Echo selection method for lidar, and device and storage medium

Abstract

An echo selection method for a LiDAR, and a device and a storage medium. The echo selection method for a LiDAR comprises: acquiring echo data at pixel points (210); determining a spatial similarity between a target pixel point and a neighboring pixel point (220); determining a temporal similarity between each echo at the target pixel point and each reference echo at the neighboring pixel point (230); on the basis of the spatial similarity, the temporal similarity, and a peak value of each echo at the target pixel point, calculating an echo score of each echo at the target pixel point (240); and on the basis of the echo score of each echo at the target pixel point, determining an object echo at the target pixel point (250). In the method, surrounding pixel points are taken into account, and the characteristic of objects exhibiting spatial continuity in terms of distances is used, thereby improving the selection accuracy of real object echoes.

Description

Echo selection method, device and storage medium for laser radar Technical Field

[0001] The present disclosure relates generally to lidar, and particularly to echo selection in lidar. Background Art

[0002] LiDAR (Light Detection and Ranging) is a distance measurement technology. The basic principle of LiDAR is that a transmitter emits a laser beam. After striking an object in front of it, the laser beam reflects and returns to a receiver. The distance to the object is determined based on the time difference between transmission and reception (i.e., time-of-flight, ToF). The receiver records the number of photons received at different times, which is called echo data. When the laser beam reflects off an object, it forms an echo on the echo data. The position of this echo can be used to determine the distance to the object.

[0003] However, in real-world scenarios, accurately identifying true object echoes is a challenge. Generally speaking, echo data contains not only echoes from the true object but also echoes from ambient light (also known as noise). Current technology typically selects the echo with the largest peak value as the true object echo. However, in the presence of strong ambient light or when measuring distant objects, the peak value of the ambient light echo may be higher than that of the true object echo. This can lead to incorrect echo selection, resulting in reduced ranging accuracy.

[0004] Therefore, developing a more accurate and robust echo selection (also known as peak selection) method is an urgent problem to be solved in current lidar technology.

[0005] Summary of the Invention

[0006] One aspect of the present disclosure relates to an echo selection method for a laser radar. According to an embodiment of the present disclosure, the method includes: acquiring echo data at a pixel point; determining the spatial similarity between a target pixel point and adjacent pixels; determining the temporal similarity between each echo at the target pixel point and each reference echo at adjacent pixels; calculating the echo score of each echo at the target pixel point based on the spatial similarity, the temporal similarity, and the peak value of each echo at the target pixel point; and determining the object echo at the target pixel point based on the echo score of each echo at the target pixel point.

[0007] Another aspect of the present disclosure relates to an electronic device. The electronic device includes one or more processors and one or more memories storing one or more instructions. When executed by the one or more processors, the one or more instructions cause the one or more processors to perform the steps of the echo selection method for a laser radar according to an embodiment of the present disclosure.

[0008] Yet another aspect of the present disclosure relates to a computer-readable storage medium having one or more instructions stored thereon, which, when executed by a processor, causes the processor to perform the steps of the echo selection method for a laser radar according to an embodiment of the present disclosure.

[0009] Yet another aspect of the present disclosure relates to a computer program product comprising one or more instructions. When executed by a processor, the one or more instructions cause the processor to perform the steps of the echo selection method for a laser radar according to an embodiment of the present disclosure.

[0010] The above summary is provided to summarize some exemplary embodiments in order to provide a basic understanding of various aspects of the subject matter described herein. Therefore, the above features are merely examples and should not be construed as narrowing the scope or spirit of the subject matter described herein in any way. Other features, aspects, and advantages of the subject matter described herein will become apparent from the detailed description described below in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] A better understanding of the present disclosure may be obtained when the following detailed description of the embodiments is considered in conjunction with the accompanying drawings. The same or similar reference numerals are used in the various drawings to represent the same or similar components. The accompanying drawings, together with the following detailed description, are incorporated into and form a part of this specification and are used to illustrate the embodiments of the present disclosure and to explain the principles and advantages of the present disclosure. In particular:

[0012] Figure 1A is a schematic diagram illustrating an application scenario of an echo selection method for a lidar according to one or more embodiments of the present disclosure; and Figure 1B is a schematic diagram illustrating echo data obtained in the application scenario of Figure 1A.

[0013] FIG2 is a flowchart illustrating the steps of an echo selection method for a laser radar according to one or more embodiments of the present disclosure.

[0014] FIG3 is a flowchart illustrating an example of sub-steps of the step of calculating an echo score according to one or more embodiments of the present disclosure.

[0015] FIG4 is a flowchart illustrating an example of sub-steps of the step of determining an object echo according to one or more embodiments of the present disclosure.

[0016] FIG5 is a schematic diagram illustrating the steps of determining an object echo according to one or more embodiments of the present disclosure.

[0017] FIG6 is a schematic diagram illustrating an echo selection device for a laser radar according to one or more embodiments of the present disclosure, which illustrates the main functional modules and information interactions that constitute the device.

[0018] Figure 7A is a schematic diagram illustrating an application scenario for performing lidar ranging and the ambient light data therein; and Figures 7B-7C are schematic diagrams illustrating the ranging effects of using the echo selection method according to one or more embodiments of the present disclosure and the echo selection method using the prior art when performing lidar ranging on the scene illustrated in Figure 7A.

[0019] FIG8 is an example block diagram illustrating an electronic device according to one or more embodiments of the present disclosure.

[0020] While the embodiments described in this disclosure may be susceptible to various modifications and alternative forms, specific embodiments thereof are shown by way of example in the drawings and are herein described in detail. However, it should be understood that the drawings and detailed description thereof are not intended to limit the embodiments to the particular forms disclosed, but on the contrary, the intent is to cover all modifications, equivalents, and alternatives falling within the spirit and scope of the claims. DETAILED DESCRIPTION

[0021] The following describes representative applications of various aspects of the apparatus and method of the present disclosure. The description of these examples is only to add context and help understand the described embodiments. Therefore, it is clear to those skilled in the art that the embodiments described below can be implemented without some or all of the specific details. In other cases, well-known process steps are not described in detail to avoid unnecessarily obscuring the described embodiments. Other applications are also possible, and the solutions of the present disclosure are not limited to these examples.

[0022] In existing lidar technologies, the echo signal with the highest peak value at each pixel is typically selected as the true object echo signal at that pixel. In other words, the peak selection for each pixel is independent of each other. However, the inventors of this application have recognized that this peak selection approach cannot cope with complex situations such as strong ambient light or when measuring distant objects.

[0023] The inventors of this application also recognized that the objects to be measured typically exhibit a certain degree of spatial continuity, so the object echoes at pixels measuring the same object will be relatively concentrated. For example, assuming the distance measured at the target pixel is D, if nearby pixels measure the same object, the distances measured by these nearby pixels will likely be close to D. In this way, the flight times of the object echoes from the target pixel and this nearby pixel will be very close, or even overlap. In contrast, the echoes from ambient light noise are temporally random.

