different sizes, for example, conv=1×1 (refer to Figure 8 ), conv=3×3 (refer to Fig. 9 ), conv=5×5 (refer to Fig.10 ), conv=3×2 (not shown in the figure), etc.

[0142] Since the convolution kernel has a certain size, the convolution operation in the boundary area of ​​the detection unit will be affected. To address this issue, the pixel data in the boundary area of ​​the detection unit can be expanded by the data processing device 130, that is, the data processing device 130 not only collects the electrical signals output by the 8×16 pixel array of the detection unit, but also collects the electrical signals output by some pixels of other adjacent detection units, so that the number of pixels corresponding to the actually collected data is greater than 8×16, so that the convolution kernel can cover some pixels in the boundary area of ​​the detection unit and the boundary area of ​​the adjacent detection unit, thereby obtaining multiple second recombined pixel arrays with the same arrangement, which is convenient for subsequent processing. The following is a specific introduction using an example of using three convolution kernels of different sizes to perform a second convolution on the first set of detection data to obtain multiple second recombined pixel arrays, where the sizes of the convolution kernels are conv=1×1 (refer to Figure 8 ), conv=3×3 (refer to Fig. 9 ), conv=5×5 (refer to Fig.10 ).

[0143] Reference Figure 8 In this embodiment, the detection unit includes an 8×16 pixel array, the size of the convolution kernel conv=1×1, and the convolution step length stride=2. In the second convolution processing process, the convolution kernel conv=1×1 moves step by step in a preset direction (for example, from left to right, from top to bottom) with a convolution step length stride=2. Each time it moves, the electrical signals (for example, echo waveforms) output by all pixels (1 pixel) covered by the convolution kernel are accumulated once to obtain a reconstructed pixel. By traversing the 8×16 pixel array of the entire detection unit, a 4×8 reconstructed pixel array (the second reconstructed pixel array) can be obtained, in which a dark grid represents a reconstructed pixel. Figure 8 The dotted box outlines one of the recombined pixels and its corresponding single pixel (original pixel). Figure 8The lower right portion shows the signal waveform of one reconstructed pixel (original pixel).

[0144] Reference Fig. 9 In this embodiment, the detection unit includes an 8×16 pixel array, the size of the convolution kernel conv=3×3, and the convolution step length stride=2. The data processing device 130 can expand 1 column and 1 row of pixels on the right and bottom sides of the detection unit, respectively. In the second convolution processing process, the convolution kernel conv=3×3 moves step by step in a preset direction (for example, from left to right, from top to bottom) with a convolution step length stride=2. For each movement, the electrical signals (for example, echo waveforms) output by all the pixels (9 pixels) covered by the convolution kernel are accumulated once to obtain a recombined pixel. By traversing the 8×16 pixel array of the entire detection unit, a 4×8 recombined pixel array (the second recombined pixel array) can also be obtained, in which a dark grid represents a recombined pixel. Fig. 9 The dotted box outlines one of the reconstructed pixels and its corresponding 9 pixels (original pixels). Fig. 9 The lower right part shows the signal waveform comparison between a reconstructed pixel and a single pixel. It can be seen that when the size of the convolution kernel is larger than 1×1, the signal intensity of the reconstructed pixel is stronger and the signal-to-noise ratio is higher.

[0145] Reference Fig.10 In this embodiment, the detection unit includes an 8×16 pixel array, the size of the convolution kernel conv=5×5, and the convolution step length stride=2. The data processing device 130 can expand 3 columns and 3 rows of pixels on the right and bottom sides of the detection unit, respectively. In the second convolution processing process, the convolution kernel conv=5×5 moves step by step in a preset direction (for example, from left to right, from top to bottom) with a convolution step length stride=2. For each movement, the electrical signals (for example, echo waveforms) output by all pixels (25 pixels) covered by the convolution kernel are accumulated once to obtain a recombined pixel. By traversing the 8×16 pixel array of the entire detection unit, a 4×8 recombined pixel array (the second recombined pixel array) can also be obtained, in which a dark grid represents a recombined pixel. Fig.10 The dotted box outlines one of the reconstructed pixels and its corresponding 25 pixels (original pixels). Fig.10 The lower right part shows the signal waveform comparison between a recombined pixel and a single pixel. It can be seen that the signal strength of the recombined pixel is stronger and the signal-to-noise ratio is higher. Fig.10 The reconstructed pixel ratio in the embodiment Fig. 9 The signal intensity of the recombined pixel in the embodiment is stronger and the signal-to-noise ratio is higher.

[0146] The above embodiment introduces the detailed process of the output processing module 132 performing the second convolution processing on the first group of detection data. The processing module 132 can use a single-sized convolution kernel to perform the second convolution processing on the first group of detection data, or can use multiple convolution kernels of different sizes to perform the second convolution processing on the first group of detection data, and the size of the convolution kernel and the convolution step size can be flexibly set according to actual conditions, which is conducive to improving the dynamic adjustment capability of the laser radar. Preferably, compared with the convolution step size in the first convolution processing, the output processing module 132 can preferentially use a smaller convolution step size in the second convolution processing process to improve the spatial resolution, and finally obtain a fine calculation result.

[0147] According to a preferred embodiment of the present invention, after the second convolution processing is completed, the output processing module 132 is further configured to perform ambient light compensation on at least part of the reconstructed pixel data in the second reconstructed pixel array to reduce the influence of ambient light on the data of the reconstructed pixels in the second reconstructed pixel array, so that the data of the reconstructed pixels in the second reconstructed pixel array is more accurate. The output processing module 132 performs ambient light compensation on at least part of the reconstructed pixel data in the second reconstructed pixel array, which is similar to the operation of the feedback processing module 131 performing ambient light compensation on at least part of the reconstructed pixel data in the first reconstructed pixel array, and will not be repeated here.

[0148] According to a preferred embodiment of the present invention, the output processing module 132 is configured to select an optimal second recombined pixel array from the plurality of second recombined pixel arrays to generate a lidar point cloud.

[0149] Assuming that the convolution kernel size is P×P, and the echoes in the field of view covered by the convolution kernel correspond to the same obstacle and are evenly distributed, the signal-to-noise ratio of the reconstructed pixels obtained by convolution kernels of different sizes can be expressed as:

[0150] SNR=P×SNR0

[0151] Among them, SNR0 is the signal-to-noise ratio of a single original pixel, and SNR is the signal-to-noise ratio of the reconstructed pixel. It can be seen that the larger the size of the convolution kernel, the stronger the signal intensity of the reconstructed pixel obtained after convolution processing, and the higher the signal-to-noise ratio, that is, the signal-to-noise ratio (or signal intensity) of the reconstructed pixel is positively correlated with the size of the convolution kernel.

[0152] It can be seen that in the actual detection process of LiDAR, in order to improve the signal-to-noise ratio, obtain better distance measurement performance and more accurate detection results, a larger convolution kernel can be used to perform convolution processing on the electrical signal output by the pixel array of the detection unit, which is especially beneficial for the detection of long-distance and weak targets.

[0153] Figures 8 to 10In the embodiment, the data processing device 130 expands the pixel data of the detection unit boundary area, and the output processing module 132 uses multiple convolution kernels of different sizes to perform the second convolution processing with the same convolution step size, so as to obtain multiple second reconstructed pixel arrays of the same size, which effectively alleviates the influence of the convolution kernel size on the convolution operation of the detection unit boundary area, and each reconstructed pixel at the same position has multiple data values. Specifically, refer to Figures 8 to 10 For example, each reconstructed pixel at the same position has three data values, including 1×1 convolution data, 3×3 convolution data, and 5×5 convolution data. Therefore, each reconstructed pixel at the same position has waveform data with different characteristics, so that multiple reconstructed pixel arrays have waveform data with different characteristics, such as different resolutions, signal-to-noise ratios, echo pulse strengths, etc. The output processing module 132 can select the optimal reconstructed pixel array according to the waveform data of the multiple reconstructed pixel arrays, and generate a laser radar point cloud according to the selected optimal reconstructed pixel array, which is conducive to further improving the adaptive ability and dynamic range of the laser radar and improving the detection performance of the laser radar. According to a preferred embodiment of the present invention, the waveform data includes a signal-to-noise ratio (SNR). The output processing module 132 is configured to determine the signal-to-noise ratio of each of the multiple second reconstructed pixel arrays, and select the signal-to-noise ratio greater than the signal-to-noise ratio threshold T from the multiple second reconstructed pixel arrays. SNR When the signal-to-noise ratio of the plurality of second recombined pixel arrays is higher than the signal-to-noise ratio threshold T SNR When , the second reorganized pixel array with the smallest convolution kernel size is selected. Specifically, for example Figure 8 , Fig. 9 and Fig.10 In the convolution process, convolution kernels with sizes of 1×1, 3×3, and 5×5 are used respectively. The signal-to-noise ratios of the signal waveforms of the three second recombined pixel arrays obtained after the convolution process are T1, T3, and T5 respectively. If T1 <T SNR , T3 <T SNR , T5>T SNR , then the convolution result (second reorganized pixel array) obtained by selecting the 5×5 convolution kernel is used to generate the lidar point cloud, and the obstacle information is calculated based on the convolution result of the 5×5 convolution kernel; if T1 <T SNR , T3>T SNR , T5>T SNR , then the convolution result obtained by selecting the 3×3 convolution kernel generates the lidar point cloud, and the obstacle information is calculated based on the convolution result of the 3×3 convolution kernel; if T1>T SNR , T3>T SNR , T5>T SNR, then the convolution result obtained by selecting a 1×1 convolution kernel (i.e., the original pixel array) is used to generate lidar point clouds, and obstacle information is calculated based on the convolution result of the 1×1 convolution kernel (i.e., the original single pixel); if T1 < T SNR , T3 < T SNR , T5 < T SNR , it indicates that the signal-to-noise ratio of the signal is too low and no target is detected, so no lidar point clouds are generated.

[0154] Therefore, when processing the detection data output by multiple pixels in the detection unit, convolution processing is respectively performed using convolution kernels with side lengths of n0, n1, n2,... (where n0 < n1 < n2 <...). Assuming that the signal-to-noise ratios of the echo signals of the multiple second recombined pixel arrays obtained are SNR(n0), SNR(n1), SNR(n2),..., and so on, a signal-to-noise ratio threshold is set as T SNR , then the finally selected convolution kernel size is:

[0155] n = argmin(SNR(ni) > T SNR )

[0156] That is, the minimum convolution kernel size with a signal-to-noise ratio exceeding the threshold T SNR is selected as the finally selected convolution kernel size, and lidar point clouds are generated based on the second recombined pixel array obtained by the convolution kernel of the selected size.

[0157] According to another preferred embodiment of the present invention, the waveform data further includes pulse intensity. The output processing module 132 is configured to determine the pulse intensity of each second recombined pixel array and select a second recombined pixel array with a pulse intensity greater than the pulse intensity threshold T PS from the multiple second recombined pixel arrays. When the pulse intensities of the multiple second recombined pixel arrays are higher than the pulse intensity threshold T PS , the second recombined pixel array with the smallest convolution kernel size is selected. Specifically, for example, convolution processing is respectively performed using convolution kernels with sizes of 1×1, 3×3, and 5×5. The pulse intensities of the signal waveforms of the three second recombined pixel arrays obtained after convolution processing are P1, P3, and P5 respectively. If P1 < T PS , P3 < T PS , P5 > T PS , then the convolution result (second recombined pixel array) obtained by the 5×5 convolution kernel is used to generate lidar point clouds, and obstacle information is calculated based on the convolution result of the 5×5 convolution kernel; if P1 < T PS , P3 > T PS , P5 > T PS, then the convolution result (the second reorganized pixel array) obtained by selecting a 3×3 convolution kernel is used to generate lidar point cloud, and obstacle information is calculated based on the convolution result of the 3×3 convolution kernel; if P1>T PS , P3>T PS , P5>T PS , then the convolution result (i.e., the original pixel array) obtained by selecting a 1×1 convolution kernel is used to generate lidar point cloud, and obstacle information is calculated based on the convolution result of the 1×1 convolution kernel; if P1<T PS , P3<T PS , P5<T PS , then no lidar point cloud is generated.

[0158] Therefore, when processing the detection data output by multiple pixels in the detection unit, convolution processing is respectively performed using convolution kernels with side lengths of n0, n1, n2, …… (where n0 < n1 < n2 < ……). Assuming that the pulse intensities of the echo signals of the reorganized pixels obtained are PS(n0), PS(n1), PS(n2) ……, and so on, a pulse intensity threshold is set as T PS , then the finally selected convolution kernel size is:

[0159] n = argmin(PS(ni)>T PS )

[0160] That is, the minimum convolution kernel size with a pulse intensity exceeding the threshold T PS is selected as the finally selected convolution kernel size, and lidar point cloud is generated based on the second reorganized pixel array obtained by the convolution kernel of the selected size.