[0024] Therefore, this disclosure proposes a new and effective method for LiDAR echo selection. Compared to existing techniques that focus solely on a single pixel, this method also considers surrounding pixels. By leveraging the spatial continuity of object distances, it improves the accuracy of real-world object echo selection.

[0025] Specifically, the echo selection method for lidar proposed in the present disclosure screens out pixel points that may measure the same object based on the lidar's angular resolution, ambient light and other information. On this basis, echoes that are close in time are aggregated together to better select the real object echo, thereby improving the accuracy of selecting the real object echo and optimizing the lidar's measurement accuracy, bringing higher efficiency and practicality to the lidar.

[0026] The solution of the present disclosure is applicable to scenarios where time-of-flight laser radar ranging is performed. For example, the solution of the present disclosure can be applied to various types of direct time-of-flight (DTOF) laser radars.

[0027] Furthermore, because it can avoid noise interference and more accurately select real object echoes, the disclosed solution is particularly suitable for noise-sensitive LiDAR ranging scenarios. For example, single-photon avalanche diode (SPAD) LiDARs with high detection sensitivity and flash LiDARs with low emission energy density. Furthermore, as mentioned above, the disclosed solution is also particularly suitable for scenarios with strong ambient light or when measuring distant objects.

[0028] FIG1A illustrates a schematic diagram of an application scenario of an echo selection method for a laser radar according to one or more embodiments of the present disclosure. FIG1B illustrates a schematic diagram of echo data obtained in the application scenario of FIG1A . Those skilled in the art will readily appreciate that the application scenarios illustrated in FIG1A and FIG1B are merely examples and the present disclosure is not limited thereto.

[0029] As shown in Figure 1A, in a LiDAR ranging solution, a transmitter emits laser light and a receiver receives the reflected signal. The receiver records the number of photons received at different times, thereby generating echo data (as shown in Figure 1B). In Figure 1A, the solid line represents the laser light emitted by the transmitter and the echo of the real object, while the dotted line represents the ambient light. Assume that the LiDAR system shown in Figure 1A has a total of three pixels P1, P2, and P3, which have the corresponding relationship with the measurement object in the field of view as shown in the figure.

[0030] Here, the meaning of a pixel in a lidar is similar to that of a pixel in an image sensor, but the data recorded for each pixel is not the RGB color components, but rather the number of received photons (echo data). Pixels correspond one-to-one with the lidar's field of view, so they can be represented and distinguished by the field of view. Furthermore, the density of pixels can reflect the lidar's angular resolution.

[0031] (1), (2), and (3) in Figure 1B show the echo data at pixel points P1, P2, and P3, respectively. The horizontal axis indicates time, so the flight time of the echo can be shown; the vertical axis indicates the number of received photons, so the echo intensity can be shown. In the echo data in Figure 1B, solid line echoes (such as a, b, and c) represent real object echoes, and dotted line echoes (such as d, e, and f) represent ambient light echoes. In addition, there are some lower peaks in the echo data, but because they do not meet the echo judgment criteria (such as intensity and duration conditions, etc.), they can be ignored.

[0032] It's worth noting that in the example shown in Figure 1B , the real object echoes a and b are not the maximum echoes at the corresponding pixels P1 and P2, while the real object echo c at pixel P3 has similar intensities to the ambient light echo f. In these cases, it would be difficult to correctly select the real object echo using existing solutions.

[0033] FIG2 illustrates a flowchart of the steps of a method 200 for selecting an echo for a laser radar according to one or more embodiments of the present disclosure. FIG3-FIG4 illustrates a flowchart of examples of sub-steps of some steps of the method for selecting an echo for a laser radar according to an embodiment of the present disclosure. FIG5 illustrates a schematic diagram of the steps of determining an object echo according to one or more embodiments of the present disclosure. The method according to an embodiment of the present disclosure can be performed by any device including a processing device. For example, the method according to an embodiment of the present disclosure can be performed by a controller in a laser radar device, any processing device that communicates with the laser radar device, or a combination thereof.

[0034] As shown in FIG2 , according to an embodiment of the present disclosure, an echo selection method 200 for a laser radar may mainly include the following steps:

[0035] In step 210, echo data at the pixel point is acquired;

[0036] In step 220, the spatial similarity between the target pixel and the adjacent pixels is determined;

[0037] At step 230 , the temporal similarity between each echo at the target pixel and each reference echo at a nearby pixel is determined;

[0038] At step 240 , an echo score of each echo at the target pixel is calculated based on the spatial similarity, the temporal similarity, and the peak value of each echo at the target pixel; and

[0039] In step 250 , the object echo at the target pixel is determined based on the echo score of each echo at the target pixel.

[0040] Here, the target pixel point may refer to the pixel point that is currently being processed to determine its object echo. In addition, a similar pixel point may refer to a pixel point that is located near the target pixel point and provides reference information for determining the object echo. Therefore, the target pixel point and the similar pixel point are relative concepts and can even be interchanged in calculations for different pixel points. As an example, all pixel points in the lidar can be target pixel points. However, the present disclosure is not limited to this. For example, in some embodiments, only pixel points that meet specific conditions can be used as target pixel points. In the following description, the case where the pixel point P2 shown in Figures 1A-1B is the target pixel point is mainly described as an example, but the present disclosure is not limited to this.

[0041] First, as shown in FIG2 , in step 210 , echo data at a pixel point is acquired.

[0042] In some embodiments, step 210 of acquiring echo data at a pixel may include acquiring all echo data used in subsequent steps. However, the present disclosure is not limited thereto. For example, in some embodiments, echo data for all pixels may be acquired in step 210. Alternatively, in some embodiments, echo data for only one or more target pixels and their adjacent pixels may be acquired in step 210.

[0043] Next, as shown in FIG2 , in step 220 , the spatial similarity between the target pixel and its adjacent pixels is determined. For example, for the example scene shown in FIG1A-FIG1B , in step 220 , the spatial similarity between the target pixel P2 and its adjacent pixels P1 and P3 is determined respectively.

[0044] Here, the spatial similarity between pixels can be used to represent the similarity of the spatial position of the field of view corresponding to the pixel or the similarity of the measured object in the field of view corresponding to the pixel, thereby indicating the possibility that the pixel measures the same object in space. Therefore, calculating the spatial similarity between the target pixel and the adjacent pixels is beneficial for screening out the pixels that measure the same object as the target pixel. It is worth noting that, generally speaking, the pixels that measure the same object will be relatively close, so the present disclosure mainly considers screening the pixels near the target pixel to reduce unnecessary calculations.

[0045] The inventors of the present application have realized that, since objects are generally continuous in space, the possibility of measuring the same object can be determined based on the spatial position of the field of view.

[0046] As an example, in some embodiments, spatial similarity can be determined by the difference in field of view angles. For example, if the difference in field of view angles corresponding to two pixels is small, the spatial similarity of the two pixels is considered to be relatively high; otherwise, the spatial similarity of the two pixels is considered to be relatively low.

[0047] Therefore, in some embodiments, determining the spatial similarity between the target pixel point and the similar pixel points in step 220 may include: based on the difference in the field of view angles corresponding to the target pixel point and the similar pixel points, using a preset spatial threshold to determine the spatial similarity between the target pixel point and the similar pixel points.

[0048] It's worth noting that the difference in field of view angle between two adjacent pixels generally corresponds to the field of view angle resolution in the direction of arrangement. For example, the difference in horizontal field of view angle between two adjacent pixels in the horizontal direction can correspond to the field of view angle resolution in the horizontal direction; and the difference in vertical field of view angle between two adjacent pixels in the vertical direction can correspond to the field of view angle resolution in the vertical direction. Therefore, for two adjacent pixels, the difference in field of view angle in each direction can be determined based on the field of view angle resolution and the pixel number interval in each direction.

[0049] Generally speaking, once the lidar is installed and debugged, the field of view (FOV) resolution is fixed. Therefore, the correspondence between the pixel count interval and the field of view angle difference can be calculated based on the FOV resolution and a predetermined spatial threshold. In some embodiments, the spatial similarity between two pixels can be determined based on the pixel count interval between the two pixels.

[0050] For example, if it is believed based on calculation or experience that the objects measured within a horizontal field of view angle of ±1° are more likely to be the same object, that is, the predetermined spatial threshold is ±1°, and the field of view angle resolution in the horizontal direction is designed to be 0.5°, then it can be determined that the spatial similarity between the target pixel point and the pixel points arranged within ±2 in the horizontal direction is high.

[0051] Here, the number of preset space thresholds may be one or more.

[0052] For example, when the number of preset spatial thresholds is one, the spatial similarity determined using the preset spatial thresholds may have two values; when the number of preset spatial thresholds is multiple, the spatial similarity determined using the preset spatial thresholds may have three or more values.

[0053] Here, the spatial similarity can be a discrete set of values. In particular, for ease of calculation, the spatial similarity can be a discrete set of values ​​falling within the interval [0, 1]. However, the present disclosure is not limited thereto. For example, the spatial similarity can be a continuous value, in particular, a continuous value falling within the interval [0, 1].

[0054] The inventors of this application also recognized that the material of objects is generally consistent, and therefore the similarity of the fields of view can be determined based on the consistency of the material of the measured objects within the field of view. Furthermore, the material of the measured object (such as a wall, a metal sign, clothing fabric, etc.) generally affects the reflectivity, and thus the corresponding ambient light intensity. Therefore, differences in ambient light intensity can, to a certain extent, reflect differences in the material of the measured object.

[0055] Therefore, in some embodiments, the material consistency of the measured object can be evaluated by the difference in ambient light intensity, thereby determining spatial similarity. For example, if the difference in ambient light intensity corresponding to two pixels is small, the spatial similarity of the two pixels is considered to be relatively high; otherwise, the spatial similarity of the two pixels is considered to be relatively low.

[0056] Therefore, in some embodiments, determining the spatial similarity between the target pixel and the similar pixel points in step 220 may include: using a similarity measurement function to calculate the spatial similarity between the target pixel and the similar pixel points based on the ambient light information in the field of view corresponding to the target pixel and the similar pixel points.

[0057] Here, the ambient light information may be any information reflecting the intensity of the ambient light. For example, the ambient light information may include but is not limited to at least one of the following: ambient light data or echo data baseline.

[0058] In some embodiments, ambient light data can be obtained by directly measuring ambient light intensity. For example, in some embodiments, ambient light intensity can be determined based on received signals during a period when laser light is not emitted, thereby obtaining ambient light data. Alternatively, in some embodiments, weak echoes in echo data during a period when laser light is emitted can be selected to represent ambient light intensity, thereby obtaining ambient light data.

[0059] In some embodiments, the echo data baseline can be obtained by averaging the collected echo data. It is worth noting that the echo data acquisition time is generally much longer than the laser pulse duration. Therefore, even if the echo data includes real object echoes, it can still be used to represent the ambient light intensity. However, because the echo data within the acquisition time period may include real object echoes, the echo data baseline is generally slightly larger than the ambient light data.

[0060] In some embodiments, the similarity measurement function used to calculate the spatial similarity between the target pixel point and the similar pixel points may include but is not limited to at least one of the following: a similarity measurement function based on mean square error distance, a similarity measurement function based on absolute value distance, a Gaussian kernel function, and a similarity measurement function obtained based on machine learning, etc.

[0061] For example, for the example scene shown in FIG1A-FIG1B, if the Gaussian kernel function is used to calculate the spatial similarity between the target pixel point P2 and the adjacent pixels P1 and P3, the spatial similarity sim between the target pixel point P2 and the pixel point P1 is s,21 And the spatial similarity sim between the target pixel P2 and pixel P3 s,23 They can be expressed as:

[0062] [Formula 1]

[0063] Among them, x1, x2, and x3 are the ambient light intensities of pixels P1, P2, and P3 respectively, and r1 is the variance parameter of the Gaussian kernel function.

[0064] Although the above examples illustrate a method of determining spatial similarity using a preset spatial threshold and a method of calculating spatial similarity using a similarity measurement function, these are merely examples and the present disclosure is not limited thereto. Moreover, the two methods illustrated above may also be used in combination.

[0065] As shown in FIG2 , in step 230, the temporal similarity between each echo at the target pixel and each reference echo at a nearby pixel is determined. For example, for the example scenario shown in FIG1A-1B , in step 230, the temporal similarity between each echo (e, b) at the target pixel P2 and each reference echo (a, d, c, f) at the nearby pixels P1 and P3 is determined, respectively.

[0066] Here, temporal similarity between echoes can be used to indicate the similarity between their corresponding times of flight, thereby indicating the likelihood that the echoes originate from the same object. Since objects are generally spatially continuous, the measured distances of the same object are usually close, so the corresponding echoes' times of flight are also close, or even identical.

[0067] In some embodiments, the temporal similarity can be determined by the difference in flight time. For example, if the difference in flight time between two echoes is small, the temporal similarity between the two echoes is considered to be relatively high; otherwise, the temporal similarity between the two echoes is considered to be relatively low.

[0068] Therefore, in some embodiments, determining the temporal similarity between each echo at the target pixel and each reference echo at a similar pixel in step 230 may include: determining the temporal similarity using a preset time threshold based on the difference between the flight time of each echo and each reference echo.

[0069] In some embodiments, the number of preset time thresholds may be one or more.

[0070] For example, when the number of preset time thresholds is one, the time similarity determined using the preset time thresholds may have two values; when the number of preset time thresholds is multiple, the time similarity determined using the preset time thresholds may have three or more values.

[0071] The temporal similarity determined based on the temporal threshold may be a discrete set of values. In particular, for ease of calculation, the temporal similarity may be a discrete set of values ​​falling within the interval [0, 1]. However, the present disclosure is not limited thereto. For example, the temporal similarity may be a continuous value, in particular, a continuous value falling within the interval [0, 1].