[0161] According to another preferred embodiment of the present invention, the size of the convolution kernel can also be comprehensively determined based on the signal-to-noise ratio and pulse intensity of the second reorganized pixel array.

[0162] Specifically, for example, the minimum convolution kernel size with a signal-to-noise ratio exceeding the threshold T SNR and a pulse intensity exceeding the threshold T PS is selected as the finally selected convolution kernel size, as follows:

[0163] n = argmin(SNR(ni)>T SNR &PS(ni)>T PS )

[0164] Another example is to select the minimum convolution kernel size with a signal-to-noise ratio exceeding the threshold T SNR or a pulse intensity exceeding the threshold T PS as the finally selected convolution kernel size, as follows:

[0165] n = argmin(SNR(ni)>T SNR|PS(ni)>T PS )

[0166] It should be noted that Figures 8 to 10 In the embodiments, the sizes of the convolution kernels are 1×1, 3×3, and 5×5, respectively, and the convolution step size is stride=2, but the present invention is not limited thereto. The size of the convolution kernel may also be other sizes, such as conv=3×2, and the convolution step size may also be other convolution step sizes, such as stride=1, etc.

[0167] Due to the above reference Figure 8 , Fig. 9 , Fig.10 The arrangement of the plurality of second recombined pixel arrays described is the same, so each recombined pixel has a plurality of data values ​​(at least two). Figure 8 , Fig. 9 , Fig.10 For example, each reconstructed pixel has three data values, including 1×1 convolution data, 3×3 convolution data, and 5×5 convolution data. For the reconstructed pixel data values ​​obtained by performing the second convolution processing using convolution kernels of other sizes, similarly, assuming that the size of the convolution kernel is P×Q, the reconstructed pixel data value is P×Q convolution data, and so on, where P and Q are both integers greater than or equal to 1, and P and Q can be the same or different.

[0168] According to a preferred embodiment of the present invention, the output processing module 132 is further configured to determine the state of each recombined pixel data in the plurality of second recombined pixel arrays. Preferably, the state includes a pulse over-threshold state. It should be noted that the over-threshold in this embodiment refers to the data value of the recombined pixel exceeding the third intensity threshold, and the present invention does not limit the relationship between the third intensity threshold and the aforementioned first intensity threshold and second intensity threshold. Specifically, the output processing module 132 can perform pulse detection on the data of the recombined pixels (such as echo waveforms) in the plurality of second recombined pixel arrays, and determine the state of each recombined pixel data based on the pulse detection result. More specifically, refer to Figure 8 , Fig. 9 , Fig.10, the output processing module 132 can respectively detect whether the 1×1 convolution data (i.e., single point data), 3×3 convolution data, and 5×5 convolution data of each recombined pixel data have crossed the threshold to determine the state of each recombined pixel data (referred to as pulse threshold crossing state). For example, if the 1×1 convolution data of a recombined pixel data has crossed the threshold, the state of the recombined pixel data is single point detection threshold crossing; if the 3×3 convolution data and / or 5×5 convolution data of a recombined pixel data have crossed the threshold, the state of the recombined pixel data is convolution detection threshold crossing; if the 1×1 convolution data, 3×3 convolution data, and 5×5 convolution data of a recombined pixel data have not crossed the threshold, the state of the recombined pixel data is not crossing the threshold. Therefore, the state of the recombined pixel data (pulse threshold crossing state) may include single point detection threshold crossing, convolution detection threshold crossing, and not crossing the threshold; or may also include single point detection threshold crossing, 3×3 convolution detection threshold crossing, 5×5 convolution detection threshold crossing, and not crossing the threshold. Preferably, the output processing module 132 can also mark and classify each recombined pixel based on the status of each recombined pixel data, for example, marking it as single point detection over-threshold, convolution detection over-threshold, and not over-threshold, for a total of 3 categories; or marking it as single point detection over-threshold, 3×3 convolution detection over-threshold, 5×5 convolution detection over-threshold, and not over-threshold, for a total of 4 categories, to facilitate subsequent processing.

[0169] like Figure 8 , Fig. 9 and Fig.10 As shown in , the dimension of the first set of detection data is 8×16, while the dimension of the second recombined pixel array is 4×8. By adopting a larger convolution step size for the second convolution processing, multiple discrete recombined pixels can be obtained to form the second recombined pixel array, thereby reducing the amount of calculation, speeding up the calculation speed, and saving system power consumption. In order to improve the resolution, preferably, the blank pixels in the middle of the discrete recombined pixels can be calculated based on the discrete recombined pixels, so as to determine the final recombined pixel array, and generate the lidar point cloud based on the final recombined pixel array. The following is a detailed introduction.

[0170] According to a preferred embodiment of the present invention, the output processing module 132 is configured to determine the final recombined pixel array according to the state of the recombined pixel data in the plurality of second recombined pixel arrays, and generate the lidar point cloud according to the final recombined pixel array. Preferably, the output processing module 132 is configured to determine the final recombined pixel array according to the pulse threshold crossing state of at least two adjacent recombined pixels in the second recombined pixel arrays. The adjacent recombined pixels include recombined pixels adjacent to each other up and down, left and right, or diagonally. Fig.11As shown in the example, recombined pixel 1 and recombined pixel 2 are recombined pixels adjacent to the middle recombined pixel C1 in the left and right, and recombined pixel 3 and recombined pixel 4 are recombined pixels adjacent to the middle recombined pixel C4 in the left and right; recombined pixel 1 and recombined pixel 3 are recombined pixels adjacent to the middle recombined pixel C2 in the top and bottom, and recombined pixel 2 and recombined pixel 4 are recombined pixels adjacent to the middle recombined pixel C3 in the top and bottom; recombined pixel 1 and recombined pixel 4 are recombined pixels adjacent to the middle recombined pixel C5 in the diagonal line, and recombined pixel 2 and recombined pixel 3 are also recombined pixels adjacent to the middle recombined pixel C5 in the diagonal line. And, Fig.11 Among them, the darkest colors, such as reconstructed pixels 1, 2, 3, and 4, are reconstructed pixels obtained by the second convolution process, and the middle reconstructed pixels C1, C2, C3, C4, and C5 are pixels that have not been processed by the second convolution process (blank pixels).