[0072] In some embodiments, determining the temporal similarity between each echo at the target pixel and each reference echo at a nearby pixel in step 230 may include calculating the temporal similarity using a similarity metric function based on the difference between the flight time of each echo and each reference echo.

[0073] In some embodiments, the similarity measurement function used to calculate the temporal similarity between each echo at the target pixel point and each reference echo at the similar pixel point includes but is not limited to at least one of the following: a similarity measurement function based on mean square error distance, a similarity measurement function based on absolute value distance, a Gaussian kernel function, and a similarity measurement function obtained based on machine learning, etc.

[0074] The inventors of this application have recognized that the accuracy of measuring an echo's time-of-flight is related to the peak value of the echo. Generally speaking, the larger the peak value of the echo, the smaller the error in the measured echo's time-of-flight. Therefore, when using a similarity metric function to calculate temporal similarity based on time-of-flight, the calculated result may be affected by the peak value of the echo.

[0075] Therefore, in some embodiments, the similarity metric function used to calculate the temporal similarity may be configured to dynamically adjust parameters according to the peak values ​​of the two echoes for which the temporal similarity is calculated.

[0076] For example, if the peak values ​​of the two echoes are higher and thus the measured flight time is more accurate, the time similarity between the two echoes is calculated with a stricter standard; otherwise, the time similarity between the two echoes is calculated with a looser standard.

[0077] Advantageously, by dynamically adjusting the parameters of the similarity metric function used to calculate the temporal similarity, the influence of the peak fluctuation of the echo on the calculation result of the temporal similarity can be reduced.

[0078] For example, for the example scenario shown in FIG1A-FIG1B, if the Gaussian kernel function with dynamically adjusted parameters is used to calculate the temporal similarity between each echo at the target pixel point P2 and each reference echo at the adjacent pixels P1 and P3, the temporal similarity simt between the echo b at the target pixel point P2 and the echo c at the pixel point P3 is ,bc It can be expressed as:

[0079] [Formula 2]

[0080] Among them, t b , t c are the flight times of echoes b and c, respectively, and z b 、z c are the peak values ​​of echoes b and c, N is the maximum photon count per unit time of the receiver, and r2 is the variance parameter of the Gaussian kernel function. As shown in Equation 2, the Gaussian kernel function used to calculate the temporal similarity is calculated based on the peak values ​​z of echoes b and c. b 、z c And dynamically adjust the parameters.

[0081] In the example, the time similarity sim between the echo b at the target pixel point P2 and the echoes a, d, and f at the pixel points P1 and P3 can be calculated based on a similar method. t,ba 、sim t,bd and sim t,bf .

[0082] In the examples given above, both the spatial similarity and the temporal similarity are calculated using a similarity measurement function, but these are merely examples and the present disclosure is not limited thereto.

[0083] In addition, although the method of determining time similarity using a preset time threshold and the method of calculating time similarity using a similarity measurement function are illustrated above, these are only examples and the present disclosure is not limited thereto. Moreover, the two methods illustrated above can also be used in combination.

[0084] Next, as shown in FIG2 , in step 240, the echo score of each echo at the target pixel is calculated based on the spatial similarity, the temporal similarity, and the peak value of each echo at the target pixel. For example, for the example scene shown in FIG1A-FIG1B , the echo score of each echo (b, e) at the target pixel P2 is calculated in step 240.

[0085] An example of calculating the echo score for each echo at the target pixel (step 240) will be described in detail below with reference to the flowchart of FIG3 . Those skilled in the art will readily appreciate that the calculation method shown in FIG3 is merely an example, and the present disclosure is not limited thereto. Those skilled in the art may also utilize various existing calculation methods in conjunction with the concepts disclosed herein to obtain the echo score for each echo at the target pixel.

[0086] As shown in FIG. 3 , in some embodiments, step 240 of calculating the echo score of each echo at the target pixel point may include the following sub-steps.

[0087] In step 242 , the inter-echo similarity between the echo and each reference echo is calculated based on the temporal similarity between the echo and each reference echo and the spatial similarity between the target pixel and the adjacent pixel to which the reference echo belongs.

[0088] In some embodiments, calculating the inter-echo similarity between the echo and each reference echo in step 242 may include: multiplying the temporal similarity between the echo and each reference echo and the spatial similarity between the target pixel point and the adjacent pixel point to which the reference echo belongs to obtain the inter-echo similarity between the echo and each reference echo.

[0089] For example, for the example scenario shown in FIG. 1A-FIG . 1B , the inter-echo similarity between the echo b at the target pixel point P2 and the echo c at the pixel point P3 can be expressed as:

[0090] [Equation 3] sim bc =sim s,23 ·sim t,bc .

[0091] In addition, the similarity sim between the echo b at the target pixel point P2 and the echoes a, d, and f at the pixel points P1 and P3 can be calculated based on a similar method. ba 、sim bd and sim bf .

[0092] However, the above method for calculating the similarity between echoes is only an example, and the present disclosure is not limited thereto. For example, the similarity between echoes may also be calculated by taking a weighted average or the like.

[0093] In step 244 , a similar echo peak value of the echo is calculated based on the inter-echo similarity between the echo and each reference echo and the peak value of each reference echo.

[0094] In some embodiments, calculating the similar echo peak value of the echo in step 244 may include: performing weighted summation on the peak values ​​of all reference echoes of the echo based on the inter-echo similarity between the echo and each reference echo to obtain the similar echo peak value of the echo.

[0095] For example, for the example scenes shown in FIG. 1A and FIG. 1B , the similar echo peak value of the echo b at the target pixel point P2 can be expressed as:

[0096] [Formula 4]

[0097] z′ b =sim ba z a +sim bd z d +sim bc z c +sim bf z f

[0098] =sim s,12 sim t,ba z a +sim s,12 sim t,bd z d +sim s,23 sim t,bc z c +sim s,23 sim t,bf z f .

[0099] In addition, the similar echo peak value z′ of the echo e at the target pixel point P2 can be calculated based on a similar method e .

[0100] At step 246 , an echo score of the echo is calculated based on the peak value of the echo and the peak values ​​of similar echoes of the echo.

[0101] In some embodiments, calculating the echo score of the echo in step 246 may include: summing the peak value of the echo and the peak values ​​of similar echoes of the echo according to a preset proportionality coefficient to obtain the echo score of the echo.

[0102] For example, for the example scene shown in FIG. 1A-FIG . 1B , the echo score of the echo b at the target pixel point P2 can be expressed as:

[0103] [Formula 5] score b =kz b +jz′ b

[0104] Among them, k and j are preset proportional coefficients.

[0105] In addition, the echo score score of the echo e at the target pixel point P2 can be calculated based on a similar method. e .

[0106] In some embodiments, the above-mentioned proportional coefficients k and j can be set to constants. For example, the above-mentioned proportional coefficients k and j can be set to satisfy 1:1.

[0107] Alternatively, in some embodiments, the proportionality factor may be set to be dynamically adjustable. For example, the proportionality factor may be set to be dynamically adjusted according to the relationship between the peak value of the echo and the amplitude of the ambient light.

[0108] In some embodiments, the proportionality coefficient between the peak value of the echo and the peak value of a similar echo may be set to satisfy: the greater the ratio of the peak value of the echo to the amplitude of the ambient light, the greater the proportionality coefficient.

[0109] Generally speaking, the larger the peak value of the echo relative to the ambient light amplitude, the easier it is to directly determine the object echo based on the peak value, and the less dependent on the reference echo from nearby pixels. Therefore, dynamically adjusting the above-mentioned proportionality factor based on the relationship between the peak value of the echo and the amplitude of the ambient light can advantageously prevent the reference echo from nearby pixels from excessively influencing the selection of the object echo from the target pixel, thereby improving ranging accuracy.

[0110] Those skilled in the art will readily appreciate that the above-described method for calculating echo scores is merely an example and the present disclosure is not limited thereto. For example, in some embodiments, the echo score of each echo may be calculated based solely on spatial similarity, temporal similarity, and the peak value of each echo at the target pixel, without considering the peak value of the reference echo at the adjacent pixel.

[0111] Then, as shown in FIG2 , in step 250, the object echo at the target pixel is determined based on the echo score of each echo at the target pixel. For example, for the example scene shown in FIG1A-FIG1B , in step 250, based on the echo score of each echo b and e at the target pixel P2, b and score e , determine the object echo at the target pixel point P2.

[0112] In some embodiments, determining the object echo at the target pixel point in step 250 may include: comparing the echo scores of all echoes at the target pixel point; and based on the comparison result, selecting the echo with the largest echo score from all echoes as the object echo at the target pixel point.

[0113] For example, for the example scenario shown in FIG. 1A-FIG . 1B , in step 250 , the echo scores of all echoes b and e at the target pixel point P2 are compared. b 、score e (As shown in FIG5 ), and based on the comparison result, the echo b with the largest echo score is selected from all the echoes as the object echo at the target pixel point P2.

[0114] The flowchart of FIG4 illustrates another example of determining the object echo at the target pixel point (step 250). This example will be described in detail below with reference to the flowchart of FIG4.

[0115] As shown in FIG. 4 , in some embodiments, step 250 of determining the object echo at the target pixel point may include the following sub-steps.

[0116] In step 252 , it is determined whether the target pixel is a valid pixel or an invalid pixel based on the echo scores of the respective echoes at the target pixel.

[0117] For example, for the example scenario shown in Figures 1A and 1B, after calculating the echo scores of each echo using pixels P1, P2, and P3 as target pixels, if the highest echo score of pixel P1 is still less than a preset echo score threshold, pixel P1 can be considered invalid. Those skilled in the art will readily appreciate that the above method for determining valid / invalid pixels is merely an example and the present disclosure is not limited thereto.

[0118] In step 253, it is determined whether the target pixel is a valid pixel or an invalid pixel.

[0119] In step 254 , in response to determining that the target pixel is an invalid pixel (“No” in step 253 ), the target pixel is deleted.

[0120] The inventors of the present application have realized that if the echo scores of all echoes at a certain pixel are very low, then the pixel may not actually detect a real object. Deleting such pixels can advantageously avoid the influence of noise on the results and reduce the amount of calculation.

[0121] In step 256 , in response to determining that the target pixel is a valid pixel (“yes” in step 253 ), the echo score of each echo at the target pixel is recalculated after deleting invalid pixels.

[0122] In some embodiments, the step of recalculating the echo score of each echo at the target pixel is similar to the above steps 220 - 240 , except that only valid pixels are used for calculation.

[0123] For example, recalculating the echo score of each echo at the target pixel point in step 256 may include: determining the spatial similarity between the target pixel point and the adjacent valid pixel points; determining the temporal similarity between each echo at the target pixel point and each reference echo at the adjacent valid pixel points; and calculating the echo score of each echo at the target pixel point based on the spatial similarity, the temporal similarity and the peak value of each echo at the target pixel point.

[0124] At step 258 , object returns are determined based on the recalculated echo scores for each return.

[0125] In some embodiments, in step 258 , the recalculated echo scores of all echoes at the target pixel are compared; and based on the comparison result, the echo with the largest recalculated echo score is selected from all echoes as the object echo at the target pixel.

[0126] The boundaries between the various steps in the method described above are merely illustrative. In actual operation, the various steps can be combined in any manner, or even combined into a single step. In addition, the order in which the various steps are performed is not limited by the order in which they are described, and some steps may be omitted. The operational steps of the various embodiments may also be combined with each other in any appropriate order to similarly implement more or fewer operations than described.

[0127] For example, in some embodiments, step 220 of determining the spatial similarity of pixels and step 230 of determining the temporal similarity of echoes may be performed in parallel. Alternatively, in some embodiments, step 220 of determining the spatial similarity of pixels and step 230 of determining the temporal similarity of echoes may be performed in a specific order.

[0128] For example, in some embodiments, the calculation of spatial similarity is performed first. If the spatial similarity of two pixels is low, for example, lower than a preset minimum spatial similarity threshold, the spatial similarity of the two pixels can be set to 0. In this case, the calculation of the temporal similarity of the echoes at the two pixels can even be omitted, and the corresponding temporal similarity can be directly set to 0, thereby advantageously improving the accuracy of peak selection and saving computing resources. For example, for the example scenario shown in Figures 1A-1B, if it is determined in step 220 that the spatial similarity between pixels P2 and P1 is low, or even lower than a preset minimum spatial similarity threshold, the spatial similarity between pixel P2 and pixel P1 can be set to 0. In addition, the calculation of the temporal similarity between the echoes of pixel P2 and pixel P1 can be directly skipped. Therefore, although the temporal similarity between echo d and echo e may be high, it will not have a negative impact on the results.

[0129] The following is an exemplary description of an echo selection device 600 for a laser radar according to an embodiment of the present disclosure in conjunction with FIG6 . For ease of understanding, FIG6 illustrates the main functional modules and some information interactions of the device 600. According to an embodiment of the present disclosure, the echo selection device 600 for a laser radar can be configured to perform the various steps of the echo selection method for a laser radar according to an embodiment of the present disclosure. The contents described above in conjunction with FIG1A-FIG1B and FIG2-FIG5 can also be applied to the corresponding features, and the description of some repeated contents will be omitted.

[0130] In an embodiment of the present disclosure, as shown in FIG6 , an echo selection device 600 for a laser radar may include:

[0131] The echo data acquisition module 602 is configured to acquire echo data at a pixel point;

[0132] The spatial similarity calculation module 604 is configured to determine the spatial similarity between the target pixel and the adjacent pixels;

[0133] a temporal similarity calculation module 606 configured to determine a temporal similarity between each echo at a target pixel and each reference echo at a nearby pixel;

[0134] an echo score calculation module 608 configured to calculate an echo score of each echo at a target pixel based on the spatial similarity, the temporal similarity, and the peak value of each echo at the target pixel; and

[0135] The object echo selection module 610 is configured to determine the object echo at the target pixel point based on the echo score of each echo at the target pixel point.