[0171] In order to determine the final recombined pixel array, according to a preferred embodiment of the present invention, the output processing module 132 is configured to determine the calculation method of the intermediate recombined pixel data according to the pulse threshold crossing state of at least two adjacent recombined pixel data in the second recombined pixel array. The intermediate recombined pixel exists between every two adjacent recombined pixels. Fig.11 As shown in the example, the intermediate recombined pixel C2 exists between the recombined pixel 1 and the recombined pixel 3 adjacent to each other; the intermediate recombined pixel C3 exists between the recombined pixel 2 and the recombined pixel 4 adjacent to each other; the intermediate recombined pixel C1 exists between the recombined pixel 1 and the recombined pixel 2 adjacent to each other; the intermediate recombined pixel C4 exists between the recombined pixel 3 and the recombined pixel 4 adjacent to each other; and the intermediate recombined pixel C5 exists between the diagonally adjacent recombined pixel 1 and the recombined pixel 4 (or the recombined pixel 2 and the recombined pixel 3). Therefore, the output processing module 132 can determine the calculation method of the intermediate recombined pixel data according to the pulse threshold crossing state of at least two adjacent recombined pixel data, and then determine the final recombined pixel array. According to a preferred embodiment of the present invention, the output processing module 132 is configured to determine the calculation method of the intermediate recombined pixel data according to the pulse threshold crossing state of at least two adjacent recombined pixel data and the distance and / or reflectivity of the corresponding obstacles. The calculation method of the intermediate recombined pixel data includes one of the following: interpolation calculation of the intermediate recombined pixel, single point calculation of the intermediate recombined pixel, and convolution calculation of the intermediate pixel.

[0172] The following description takes the calculation method of determining the data of the intermediate recombined pixel C1 according to the pulse threshold crossing states of the adjacent recombined pixels 1 and 2 as an example. It should be understood that the calculation method of determining the data of the intermediate recombined pixel is similar when the pulse threshold crossing states of a larger number of adjacent recombined pixels are used.

[0173] Reference Fig.11, assuming that the pulse threshold crossing states of the recombined pixel 1 and the recombined pixel 2 are both single-point detection threshold crossings, in this case, it indicates that the echo pulse intensities of the two recombined pixels are relatively strong, but the two adjacent recombined pixels may correspond to the same obstacle or may not correspond to the same obstacle. In this regard, the output processing module 132 can respectively determine the distance and / or reflectivity of the obstacles corresponding to the two recombined pixels, and use the first judgment condition to judge the distance and / or reflectivity to avoid or reduce false targets, so as to make the result more accurate. If the difference in distance and / or reflectivity of the obstacles corresponding to the two recombined pixels is less than a preset value, it is considered that the recombined pixel 1 and the recombined pixel 2 meet the first judgment condition and correspond to the same obstacle. At this time, the data of the recombined pixel 1 and / or the recombined pixel 2 can be used to perform interpolation calculation on the middle pixel C1; conversely, if the difference in distance and / or reflectivity of the obstacles corresponding to the two recombined pixels is not less than a preset value, it is considered that the recombined pixel 1 and the recombined pixel 2 do not meet the first judgment condition and do not correspond to the same obstacle. At this time, a single-point calculation can be performed on the middle recombined pixel C1, that is, the data at the corresponding position in the first set of detection data is used as the data of the middle recombined pixel C1, which can achieve high-resolution detection and reduce or even avoid the problem of target expansion, which is beneficial to reducing the amount of calculation, saving the power consumption of the laser radar system, and taking into account the accuracy of the detection results. It is particularly suitable for the detection of close-range or high-reflectivity targets.

[0174] Reference Fig.11 , assuming that the pulse threshold state of the recombined pixel 1 is a single-point detection threshold state, and the pulse threshold state of the recombined pixel 2 is a convolution detection threshold state, in this case, the output processing module 132 can respectively determine the distance and / or reflectivity of the obstacles corresponding to the two recombined pixels, and use the second judgment condition to judge the distance and / or reflectivity. If the second judgment condition is met, the data of the recombined pixel 1 and / or the recombined pixel 2 are used for interpolation calculation for the middle pixel C1; if the second judgment condition is not met, a single-point calculation is performed on the middle recombined pixel C1, that is, the data at the corresponding position in the first set of detection data is used as the data of the middle recombined pixel C1. It should be understood that the first judgment condition can be more stringent than the second judgment condition (that is, the judgment is also made by the difference in the distance and / or reflectivity of the obstacles corresponding to the two recombined pixels, but the preset value adopted in the first judgment condition is smaller), and it can be understood that the two can also be equal.

[0175] Reference Fig.11 , assuming that the pulse threshold crossing state of the recombined pixel 1 is a single-point detection threshold crossing, and the pulse threshold crossing state of the recombined pixel 2 is not a threshold crossing, in this case, the output processing module 132 can perform a single-point calculation on the intermediate recombined pixel C1, that is, perform a 1×1 convolution calculation, to obtain the data of the intermediate recombined pixel C1.

[0176] For weaker targets, refer to Fig.11 , assuming that the pulse threshold-crossing states of the recombined pixel 1 and the recombined pixel 2 are both convolution detection threshold-crossing states, in this case, the output processing module 132 can respectively determine the distance and / or reflectivity of the obstacles corresponding to the two recombined pixels, and use the third determination condition to determine the distance and / or reflectivity. If the third determination condition is met, that is, the distance and / or reflectivity of the two are slightly different, the output processing module 132 uses the data of the recombined pixel 1 and / or the recombined pixel 2 (for example, 3×3 convolution data or 5×5 convolution data) to interpolate the middle pixel C1; if the third determination condition is not met, that is, the distance and / or reflectivity of the two are significantly different, the data of the middle pixel C1 is output as 0. It should be noted that the third determination condition may be the same as or different from the first determination condition and / or the second determination condition.

[0177] Reference Fig.11 , assuming that the pulse threshold state of the recombined pixel 1 is the convolution detection threshold, and the pulse threshold state of the recombined pixel 2 is not over the threshold, in this case, the signal strength of the target echo is weak, and the output processing module 132 can use a large-size convolution kernel to perform convolution calculation on the middle pixel C1, for example, use a 5×5 convolution kernel to perform convolution calculation to obtain data for the middle pixel C1; or the convolution data of the large-size convolution kernel of the recombined pixel 1 can be used as the data of the middle pixel C1; in this way, a larger convolution kernel can be used for convolution calculation for the weaker echo signal to enhance the signal strength of the middle pixel C1, thereby improving the signal-to-noise ratio of the weak target echo signal.