[0136] In some embodiments, the spatial similarity calculation module 604 may be configured to determine the spatial similarity between the target pixel and the adjacent pixel using a preset spatial threshold based on the difference in field of view angles corresponding to the target pixel and the adjacent pixel.

[0137] Alternatively, in some embodiments, the spatial similarity calculation module 604 can be configured to use a similarity measurement function to calculate the spatial similarity between the target pixel and the adjacent pixels based on the ambient light information in the field of view corresponding to the target pixel and the adjacent pixels.

[0138] Although only two examples of the spatial similarity calculation module 604 performing corresponding functions are described here, these are merely examples and the present disclosure is not limited thereto. Moreover, the above examples may also be combined.

[0139] In some embodiments, the time similarity calculation module 606 may be configured to determine the time similarity using a preset time threshold based on the difference between the flight time of each echo and each reference echo.

[0140] Alternatively, in some embodiments, the temporal similarity calculation module 606 may be configured to calculate the temporal similarity using a similarity metric function based on the difference between the flight time of each echo and each reference echo.

[0141] Optionally, in some embodiments, the time similarity calculation module 606 may be configured to dynamically adjust parameters of the similarity metric function according to the peak values ​​of the two echoes for which the time similarity is calculated.

[0142] Although only two examples of the time similarity calculation module 606 performing corresponding functions are described here, these are only examples and the present disclosure is not limited thereto. Moreover, the above examples may also be combined.

[0143] In some embodiments, the echo score calculation module 608 may include an inter-echo similarity calculation submodule. The inter-echo similarity calculation submodule may be configured to calculate the inter-echo similarity between the echo and each reference echo based on the temporal similarity between the echo and each reference echo and the spatial similarity between the target pixel and the adjacent pixel to which the reference echo belongs.

[0144] Optionally, in some embodiments, the inter-echo similarity calculation submodule can be configured to multiply the temporal similarity between the echo and each reference echo and the spatial similarity between the target pixel point and the adjacent pixel points to which the reference echo belongs to obtain the inter-echo similarity between the echo and each reference echo.

[0145] In some embodiments, the echo score calculation module 608 may further include a similar echo peak calculation submodule configured to calculate the similar echo peak value of the echo based on the inter-echo similarity between the echo and each reference echo and the peak value of each reference echo.

[0146] Optionally, in some embodiments, the similar echo peak value calculation submodule may be configured to perform weighted summation on the peak values ​​of all reference echoes of the echo based on the inter-echo similarity between the echo and each reference echo to obtain the similar echo peak value of the echo.

[0147] In some embodiments, the echo score calculation module 608 may further include an echo score acquisition submodule. The echo score acquisition submodule may be configured to calculate the echo score of the echo based on the peak value of the echo and the peak value of similar echoes of the echo.

[0148] Optionally, in some embodiments, the echo score acquisition submodule may be configured to sum the peak value of the echo and the peak values ​​of similar echoes of the echo according to a preset proportional coefficient to obtain the echo score of the echo.

[0149] Optionally, in some embodiments, the above-mentioned proportional coefficient is set to a constant.

[0150] Optionally, in some embodiments, the proportionality coefficient is set to be dynamically adjustable. For example, the proportionality coefficient can be set to be dynamically adjusted according to the relationship between the peak value of the echo and the amplitude of the ambient light.

[0151] In some embodiments, the object echo selection module 610 may include a comparison submodule and a selection submodule. The comparison submodule may be configured to compare the echo scores of all echoes at the target pixel. The selection submodule may be configured to select the echo with the largest echo score from all echoes based on the comparison result as the object echo at the target pixel.

[0152] Alternatively, in some embodiments, the object echo selection module 610 may include a validity determination submodule configured to determine whether a target pixel is a valid pixel or an invalid pixel based on the echo scores of each echo at the target pixel.

[0153] In some embodiments, the object echo selection module 610 may further include a pixel deletion submodule. The pixel deletion submodule may be configured to delete a target pixel in response to determining that the target pixel is an invalid pixel.

[0154] In some embodiments, the object echo selection module 610 may further include a recalculation submodule. The recalculation submodule may be configured to, in response to determining that the target pixel is a valid pixel, recalculate the echo score of each echo at the target pixel after deleting invalid pixels. The object echo selection module 610 may further include a determination submodule. The determination submodule may be configured to determine the object echo based on the recalculated echo score of each echo.

[0155] As analyzed above, the echo selection method and device for laser radar proposed in the present disclosure can better select the echo of the real object by referring to similar echoes at pixel points that may measure the same object, thereby improving the accuracy of selecting the echo of the real object, improving the ranging performance of the laser radar in strong ambient light, low laser energy and long distance, optimizing the measurement accuracy of the laser radar, and bringing higher efficiency and practicality to the laser radar.

[0156] Figure 7A illustrates a schematic diagram of an application scenario for performing lidar ranging and the ambient light data therein, and Figures 7B-7C respectively illustrate schematic diagrams of the ranging effects (point clouds) when performing lidar ranging on the scene illustrated in Figure 7A using the echo selection method according to one or more embodiments of the present disclosure and the echo selection method using the prior art.

[0157] In the example, a SPAD laser radar can be used to measure the distance of the scene shown in Figure 7A, and a point cloud can be drawn based on the distance measured for each pixel. For ease of comparison, a reflectivity plate is set in the scene shown in Figure 7A. Figure 7B illustrates a schematic diagram of the ranging effect (point cloud) using the echo selection method of the prior art. As shown in Figure 7B, due to the strong ambient light, the real object echo of the laser is difficult to distinguish from the random ambient light echo, resulting in the peak selection method based on the maximum peak value selecting a large number of erroneous echoes, which in turn affects the final ranging and point cloud effects. For example, the 10% reflectivity plate is almost invisible on the point cloud. Figure 7C illustrates a schematic diagram of the ranging effect (point cloud) using the echo selection method according to one or more embodiments of the present disclosure. As shown in Figure 7C, the echo selection method according to one or more embodiments of the present disclosure can significantly improve the peak selection effect, thereby improving the ranging and point cloud quality. For example, the point cloud of the 10% reflectivity plate is basically correct.

[0158] The present disclosure also provides an electronic device. The electronic device includes one or more processors and one or more memories storing one or more instructions. When executed by the one or more processors, the one or more instructions cause the one or more processors to perform the steps of the echo selection method for a laser radar according to the present disclosure.

[0159] The present disclosure also provides a computer-readable storage medium having one or more instructions stored thereon. When executed by a processor, the one or more instructions cause the processor to perform the steps of the echo selection method for a laser radar according to the present disclosure.

[0160] The present disclosure also provides a computer program product comprising one or more instructions. When executed by a processor, the one or more instructions cause the processor to perform the steps of the echo selection method for a laser radar according to the present disclosure.

[0161] It should be understood that the instructions in the computer-readable storage medium according to the embodiments of the present disclosure can be configured to perform operations corresponding to the above-mentioned system and method embodiments. When referring to the above-mentioned system and method embodiments, the embodiments of the computer-readable storage medium are clear to those skilled in the art and are therefore not described again. Computer-readable storage media for carrying or including the above-mentioned instructions also fall within the scope of the present disclosure. Such computer-readable storage media may include, but are not limited to, floppy disks, optical disks, magneto-optical disks, memory cards, memory sticks, and the like.

[0162] The embodiments of the present disclosure also provide various devices including components or units for executing the steps of the echo selection method for laser radar in the above embodiments.

[0163] It should be noted that the above-mentioned various components or units are only logical modules divided according to the specific functions implemented by them, rather than being used to limit specific implementation methods, such as can be implemented in the form of software, hardware or a combination of software and hardware. In actual implementation, the above-mentioned various components or units can be implemented as independent physical entities, or can also be implemented by a single entity (for example, a processor (CPU or DSP, etc.), an integrated circuit, etc.). For example, a plurality of functions included in a unit in the above embodiments can be implemented by separate devices. Alternatively, a plurality of functions implemented by a plurality of units in the above embodiments can be implemented by separate devices respectively. In addition, one of the above functions can be implemented by a plurality of units.

[0164] Additionally, it should be understood that the aforementioned series of processes and devices can also be implemented via software and / or firmware. When implemented via software and / or firmware, the programs constituting the software are installed from a storage medium or network onto a computer with dedicated hardware, such as the general-purpose computer 800 shown in FIG8 . When the various programs are installed, the computer can perform various functions, among others. FIG8 illustrates an example block diagram of a computer that can be implemented as an echo selection device, application device, and system for a lidar according to an embodiment of the present disclosure.

[0165] 8 , a central processing unit (CPU) 801 executes various processes according to a program stored in a read-only memory (ROM) 802 or a program loaded from a storage section 808 to a random access memory (RAM) 803. In the RAM 803, data required when the CPU 801 executes various processes and the like is also stored as needed.

[0166] The CPU 801, the ROM 802, and the RAM 803 are connected to one another via a bus 804. An input / output interface 805 is also connected to the bus 804.

[0167] The following components are connected to the input / output interface 805: an input section 806 including a keyboard, a mouse, etc.; an output section 807 including a display such as a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 808 including a hard disk, etc.; and a communication section 809 including a network interface card such as a LAN card, a modem, etc. The communication section 809 performs communication processing via a network such as the Internet.

[0168] A drive 810 is also connected to the input / output interface 805 as needed. A removable medium 811 such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc. is mounted on the drive 810 as needed so that a computer program read therefrom is installed in the storage section 808 as needed.

[0169] In the case of realizing the above-described series of processing by software, a program constituting the software is installed from a network such as the Internet or a storage medium such as the removable medium 811 .

[0170] Those skilled in the art will appreciate that such storage media are not limited to the removable medium 811 shown in FIG8 , which stores the program and is distributed separately from the device to provide the program to the user. Examples of the removable medium 811 include magnetic disks (including floppy disks (registered trademark)), optical disks (including compact disk read-only memories (CD-ROMs) and digital versatile disks (DVDs)), magneto-optical disks (including minidiscs (MDs) (registered trademark)), and semiconductor memories. Alternatively, the storage medium may be a ROM 802, a hard disk included in the storage section 808, or the like, in which the program is stored and distributed to the user together with the device containing the program.

[0171] The exemplary embodiments of the present disclosure are described above with reference to the accompanying drawings, but the present disclosure is certainly not limited to the above examples. Those skilled in the art may obtain various changes and modifications within the scope of the appended claims, and it should be understood that these changes and modifications will naturally fall within the technical scope of the present disclosure.

[0172] Although the present disclosure and its advantages have been described in detail, it should be understood that various changes, substitutions and transformations can be made without departing from the spirit and scope of the present disclosure as defined by the appended claims. Moreover, the terms "comprises," "comprising," or any other variations thereof in the embodiments of the present disclosure are intended to cover non-exclusive inclusions, such that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article, or device. In the absence of further restrictions, an element defined by the statement "comprising a ..." does not exclude the presence of additional identical elements in the process, method, article, or device comprising the element.

[0173] The embodiments of the present disclosure also include the following.

[0174] 1. A method for selecting echoes for a laser radar, comprising:

[0175] Acquire echo data at the pixel point;

[0176] Determine the spatial similarity between the target pixel and the adjacent pixels;

[0177] Determining the temporal similarity between each echo at the target pixel and each reference echo at a nearby pixel;

[0178] Calculating an echo score of each echo at the target pixel based on the spatial similarity, the temporal similarity, and the peak value of each echo at the target pixel; and

[0179] An object echo at the target pixel is determined based on the echo score of each echo at the target pixel.

[0180] 2. The method according to item 1, wherein determining the spatial similarity between the target pixel and the adjacent pixels comprises:

[0181] Based on the difference in the field of view angles corresponding to the target pixel and the adjacent pixel, a preset spatial threshold is used to determine the spatial similarity between the target pixel and the adjacent pixel.

[0182] 3. The method according to item 1, wherein determining the spatial similarity between the target pixel and the adjacent pixels comprises:

[0183] Based on the ambient light information in the field of view corresponding to the target pixel and the adjacent pixels, a similarity measurement function is used to calculate the spatial similarity between the target pixel and the adjacent pixels.

[0184] 4. The method according to item 3, wherein the ambient light information includes at least one of the following:

[0185] Ambient light data; or

[0186] Echo data baseline.

[0187] 5. The method according to item 1, wherein determining the temporal similarity between each echo at the target pixel and each reference echo at a nearby pixel comprises:

[0188] The temporal similarity is determined using a preset time threshold based on the difference between the flight time of each echo and each reference echo.

[0189] 6. The method according to item 1, wherein determining the temporal similarity between each echo at the target pixel and each reference echo at a nearby pixel comprises:

[0190] The temporal similarity is calculated using a similarity metric function based on the difference between the flight time of each echo and each reference echo.

[0191] 7. The method according to item 6, wherein the similarity measurement function is configured to dynamically adjust parameters according to the peak values ​​of the two echoes for calculating the temporal similarity.

[0192] 8. The method according to item 1, wherein calculating the echo score of each echo at the target pixel comprises:

[0193] Calculating the inter-echo similarity between the echo and each of the reference echoes based on the temporal similarity between the echo and each of the reference echoes and the spatial similarity between the target pixel and the adjacent pixel to which the reference echo belongs;

[0194] Calculating a similar echo peak value of the echo according to the inter-echo similarity between the echo and each reference echo and the peak value of each reference echo;

[0195] An echo score of the echo is calculated according to a peak value of the echo and a peak value of a similar echo of the echo.

[0196] 9. The method according to item 8, wherein the calculation of the inter-echo similarity between the echo and each reference echo includes: multiplying the temporal similarity between the echo and each reference echo and the spatial similarity between the target pixel point and the adjacent pixel point to which the reference echo belongs, so as to obtain the inter-echo similarity between the echo and each reference echo.