[0178] Reference Fig.11 , assuming that the pulse threshold crossing states of the recombined pixel 1 and the recombined pixel 2 are both not crossing the threshold, in this case, it indicates that no obstacle is detected, and the data output of the middle pixel C1 is 0.

[0179] The above embodiment introduces how to determine the calculation method of the intermediate recombined pixel C1 data according to the pulse threshold crossing state of the adjacent recombined pixel 1 and the recombined pixel 2. Similarly, for the intermediate recombined pixel C2 data, it can be determined according to the pulse threshold crossing state of the adjacent recombined pixel 1 and the recombined pixel 3; for the intermediate recombined pixel C3 data, it can be determined according to the pulse threshold crossing state of the adjacent recombined pixel 2 and the recombined pixel 4; for the intermediate recombined pixel C4 data, it can be determined according to the pulse threshold crossing state of the adjacent recombined pixel 3 and the recombined pixel 4; for the intermediate recombined pixel C5 data, it can be determined according to the pulse threshold crossing state of the adjacent recombined pixels 1, 2, 3, 4. In short, the intermediate recombined pixel data can be determined according to the pulse threshold crossing state of at least two adjacent recombined pixels, and then the final recombined pixel array can be determined, and the lidar point cloud can be generated according to the final recombined pixel array. By obtaining intermediate recombinant pixel data (pixel data that has not undergone the second convolution processing) based on the data of adjacent recombinant pixels, the amount of convolution calculation can be reduced, the calculation speed can be improved, the system power consumption can be saved, and the resolution and accuracy of the detection results will not be lost; at the same time, based on the judgment of distance and / or reflectivity, problems such as the expansion of high-reflectivity objects can be reduced. Compared with using a convolution kernel of the same size, the convolution calculation is performed on each intermediate recombinant pixel, and the intermediate recombinant pixels and the final recombinant pixel array obtained have better resolution and distance and reflectivity accuracy. At the same time, the resolution and signal-to-noise ratio of the final recombinant pixel array of the entire detection unit can be dynamically distributed according to the signal strength, so that the laser radar can dynamically maintain better detection performance throughout the field of view.

[0180] In addition, it should be noted that, in the present invention, the feedback operation is implemented by the feedback processing module 131 for the purpose of reducing the power consumption of the laser radar system. In some cases, especially when the laser radar has a large power margin, the feedback operation of the feedback processing module 131 can also be skipped, that is, the preset total number of scans Nsweep is continuously executed, and then the detection data obtained according to the Nsweep scans is processed, and the subsequent data processing operation is performed by the output processing module 132. For example, for some laser radars dedicated to measuring close-range obstacles, the power required for each detection is relatively low, so the above-mentioned feedback operation can be skipped or omitted, and Nsweep detection pulses are directly emitted, and the relevant information of the obstacle is calculated based on the Nsweep detection results.

[0181] The laser radar of the present invention can realize the dynamic resolution and signal-to-noise ratio of the entire field of view as a whole. It has a higher spatial resolution for the strong echo signal area, and a higher signal-to-noise ratio and a stronger distance measurement capability for the weak echo signal area. It can detect long-distance, low-reflectivity, small target signals, and maintain a high spatial resolution for the detection of close-range, strong echo (high-reflectivity) targets. It can also reduce the expansion problem of high-reflectivity objects, which is conducive to improving the accuracy of the detection results, so that the laser radar can dynamically maintain better detection performance in the entire field of view.

[0182] The present invention also provides a detection method 10 of a laser radar 100, referring to Figure 2 The laser radar 100 in the figure includes a transmitting device 110 and a detecting device 120, wherein the transmitting device 110 includes at least one light-emitting unit 111, wherein the light-emitting unit 111 transmits a plurality of detection pulses L during one detection process, and the detecting device 120 includes at least one detection unit 121, wherein each detection unit 121 includes a pixel array; the detection method 10 includes: during one detection process, including steps S11-S13:

[0183] In step S11, a first group of detection pulses is first transmitted, wherein the first group of detection pulses includes one or more detection pulses;

[0184] In step S12, the pixel array of the detection unit 121 responds to the echo of the first group of detection pulses reflected on the obstacle OB and generates a first group of detection data; and

[0185] In step S13, a subsequent transmission strategy is determined according to the first set of detection data.

[0186] According to a preferred embodiment of the present invention, step S13 comprises:

[0187] performing a first convolution process on the first set of detection data to obtain a first reconstructed pixel array; and

[0188] A subsequent emission strategy of the emission device is determined according to the first recombined pixel array.

[0189] According to a preferred embodiment of the present invention, step S13 comprises:

[0190] Taking statistics on the data of the reconstructed pixels in the first reconstructed pixel array; and

[0191] According to the statistical results, the subsequent transmission strategy of the transmitting device 110 is determined.

[0192] According to a preferred embodiment of the present invention, the step of performing statistics on the data of the reconstructed pixels in the first reconstructed pixel array comprises:

[0193] Counting whether there is an over-threshold pulse in the data of the recombined pixels in the first recombined pixel array and obtaining a pulse statistical result, wherein the over-threshold pulse is a pulse exceeding a first intensity threshold;

[0194] The subsequent transmission strategy of the transmitting device is determined according to the pulse statistical result.

[0195] According to a preferred embodiment of the present invention, when the over-threshold pulse exists in the data of the reconstructed pixels in the first reconstructed pixel array, the step of performing statistics on the data of the reconstructed pixels in the first reconstructed pixel array further comprises:

[0196] Further counting one or more of the following parameters: the intensity of each of the over-threshold pulses, the obstacle distance corresponding to each of the over-threshold pulses, the number of reconstructed pixels having the over-threshold pulses, and the ratio of the reconstructed pixels having the over-threshold pulses to all the reconstructed pixels in the first reconstructed pixel array;

[0197] According to the further statistical results, the subsequent transmission strategy of the transmitting device 110 is determined.

[0198] According to a preferred embodiment of the present invention, the step S13 comprises: performing ambient light compensation on at least part of the reorganized pixel data in the first reorganized pixel array.

[0199] According to a preferred embodiment of the present invention, the transmission strategy includes one of the following:

[0200] continuing to transmit a second set of detection pulses, wherein the second set of detection pulses includes one or more detection pulses; and

[0201] Stop sending detection pulses and end this detection.