[0197] 10. The method according to item 8, wherein the calculating of the similar echo peak value of the echo comprises: performing weighted summation on the peak values ​​of all reference echoes of the echo based on the inter-echo similarity between the echo and each reference echo to obtain the similar echo peak value of the echo.

[0198] 11. The method according to item 8, wherein calculating the echo score of the echo comprises: summing the peak value of the echo and the peak value of similar echoes of the echo according to a preset proportional coefficient to obtain the echo score of the echo.

[0199] 12. The method according to item 11, wherein the proportionality factor is set to a constant.

[0200] 13. The method according to item 11, wherein the proportional coefficient is set to be dynamically adjustable.

[0201] 14. The method according to item 13, wherein the proportional coefficient is configured to be dynamically adjusted according to a magnitude relationship between a peak value of the echo and an amplitude of ambient light.

[0202] 15. The method according to item 1, wherein determining the object echo at the target pixel point comprises:

[0203] Comparing the echo scores of all echoes at the target pixel; and

[0204] Based on the comparison results, the echo with the largest echo score is selected from all echoes as the object echo at the target pixel.

[0205] 16. The method according to item 1, wherein determining the object echo at the target pixel comprises:

[0206] According to the echo scores of each echo at the target pixel point, it is determined whether the target pixel point is a valid pixel point or an invalid pixel point.

[0207] 17. The method according to item 16, wherein, in response to determining that the target pixel is an invalid pixel, the target pixel is deleted.

[0208] 18. The method according to item 17, wherein, in response to determining that the target pixel is a valid pixel,

[0209] After deleting invalid pixels, recalculating the echo score of each echo at the target pixel; and

[0210] The object echo is determined based on the recalculated echo score of each echo.

[0211] 19. An electronic device comprising:

[0212] one or more processors; and

[0213] one or more memories having one or more instructions stored thereon;

[0214] When the one or more instructions are executed by the one or more processors, the one or more processors are caused to perform the steps of the method according to any one of items 1-18.

[0215] 20. A computer-readable storage medium having one or more instructions stored thereon, which, when executed by a processor, cause the processor to perform the steps of the method according to any one of items 1-18.

[0216] 21. A computer program product comprising one or more instructions, which, when executed by a processor, cause the processor to perform the steps of the method according to any one of items 1-18.

Claims

1. A method for selecting echoes for a laser radar, comprising: Acquire echo data at a pixel point; Determine the spatial similarity between the target pixel and the adjacent pixels; Determine the temporal similarity between each echo at the target pixel and each reference echo at a nearby pixel; Calculating an echo score of each echo at the target pixel based on the spatial similarity, the temporal similarity, and the peak value of each echo at the target pixel; as well as The object echo at the target pixel is determined based on the echo score of each echo at the target pixel.

2. The method of claim 1, wherein: Determining the spatial similarity between the target pixel and the adjacent pixels includes: Based on the difference in the field of view angles corresponding to the target pixel and the adjacent pixel, a preset spatial threshold is used to determine the spatial similarity between the target pixel and the adjacent pixel.

3. The method of claim 1, wherein: Determining the spatial similarity between the target pixel and the adjacent pixels includes: Based on the ambient light information in the field of view corresponding to the target pixel and the adjacent pixels, a similarity measurement function is used to calculate the spatial similarity between the target pixel and the adjacent pixels.

4. The method of claim 3, wherein: The ambient light information includes at least one of the following: Ambient light data; or Echo data baseline.

5. The method of claim 1, wherein: Determining the time similarity between each echo at the target pixel and each reference echo at a similar pixel includes: The temporal similarity is determined using a preset time threshold based on the difference between the flight time of each echo and each reference echo.

6. The method of claim 1, wherein: Determining the time similarity between each echo at the target pixel and each reference echo at a similar pixel includes: The time similarity is calculated using a similarity metric function based on the difference between the flight time of each echo and each reference echo.

7. The method of claim 6, wherein: The similarity metric function is configured to dynamically adjust parameters according to the peak values ​​of the two echoes for calculating the temporal similarity.

8. The method of claim 1, wherein: The calculating the echo score of each echo at the target pixel point comprises: Calculating the inter-echo similarity between the echo and each of the reference echoes according to the temporal similarity between the echo and each of the reference echoes and the spatial similarity between the target pixel and the adjacent pixel to which the reference echo belongs; Calculating a similar echo peak value of the echo according to the echo-echo similarity between the echo and each reference echo and the peak value of each reference echo; An echo score of the echo is calculated according to a peak value of the echo and a peak value of a similar echo of the echo.

9. The method of claim 8, wherein: Calculating the inter-echo similarity between the echo and each reference echo includes: multiplying the temporal similarity between the echo and each reference echo and the spatial similarity between the target pixel and the adjacent pixel to which the reference echo belongs, so as to obtain the inter-echo similarity between the echo and each reference echo.

10. The method of claim 8, wherein: The calculating the similar echo peak value of the echo comprises: performing weighted summation on the peak values ​​of all reference echoes of the echo based on the inter-echo similarity between the echo and each reference echo, so as to obtain the similar echo peak value of the echo.

11. The method of claim 8, wherein: The calculating the echo score of the echo includes: summing the peak value of the echo and the peak value of a similar echo of the echo according to a preset proportionality coefficient to obtain the echo score of the echo.

12. The method of claim 11, wherein: The proportionality factor is set to a constant.

13. The method of claim 11, wherein: The proportionality factor is configured to be dynamically adjustable.

14. The method of claim 13, wherein: The proportionality coefficient is configured to be dynamically adjusted according to the magnitude relationship between the peak value of the echo and the amplitude of the ambient light.

15. The method according to claim 1, wherein determining the object echo at the target pixel point comprises: Compare the echo scores of all echoes at the target pixel; as well as Based on the comparison result, the echo with the largest echo score is selected from all echoes as the object echo at the target pixel.

16. The method according to claim 1, wherein determining the object echo at the target pixel point comprises: According to the echo scores of each echo at the target pixel point, it is determined whether the target pixel point is a valid pixel point or an invalid pixel point.

17. The method of claim 16, wherein: In response to determining that the target pixel is an invalid pixel, the target pixel is deleted.

18. The method of claim 17, wherein: In response to determining that the target pixel is a valid pixel, After deleting the invalid pixels, the echo score of each echo at the target pixel is recalculated; as well as The object echo is determined based on the recalculated echo score of each echo.

19. An electronic device comprising: one or more processors; as well as One or more memories having one or more instructions stored thereon; When the one or more instructions are executed by the one or more processors, the one or more processors are caused to perform the steps of the method according to any one of claims 1-18.

20. A computer-readable storage medium having one or more instructions stored thereon, which, when executed by a processor, cause the processor to perform the steps of the method according to any one of claims 1-18.

21. A computer program product comprising one or more instructions which, when executed by a processor, cause the processor to perform the steps of the method according to any one of claims 1 to 18.

Images