[0202] According to a preferred embodiment of the present invention, the step of continuing to transmit the second group of detection pulses comprises:

[0203] continuing to transmit the second set of detection pulses at a reduced transmission intensity; or

[0204] The second set of detection pulses continues to be emitted at a constant emission intensity.

[0205] According to a preferred embodiment of the present invention, step S13 comprises:

[0206] When it is determined through the first group of detection data that an obstacle is detected based on a preset standard, the emission of detection pulses is stopped; otherwise, the emission of the second group of detection pulses continues.

[0207] According to a preferred embodiment of the present invention, the detection method 10 further includes:

[0208] Feedback the transmission strategy to the transmission device 110; and

[0209] The transmitting device 110 is controlled to execute the fed-back transmitting strategy.

[0210] According to a preferred embodiment of the present invention, during a detection process, when the second group of detection pulses continues to be emitted, the detection method 10 further includes:

[0211] Acquire a second set of detection data, where the second set of detection data is detection data generated after the pixel array responds to echoes reflected by the second set of detection pulses on the obstacle OB; and

[0212] The subsequent transmission strategy is determined based on the second set of detection data.

[0213] According to a preferred embodiment of the present invention, when it is determined that the emission strategy is to stop emitting detection pulses, the detection method 10 further includes: performing a second convolution processing on the detection data generated by at least one group of detection pulses emitted by the transmitting device; obtaining a second recombinant pixel array, and generating a lidar point cloud based on the second recombinant pixel array, wherein the detection data generated by the at least one group of detection pulses includes the first group of detection data.

[0214] According to a preferred embodiment of the present invention, the step of performing a second convolution processing on the detection data includes: performing convolution processing on the detection data using a plurality of convolution kernels of different sizes to obtain a plurality of second recombined pixel arrays, and generating a lidar point cloud based on the plurality of second recombined pixel arrays.

[0215] According to a preferred embodiment of the present invention, the detection method 10 further comprises: performing ambient light compensation on at least part of the reconstructed pixel data in the second reconstructed pixel array.

[0216] According to a preferred embodiment of the present invention, the detection method 10 further includes: determining a final recombined pixel array according to the state of the recombined pixel data in the plurality of second recombined pixel arrays, and generating a lidar point cloud according to the final recombined pixel array.

[0217] According to a preferred embodiment of the present invention, the step of determining the final recombined pixel array according to the states of the recombined pixel data in the plurality of second recombined pixel arrays comprises: determining the final recombined pixel array according to the pulse threshold crossing states of at least two adjacent recombined pixels in the second recombined pixel arrays.

[0218] According to a preferred embodiment of the present invention, the step of determining the final recombined pixel array according to the states of the recombined pixel data in the plurality of second recombined pixel arrays further comprises: determining a calculation method for the intermediate recombined pixel data according to the pulse over-threshold states of at least two adjacent recombined pixel data in the second recombined pixel arrays.

[0219] According to a preferred embodiment of the present invention, the step of determining the final recombined pixel array based on the states of the recombined pixel data in the plurality of second recombined pixel arrays further comprises: determining a calculation method for the intermediate recombined pixel data based on the pulse over-threshold states of the at least two recombined pixel data and the distance and / or reflectivity of the obstacles to which they correspond.

[0220] According to a preferred embodiment of the present invention, the calculation method of the intermediate recombined pixel data includes one of the following: interpolation of the intermediate pixels, single-point calculation of the intermediate pixels, and convolution calculation of the intermediate pixels.

[0221] The present invention also provides a computer-readable storage medium, comprising computer-executable instructions stored thereon, wherein the executable instructions, when executed by a processor, implement the detection method 10 of the laser radar 100 as described above.

[0222] In some preferred embodiments, the computer-readable storage medium can be any tangible medium containing or storing a program, which can be used by an instruction execution system, device or device or used in combination with it. The computer-readable storage medium can adopt any combination of one or more computer-readable media. The computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, or semiconductor form or device, and more specific examples (non-exhaustive list) include: an electrical connection with one or more wires, a portable computer hard disk, a hard disk, a random access memory (RAM), a non-volatile random access memory (NVRAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0223] In addition, the present invention also provides a computer device (not shown), comprising a memory and a processor, wherein the memory is configured to store computer executable instructions, and the executable instructions implement the detection method 10 described above when executed by the processor.

[0224] In some preferred embodiments, the processor / data processing device 130 / feedback processing module 131 / output processing module 132 may include a central processing unit (CPU), and may also include other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The present invention is not limited to this and it depends on the specific circumstances.

[0225] In summary, the scheme of the present invention has been introduced in detail. The laser radar of the present invention works based on a feedback mechanism. In a detection process, the transmitting device first transmits a first group of detection pulses, and the feedback processing module determines the subsequent emission strategy of the transmitting device based on the first group of detection data. The subsequent emission strategy may be to stop emitting light, or to continue to emit detection pulses with a reduced emission intensity, or to continue to emit detection pulses with an unchanged emission intensity. Thereby, the overall number of scans can be reduced, greatly saving the system power consumption of the laser radar. At the same time, each of the multiple light-emitting units can decide the total number of scans corresponding to the light-emitting unit according to its real-time detection results, thereby realizing adaptive adjustment of the number of emitted detection pulses, which can reduce the system power consumption while ensuring high detection accuracy within the entire field of view.

[0226] Furthermore, the feedback processing module performs the first convolution processing on the first group of detection data by adopting a larger step size and a smaller convolution kernel, so that the calculation amount of the first convolution processing is lower and the calculation speed is faster, which can reduce the system power consumption of the laser radar.

[0227] Furthermore, during the second convolution processing, the output processing module preferentially adopts a smaller convolution step size for the second convolution processing than the convolution step size in the first convolution processing. By adopting a coarse-first-then-fine approach, the amount of calculation can be reduced, the system power consumption of the lidar can be reduced, and ultimately a precise calculation result can be obtained.

[0228] Furthermore, the feedback processing module / output processing module can reduce the interference of ambient light by performing ambient light compensation on the data of at least part of the recombined pixels in the first / second recombined pixel array, thereby making the result more accurate.

[0229] In addition, for different detection targets, this solution can also adjust the total number of light emissions through a feedback mechanism. For close-range, high-reflectivity (strong echo signal) areas, fewer light emissions are used, while for long-range, low-reflectivity (weak echo signal) areas, more light emissions are used, thereby achieving adaptive adjustment of the number of light emissions based on real-time echo signals and reducing the system power consumption of the entire lidar.

[0230] In addition, the laser radar of the present invention can adjust the spatial resolution and signal-to-noise ratio. For close-range, high-reflectivity targets, it can achieve high spatial resolution; for long-range, weak-reflectivity targets, it can maintain a high signal-to-noise ratio and achieve longer detection capabilities. It can also alleviate the expansion problem of high-reflectivity objects and the problem of false targets, which is beneficial to improving the accuracy of the detection results and enabling the laser radar to dynamically maintain better detection performance throughout the entire field of view.

[0231] It should be noted that, although several modules or submodules of the laser radar 100 are mentioned in the above detailed description, such division is not mandatory. In fact, according to an embodiment of the present invention, the features and functions of two or more modules described above can be implemented in one module. Conversely, the features and functions of one module described above can be further divided into multiple modules to be embodied.

[0232] It should be noted that this specification provides method operation steps as described in the embodiments or schematic diagrams, but more or fewer operation steps may be included based on conventional or non-creative labor. The order of steps listed in the embodiments is only one way of executing the order of many steps and does not represent the only execution order. When the actual system or device product is executed, it can be executed in the order or in parallel according to the method shown in the embodiments or flowcharts.

[0233] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A laser radar, comprising: A transmitting device, comprising at least one light emitting unit, wherein the light emitting unit is configured to emit a plurality of detection pulses for detecting obstacles during one detection process; A detection device, comprising at least one detection unit, wherein each detection unit comprises a pixel array, wherein the pixel array can respond to an echo reflected by the detection pulse on an obstacle and generate detection data; and a data processing device, coupled to the detection device and configured to acquire the detection data; The data processing device comprises a feedback processing module, and the feedback processing module is coupled to the transmitting device; Wherein, in one detection process, the transmitting device first transmits a first group of detection pulses, wherein the first group of detection pulses includes one or more detection pulses; The data processing device acquires a first set of detection data, where the first set of detection data is detection data generated after the pixel array responds to the echo reflected by the first set of detection pulses on the obstacle; The feedback processing module is configured to determine a subsequent transmission strategy of the transmitting device according to the first set of detection data.

2. The laser radar according to claim 1, wherein the feedback processing module is configured to: performing a first convolution process on the first set of detection data to obtain a first reconstructed pixel array; and A subsequent emission strategy of the emission device is determined according to the first recombined pixel array.

3. The laser radar according to claim 2, wherein the feedback processing module is configured to: Taking statistics on the data of the reconstructed pixels in the first reconstructed pixel array; and According to the statistical results, the subsequent launch strategy of the launch device is determined.

4. The laser radar according to claim 3, wherein the feedback processing module is configured to: Counting whether there is an over-threshold pulse in the data of the recombined pixels in the first recombined pixel array and obtaining a pulse statistical result, wherein the over-threshold pulse is a pulse exceeding a first intensity threshold; The subsequent transmission strategy of the transmitting device is determined according to the pulse statistical result.

5. The laser radar according to claim 4, wherein the feedback processing module is configured to: When the cross-threshold pulse exists in the data of the recombined pixels in the first recombined pixel array, further counting one or more of the following parameters: the intensity of each of the cross-threshold pulses, the obstacle distance corresponding to each of the cross-threshold pulses, the number of recombined pixels having the cross-threshold pulse, and the ratio of the recombined pixels having the cross-threshold pulse to all the recombined pixels in the first recombined pixel array; According to the further statistical results, the subsequent launch strategy of the launch device is determined. 6 . The laser radar according to claim 2 , wherein the feedback processing module is configured to perform ambient light compensation on data of at least part of the recombined pixels in the first recombined pixel array.

7. The laser radar according to any one of claims 1 to 6, wherein the emission strategy comprises one of the following: continuing to transmit a second set of detection pulses, wherein the second set of detection pulses includes one or more detection pulses; and Stop sending detection pulses and end this detection.

8. The laser radar according to claim 7, wherein the continuing to transmit the second set of detection pulses comprises: continuing to transmit the second set of detection pulses at a reduced transmission intensity; or The second set of detection pulses continues to be emitted at a constant emission intensity.

9. The laser radar according to claim 7, wherein the feedback processing module is configured to: When it is determined through the first group of detection data that an obstacle is detected based on a preset standard, the emission of detection pulses is stopped; otherwise, the emission of the second group of detection pulses continues.

10. The laser radar according to claim 7, wherein the feedback processing module is configured to: feed back the emission strategy to the emission device, and the emission device is configured to execute the emission strategy.

11. According to the laser radar of claim 10, during a detection process, when the transmitting device continues to transmit a second group of detection pulses, the data processing device is configured to: obtain a second group of detection data, wherein the second group of detection data is detection data generated after the pixel array responds to the echo reflected by the second group of detection pulses on the obstacle; the feedback processing module is configured to: determine the subsequent emission strategy of the transmitting device based on the second group of detection data.

12. The laser radar according to claim 7, wherein the data processing device further comprises: An output processing module is coupled to the feedback processing module. When the feedback processing module determines that the emission strategy is to stop emitting detection pulses, the output processing module is configured to: perform a second convolution processing on the detection data generated according to at least one group of detection pulses emitted by the transmitting device to obtain a second recombined pixel array, and generate a lidar point cloud according to the second recombined pixel array, wherein the detection data generated according to at least one group of detection pulses includes the first group of detection data.

13. The laser radar according to claim 12, wherein the second convolution process comprises: The detection data is convolved using a plurality of convolution kernels of different sizes to obtain a plurality of second recombined pixel arrays, and a laser radar point cloud is generated according to the plurality of second recombined pixel arrays.

14. The laser radar according to claim 12, wherein the output processing module is configured to: perform ambient light compensation on at least a portion of the recombined pixel data in the second recombined pixel array.

15. The laser radar according to claim 13, wherein the output processing module is configured to: determine a final recombined pixel array according to the state of the recombined pixel data in the plurality of second recombined pixel arrays, and generate a laser radar point cloud according to the final recombined pixel array.

16. The laser radar according to claim 15, wherein the output processing module is configured to determine the final recombined pixel array based on the pulse threshold crossing states of at least two adjacent recombined pixels in the second recombined pixel array.

17. The laser radar according to claim 16, wherein the output processing module is configured to determine a calculation method for intermediate recombined pixel data based on a pulse threshold crossing state of at least two adjacent recombined pixel data in the second recombined pixel array.

18. The laser radar according to claim 17, wherein the output processing module is configured to determine a calculation method for the intermediate recombined pixel data based on the pulse threshold crossing state of the at least two adjacent recombined pixel data and the distance and / or reflectivity of the corresponding obstacle.

19. The laser radar according to claim 17, wherein the calculation method of the intermediate recombined pixel data includes one of the following: interpolation calculation on the intermediate recombined pixels, single point calculation on the intermediate recombined pixels, and convolution calculation on the intermediate pixels.

20. A detection method of a laser radar, the laser radar comprising a transmitting device and a detecting device, the transmitting device comprising at least one light emitting unit, wherein the light emitting unit can emit a plurality of detection pulses in one detection process, the detecting device comprising at least one detection unit, wherein each detection unit comprises a pixel array; The detection method comprises: in a detection process, S11: firstly transmitting a first group of detection pulses, wherein the first group of detection pulses includes one or more detection pulses; S12: Responding to the echoes of the first group of detection pulses reflected on the obstacles through the pixel array of the detection unit and generating a first group of detection data; and S13: Determine a subsequent transmission strategy according to the first set of detection data.

21. The detection method according to claim 20, wherein the step S13 comprises: performing a first convolution process on the first set of detection data to obtain a first recombined pixel array; and A subsequent emission strategy of the emission device is determined according to the first recombined pixel array.

22. The detection method according to claim 21, wherein the step S13 comprises: Taking statistics on the data of the reconstructed pixels in the first reconstructed pixel array; and According to the statistical results, the subsequent launch strategy of the launch device is determined.

23. The detection method according to claim 22, wherein the step of performing statistics on the data of the recombined pixels in the first recombined pixel array comprises: Counting whether there is an over-threshold pulse in the data of the recombined pixels in the first recombined pixel array and obtaining a pulse statistical result, wherein the over-threshold pulse is a pulse exceeding a first intensity threshold; The subsequent transmission strategy of the transmitting device is determined according to the pulse statistical result.

24. The detection method according to claim 23, when the over-threshold pulse exists in the data of the recombined pixels in the first recombined pixel array, the step of taking statistics on the data of the recombined pixels in the first recombined pixel array further comprises: Further counting one or more of the following parameters: the intensity of each of the over-threshold pulses, the obstacle distance corresponding to each of the over-threshold pulses, the number of reconstructed pixels having the over-threshold pulses, and the ratio of the reconstructed pixels having the over-threshold pulses to all the reconstructed pixels in the first reconstructed pixel array; According to the further statistical results, the subsequent launch strategy of the launch device is determined.

25. The detection method according to claim 21, wherein the step S13 comprises: Ambient light compensation is performed on at least a portion of the reconstructed pixel data in the first reconstructed pixel array.

26. The detection method according to any one of claims 20 to 25, wherein the transmission strategy comprises one of the following: continuing to transmit a second set of detection pulses, wherein the second set of detection pulses includes one or more detection pulses; and Stop sending detection pulses and end this detection.

27. The detection method according to claim 26, wherein the step of continuing to transmit the second group of detection pulses comprises: continuing to transmit the second set of detection pulses at a reduced transmission intensity; or The second set of detection pulses continues to be emitted at a constant emission intensity.

28. The detection method according to claim 26, wherein the step S13 comprises: When an obstacle is detected based on the preset criteria through the first set of detection data, stopping the emission of detection pulses; Otherwise, the second group of detection pulses continues to be transmitted.

29. The detection method according to claim 26, further comprising: Feedback the launch strategy to the launch device; and The transmitting device is controlled to execute the fed-back transmitting strategy.

30. The detection method according to claim 29, during a detection process, when the second group of detection pulses continues to be emitted, the detection method further comprises: Acquire a second set of detection data, where the second set of detection data is detection data generated after the pixel array responds to echoes reflected by the second set of detection pulses on an obstacle; and The subsequent transmission strategy is determined based on the second set of detection data.

31. The detection method according to claim 26, when it is determined that the emission strategy is to stop emitting the detection pulse, the detection method further comprises: performing a second convolution process on the detection data generated according to at least one group of detection pulses transmitted by the transmitting device; A second recombined pixel array is obtained, and a lidar point cloud is generated according to the second recombined pixel array, wherein the detection data generated according to at least one group of detection pulses includes the first group of detection data.

32. The detection method according to claim 31, wherein the step of performing a second convolution process on the detection data generated according to at least one group of detection pulses transmitted by the transmitting device comprises: The detection data is convolved using a plurality of convolution kernels of different sizes to obtain a plurality of second reorganized pixel arrays, and a laser radar point cloud is generated according to the plurality of second reorganized pixel arrays.

33. The detection method according to claim 31, further comprising: Ambient light compensation is performed on at least a portion of the reconstructed pixel data in the second reconstructed pixel array.

34. The detection method according to claim 32, further comprising: A final recombined pixel array is determined according to the states of the recombined pixel data in the plurality of second recombined pixel arrays, and a laser radar point cloud is generated according to the final recombined pixel array.

35. The detection method according to claim 34, wherein the step of determining the final reconstructed pixel array according to the states of the reconstructed pixel data in the plurality of second reconstructed pixel arrays comprises: The final reorganized pixel array is determined according to pulse threshold crossing states of at least two adjacent reorganized pixels in the second reorganized pixel array.

36. The detection method according to claim 35, wherein the step of determining the final reconstructed pixel array according to the states of the reconstructed pixel data in the plurality of second reconstructed pixel arrays further comprises: The calculation method of the intermediate recombined pixel data is determined according to the pulse threshold crossing state of at least two adjacent recombined pixel data in the second recombined pixel array.

37. The detection method according to claim 36, wherein the step of determining the final reconstructed pixel array according to the states of the reconstructed pixel data in the plurality of second reconstructed pixel arrays further comprises: A calculation method of the intermediate recombined pixel data is determined according to the pulse threshold crossing states of the at least two recombined pixel data and the distances and / or reflectivities of the obstacles corresponding thereto.

38. The detection method according to claim 36, wherein the calculation method of the intermediate recombined pixel data comprises one of the following: interpolation of the intermediate pixels, single-point calculation of the intermediate pixels, and convolution calculation of the intermediate pixels.

Images